Sleep Quality in Elite Sport: Measurement, Management, and Performance

Sleep quality in elite sport: measurement, management, and performance

Luke Gupta

Doctoral Thesis

Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University

May 2020

Abstract

Little research to date, has explored sleep ‘quality’-sport interactions within the sport science and medicine literature and, as a result, information on insomnia symptomatology and its implications for sports performance among elite athletes remains unexplored. Broadly, the aim of the thesis was, therefore, to explore the construct of sleep quality in the context of elite sport performance.

The first chapter systematised the literature on sleep quality among elite athletic populations and found that levels of poor sleep quality and insomnia symptomatology were generally high. The most influential source of sleep disturbance identified in the systematic review was competitions, with the literature further identifying nights immediately prior to competitions as characterised by longer sleep latencies. Because sleep is repeatedly challenged in elite sport, and night-time sleep loss is likely among athletes, the ability to compensate through daytime napping takes on a special importance. The finding in Chapter 3, however, found napping was unrelated to night-time sleep quality and introduced the possibility that napping in elite sport may, under some circumstances, be less related to homeostatic sleep need, and more related to the construct of ‘sleepability’ (a capacity to nap on demand that is only weakly related to homeostatic sleep pressure). Again, the construct of arousal is relevant, since an ability to nap ‘on demand’ requires also an ability to effectively manage the arousal mechanisms which promote daytime wakefulness. These assumptions were confirmed in Chapter 4, which also supplied evidence for the influence of arousal mechanisms in predicting first night effects in elite athletes. Correlations between sleep latency and Ford Insomnia Response Test (FIRST) scores (which are presumed to reflect a propensity for hyperarousal) for the adaptation trial suggest that hyperarousal may identify those athletes whose sleep is most influenced by first-night effects.

which, importantly, emphasised strategies to combat and manage pre-sleep arousal. That the programme reduced Dysfunctional Beliefs and Attitudes about Sleep (DBAS) scores and increased the utility of cognitive-behavioural techniques in the pilot study (Chapter 6) offers evidence for the successful transfer of ‘de-arousal’ strategies. The acceptance of the sleep management programme during a period of competition (Chapter 7) further supports the relevance of cognitive-behavioural techniques, however modest utility and efficacy suggests further modification may be required in the context of elite competition.

It is important that future research continues to connect, rather than separate, the behavioural and physiological sleep phenomenology of elite athletes and their non-athletic counterparts to accelerate the evolution of complex theoretical and conceptual sleep constructs, and the development of subsequent protocols of sleep management delivery and applied practice. It is a reasonable overall conclusion from this programme of research that the systematic assessment and control of arousal-related mechanisms can improve the ability of sport management to predict and protect sleep outcomes among elite athletes.

Key words: sleep quality; insomnia, elite athletes; sleep management.

Acknowledgements

I would like to acknowledge the support, guidance and continued mentorship from my supervisory team, Professor. Kevin Morgan and Dr. Sarah Gilchrist. I am grateful for their collective willingness to openly share experiences and knowledge in fields of both behavioural sleep medicine, and elite sport. Kevin, I would like to thank you for helping navigate my way through the world of applied research and sleep science. Your general wisdom, coupled with your ability to rapidly understand the unique, and sometimes challenging, world of elite sport has been integral to my development as a researcher. Sarah, I am forever grateful for your belief and trust you have showed in me from when I was first employed by the English Institute of Sport 8 years ago. Your professional and personal guidance has formed the foundation for my development as a researcher, a practitioner, and a person.

I would like to thank my sponsors, the English Institute of Sport (EIS) and the Clinical Sleep Research Unit within the School of Health, Exercise and Health Sciences at Loughborough University, for the opportunity to complete a research programme. I would like to thank my employers, the EIS and the physiology leadership team, for giving me the trust and confidence to work flexibly to complete my research, alongside my full-time role. A special thank you to Dr. Sarah Gilchrist and Dr. Steven Ingham for the conception of a research programme in sport sleep science. A special thank you to Dr. Wai Yeung and Dr. Iuliana Hartescu for their technical support throughout my PhD.

I would like to thank the sports, support staff and athletes who opted to take part in the research programme. I would like to acknowledge the involvement of Andy Hudson from GB and England Hockey, and Katie James from the England Rugby Football Union and their commitment and open mindedness to take part in applied research.

I would like to thank my friends and colleagues for their continued support over the past 7 years. A special thank you to fellow (former) PhD students Dr. David Green, Dr. Mehdi Kordi and Dr. Gareth Turner for their openness to share challenges from their research experiences, and to generally ‘chew the fat’.

iii

I would like to thank my family for their continued love and support, and yes, mum; I have finished my PhD. A special thank you to my wife, Sophia, who has provided me with unconditional love, support, and encouragement over the past 10 years (despite promising you I would complete my PhD before we were married!). Sophia, you have sacrificed half a decade of your life while I complete my PhD and have put up with many of my ups-and-downs. Saying that, in that time we have achieved so much together, and, finally, no I will not be embarking on any further education!

Preface

The findings within this thesis have been peer-reviewed and published as follows:

1. Gupta, L; Morgan, K; & Gilchrist, S (2017). Does elite sport degrade sleep quality? A systematic review. Sports Medicine. 47(7):1317-1333. Doi: 10.1007/s40279-016-0650-6. 2. Gupta, L; Morgan, K & Gilchrist (2017). Author’s reply to Bender and Samuels “Does elite sport degrade sleep quality? A systematic review. Sports Medicine. 47(7): 1455-1456. doi: 10.1007/s40279-017-0710-6. 3. Gupta, L; Morgan, K; North, C; & Gilchrist, S (2020). Napping in high- performance athletes: sleepiness or sleepability? European Journal of Sport Science. Doi: 1080/17461391.2020.1743765

Abstracts from the following studies have been peer-reviewed and accepted for conference presentations as follows:

1. Hartescu, I & Gupta, L (2020). Competitive sport and sleep. British Sleep Society (Oral presentation, British Sleep Society meeting, Birmingham 2020) 2. Gupta, L; North; C; Morgan, K; Gilchrist (2017) Physiological sleep tendency and sleep reactivity among elite athletes. SLEEP. 40(1): A70-71. Poster presentation. Poster presentation. SLEEP 2017 – APSS annual meeting, Boston (3rd – 7th June 2017). 3. Gupta, L; Morgan, K; Gilchrist, S. (2017). Between sport differences in sleep quality and insomnia symptomatology: A national survey of elite British athletes. SLEEP. 40(1): A123. Poster presentation. SLEEP 2017 – APSS annual meeting, Boston (3rd – 7th June 2017). 4. Gupta, L; & Morgan, K. (2015). Training schedules and sleep quality among elite athletes. Poster presentation. World Sleep 2015 – World Congress of the World Sleep Foundation, Istanbul (Oct 31 – Nov 3, 2015). 5. Gupta, L; & Morgan, K. (2015). Chronotype and sleep quality in elite athletes. Poster presentation. World Sleep 2015 – World Congress of the World Sleep Foundation, Istanbul (Oct 31 – Nov 3, 2015).

6. Gupta, L; North, C; Morgan, K. (2015). Day-time sleep pressure is greater in elite athletes. Poster presentation. World Sleep 2015 – World Congress of the World Sleep Foundation, Istanbul (Oct 31 – Nov 3, 2015). 7. Gupta, L; Grant, G; Morgan, K; Gilchrist, S. (2014). Sleep patterns and sleep quality among elite multi-sport athletes: a national pilot study. Poster presentation. SLEEP 2014 - APSS annual meeting, Minneapolis, Minnesota (May 31st – 4th June 2014).

Contents

Abstract ...... i

Acknowledgements ...... iii

Preface...... v

List of figures ...... xvi

List of tables ...... xix

List of equations ...... xxiii

List of abbreviations ...... xxiv

1. Sleep quality in elite sport: a review and research agenda ...... 1

1.1 Introduction ...... 1

1.1.1 Insomnia and elite sport ...... 1

1.1.2 Aims of the review ...... 4

1.2 Methods ...... 4

1.2.1 Search strategy ...... 4

1.2.2 Eligibility criteria ...... 5

1.2.3 Study selection and data extraction ...... 5

1.2.4 Study quality appraisal ...... 5

1.2.5 Definitions of sleep terminology ...... 6

1.3 Results ...... 8

1.3.1 Participant ‘eliteness’ assessment ...... 9

1.3.2 Evidence quality appraisal ...... 13

1.3.3 Sleep structure and patterns ...... 16

1.3.3.1 Summary of sleep structure and patterns ...... 16

1.3.4 Sleep quality and insomnia symptomatology ...... 21

1.3.4.1 Summary of sleep quality and insomnia symptomatology ...... 23

1.3.5 Risk factors for sleep disturbance ...... 26

1.3.5.1 Competition ...... 26

vii

1.3.5.2 Travel ...... 27

1.3.5.3 Training ...... 28

1.3.5.4 Summary of risk factors ...... 28

1.4 Discussion ...... 37

1.4.1 Sleep patterns (hypothesis 1) ...... 37

1.4.2 Sleep quality and insomnia symptomatology (hypothesis 1) ...... 38

1.4.3 Risk factors for sleep disturbance (hypothesis 2) ...... 39

1.4.4 Between-sport differences ...... 40

1.4.5 Within-sport differences ...... 41

1.5 Conclusion and research agenda ...... 42

2. General methods used in this research programme ...... 44

2.1 Introduction ...... 44

2.2 Taxonomy of eliteness ...... 44

2.3 Newcastle Ottawa Scale (adapted from cross sectional studies) ...... 46

2.4 Psychometric sleep assessments ...... 49

2.4.1 Pittsburgh sleep quality index ...... 49

2.4.2 Ford Insomnia Response to Stress Test ...... 49

2.4.3 Dysfunctional Beliefs and Attitudes about Sleep ...... 50

2.4.4 Pre-sleep Arousal Scale ...... 50

2.5 Instrumental measures ...... 51

2.5.1 Heart rate recovery run ...... 51

2.5.1.1 Heart rate recovery run description ...... 51

2.5.1.2 Heart rate recovery run test-retest reliability ...... 51

2.5.2 Wrist actigraphy ...... 54

2.5.2.1 Actigraphy protocol ...... 54

2.5.2.2 Levels of agreement ...... 56

2.6 The Athlete Sleep Management Programme ...... 59

2.6.1 Programme development ...... 60

viii

2.6.2 Programme design ...... 60

2.6.3 Programme content ...... 61

2.6.3.1 Component 1 ...... 61

2.6.3.2 Component 2 ...... 61

2.6.3.3 Component 3 ...... 62

2.6.3.4 Component 4 ...... 63

2.6.4 Workshop materials ...... 64

2.6.5 Programme evaluation ...... 64

3. A comparison of between-sport differences in sleep quality and insomnia symptomatology among elite British athletes ...... 67

3.1 Introduction ...... 67

3.1.2 The ‘3-factor’ model of insomnia development ...... 68

3.1.3 Between sport differences ...... 69

3.1.4 Research aims and hypotheses ...... 70

3.2 Methodology ...... 71

3.2.1 Athlete demographics and training characteristics ...... 71

3.2.2 Sleep quality and insomnia symptomatology ...... 72

3.2.3 Stress induced sleep disturbances ...... 74

3.2.4 Sleep management ...... 74

3.2.5 The sample ...... 74

3.2.6 Data analysis ...... 75

3.3 Results ...... 75

3.3.1 Sleep quality and insomnia symptomatology ...... 76

3.3.2 Stress-related sleep disturbances ...... 78

3.3.3 Sleep management ...... 78

3.4 Discussion ...... 97

3.4.1 Between-sport differences in sleep quality (hypothesis 1) ...... 97

3.4.2 Risk factors for poor sleep quality (hypothesis 2)...... 99

3.4.3 Sleep management (hypothesis 3) ...... 101

3.4.4 Limitations ...... 103

3.5 Conclusion ...... 104

4. Napping in high performance athletes: sleepiness or sleepability? ...... 106

4.1 Introduction ...... 106

4.1.1 Insomnia and daytime sleepiness ...... 108

4.1.2 Sleep Reactivity and ‘first night effects’ ...... 108

4.1.3 Research aims and hypotheses ...... 109

4.2 Methods ...... 110

4.2.1 Participant selection ...... 110

4.2.2 Sample size ...... 112

4.2.3 Sleep latency recordings ...... 112

4.2.4 Sleepiness assessment ...... 113

4.2.5 Statistical analysis ...... 113

4.3 Results ...... 114

4.3.1 Sleep scoring reliability ...... 114

4.3.2 Pre-trial sleepiness ...... 114

4.3.3 Sleep tendency measures ...... 118

4.3.3.1 Adaptation Trial ...... 118

4.3.3.2 Experimental Trial ...... 122

4.4 Discussion ...... 124

4.4.1 Sleep tendency (hypothesis 1) ...... 125

4.4.2 Sleepability ...... 126

4.4.3 Limitations ...... 126

4.4.4 Practical application ...... 127

4.5 Conclusions ...... 128

5. Sleep reactivity during a simulated Olympic tournament in elite British field hockey players ...... 129

5.1. Introduction ...... 129

5.1.1 Elite competition sleep ...... 130

5.1.2 Pre-sleep arousal...... 130

5.1.3 Research context, aims and hypotheses ...... 131

5.2 Methodology ...... 133

5.2.1 Participants ...... 133

5.2.2 Study design ...... 133

5.2.2.1 Olympic tournament format ...... 134

5.2.3 Questionnaire assessment ...... 136

5.2.3.1 Sleep questionnaires ...... 136

5.2.3.2. Self-reported sleep symptoms ...... 136

5.2.4 Actigraphy ...... 137

5.2.5 Match-play demands ...... 137

5.2.6 Statistical analysis ...... 138

5.3 Results ...... 139

5.3.1 Match-play demands ...... 139

5.3.2 Participant sleep characteristics ...... 140

5.3.3 Sleep symptoms during an Olympic field Hockey tournament ..... 143

5.3.3.1 Sleep assessments ...... 143

5.3.3.2 Self-reported insomnia symptoms: ...... 146

5.4 Discussion ...... 148

5.4.1 Sleep reactivity (hypothesis 1 and 2) ...... 149

5.4.3 Practical applications ...... 151

5.4.1 Limitations ...... 151

5.5 Summary and conclusion ...... 152

6. Sleep management in elite team sport athletes using cognitive- behavioural principles: A pilot study ...... 154

6.4 Introduction ...... 154

6.1.1 Sleep management in elite sport ...... 154

6.1.2 Cognitive behavioural principles in sleep management ...... 160

6.1.3 Development of the ASMP ...... 161

6.1.5. Research aims and hypotheses ...... 162

6.2 Method ...... 163

6.2.1. Participants ...... 163

6.2.2 Study Design ...... 163

6.2.3 Delivery of the ASMP ...... 164

6.2.4 Baseline and follow-up Questionnaires ...... 165

6.2.4.1 Sleep questionnaires ...... 165

6.2.4.2 Self-reported sleep symptoms ...... 165

6.2.4.3 Programme Evaluation ...... 166

6.2.5 Physiological assessment ...... 166

6.2.5.1 Saliva alpha amylase ...... 166

6.2.5.2 Saliva collection ...... 167

6.2.5.3 Heart rate recovery run ...... 168

6.2.5.4 Training load ...... 168

6.2.6 Actigraphy ...... 168

6.2.7 Statistical analyses ...... 168

6.3 Results ...... 169

6.3.1 Sample ...... 169

6.3.2 Training load ...... 170

6.3.3 Baseline sleep outcomes ...... 170

6.3.4 Follow up sleep outcomes ...... 174

6.3.4.1 Sleep questionnaires ...... 174

6.3.4.2 Actigraphy ...... 178

6.3.4.3 Self-reported sleep symptoms ...... 180

6.3.4.4 Physiological assessment ...... 181

xii

6.3.5 Programme Evaluation...... 183

6.3.6 Programme utility ...... 184

6.4 Discussion ...... 185

6.4.1 Cognitive behavioural approaches (hypothesis 1 and 2) ...... 186

6.4.2 Sleep beliefs and attitudes ...... 187

6.4.3 Sleep programme utility and value (hypothesis 3) ...... 187

6.4.4 Limitations ...... 188

6.5 Conclusion ...... 190

7. Delivering group-based psychoeducational sleep management to elite team sport athletes: an open label trial ...... 191

7.1 Introduction ...... 191

7.1.1 Sleep management during competition ...... 191

7.1.2 Sleep and competition performance ...... 192

7.1.3 Research questions and hypotheses ...... 193

7.2 Methods ...... 193

7.2.1 Competition format ...... 193

7.2.2 Competition samples ...... 194

7.2.3 Participants ...... 194

7.2.4 Study design ...... 195

7.2.5 Baseline and follow up Questionnaires ...... 195

7.2.5.1 Sleep questionnaires ...... 195

7.2.5.2 Wellbeing Diary ...... 197

7.2.5.3 Programme evaluation questionnaire ...... 198

7.2.6 Actigraphy ...... 199

7.2.7 Tournament and training running patterns ...... 199

7.2.8 The Athlete Sleep Management Programme ...... 202

7.2.9 Statistical analysis ...... 202

7.3 Results ...... 203

xiii

7.3.1 Sample ...... 203

7.3.2 Running patterns ...... 204

7.3.3 Baseline sleep characteristics ...... 205

7.3.4 Sleep outcomes Competition 1 ...... 207

7.3.4.1 Actigraphy ...... 207

7.3.2.2 Questionnaire assessments ...... 210

7.3.5 Pre-post assessment of the Athlete Sleep Management Programme 211

7.3.5.1 Questionnaires ...... 211

7.3.5.2 Actigraphy ...... 216

7.3.6 Sleep programme evaluation ...... 217

7.3.6.1 Programme utility ...... 217

7.4 Discussion ...... 219

7.4.1 Insomnia symptoms and sleep quality (hypothesis 1) ...... 219

7.4.2 Competition sleep (hypothesis 2) ...... 220

7.4.3 Sleep management (hypothesis 3 and 4) ...... 221

7.4.4 Sleep and athletic performance ...... 224

7.4.5 Limitations ...... 224

7.5 Conclusion ...... 225

8 A general discussion of findings ...... 227

8.1 Introduction ...... 227

8.2 Connecting findings across chapters ...... 227

8.2. Practical application of findings ...... 233

8.2.1 Measurement ...... 234

8.2.2 Sleep management ...... 238

8.3. Future research directions ...... 241

8.4 Summary ...... 245

9. Appendices ...... 247

xiv

9.1 Appendix A - ASMP infographics...... 248

9.1.1 Workshop one ...... 248

9.1.2 Workshop two ...... 249

9.1.3 Workshop three ...... 250

9.1.4 Workshop four ...... 251

9.2 Appendix B - Sleep diary ...... 252

9.3 Appendix C - Sport Sleep Survey landing page ...... 253

10. References ...... 256

List of figures

Figure 1.1: Study selection PRISMA flow diagram ...... 9 Figure 2.1: Bland-Altman plots for 13 hockey players: A) heart rate during exercise (HRex); B) during rest immediately following the heart rate recovery (HRrest) run; C) the difference between HRex and HRrest (HRR); and D) the difference between HRex and HRR expressed as a percentage (HRR%) ...... 53 Figure 2.2: Images of: A) the Actiwatch 2 (Philips Resperonics, USA); and B) the Motionwatch 8 (Camntech, Cambridge, UK). Both devices are fitted with a photometer and an event marker...... 55 Figure 2.3: Bland-Altman plots for A) sleep efficiency (SE); B) sleep onset latency (SOL); C) total sleep time (TST) between Actiwatch 2 (AW2) and Motionwatch 8 (MW8). Grey line = mean bias; dashed line(s) = 95% confidence intervals ...... 58 Figure 3.1: Mean Pittsburgh Sleep Quality (PSQI) Index global scores across sport types...... 77 Figure 3.2: Pittsburgh sleep quality index (PSQI) global scores between elite athletes who felt their sleep quality has ‘deteriorated’, ‘remained unchanged’ or ‘improved’...... 96 Figure 4.1: Mean Karolinska Sleepiness Scale (KSS) scores for non-athletes (○) sub- elite athletes (Δ) and elite athletes (□) prior to: the adaptation trial (above); and the experimental trial (below). Bars indicate 1 standard deviation...... 117 Figure 4.2: Mean adaptation trial sleep latency scores for a single nap opportunity (n = 10/group). Bars indicate 1 standard deviation...... 121 Figure 4.3: Levels of high sleep tendency (sleep latency scores ≤8 minutes) for the adaptation trial (n = 10/group)...... 122 Figure 4.4: Mean experimental trial sleep latency scores for a single nap opportunity (n = 10/group). Bars indicate 1 standard deviation...... 123 Figure 4.5: Levels of high sleep tendency (sleep latency scores ≤8 minutes) for the experimental trial (n = 10/group)...... 124 Figure 5.1: A schematic of the study design, and simulated tournament schedule ...... 135 Figure 5.2: Likelihood levels of sleep being disturbed after losing a match or playing badly within groups of reactive and nonreactive sleepers (n=8/ group) ...... 143

xvi

Figure 5.3: Self-reported TST (STST) values across baseline week (BW), simulated Olympic tournament week 1 (TW1) and week 2 (TW2), and a recovery week (RW) (n=16)...... 145 Figure 5.4: Changes from baseline values in actigraphic total sleep time during the tournament week 1 (TW1) and tournament week 2 (TW2), and recovery week (RW) among reactive (□) and nonreactive (○) sleepers (n=8/ group) ...... 145 Figure 5.5: Feelings of freshness on awakening (FROA) values across baseline week (BW), tournament week 1 (TW1), tournament week 2 (TW2), and recovery week (RW) (n=16)...... 147 Figure 5.6: Ease to fall asleep (ETFA) values across baseline week (BW), tournament week 1 (TW1), tournament week 2 (TW2), and recovery week (RW) (n=16)...... 147 Figure 5.7: Sleep quality (SQ) values across baseline week (BW), tournament week 1 (TW1), tournament week 2 (TW2), and recovery week (RW) (n=16). ... 148 Figure 6.1: Study design for a pre-post-test of the English Institute of Sport - Sleep Management Programme (delivered during the ‘Implementation Period’) .... 164 Figure 6.2: Distribution of poor sleepers (defined by scores of ≥5 on Pittsburgh Sleep Quality Index) at Baseline, and 6 weeks (Follow-up 1) and 10 weeks (Follow- up 2) post baseline (n=18) ...... 174 Figure 6.3: Mean Dysfunctional Beliefs and Attitudes about Sleep (DBAS) scores at Baseline, and 6 weeks (Follow-up 1) and 10 weeks (Follow-up 2) post baseline (n=18)...... 176 Figure 6.4: Sleep efficiency at Baseline and 6 weeks (Follow-up 1) post baseline (n=18)...... 178 Figure 6.5: Sleep onset latency at Baseline and 6 weeks (Follow-up 1) post baseline (n=18)...... 179 Figure 6.6: Wake time after sleep onset at Baseline and 6 weeks (Follow-up 1) post baseline (n=18)...... 180 Figure 6.7: Ratings of mood (MD) for the Baseline and Follow-up 1 weeks...... 181 Figure 6.8: Evaluation of sleep programme ...... 183 Figure 6.9 distribution of players performing sleep promoting activities prior to sleep onset (n=18) ...... 184 Figure 6.10 distribution of players performing techniques to help fall asleep (n=18) ...... 185

xvii

Figure 7.1 Structure of the 10-week study period covering 2 competitions in the 2016 Women's World Rugby Sevens Series (Team GB)’ ...... 200 Figure 7.2 Outline of the 10 week study protocol ...... 201 Figure 7.3 Sleep Onset Latency in competing (COMP; n=14) and training (TRAIN; n=8) players during Competition 1...... 209 Figure 7.4 Total Sleep Time in competing (COMP; n=14) and training (TRAIN; n=8) players during Competition 1...... 209 Figure 7.5 Pittsburgh Sleep Quality Index (PSQI) global scores before (Competition 1) and after intervention (Competition 2) in players scoring >5 on PSQI at Baseline (n = 6)...... 211 Figure 7.6 Dysfunctional Beliefs and Attitudes about Sleep (DBAS) consequences component before (PRE) and after (POST) the sleep management intervention (n = 12)...... 212 Figure 7.7 Rating of agreement to statement “I am concerned poor sleep will impact my playing performance on the pitch” (higher scores = greater concern; n = 12)...... 213 Figure 7.8 Sleep quality ratings between competition 1 and competition 2 .... 215 Figure 7.9 Stress ratings between competition 1 and competition 2 ...... 216 Figure 7.10 total sleep time (TST) between the competition before the intervention (competition 1) and after (competition 2), n = 12...... 217 Figure 7.11 Evaluation of the sleep management programme. n = 12 ...... 218 Figure 8.1 a conceptual diagram showing the application of Spielman’s (Spielman et al., 1987) model of insomnia development in elite sport. Black arrows = evidenced in the current research programme, Grey arrows = hypothesised link ...... 242

xviii

List of tables

Table 1.1 sleep related terms used in this review ...... 7 Table 1.2 sleep related terms used in this review (continued) ...... 8 Table 1.3 the modified taxonomy of ‘eliteness’ (Swann, Moran and Piggott, 2015) (mean scores in Table 1.5) ...... 10 Table 1.4 the modified taxonomy of ‘eliteness’ (Swann, Moran and Piggott, 2015) (continued; mean score in Table 1.5) ...... 11 Table 1.5 the modified taxonomy of ‘eliteness’ (Swann, Moran and Piggott, 2015) (continued) ...... 12 Table 1.6 Evidence quality appraisal using the adapted Newcastle-Ottawa scale (NOS) for cross-sectional studies (see Table 1.7 for total mean scores) ...... 14 Table 1.7 Evidence quality appraisal using the adapted Newcastle-Ottawa scale (NOS) for cross-sectional studies (continued) ...... 15 Table 1.8: Sleep characteristics of elite athletes using actigraphy (See Table 1.9 for total mean scores) ...... 17 Table 1.9 Sleep characteristics of elite athletes using actigraphy (continued) .18 Table 1.10 Sleep characteristics of elite athletes using sleep diaries and polysomnography ...... 19 Table 1.11 Sleep characteristics of elite athletes using questionnaires ...... 20 Table 1.12: Characteristics of subjective sleep quality in elite athletes...... 24 Table 1.13: Sleep assessments in elite athletes using the Pittsburgh Sleep Quality Index ...... 25 Table 1.14: Prevalence of insomnia symptomatology and changes in sleep patterns pre-competition ...... 30 Table 1.15 Prevalence of insomnia symptomatology and changes in sleep patterns pre-competition (continued) ...... 31 Table 1.16: Prevalence of insomnia symptomatology and changes in actigraphy and sleep diary derived sleep patterns post-competition ...... 32 Table 1.17 Prevalence of insomnia symptomatology and changes in Polysomnography and questionnaire derived sleep patterns post-competition33 Table 1.18: Changes in sleep patterns, sleep quality and jet lag following long- and short-haul travel...... 34 Table 1.19: Prevalence of insomnia symptoms and changes in sleep patterns during training days v rest days ...... 35

xix

Table 1.20 Prevalence of insomnia symptoms and changes in sleep patterns during intensified training v normal training...... 36 Table 2.1: Terms used in Swann, Moran and Piggott’s (2015) equation for a taxonomy of eliteness ...... 45 Table 2.2 Newcastle Ottawa Scale (adapted for cross sectional studies) component (selection, comparability and outcome) items and ratings ...... 48 Table 2.3: Estimates of test-retest reliability of heart rate during exercise (HRex), heart rate during rest (HRrest), heart rate recovery (HRR), and heart rate recovery expressed as a percentage (HRR%) for 13 hockey players...... 52 Table 2.4: Specifications for Actiwatch 2 (AW2) and Motionwatch 8 (MW8) .....55 Table 2.5: Agreement of sleep onset latency, sleep efficiency, and total sleep time between Actiwatch 2 and Motionwatch 8 actigraphy devices ...... 57 Table 2.6: Overview of sleep management programme content ...... 66 Table 3.1: Sport categorisations used in the EIS online sleep survey: Sport (n of athletes) ...... 73 Table 3.2: Athlete demographics and training and competition characteristics between sport-types...... 80 Table 3.3: Prevalence and relative risk of poor sleep quality between sport-types, ability, gender and age groups...... 81 Table 3.4 Prevalence and relative risk of poor sleep quality and sport sleep challenges ...... 82 Table 3.5: Sleep quality and insomnia symptoms between sport types ...... 83 Table 3.6 Sleep quality, insomnia symptoms and sleep management between sport types ...... 84 Table 4.1: Participant characteristics ...... 115 Table 4.2: Mean pre-trial sleep patterns (for the nights immediately preceding laboratory visits) ...... 116 Table 4.3: Pre-trial sleep assessments and unadjusted trial sleep latency outcomes for adaptation and experimental trials ...... 119 Table 4.4 adjusted (for pre-trial TST and 15:00 KSS) trial sleep latency scores for adaptation and experimental trials ...... 120 Table 5.1: Comparisons of match play demands across the first (TW1) and second (TW2) weeks of a simulated Olympic field hockey tournament ...... 140

Table 5.2: Comparisons of match play demands between reactive (FIRST score ≤ 18) and nonreactive (FIRST score >18) sleepers during a 2-week simulated Olympic field hockey tournament ...... 141 Table 5.3: Comparisons of demographic characteristics, sleep patterns, and sleep symptoms between reactive (FIRST score ≤ 18) and nonreactive (FIRST score >18) sleepers prior to a 2-week simulated Olympic field hockey tournament ...... 142 Table 6.1: Sleep extension studies conducted among high-performance athletes ...... 157 Table 6.2 Novel interventions conducted among elite athletes among high performance athletes ...... 158 Table 6.3 Sleep hygiene, psychoeducation and counselling interventions ...... 159 Table 6.4: Sleep characteristics of 18 elite field hockey players participating in a 10-week training period...... 171 Table 6.5 Actigraphy derived sleep patterns of 18 elite field hockey players participating in a 10-week training period...... 172 Table 6.6: Pearson correlation coefficients between age, sleep questionnaire and actigraphy variables ...... 173 Table 6.7: Sleep questionnaire scores at Baseline, Follow-up 1, and Follow-up 2 intervention: score values are mean (SD) ...... 177 Table 6.8: actigraphy, sleep diary and day-time arousal variables pre-post intervention ...... 182 Table 7.1 Baseline sleep characteristics for the selected (COMP, n = 14) and non- selected (TRAIN, n = 9) players for a world rugby sevens series competition .... 206 Table 7.2 Baseline sleep characteristics for the entire squad (n = 23), selected players (COMP, n = 14) and non-selected (TRAIN, n = 9) for a world rugby sevens series competition (Continued)...... 207 Table 7.3 Intercorrelations (Pearson product moment correlation coefficients) between sleep variables among a squad of female rugby sevens players (n=23) ...... 208 Table 7.4 Responses to Competition, Sport, Sleep, Dreams Questionnaire (CSSDQ) during world series tournaments in Atlanta, USA and Langford, Canada (Competition 1) and Clermont, France (Competition 2) ...... 214 Table 8.1 null hypotheses presented in Chapters 2-5 ...... 230

xxi

Table 8.2 null hypotheses presented in Chapters 6-7 ...... 231

xxii

List of equations

Equation 2.1 Taxonomy of eliteness ...... 46 Equation 2.2: Modified Taxonomy of eliteness ...... 46 Equation 2.3 Newcastle Ottawa Scale scoring ...... 47

xxiii

List of abbreviations

AASM America academy of sleep medicine ANCOVA Analysis of covariance ANOVA Analysis of variance ARF Australian rules football ASSQ Athlete sleep screening questionnaire AW2 Actiwatch 2 BW Baseline week CBT-I Cognitive behavioural therapy for insomnia CV Coefficient of variation COMP Selected players DBAS Dysfunctional beliefs about sleep E East EDS Excessive daytime sleepiness EEG Electroencephalography ASMP Athlete sleep management programme

EMG Electromyogram EOG Electrooculogram ETFA Ease to fall asleep F Female FET Fisher's exact test FIRST Ford's Insomnia Response to Stress Test FROA Freshness on awakening GPS Global positioning system HR Heart rate Hrex Exercising heart rate HRR Heart rate recovery HRR% Heart rate recovery (%) Hrrest Heart rate at rest HSR High speed running ICC Interclass correlation coefficient ISI Insomnia severity index

xxiv

JL Jet lag KSS Karolinska sleepiness scale M Male MD Mood MEQ Morningness--eveningness questionnaire MSLT Multiple sleep latency test MW8 Motion watch 8 PMR Progressive muscle relaxation PSAScog Pre-sleep arousal scale - cognitive PSASsom Pre-sleep arousal scale - somatic PSG Polysomnography PSQI Pittsburgh sleep quality index RW Recovery week sAA Saliva alpha amylase SE Sleep efficiency SLT Sleep latency test SOL Sleep onset latency SPT Sleep period time SQ Sleep quality sRPE Session rating of perceived exertion SS Synchronised swimming STST Self-reported total sleep time TEM Technical error of measurement TIB Time in bed TRAIN Non-selected players TST Total sleep time TW Training week W West WASO Wake time after sleep onset

xxv

Chapter 1

1. Sleep quality in elite sport: a review and research agenda The review conducted as part of this chapter was published as: Gupta, L., Morgan, K. and Gilchrist, S. (2017) ‘Does Elite Sport Degrade Sleep Quality? A Systematic Review’. Sports Medicine; 47(7). pp.1317–1333.

1.1 Introduction While reviews of the sport (Chennaoui, Arnal, Sauvet, & Léger, 2015; Cummiskey, Natsis, & Papathanasiou, 2013; Davenne, 2009; Halson, 2008, 2014b; Leger, Metlaine, & Choudat, 2005; Samuels, 2008; Shapiro, 1981; Venter, 2012) and exercise (Chennaoui et al., 2015; Driver & Taylor, 2000; S. Driver & Taylor, 1996; Kubitz, Landers, Petruzzello, & Han, 1996; Youngstedt, O’Connor, & Dishman, 1997) literature suggest a reciprocal relationship between sleep and athletic performance, this evidence provides an incomplete indication of relevant sleep- sport interactions. To date, most attention has focussed on relationships between athletic performance and either the electrophysiological composition and duration of sleep or, more recently, actigraphic measurements of sleep-wake patterns (Halson, 2014a). Much less attention, however, has been paid to interactive associations between sleep quality (i.e. the subjective experience and perceived adequacy of sleep), and the demands of elite sport participation. As a result, information on insomnia symptomatology and its implications for sports performance among elite athletes remains poorly explored, and poorly systematised.

1.1.1 Insomnia and elite sport Insomnia disorder is defined by the cardinal symptoms of difficulty initiating or maintaining sleep, or early morning awakening at a frequency of three or more nights per week over a duration of least 3 months. Symptoms occur despite adequate opportunity to sleep, and are associated with significant distress or impaired social, personal, or occupational functioning (American Academy of Sleep Medicine, 2014; American Psychiatric Association, 2013). Daytime impairments can range from manifest fatigue (American Academy of Sleep Medicine, 2014) and emotional dysregulation (Kahn, Sheppes, & Sadeh, 2013) to

more subtle deficits in psychomotor (Riedel & Lichstein, 2000) and neuropsychological (Bastien, 2011) performance. Importantly, the symptom of excessive daytime sleepiness (EDS), as opposed to fatigue, is not a typical characteristic of insomnia; objective (e.g. multiple sleep latency tests, which directly assess daytime sleep pressure, Sateia et al. (1993) and Lichstein et al. (1994)), and subjective measures of daytime sleepiness e.g. Lichstein et al. (1994), poorly discriminate between those with insomnia, and those without. Research evidence supports the view that insomnia is a disorder of hyperarousal, where the healthy transition from wake to sleep is substantially inhibited by two processes:1) ‘cognitive arousal’ (Espie, Broomfield, MacMahon, Macphee, & Taylor, 2006; Harvey, 2002), engagement in seemingly uncontrollable pre-sleep cognitive activity which ultimately triggers physiological (autonomic, cortical, metabolic) responses inconsistent with pre-sleep de-arousal; and 2) ‘attentional bias’ (Espie et al., 2006) a tendency to focus excessively on (Barclay & Ellis, 2013), or a difficulty in switching attention from (Jansson-Fröjmark, Bermås, & Kjellén, 2013) sleep related problems. The research evidence also suggests that personality attributes, particularly those reflecting ‘anxious concerns’ or traits associated with perfectionism (van de Laar et al., 2010), and difficulties in regulating arousal, can combine in certain ‘at risk’ phenotypes, making some people inherently more predisposed to insomnia than others (Harvey et al., 2014).

In the context of elite sport, these characteristics of insomnia symptomology introduce a range of clinical, empirical and theoretical factors which justify specific attention to sleep quality (in addition to sleep structure and patterns) when reviewing sleep and sport interactions:

1. In previous reviews (Chennaoui et al., 2015; Driver & Taylor, 2000; Driver & Taylor, 1996; Kubitz et al., 1996; Youngstedt et al., 1997) the impact of exercise on sleep has most frequently been expressed in terms of post exercise changes in sleep stage organisation, with increasing levels of exercise leading to greater duration and intensity of stage 3 sleep (Chennaoui et al., 2015; Driver & Taylor, 2000; Driver & Taylor, 1996; Kubitz et al., 1996; Youngstedt et al., 1997). It has long been recognised, however, that polysomnographic macrostructure (as reflected in standard sleep stages) poorly discriminates between those who report good quality sleep

and those who report symptoms of insomnia (e.g. Edinger et al. (2013), with evidence suggesting that any differences may reside more in polysomnographic microstructure (Bastien, 2011) or, where subjective estimates of sleep duration are concerned, in psychological characteristics which mediate sleep experience (Edinger et al., 2000). Previous emphases on sleep structure and exercise, therefore, inadequately address issues of sleep quality.

2. Sleep and sport (Chennaoui et al., 2015; Cummiskey et al., 2013; Davenne, 2009; Halson, 2008; Leger et al., 2005; Samuels, 2008; Shapiro, 1981; Venter, 2012) reviews have typically equated exercise training related sleep disturbance with sleep loss, exploring the impact of disordered sleep on athletic performance through sleep deprivation models (e.g. see Fullagar et al. (2015)). However, since insomnia per se is not characterised by EDS (Sateia et al., 1993), such models may provide limited insights into sleep quality-performance relationships -a point recently emphasised in the sports literature by Dickinson and Hanrahan (2009).

3. Elite athletes are rigorously selected on the basis of not only physiological but also psychological (Allen, Greenlees, & Jones, 2013) attributes. It is possible that some of the personality characteristics which militate towards success within elite sport (i.e. perfectionism and anxious concern) also predispose individuals to insomnia (see Harvey et al. (2014)).

4. For elite athletes the multifaceted demands of elite sport, including the heightened frequency, intensity and volume (and also scheduling (Sargent, Halson, & Roach, 2014; Sargent, Lastella, Halson, & Roach, 2014)) of training sessions (Hausswirth et al., 2014; Killer, Svendsen, Jeukendrup, & Gleeson, 2015; Sargent, Halson, et al., 2014; Taylor, Rogers, & Driver, 1997), pre-competition anxiety (a type of cognitive arousal) (Erlacher, Ehrlenspiel, Adegbesan, & El-Din, 2011; Juliff, Halson, & Peiffer, 2015), and the relocation required by national (Richmond, Dawson, Hillman, & Eastwood, 2004; Richmond et al., 2007) and international competitions (P Fowler, Duffield, Howle, Waterson, & Vaile, 2015; Fowler, Duffield, Lu, Hickmans, & Scott, 2016), can all be expected to precipitate (or perpetuate) episodes of sleep disruption. Chronic sleep disturbances (Fallon, 2007; Reid,

Gleeson, Williams, & Clancy, 2004), in addition to restricted sleep times (Sargent, Halson, et al., 2014; Sargent, Lastella, et al., 2014), may contribute to the high levels of daytime fatigue typically reported by competitive athletes and thus may impair training quality and adherence. The degree of sleep disruption and day-time symptoms (i.e. manifest fatigue) will be amplified in predisposed individuals.

1.1.2 Aims of the review To date no attempt has been made to systematize and critique the literature on sleep quality in relation to elite sport. Given this, the present review was broadly designed to systematize the research evidence describing sleep patterns and sleep quality among elite athletes; and to consider specific elite sport risk factors for poor sleep quality within elite sport. The review, therefore, set out to answer the following research questions

1. What is the structure, pattern, and quality of sleep among elite athletes? And; 2. What are the specific risk factors for sleep disturbance arising from the demands of elite sport?

Considering points 1-4 (see Para. 1.1.1), the null hypotheses were:

1. The prevalence of poor sleep quality and insomnia symptomatology will be low in elite athletes; and 2. Elements of elite sport will not negatively impact on sleep quality in elite athletes.

1.2 Methods The review methodology adopted the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The search strategy, together with the number of hits at each stage, is shown in Figure 1.1.

1.2.1 Search strategy Four electronic databases (SPORTDiscus, PubMed, Science Direct and Google Scholar) were systematically searched up to April 2016 using combinations of the following key words with appropriate truncation and medical subject headings

(MeSH): sleep; sleep quality; insomnia; elite athletes; high-performance athletes; training; travel; competition; and recovery.

1.2.2 Eligibility criteria The eligibility of retrieved records was then independently assessed by the programme supervisor based on title and abstract. If the information was unclear, the full-text article was screened. Studies were required to meet the following inclusion criteria: 1) data were reported for participants competing at the elite level (defined as Olympic, international, professional or national; internationally agreed criteria for ‘elite’ status are further discussed in Chapter; 2) data were reported for participants aged >17 y; 3) the study reported quantitative data on sleep outcomes; and 4) the study was published in an English language peer reviewed journal as a full text article.

1.2.3 Study selection and data extraction Titles and abstracts of potentially relevant articles were screened independently by two reviewers. Duplicates were removed and articles which did not meet the inclusion criteria were excluded. Full text articles were assessed for eligibility by two reviewers (the present author and the principal programme supervisor). A pre-designed data extraction form was employed to collate data from individual studies including study design; participant ‘eliteness’; n size; participant sex; methodology of sleep assessment; and key outcomes and findings.

1.2.4 Study quality appraisal Participant ‘eliteness’ for included studies was judged by applying the taxonomy of Swann, Moran and Piggot (2015) which ranks participants on a continuum (score range 1-16), allowing categorisations from ‘semi-elite’, through ‘competitive elite’, and ‘successful elite’, to ‘world class elite’ (see Tables 1.3, 1.4 and 1.5). Studies were critically appraised for evidence quality (principally ratings of participant selection, comparability, and outcomes) using the Newcastle- Ottawa scale (NOS) adapted for cross-sectional studies (Wells et al., 2016). Studies were appraised on a 0-10 rating scale, and categorised evidence quality as low (score <7) moderate-high (score range = 5-7) and high (score >7) (See Tables 1.6 and 1.7). This scale has been employed in systematic reviews of elite athletes (Ashley, Di Iorio, Cole, Tanday, & Needleman, 2015) and has established content and inter-rater reliability (see Luchini et al. (2017)). Full descriptions of

both the taxonomy for eliteness (Swann et al., 2015) and NOS (adapted for cross sectional studies) (Wells et al., 2016) can be seen in Chapter 2, Para 2.2 and 2.3, respectively.

1.2.5 Definitions of sleep terminology Definitions of the key sleep-related terms used in this review are shown in table 1.1. In brief, “sleep quality” refers to subjective sleep experience quantified using global items or formal psychometric assessments; “sleep pattern”, on the other hand, refers to serial instrumental measurements of 24-hour sleep-wake distributions (typically using actigraphy). For polysomnographic (PSG) studies “sleep structure” refers to the organisation of sleep stages within a recorded sleep phase. It should be noted that scoring criteria for classifying sleep stages changed in 2004 when the five-fold (stage 1; stage 2; stage 3; stage 4; and REM – rapid eye movement sleep) classification of Rechtschaffen and Kales (1968) was replaced by the American Academy of Sleep Medicine classification: stage N1 (stage 1); stage N2 (stage 2); stage N3 (stages 3 and 4); and stage R (REM) (Silber et al., 2007).

Table 1.1 sleep related terms used in this review

Term Definition

The electrophysiological composition and organisation of sleep typically described in terms of the duration of defined sleep stages, the duration of the awake stage, the aggregated time spent in all sleep stages (i.e.TST), or latencies to sleep onset (see below) and other (e.g. rapid Sleep structure eye movement sleep onset latency) stages. When combined with measures of TIB, electrophysiogical measures can provide a reliable indication of sleep efficiency (see below). Such electrophysiological measures require polysomnography (PSG), and are rarely used for 24 h or serial measurements.

As used in this review ‘sleep patterns’ refer to sleep-wake distributions typically assessed over units of 24 h using wrist actigraphy (and often repeated for multiple days). When Sleep patterns combined with measures of TIB, actigraphy can provide a reliable indication of SOL and sleep efficiency (see below).

An individual’s subjective experience of sleep typically focussing on problems initiating or maintaining sleep, or early morning awakening. Assessed through single items or formal psychometric evaluations, these experiences represent cardinal symptoms of insomnia. Combined Sleep quality with information on symptom frequency/duration and daytime symptoms (e.g. fatigue), these experiences contribute to diagnostic judgments of insomnia disorder as defined in Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-V) (American Psychiatric Association, 2013).

Sleep profile Broadly, SOL, TST, and SE reported.

The time elapsed between first getting into bed (with the Time in bed (TIB) intention of sleeping), to the final arising.

Sleep period time The time elapsed between the first onset of sleep and the (SPT) final awakening.

Table 1.2 sleep related terms used in this review (continued)

Term Definition

Wake after sleep onset (WASO) The amount of wakefulness accumulated between the first onset of sleep and the final awakening.

The total amount of time spent asleep whilst in bed (i.e: Total sleep time (TST) SPT-WASO).

TST expressed as a percentage of TIB: TST/TIBx100. Whether derived from instrumental measures or subjective estimates (of TST), SE provides a sensitive metric Sleep efficiency (SE) for estimating sleep quality. A SE below 85% is indicative of disorder.

Sleep onset latency The time elapsed between getting into bed or ‘lights out’ (SOL) or sleep to sleep onset. latency (SL)

A measure of the extent to which continuous sleep is interrupted by episodes of wakefulness. Sleep Fragmentation index fragmentation is reflected in the duration, and/or frequency of episodes of WASO.

1.3 Results The study selection PRISMA flow diagram can be seen in Figure 1.1. The search strategy returned 1,676 records. Of the 91 studies retained for full text screening, we excluded 54 which did not meet the set criteria. Thirty-seven studies were therefore eligible for review.

The 37 studies were published between 2001-2016 and included Olympic/Commonwealth (n=21, 57%), Paralympic (n=4, 11%), and professional (n=12, 32%) sports. The number of participants in the included studies ranged from 6-2067, with an overall age range of 18-30 y. Female athletes were under-

represented generally, with only eight (21%) of the studies reporting values exclusively for women.

1,676 records identified through database searching

968 duplicate records

removed from database 708 records present after search Identification duplicates removed

647 records excluded after the initial criteria assessment

lity 91 full text records identified for criteria assessment

Screening

andEligibi 54 full text records excluded after criteria

assessment

37 full text records identified for review

Included Screening

Figure 1.1: Study selection PRISMA flow diagram

1.3.1 Participant ‘eliteness’ assessment An assessment of study sample ‘eliteness’ is shown in Tables 1.3, 1.4 and 1.5. The application of the full taxonomy of ‘eliteness’ (Swann et al., 2015) was limited by participant descriptions within the included studies. Only a minority of studies (n=6, 16%) reported ‘athlete level of experience’, with only one study reporting ‘athletes’ level of success’. As a result, participants were categorised using a modified taxonomy within which only ‘semi-elite or ‘competitive elite’ categories could be judged (see Chapter 2, Para. 2.2). Accordingly, 20 studies (54%) were judged to have recruited ‘competitive elite’ participants.

Table 1.3 the modified taxonomy of ‘eliteness’ (Swann et al., 2015) (mean scores in Table 1.5) With-in sport Between sport Statistics A B C D E Global Standard of Success Experience In-country Scores Competitive Study competitiveness performance at level at level competiveness (out of 16) Elite (>8) of sport Leeder et al. (2012) 4 NR NR 2 3 10 1 Lastella et al. (2015a) 4 NR NR 2 3 10 1 Richmond et al. (2007) 4 NR NR 4 2 12 1 Richmond et al. (2004) 4 NR NR 4 2 12 1 Romyn et al. (2016) 2 NR NR 2 3 5 0 Schaal et al. (2015) 4 NR 3 2 2 8 0 Sargent et al. (2014) 4 NR NR 2 3 10 1 Kölling et al. (2016) 3 NR NR 1 4 8 0 Fowler et al. (2015) 4 NR NR 3 2 10 1 Fullagar et al. (2016a) 4 NR NR 4 4 16 1 Fullagar et al. (2016b) 4 NR 3 4 4 16 1 Shearer et al. (2015) 4 NR NR 4 4 16 1 Robey et al. (2014) 2 NR NR 2 4 6 0 Sargent et al. (2016) 2 NR NR 1 4 5 0 Netzer et al. (2001) 2 NR NR 1 4 5 0 Juliff, Halson and Peiffer 4 NR 4 2 4 12 1 (2015) Tsunoda et al. (2015) 4 NR 4 1 3 8 0 Schaal et al. (2011) 4 NR 2 2 2 8 0 See table Table 1.5 for notes

Table 1.4 the modified taxonomy of ‘eliteness’ (Swann et al., 2015) (continued; mean score in Table 1.5)

With-in sport Between sport Statistics A B C D E Global Standard of Success Experience In-country Scores Competitive Study competitiveness performance at level at level competiveness (out of 16) elite (>8) of sport Lucidi et al. (2007) 4 NR NR 2 3 10 1 Silva et al. (2012) 4 NR NR 1 3 8 0 Rodrigues et al. (2015) 4 NR NR 1 3 8 0 Samuels et al. (2016) 4 NR NR 4 3 14 1 Dickinson and Hanrahan 4 NR NR 2 2 8 0 (2009) Samuels (2008) 4 NR NR 4 3 14 1 Venter et al. (2010) 4 NR NR 4 4 15 1 Swinbourne et al. (2016) 4 NR NR 4 4 15 1 Bleyer et al. (2015) 3 NR NR 1 3 6 0 Chennaoui et al. (2016) 4 3 NR 1 4 10 1 See Table 1.5 for notes

Table 1.5 the modified taxonomy of ‘eliteness’ (Swann et al., 2015) (continued) With-in sport Between sport Statistics A B C D E Global Study Standard of Success Experience In-country Scores Competitive competitiveness of performance at level at level competiveness (out of 16) elite (>8) sport Silva and Paiva 4 NR 4 3 4 14 1 (2016) Fowler et al. (2014) 4 NR NR 4 2 12 1 Erlacher et al. (2011) 4 NR NR 2 2 8 0 Fowler et al. (2017) 4 NR NR 3 2 10 1 Dekker et al. (2014) 2 NR NR 1 3 4 0 Lastella et al. (2014) 3 NR NR 3 4 11 1 Durán et al. ( 2015) 4 NR NR 1 3 8 0 Elbayoumy and 3 NR NR 1 3 6 0 Elbayoumy (2015) Sargent, Halson and 2 NR NR 1 3 4 0 Roach (2014) Mean 3 NA NA 2 3 10 *20 SD 1 NA NA 1 1 4 NA Notes: Modified equation,[A x ((D + E)/2)]; a score >8 is judged to be a study that has recruited ‘competitive elite’ athletes; NR, not reported; NA, not applicable; SD, standard deviation; * total sum of studies

1.3.2 Evidence quality appraisal Given the diversity of study designs, reporting standards, outcome variables, and the very limited number of control participants, a meaningful calculation of risk (as odds ratios) and a subsequent meta-analysis was not possible. However, where studies reporting sleep quality outcomes used similar instruments, pooled estimates of prevalence were calculated. Relative to the objectives of this review, studies fell into three categories: 1) studies describing sleep structure and patterns; 2) studies describing sleep quality and insomnia symptomatology; and 3) studies exploring sport related risk for sleep disturbances.

A formal appraisal of the evidence quality for the 37 studies is shown in Tables 1.6 and 1.7. Overall, the evidence quality of the selected studies was generally ‘low’ (mean NOS score=5, SD=2), with 24 studies (65%) scoring <5 (low quality) and only two (5%) scoring >7 (good quality). Study designs employed were generally observational (n=34, 92%), with only three (8%) of this sample employing control-group designs. Of the observational studies 18 (49% of all studies) were cross-sectional and 14 (38% of all studies) were longitudinal. Very few studies (14%, n=5), adequately reported participant level of performance, age, sex, sport, and level of experience. Just under half of the studies (49%, n=18) provided a clear description of the protocol employed to measure sleep and used validated instruments, whilst adhering to measurement standards.

Table 1.6 Evidence quality appraisal using the adapted Newcastle-Ottawa scale (NOS) for cross-sectional studies (see Table 1.7 for total mean scores) Selectiona Comparabilitya Outcomeb Statistics

sample the ness of Representive size Sample respondents Non athletes of elite Description Subtotal (athletism) factor important most for Control age) (Sex and other factors for Control Subtotal of outcome Assessment test Statistical Subtotal (>7) High 7) Mod Low (<5)

Total (out of Total (out

10)

high (5 high

Study

Leeder et al. (2012) 0 1 0 1 2 1 1 2 1 1 2 6 0 1 0 Lastella et al. (2015a) 1 1 0 1 3 0 0 0 2 1 3 6 0 1 0 Richmond et al. (2007) 0 1 0 1 2 0 0 0 0 1 1 3 0 0 1 Richmond et al. (2004) 0 0 0 1 1 0 0 0 0 1 1 2 0 0 1 Romyn et al. (2016) 0 0 0 1 1 0 0 0 2 1 3 4 0 0 1 Schaal et al. (2015) 0 1 0 2 3 0 0 0 2 1 3 6 0 1 0 Sargent et al. (2014) 1 1 0 1 3 0 0 0 2 1 3 6 0 1 0 Kölling et al. (2016) 0 1 1 1 3 0 0 0 2 1 3 6 0 1 0 Fowler et al. (2015) 0 0 0 2 2 0 0 0 1 1 2 4 0 0 1 Fullagar et al. (2016a) 0 1 0 1 2 0 0 0 1 1 2 4 0 0 1 Fullagar et al. (2016b) 0 1 1 1 3 0 0 0 0 1 1 4 0 0 1 Shearer et al. (2015) 0 1 0 1 2 0 0 0 0 1 1 3 0 0 1 Robey et al. (2014) 0 0 0 1 1 0 0 0 1 1 2 3 0 0 1 Sargent al. (2016) 0 0 1 1 2 0 0 0 2 1 3 5 0 0 1 Netzer et al. (2001) 0 0 0 1 1 0 0 0 1 1 2 3 0 0 1 Juliff, Halson and Peiffer 1 1 0 2 4 0 1 1 0 1 1 6 0 1 0 (2015) Tsunoda et al. (2015) 0 0 0 2 2 1 0 1 2 1 3 6 0 1 0 Schaal et al. (2011) 1 1 1 2 5 0 1 1 2 1 3 9 1 0 0 Lucidi et al. (2007) 1 1 0 1 3 1 1 2 2 1 3 8 1 0 0 Silva et al. (2012) 0 1 0 1 2 0 0 0 2 0 2 4 0 0 1 Rodrigues et al. ( 2015) 0 1 0 0 1 0 0 0 1 0 1 2 0 0 1 See Table 1.7 for notes

Table 1.7 Evidence quality appraisal using the adapted Newcastle-Ottawa scale (NOS) for cross-sectional studies (continued) Selectiona Comparabilitya Outcomeb Statistics

sample the ness of Representive size Sample respondents Non elite of Description Subtotal (athletism) factor important most for Control age) (Sex and other factors for Control Subtotal outcome of Assessment test Statistical Subtotal 10) of Total (out (>7) High Mod Low (<5)

athletes

high (5 high

Study

Samuels et al. (2016) 1 1 0 0 2 0 1 1 0 0 0 3 0 0 1 Dickinson and Hanrahan (2009) 0 0 0 1 1 0 0 0 2 1 3 4 0 0 1 Samuels (2008) 0 0 0 1 1 0 0 0 2 0 2 3 0 0 1 Venter et al. (2010) 1 1 1 1 4 0 0 0 0 1 1 5 0 0 1 Swinbourne et al. (2016) 1 1 0 0 2 0 0 0 2 0 2 4 0 0 1 Bleyer et al. (2015) 1 1 1 1 4 0 0 0 2 1 3 7 0 1 0 Chennaoui et al. (2016) 0 0 0 1 1 0 0 0 1 1 2 3 0 0 1 Silva and Paiva (2016) 0 1 0 1 2 0 0 0 1 1 2 4 0 0 1 Fowler et al. (2014) 0 1 0 1 2 0 0 0 1 1 2 4 0 0 1 Erlacher et al. (2011) 1 1 0 1 3 0 0 0 1 1 2 6 0 1 0 Fowler et al. (2017) 0 0 0 1 1 0 1 1 2 1 3 5 0 1 0 Dekker et al. (2014) 0 0 0 1 1 0 0 0 1 1 2 3 0 0 1 Lastella et al. (2014) 0 1 0 1 2 0 0 0 1 1 2 4 0 0 1 Durán et al. (2015) 0 1 0 1 2 0 1 1 2 1 3 6 0 1 0 Elbayoumy and Elbayoumy 0 1 0 0 1 0 0 0 2 1 3 4 0 0 1 (2015) Sargent, Halson and Roach 0 0 0 1 1 0 0 0 2 1 3 4 0 0 1 (2014) Mean 0 1 0 1 2 0 0 0 1 1 2 5 *2 *11 *24 SD 0 0 0 0 1 0 0 1 1 0 1 2 NA NA NA Notes: a subscale items rated 0-1; b subscale items rated 0-2; NA, not applicable; SD, standard deviation. Total score = selection subtotal + comparability subtotal + Outcome subtotal. *total sum of studies

1.3.3 Sleep structure and patterns A total of 20 studies, published between 2001-2015, describing typical sleep profiles, assessed using wrist actigraphy (n=11) (see Tables 1.8 and 1.9), PSG (n=2) (see Table 1.9), sleep diaries (n=4) (see Table 1.10), and questionnaires (n=4) (see Table 1.10), for mainly male elite athletes engaged in normal training were identified. Earlier studies showed a preference for Polysomnography (PSG), while the more recently published studies utilised wrist actigraphy and self-report inventories principally focussing on total sleep time (TST), sleep efficiency (SE) and sleep onset latency (SOL). Only one study compared athletes with controls. Leeder and colleagues (2012) found no significant difference in TST between 46 Summer Olympic athletes and 20 age-matched non-athletes, but did report a significantly lower SE, and significantly higher time in bed (TIB), wake time after sleep onset (WASO), SOL, and sleep fragmentation in athletes. Differences between sports were reported in three actigraphy studies. In a comparison of team and individual sports (Lastella et al., 2015a), individual competitors showed significantly lower TST and SE and longer SOLs. On the other hand, a comparison of canoeing, diving, rowing and skating athletes reported the lowest TST, the shortest SOL, but the highest SE for rowers (Leeder et al., 2012). Consistent with these findings, individual sports were reported to have shorter TSTs than team sports and napped more frequently in the day (15% of 754 days) than team sports (11% of 613 days) (Lastella et al., 2015a). In the only study to compare sex (Leeder et al., 2012), the TIB of male athletes was reported to be 54 min longer than that for female athletes.

1.3.3.1 Summary of sleep structure and patterns Instrumental measurements indicate that while the typical sleep duration of elite athletes may be like that of non-athletes, patterned differences suggest a more fragmented, lower quality sleep among athletes, with most actigraphy studies reporting athlete SE below 90% (see Tables 1.8 and 1.9). Subjective estimates of sleep duration among athletes are broadly consistent with instrumental measures. Both the duration and structure of sleep showed between-sport and sex differences; TSTs are shorter in individual (versus team) sports, and shorter in women.

Table 1.8: Sleep characteristics of elite athletes using actigraphy (See Table 1.9 for total mean scores)

Level of Mean values (SD) Nights Study Sport performance n Sex recorded description Age: y TST: h SE: % SOL: min

Actigraphy

Leeder et al. (2012) Multi-sports GB squad 46 M+F 4 NR 6.9 (0.7) 81 (6) 18 (17)

Lastella et al. (2015a) Multi-sports Elite 124 M+F 12 22.2 (3.0) 6.8 (1.2) 86 (2) 19 (24)

Richmond et al. ARF Professional 19 M 4 24.1 (3.3) 8.9 (0.1) 93 (1) NR (2007)

Richmond et al.(2004) ARF Professional 10 M 7 23.2 (2) 8.4 (0.3) 88 (4) NR

Romyn et al. (2016) Netball State level 8 F 7 19.6 (1.5) 8.2 (0.5) 85 (4) 28 (26)

Schaal et al. (2015) SS International 10 F 7 20.4 (0.4) 7.2 (0.1) 85 (1) 17 (2)

See Table Table 1.9 for notes

Table 1.9 Sleep characteristics of elite athletes using actigraphy (continued)

Level of Nights Mean values (SD) Study Sport performance n Sex recorded description Age: y TST: h SE: % SOL: min

Sargent et al. (2014) Multi-sports National 70 M+F 14 20.3 (2.9) 6.8 (1.5) 86 (7) NR

Robey et al. (2014) Football Elite youth 12 M NR 18.0 (1.4) 7.2 (0.7) 89 (6) 21 (11)

Sargent et al. (2016) Cycling National 16 M 8a,b 19.3 (1.5) 7.6 (0.6) 85 (5) 17 (14)

Rugby Shearer et al. (2015) Elite 28 M 24.4 (2.9) 7.1 (1.0) 79 (9) 34 (40) Union 1

Lastella et al. (2014) Football Elite 16 M 3 18.8 (0.9) 7.5 (1.3) 85 (NR) NR

Mean 7 (4) 21.0 (2.2) 7.5 (0.5) 86 (5) 22 (19)

Notes: SD, standard deviation; NA, not applicable; NR, not reported; M, male; F, female; GB, Great Britain; ARF, Australian rules football; SS, synchronised swimming TST, total sleep time; SOL, sleep onset latency; SE, sleep efficiency; a average number of nights reported; b ‘medium threshold’ selected to compute sleep outcomes

Table 1.10 Sleep characteristics of elite athletes using sleep diaries and polysomnography

Level of performance Nights Mean values (SD) Study Sport n Sex description recorded Age: y TST: h SE: % SOL: min

Sleep diary

Fullagar et al. (2016b) Football Elite 15 M 3 25.5 (4.9) 8.5 (1.2) 92 (4) 20 (17)

Fullagar et al. (2016a) Football Elite 16 M NR 25.9 (7.5) 8.7 (0.7) 96 (NR) 16 (7)

Kölling et al. (2016) Rowing National 55 M+F 6 17.7 (0.6) 6.9 (0.3) 93 (4) 26 (17)

Fowler et al. (2016) Rugby league Professional 18 M 1 24.2 (3.3) 7.9 (1.0) NR NR

Mean 23.3 (4.1) 8.0 (0.8) 94 (4) 21 (14)

Polysomnography

Sargent et al. (2016) Cycling National 16 M 1 19.3 (1.5) 8.5 (0.4) 90 (5) 18 (13)

Netzer et al. (2001) Cycling National 15 M 1 23.9 (NR) NR 93 (3) 19 (16)

Mean 21.6 (2.9) 8.5 (0.4) 92 (4) 19 (15)

Notes: SD, standard deviation; NA, not applicable; NR, not reported; M, male; F, female; GB, Great Britain; TST, total sleep time; SOL, sleep onset latency; SE, sleep efficiency

Table 1.11 Sleep characteristics of elite athletes using questionnaires

Level of Mean values (SD) Nights Study Sport performance n Sex recorded description Age: y TST: h SE: % SOL: min

Questionnairec

Tsunoda et al. ( 2015) WCB Elite 14 M NA 29.5 (5.2) 6.5 (0.9) 88 (9) 25 (22)

Swinbourne et al. (2016) Team sport National 175 M+F NA NR 7.9 (1.3) NR NR

Bleyer et al. (2015) Multi-sport Elite 452 M+F NA 21.9 (2.6) 7.9 (1.5) NR NR

Multi-sport Durán et al. (2015) Elite 33 M+F NA 26.4 (9.8) 6.9 (1.4) (Paralympic) 83 (NR) 44 (8)

Mean 25.7 (6.9) 7.3 (1.3) 86 (9) 35 (15)

Notes: SD, standard deviation; NA, not applicable; NR, not reported; M, male; F, female; WCB, wheelchair basketball; TST, total sleep time; SOL, sleep onset latency; SE, sleep efficiency; c components of the Pittsburgh Sleep Quality Index (PSQI) used to report TST, SOL and SE

1.3.4 Sleep quality and insomnia symptomatology A total of 13 studies published between 2008-2016 reported data on athlete subjective sleep quality and general insomnia symptoms during normal training (see Tables 1.12 and 1.13). While these studies utilised a range of subjective metrics, six used the Pittsburgh Sleep Quality Index (PSQI) (Samuels, 2008; Dekker et al., 2014; Bleyer et al., 2015; Durán et al., 2015; Tsunoda, 2015; Swinbourne et al., 2016), a 19-item scale which assesses seven ‘components’ of sleep (sleep quality, sleep efficiency, sleep onset latency, sleep duration, sleep disturbance, daytime dysfunction and sleep medication use), summing the ‘component scores’ to deliver an overall ‘global’ score; global scores >5 indicate ‘poor sleepers’ (Buysse, Reynolds, Monk, Berman, & Kupfer, 1989). Studies reporting general characteristics of sleep quality or PSQI values are presented in Tables 1.12 and 1.13, respectively. Data in Table 1.12 are summarised as prevalence rates of sleep symptoms. Studies administering the PSQI adopted different reporting conventions; both mean scores and threshold (>5) prevalence rates are therefore shown in Table 1.13.

Overall, the sleep assessments shown in Table 1.12 show a relatively high level of sleep complaints, with reports of sleep disturbance ranging from 28-70%. The low prevalence of “abnormal sleep” reported by Samuels et al. (et al. 2016) involved an arbitrary cut-off which was not based on empirical data and may not, therefore, represent a robust estimate. In the two studies which reported the prevalence of sleep disturbance by sex (Lucidi et al., 2007; Schaal et al., 2011), rates were highest among women. One study (Schaal et al., 2011) explored this further, reporting that female athletes experienced more problems both initiating and maintaining sleep when compared to their male counterparts. Sleep quality differences between sports were also identified in this study (Schaal et al., 2011), with elite French athletes from aesthetic sports reporting a significantly higher prevalence of insomnia symptoms (33%) compared to all other sports (26% for the sample overall).

Formal (PSQI) assessments of sleep quality suggest similarly high levels of insomnia-type symptoms (Table 1.13), with mean values at (Bleyer et al., 2015) or above the threshold of >5 (Samuels, 2008; Dekker et al., 2014; Durán et al., 2015;

Tsunoda, 2015; Swinbourne et al., 2016). This assumption is supported by the prevalence rates reported, with levels of significant sleep disturbance ranging from 38% of multi-sport athletes (Bleyer et al., 2015) to 54% of bobsleigh competitors (Samuels, 2008). The more conservative PSQI threshold of >8, indicative of highly disturbed sleep, also showed a relatively high prevalence, ranging from 22-26% (Samuels, 2008; Swinbourne et al., 2016). However, the possibility that the higher thresholds shown in Table 1.13 may mask more severe symptoms is suggested by Swinbourne et al. (2016) who reported that 9% of elite New Zealand team sport athletes scored >10. Only one of the studies shown in Table 1.13 included control comparisons. Tsunoda et al. (2015) compared the PSQI global scores of 14 international wheel chair basketball athletes (mean=6; SD=3) with 103 non-athletes (mean=5; SD=2), and found the difference significant (P<0.05). In the same study, PSQI component score data also showed that athletes reported significantly lower subjective sleep quality and SE, even though reported total sleep time showed no significant difference between the groups.

While the PSQI is not a diagnostic tool (Buysse et al., 1989), three studies in this section used instruments validated against insomnia diagnostic criteria (Dickinson & Hanrahan, 2009; Lucidi et al., 2007; Schaal et al., 2011) which allow inferences to be drawn regarding the prevalence of insomnia caseness in elite athlete populations. Using the Athens Insomnia Scale (AIS, an instrument validated against 10th revision of the International Statistical Classification of Diseases and Related Health Problems criteria for insomnia (Soldatos, Dikeos, & Paparrigopoulos, 2000), Dickinson and Hanrahan (2009) reported a mean score for elite multisport athletes of 5 (range 0–16). Since scores of ≥6 indicate clinically significant insomnia symptoms, and since the reported score distribution from this study showed no significant skewness or kurtosis (Dickinson & Hanrahan, 2009), then it can be assumed that, while the study does not report the prevalence of ≥6 scores, a high proportion of athletes must nevertheless have experienced serious insomnia symptoms. Consistent with this assumption, Dickinson and Hanrahan (2009) also reported relatively high levels (for this age group) of daytime fatigue among athletes, together with consistent reports (from qualitative interviews) of nonrestorative sleep despite apparently adequate sleep durations. Using the Sleep Disorders Questionnaire (SDQ; a brief questionnaire validated against DSM-IV criteria for insomnia (Violani, Devoto,

Lucidi, Lombardo, & Russo, 2004), Lucidi et al. (2007) reported that 4% of Italian Olympic athletes met diagnostic criteria for insomnia. Details of symptom chronicity, closely related to insomnia diagnosis, were provided by Schaal et al. (2011) who reported a six month prevalence of insomnia symptoms of 22%, but a lifetime prevalence of insomnia symptoms of 30%, strongly indicating very high levels of sleep pathology within this nationally representative sample of elite French athletes.

1.3.4.1 Summary of sleep quality and insomnia symptomatology The general pattern of results indicates high levels of subjective sleep disturbance and insomnia symptomatology within elite sport, with the evidence suggesting that, within athlete populations, levels of sleep disturbance are higher among women, and among aesthetic athletes. Formal measurements of subjective sleep in athletes also show findings which accord with the objective data, with similarities reported in the TST of athletes and non-athletes, but significantly lower levels of sleep quality reported by athletes. Such evidence, together with that presented in the preceding section also show that levels of daytime fatigue in athletes can be directly related to degraded night-time sleep.

Table 1.12: Characteristics of subjective sleep quality in elite athletes

Level of Mean Prevalence (%) Symptoms assessed Study Sport performance n Sex age, y of symptoms description description (SD) Total M F

Experience of sleep Venter et al. (2010) Team sports National 890 M+F 22.3 (3.4) 41 NR NR problems

Ongoing sleep Schaal et al. (2011) Multi-sports National 2067 M+F 23.5 (NR) 22 20 24 problemsa

Occasional sleep Lucidi et al. (2007) Multi-sports Olympic 103 M+F 23.9 (4.1) 60 59 62 disturbancesb

Rodrigues et al. (2015) Sleep dissatisfactionc Para-athletics Paralympic 40 M+F 30.1 (7.1) 46 NR NR

General sleep Juliff, Halson and disturbanced after a Multi-sports Elite 283 M+F 24.1 (5.1) 28 NR NR Peiffer (2015) rest period

Multi-sport Durán et al. (2015) Insomnia symptomse Elite 33 M+F 26.4 (9.8) 70 NR NR (Paralympic)

Samuels et al. ( 2016) Abnormal sleepf Multi-sports Elite 349 NR NR 13 NR NR

Notes: SD, standard deviation; NR, not reported; M, male; F, female; a any of sleep onset, sleep maintenance and daytime sleepiness problems in last 6 months; b as categorised by the Sleep Disorders Questionnaire (Montserrat Sánchez-Ortuño, Edinger, Means, & Almirall, 2010); c as measured by the Federal University of Sâo Paulo (UNIFESP) Sleep Questionnaire (Pires et al., 2007); d any of sleep onset , sleep maintenance, early morning awakening, unrefreshing sleep, or disturbing dreams; e Insomnia Severity Index (ISI); f as measured by the Athlete Sleep Screening Questionnaire (ASSQ) (Samuels et al., 2016).

Table 1.13: Sleep assessments in elite athletes using the Pittsburgh Sleep Quality Index

Level of Mean Prevalence Mean global Study Sport performance n Sex age, y score (SD)a description (SD) %≥ 5 % > 5 % > 8

Dekker et al. (2014) Gymnastics National 12 M+F 22.9 (3.5) 6 (NR) NR NR NR

Samuels et al. Bobsleigh Elite 24 M+F 27.0 (NR) 6 (1) 78 54 26 (2008)

Tsunoda et al. WCB Elite 14 M 29.5 (5.2) 6 (3) NR 57 NR (2015)

Swinbourne et al. Team-sport National 175 M+F 21.9 (2.6) 6 (3) 65 50 22 (2016)

Bleyer et al. (2015) Multi-sport Elite 452 M+F 21.2 (5.8) 5 (3) NR 38 NR

Multi-sport Durán et al. (2015) Elite 33 M+F 26.4 (9.8) 11 (8) 79 NR NR (Paralympic)

Mean 24.8 (5.4) 7 (4) 74 50 24

Notes: SD, standard deviation; NR, not reported; M, male; F, female; WCB, wheelchair basketball; a scores of >5 are indicative of clinical sleep disturbance.

1.3.5 Risk factors for sleep disturbance Studies reporting sleep quality, insomnia symptomatology and changes in sleep patterns broadly focussed on three challenges to athlete sleep: 1) competition (see Tables 1.14, 1.15 and 1.16); 2) travel (see Tables 1.17 and 1.18); and 3) training (see Tables 1.19 and 1.20).

1.3.5.1 Competition Of the studies assessing subjective sleep quality pre-competition (n=7), most employed the PSQI, with 5 reporting an increased prevalence of complaints (Table 1.14 and 1.15). Silva and Paiva (2016) found that 78% of international female gymnasts scored >5 (indicative of ‘poor’ sleep) on the PSQI prior to an international competition. Gymnasts who scored more ‘competition points’, however, reported significantly worse sleep quality (mean PSQI=8) than those who scored less (mean PSQI=6). Sex and inter-sport differences were considered in two studies. Using the Competitive Sports, Sleep and Dreams Questionnaire (CSSDQ), a metric designed to assess sleep habits and disturbances prior to competition (Erlacher et al., 2011), two studies reported high prevalence rates of pre-competition sleep disturbance (64-66%), but found no differences between male and female athletes (Erlacher et al., 2011; LE Juliff et al., 2015). When comparing sports however, Erlacher et al. (2011) reported a significantly greater frequency of sleep disturbances in individual (69%) when compared to team sport athletes (60%). Such differences were not supported by Juliff, Halson and Peiffer (2015), who reported similar levels of sleep disturbance between sports.

Of six studies which used wrist actigraphy to assess sleep patterns prior to competition, most reported no significant changes in SE and SOL when compared to normal training (see Table1.14). Three studies, however, reported a significant increase in pre-competition TST (Chennaoui et al., 2016; Richmond et al., 2004; Richmond et al., 2007), while one study (Romyn et al., 2016) reported a significant increase in SE. Again, however, there was evidence of sleep- performance relationships. Chennaoui et al. (2016) reported that elite swimmers who finished within the top 4 at the French national championships exhibited more consistent TSTs across the competition when compared to swimmers who finished outside the top 4 betterr, with this latter group reporting significantly longer TSTs the night before the final race.

The six actigraphy and one sleep diary study shown in Tables 1.16 and 1.17 all reported a significant decrease in TST, and a significantly delayed bedtime, following night competitions, with no study reporting significant changes in SE or SOL. Actigraphy studies assessing post competition sleep quality more generally were equivocal, with three studies showing a significant decrease (Richmond et al., 2004; Fowler et al., 2015; Fullagar et al., 2016a) and two studies showing no change (Fowler, Duffield and Vaile, 2014; Fullagar et al., 2016b). Fullagar et al. (2016a) also reported a decrease in sleep ‘restfulness’ in elite football players following a night competition, compared to day-matches and training days. In the only study assessing PSG measured sleep structure (see Table 1.17), Netzer et al. (2001) reported a significant increase in stage 3 sleep following competition (compared to rest days) and a significantly increased REM sleep onset latency. However, no changes in SOL or SE were reported.

1.3.5.2 Travel Studies investigating the impact of both long- (n=4) and short-haul (n=3) travel on athlete sleep are shown in Table 1.18. Overall, studies used a range of designs and methodological approaches. Three of the actigraphy studies investigating long-haul travel reported a rating of jet lag in addition to sleep outcomes (Fowler et al., 2015, 2016; Fullagar et al., 2016b). Time zone changes ranged from 2-11 h, with the maximum time zone change when travelling east being 8 h (Michele Lastella et al., 2014) and the maximum when travelling west being 11 h (PM Fowler et al., 2016). Across all studies, however, no change in SOL or SE post travel was reported following travel compared to pre-travel assessments. Changes in TST and sleep quality (SQ) reported were mixed, with one study showing a significant decrease in TST following long-haul eastward travel (Lastella et al., 2014) and another showing an increase following long-haul westward travel (Fowler et al., 2016). The majority of studies reported no change in sleep quality; however Richmond et al. (2007) reported a significant decrease in sleep quality prior to away matches following a 2 h eastward time-zone change (mean score=3.4), when compared to home matches (mean score=3.8) in elite Australian Rules Football players. Ratings of jet lag following westward travel (no assessments of jet-lag were reported following eastward travel) showed a positive trend with increasing time-zone change, with a 1 (Fowler et al., 2015), 4

(Fullagar et al., 2016b) (both P<0.05) and 11 h time zone (Fowler et al., 2016) (P<0.01) change showing a significant increase from pre-travel assessments.

1.3.5.3 Training Studies investigating the impact of training on sleep also showed methodological differences, either comparing training days or rest days, or comparing intensified training with normal training. Most studies reported instrumental measures to assess changes in sleep patterns with only one study using a questionnaire (See Tables 1.19 and 1.20). In studies comparing training and rest days all studies reported significantly earlier rise times and decreased TST on training days (P<0.01). Sargent et al. (2014) reported a significant total sleep time gradient relative to training start-times across different sports, with earlier start times associated with lower TSTs and greater pre-training levels of fatigue. Only two studies assessed sleep quality (Kölling et al., 2016; Schaal et al., 2015), with one of these reporting a reduction in ‘sleep restfulness’ on a training day compared to a rest-day (Kölling et al., 2016). Two studies quantified levels of daytime sleep (Kölling et al., 2016; Sargent, Lastella, et al., 2014). Sargent et al. (2014) found elite athletes to nap at similar frequencies on training (15% of 14 days) and rest days (16% of 14 days). However, Kölling et al. (2016) reported that the proportion of elite rowers who napped on training days (43%, n=24) was greater than that for rest-days (16%, n=9).

In the studies reporting comparisons between intensified and normal training, significant decreases in TST were observed in all studies (see Table 1.20) (Juliff et al., 2015; Kölling et al., 2016; Schaal et al., 2015). However, changes in rise time, SOL, SE and sleep quality were equivocal. Schaal et al. (2015) reported a significant decrease in SE and increase in sleep onset latency, but no change in sleep quality during two weeks of intensified training when compared to baseline values in elite synchronised swimmers. Consistent with this Juliff, Halson and Peiffer (2015), using questionnaire assessments, reported that 55% of elite athletes experienced sleep disturbances during periods of heavy training.

1.3.5.4 Summary of risk factors Among elite athletes, predictable events in the training/competition cycle are associated with an increased risk of insomnia symptomatology and disturbed sleep patterns: competition; long- and short-haul travel; and training. Typically,

sleep quality significantly declines prior to competition for both men and women. Following competitions, the impact on sleep is related to the timing of events, with late-evening competitions delaying bedtimes and reducing TST. While the circadian de-synchrony (jet lag) associated with long-haul travel significantly affects sleep patterns, it appears that sleep quality, and instrumental indices of sleep quality, such as SOL and SE, are more resilient. Nevertheless, few studies of jet lag or travel fatigue in elite athletes have used formal assessments of insomnia symptoms. Finally, training days can require earlier rise times, with consequent reductions in TST, increased daytime fatigue, and an increased likelihood of daytime napping in some sports.

Table 1.14: Prevalence of insomnia symptomatology and changes in sleep patterns pre-competition

Prevalence PSQI Δ Sleep patterns and Competition Mean (%) of global sleep quality Study Sport description n Sex age: y insomnia mean (home/ away) (SD) symptoms (SD) SE SOL TST SQ

Actigraphy

National Romyn et al. (2016) Netball 8 F 19.6 (1.5) NR NR ↑ ↔ ↔ ↔ Championships

Richmond et al. AFL match ARF 19 M 24.1 (3.3) NR NR ↔ NR ↑ (2007) (home) ↔

Richmond et al. AFL match ARF 10 M 23.0 (2.0) NR NR ↔ NR ↑ (2004) (home) ↔

Rugby Celtic League Shearer et al. (2015) 28 M 24.4 (2.9) NR NR ↔ ↔ ↔ ↔ Union match (home)

Fowler, Duffield and A-League match ↔ ↔ ↔ ↔ Vaile (2014) Football (home) 6 M 23.4 (NR) NR NR

Channoui et al. National Swimming 9 M+F 22.0 (3.0) NR NR ↔ NR ↔ NR (2016)a championships

See Table 1.15 for notes

Table 1.15 Prevalence of insomnia symptomatology and changes in sleep patterns pre-competition (continued)

Prevalence PSQI Δ Sleep patterns and Competition Mean (%) of global sleep quality Study Sport description n Sex age: y insomnia mean (home/ away) (SD) symptoms (SD) SE SOL TST SQ

Questionnaire

Elbayoumy and National Swimming 40 M 19.0 (1.0) NR 5 (1) NR NR NR NR Elbayoumy (2015)b, c Championships

Swinbourne et al. (2016)b, c Team-sport In-competition 75 M+F NR NR 6 (3) NR NR NR NR

Silva and Paiva (2016)b, c, d Gymnastics FIG World Cup 67 F 18.7 (2.9) 78 7 (3) NR NR NR ↓

Para- Paralympic Rodrigues et al. (2015)b, e 40 M+F 30.1 (7.1) 37 NR NR NR NR NR athletics games

Para- Paralympic Silva et al. (2012)b, e 27 M+F 28.0 (6.0) 70 NR NR NR NR NR athletics games

Important Erlacher et al. (2011)f Multi-sport 632 M+F 21.9 (6.8) 66 NR NR NR NR NR competition

Juliff, Halson and Peiffer Multi-sport Olympic Games 283 M+F 24.0 (5.0) 64 NR NR NR NR NR (2015)b, f

Notes: SD, standard deviation; M, male; F, female; SE, sleep efficiency; SOL, sleep onset latency; TST, total sleep time; SQ, sleep quality (subjective rating); ARF, Australian Rules Football; AFL, Australian Football League; FIG, International Federation of Gymnastics; NR, not reported; ↑ significant increase; ↔ no significant change; ↓ significant decrease (all p<0.05); a sleep pattern changes in successful athletes during competition reported only; b Pittsburgh Sleep Quality Index with c poor sleep quality threshold >5 employed (Buysse et al., 1989); d comparisons made between successful and unsuccessful athletes during competition; e Pittsburgh Sleep Quality Index with poor sleep quality ≥5 employed; f Competitive Sports, Sleep, and Dreams questionnaire employed (Erlacher et al., 2011).

Table 1.16: Prevalence of insomnia symptomatology and changes in actigraphy and sleep diary derived sleep patterns post-competition

Prevalence Δ Sleep patterns and Competition Mean of insomnia sleep quality Study Sport description n Sex age, y symptoms (home/ away) (SD) (%) BT SE SOL TST SQ

Actigraphy

Rugby Celtic League Shearer et al. (2015) 28 M 24.4 (2.9) NR ↑ ↔ ↔ ↓ NR Union match (home)

Fowler, Duffield and Vaile, A-League (home) ↑ ↔ ↔ ↓ (2014) Football 6 M 23.4 (NR) NR ↓

Pre-season tour Fowler et al. (2015) ↓ ↓ Football (away) 16 M 27.0 (NR) NR NR NR NR

Pre-FIFA World Fullagar et al. (2016b) ↑ ↓ Football Cup (away) 15 M 25.5 (4.9) NR ↔ ↔ ↔

Richmond et al. (2004) ARF AFL (home) 10 M 23.0 (2.0) NR NR ↔ NR ↓ ↓

Notes: SD, standard deviation; M, male; F, female; BT, Bed-time; SE, sleep efficiency; SOL, sleep onset latency; TST, total sleep time; SQ, sleep quality (subjective rating); ARF, Australian Rules Football; AFL, Australian Football League; NR, not reported; ↑ significant increase; ↔ no significant change; ↓ significant decrease (all p<0.05); a ad-hoc question employed - “If you have a late training session or game do you find it hard to sleep after?”

Table 1.17 Prevalence of insomnia symptomatology and changes in Polysomnography and questionnaire derived sleep patterns post-competition

Prevalence Competition Mean of insomnia Δ Sleep patterns and sleep Study Sport description n Sex age, y symptoms quality (home/ away) (SD) (%)

Sleep diaries BT SE SOL TST SQ

Bundesliga/ Fullagar et al. (2016a) Football Eredevisie (home 16 M 25.9 (7.5) NR ↑ NR ↑ ↓ ↓ and away)

Polysomnography

German First Netzer et al. (2001) Cycling Division 15 M 23.9 (NR) NR NR ↔ ↔ NR NR

Questionnaire

Juliff, Halson and Peiffer, Multi- Olympic Games 283 M+F 24.0 (5.0) 53 NR NR NR NR NR (2015)a sport

Notes: SD, standard deviation; M, male; F, female; BT, Bed-time; SE, sleep efficiency; SOL, sleep onset latency; TST, total sleep time; SQ, sleep quality (subjective rating); ARF, Australian Rules Football; AFL, Australian Football League; NR, not reported; ↑ significant increase; ↔ no significant change; ↓ significant decrease (all P<0.05); a ad-hoc question employed - “If you have a late training session or game do you find it hard to sleep after?”

Table 1.18: Changes in sleep patterns, sleep quality and jet lag following long- and short-haul travel

Mean Δ Sleep patterns, sleep quality Flight type (Δ time Study Sport n Sex age, y and jet lag zone W/E) (SD) SE SOL TST SQ JL

Long haul

Actigraphy

Fullagar et al. (2016b)a Football International (4 h W) 15 M 25.5 (4.9) ↔ ↔ ↔ ↔ ↑

Fowler et al. (2015)a Football International (1 h W) 16 M 27.0 (NR) NR NR ↔ ↔ ↑

Lastella et al. (2014)a, b Football International (8 h E) 16 M 18.8 (0.9) ↔ NR ↓ ↔ NR

Sleep diary

Fowler et al. (2016)a Rugby League International (11 h W) 18 M 24.2 (3.3) NR ↔ ↑ ↔ ↑↑

Short haul

Actigraphy

Fowler, Duffield and Domestic (2 h E and Vaile (2014)c Football W) 6 M 23.4 (NR) ↔ ↔ ↔ ↔ NR

Richmond et al. (2004)c ARF Domestic (2 h E) 10 M 23.0 (2.0) ↔ NR ↔ ↔ NR

Richmond et al. (2007)c ARF Domestic (2 h E) 19 M 24.1 (3.3) ↔ NR ↔ ↓ NR

Notes: SD, standard deviation; M, male; W, westward travel; E, eastward travel; SE, sleep efficiency; SOL, sleep onset latency; TST, total sleep time; SQ, sleep quality (subjective rating); JL, Jet lag; ARF Australian Rules football; NR, not reported; ↑, sig. increase; ↔, no sig change; ↓ sig. decrease (all P<0.05); ↑↑ sig. increase (P<0.01); a sleep patterns assessed days 1-2 post travel in comparison to pre-travel assessments; b assessments made at low (1600m) altitude following travel; c assessments made at away matches and compared to home match responses.

Table 1.19: Prevalence of insomnia symptoms and changes in sleep patterns during training days v rest days

insomnia Δ Sleep patterns and sleep Training Mean Study Sport n Sex symptoms quality description age (SD) (%) RT SE SOL TST SQ

Actigraphy

Sargent, Halson and Roach Olympic 7 M+F 22.5 (1.7) NR ↓↓ ↔ ↔ ↓↓ NR (2014) Swimming preparation

Sargent et al. (2014) Multi-sport Normal training 70 M+F 20.3 (2.9) NR ↓↓ ↔ NR ↓↓ NR

World Cup Kölling et al. (2016) 18 M+F 17.7 (0.6) NR ↓↓ ↔ ↔ ↓↓ ↓↓ Rowing preparation

Questionnaire

Juliff, Halson and Peiffer Following a rest Multi-sport 283 M+F 24.0 (5.0) 28 NR NR NR NR NR (2015)a day

Notes: SD, standard deviation; M, Male; F, Female; BT, Bed-time; SE, sleep efficiency; SOL, sleep onset latency; TST, total sleep time; SQ, sleep quality (subjective rating); SS, Synchronised Swimming; NR, not reported; ↑ significant increase; ↔ no significant change; ↓ significant decrease (all p<0.05); ↑↑ significant increase; ↓↓ significant decrease (all p<0.01) a ad-hoc question employed.

Table 1.20 Prevalence of insomnia symptoms and changes in sleep patterns during intensified training v normal training

Δ Sleep patterns and sleep insomnia Training Mean quality Sport n Sex symptoms description age (SD) Study (%) RT SE SOL TST SQ

Actigraphy

Olympic Schaal et al. (2015) SS 14 F 20.4 (0.4) NR ↑↑ ↓ ↑ ↓ ↔ preparation

World Cup 18 M+F 17.7 (0.6) NR ↔ ↔ ↔ ↓ ↓ Kölling et al. (2016) Rowing preparation

Questionnaire

Juliff, Halson and Peiffer Heavy training Multi-sport 283 M+F 24.0 (5.0) 28 NR NR NR NR NR (2015)a period

1.4 Discussion This review aimed: 1) to systematize the research evidence describing sleep patterns and sleep quality among elite athletes; and 2) to consider specific risk factors for sleep quality within elite sport. The studies reviewed here broadly support a conclusion that elite athletes experience high levels of sleep disturbance, and that such disturbances are characterised by the symptoms of longer SOLs, greater sleep fragmentation, non-restorative sleep, and excessive daytime fatigue. Within elite sport environments the evidence also suggests that periods of competition, travel and training are likely to precipitate experiences of diminished sleep quality. As a result, both null hypotheses can be rejected.

1.4.1 Sleep patterns (hypothesis 1) This pattern of degraded sleep quality is clearly illustrated by the composite measure of SE derived from the actigraphy studies shown in Tables 1.8 and 1.9. The pooled average sleep efficiency for athletes (mean=86; SD=5%), is close to, and for many overlaps the threshold value, of 85%, below which insomnia symptoms are indicated (Montserrat Sánchez-Ortuño et al., 2010). These experiences are not necessarily associated with shorter objectively measured sleep durations, with some studies indicating similarly recorded total sleep TSTs in athletes and controls (Leeder et al., 2012; Tsunoda, 2015). Despite indications of degraded sleep quality among elite athletes, secondary analyses revealed no significant differences in SE in a study comparing full-time female dancers and Leeder et al.’s (2012) Olympic athletes during normal training/ practice (McCloughan, Hanrahan, Anderson, & Halson, 2016). Such comparisons suggest that low levels of sleep quality may extend to other populations with high levels of expertise and may not necessarily be exclusive to elite athletes. To some degree, however, differences observed between studies in Tables 1.8 and 1.9 are likely to reflect methodological inconsistencies in instrumental measurements of sleep; in particular lengths of recording periods selected to establish normal sleep patterns, with studies ranging from 1-14 nights (see Tables 1.8. and 1.9), and pre- selected thresholds for scoring sleep (e.g. low, medium, or high).

1.4.2 Sleep quality and insomnia symptomatology (hypothesis 1) Nevertheless, using the single PSQI sleep quality metric, approximately one third to one half of all elite athletes can be categorised as ‘poor sleepers’ (Table 1.13), with higher levels of insomnia symptoms (up to 70%) reported by Paralympic athletes (Table 1.12). Despite these apparently high levels of sleep complaints, the evidence does not unequivocally support a conclusion that elite sport per se either degrades sleep quality or drives high levels of insomnia diagnosis. Only three studies directly assessed the diagnosis of insomnia. The reported prevalence rate for clinical insomnia among Italian Olympic athletes (4%) Lucidi et al., (2007) falls comfortably within the prevalence range of 3-6% for the general population aged 15-34 y in Southern Europe (M. M. Ohayon & Sagales, 2010). Moreover, values reported for the AIS among Australian elite athletes (0-16) (Dickinson & Hanrahan, 2009), while greater than those for healthy adults (0-11) were less than those for diagnosed adults with insomnia(1-24) (Soldatos et al., 2000). The one study which compared athletes and non-athletes found no between-group differences in TST, but did report significantly superior SE and sleep quality among non-athletes (Tsunoda, 2015). To an unknown extent, these outcomes may be influenced by methodological differences. However, while the demands of training, competition and travel certainly contribute to sleep disruption in elite sport, it is also likely that high-performance competitors share sleep vulnerabilities with their non-athlete peers. In general, athletes are drawn from younger high-achieving populations. Epidemiological studies show similarly high levels of sleep disturbance among young people in general, and university students. In a community-based random sample of young adults (aged 18-29), for example, Wong and Fielding (2011) report 34% scoring >5 on the PSQI, while both Lund et al. (2010) and Ye et al. (2015) found that 60% of university students were similarly classified by the PSQI as poor sleepers. Given such findings, the expectation, a priori, reflected in the sports science literature that “...poor sleep quality would not be likely in a young, healthy athletic population” (Samuels et al., 2016, pp. 4) is not supported by the epidemiological evidence. Even so, cross- study comparisons do not allow for a definitive judgement to be made on whether elite athletes experience disproportionately higher overall levels of insomnia symptoms for their age. Such a judgement requires additional controlled (athlete vs. non-athlete) comparisons in the sports science and

medicine literature. Furthermore, it should also be acknowledged that the PSQI provides a broad measure of sleep quality (Buysse et al., 1989) and scores can be influenced by a range of sleep disorders. It cannot, therefore, be interpreted as a measure specific to insomnia symptoms. For example, in the study by Swinbourne et al. (2016) while 50% of team sport athletes scored >5 on PSQI, 28% were also ‘excessively’ sleepy (as measured by the Epworth Sleepiness Scale) and 8-38% reported symptoms of obstructive sleep apnoea.

1.4.3 Risk factors for sleep disturbance (hypothesis 2) The present results broadly identified periods of elevated insomnia symptom risk within a ‘trio’ of sleep challenges – 1) competition (See Tables 1.14, 1.15, 1.16 and 1.17), 2) travel (See Table 1.18) and 3) training (See Table 1.19 and 1.20). However, when considered in relation to other achievement-focussed cohorts of younger adults, it does not necessarily follow that “professional sportspeople (both players and officials) face unique challenges relative to their ability to achieve sufficient sleep” (Taylor et al., 2016, pp. 1). For example, the high prevalence of poor sleep reported prior to competition in elite athletes is similar to that reported by professional ballet dancers prior to a premiere (Fietze et al., 2009) and university students during an exam period (Ahrberg, Dresler, Niedermaier, Steiger, & Genzel, 2012). Similarly, the relationship between rise times and TST seen here for elite athletes (e.g. Sargent et al. (2014)) is also seen among college students when comparing the impact of earlier (07:00) and later (10:00) class start times (Lima, Medeiros, & Araujo, 2002). Furthermore, longer TSTs reported for college students at weekends, when classes were not scheduled (Lima et al., 2002) are also consistent with studies reporting longer TSTs on rest days among elite athletes (see Table 1.13). The rigorous, physical training regimes adhered to by elite athletes is a feature of elite sport which is absent among non-athlete populations. It is also relevant to note, however, that other populations which undertake high levels of physical training or practice, such as performing artists (Fietze et al., 2009; McCloughan et al., 2016), or active-duty military personnel (Mysliwiec et al., 2013), also exhibit high levels of disrupted sleep. The significant reductions in TST and SE that were reported during a short-term intensified training period among elite athletes in preparation for an Olympic games (Schaal et al., 2015) were similar to those reported during an extended rehearsal period over three months in professional male and female ballet dancers in preparation for a premiere

(Fietze et al., 2009). Given these similarities it is reasonable to suggest that the prevalence of insomnia symptoms in both elite athletes and other groups of younger adults may be mediated by common mechanisms.

States of hyperarousal (manifesting as pre-sleep cognitions and ruminations, stress or worry), were reported to be highly influential in mediating student sleep difficulties (Lund et al., 2010), and have also been identified as a prime cause of sleep disturbance in elite athletes prior to competition (Erlacher et al., 2011; LE Juliff et al., 2015). Lund et al. (2010), for example, reported that delayed sleep onset (>30 min) was one of the most commonly reported sleep problems among students, with psychometrically assessed “stress” the strongest single predictor of insomnia symptomatology. Similarly, in athlete populations both prior to competition (Erlacher et al., 2011; LE Juliff et al., 2015) and in general (Schaal et al., 2011) sleep onset problems associated with stress and pre-sleep mentation predominate. Juliff, Halson and Peiffer (2015) found that the main reasons for delayed sleep onset prior to competitions were “thoughts about the competition” and “nervousness”, while Schaal et al. (2011) found that aesthetic athletes, who may be especially concerned with body-image and ‘perfection’ in performance, reported the highest levels of stress and sleep onset problems. It is reasonable to conclude, therefore, that cognitive and physiological arousal presents both an explanation for sleep disturbance, and a target for sleep management, among elite athletes.

1.4.4 Between-sport differences Notwithstanding similarities between overall patterns of insomnia symptoms among elite athletes and younger non-athlete populations, it is also clear that some sports impact sleep patterns and quality more than others. Again, levels of pre-sleep arousal appear to play a strong role, here, with typical average sleep latencies reported to vary from 40 min (for swimmers) to 8 min (for rugby players) (Lastella et al., 2015a). For example, Schaal et al. (2011) described higher levels of sleep complaints to be reported in aesthetic sports as a result of the “psychological toll” experienced in athletes where success is based upon judgement by others e.g. judges and coaches. Conversely, Erlacher et al. (2011) suggested team sport athletes may experience less pre-competition anxiety, when compared to individual athletes, largely due to a diffusion of responsibility

among teams for their competitive outcomes. It is also relevant to note that sex composition could also explain some of the between-sport differences reported for sleep quality. It is a robust finding in sleep epidemiology that, for all adult age groupings, women tend to report higher levels of insomnia symptomatology than men (Lund et al., 2010; Ohayon, 2002; Wong & Fielding, 2011). Those sports showing higher proportions of female athletes (e.g. aesthetic sports) might be expected to reflect this trend. The evidence reviewed here further suggests that training schedules can contribute substantially to inter-sport differences. Where sports adopt very early training start times (e.g. swimming - Sargent, Halson and Roach (2014), triathlon, rowing - Leeder et al., 2012) athletes adopt correspondingly earlier bed-times (Kölling et al., 2016; Sargent, Halson, et al., 2014; Sargent, Lastella, et al., 2014). It appears, however, that such adjustment does not always compensate for truncated sleep. As a result TST tends to be lower and levels of pre-training fatigue and the frequency of daytime napping tend to be higher in sports demanding earlier rise-times (Sargent et al., 2014; Lastella et al., 2015). Training schedules which reduce TST can also have an important, and seemingly paradoxical, impact on sleep structure. Therapeutically, sleep restriction (the induction of a mild state of sleep deprivation) is used in the treatment of insomnia to increase sleep need and thereby reduce SOLs and increase SE (see Miller et al. (2014)). In Leeder’s actigraph study (2012) rowers, who have very early training start times, were also reported to have the lowest TSTs when compared with canoers, divers and skaters. It is unsurprising, therefore that rowers were also found to have the shortest SOLs, and the highest levels of SE. Although interpreted as superior sleep quality (Leeder et al., 2012) these latter findings are more likely attributable to restricted sleep.

1.4.5 Within-sport differences Overall sleep quality and the impact of specific sleep challenges do not appear uniform across athletes. Individual responses to pre-competitive stress, circadian challenges, and late night and early morning scheduling all demonstrate similarly high levels of variance, with some athletes experiencing severe sleep disturbances, while others appear unchallenged. Fullagar et al. (Fullagar et al., 2016a), for example, reported that, at a squad level, male elite football players

experienced diminished total sleep times following late-evening league matches; however, within squad comparisons also revealed wide individual variations in the degree of resultant sleep deprivation experienced. In this instance, it is possible that, to an unknown extent, these variations could reflect individual differences in chronotype (i.e. positions on the morningness–eveningness continuum) (Ferrara & De Gennaro, 2001), or sleep need (i.e. position on the short-sleeper–long sleeper continuum) (Ferrara & De Gennaro, 2001). In athletes, recent research has indicated a skew towards morning types (Samuels, 2008; A. Silva et al., 2012). However, given the large variations in training schedules adopted across sports, it remains possible that some individual athletes may be disadvantaged by, say, very early training times. Despite the wide range of metrics and instruments employed to identify individuals with sleep pathologies in elite sport, the development of methodologies to explain individual differences in response to sleep challenges, or identification of ‘at risk’ sleep phenotypes remains under researched. For example, in a single study, Juliff, Halson and Peiffer (2015) suggested that measures of general sleep quality (as measured by the PSQI) were not associated with sleep disturbances experienced during competition periods. This indicates that metrics employed to assess general levels of sleep quality may not necessarily highlight ‘at risk’ phenotypes or inform targeted sleep management in scenarios when individuals are placed under stress (e.g. competition, travel, and training).

1.5 Conclusion and research agenda While acknowledging the limited number of high-quality studies reviewed here, the current literature consistently reports that elite athletes generally show a high overall prevalence of insomnia symptoms characterised by longer SOLs, greater sleep fragmentation, non-restorative sleep, and excessive daytime fatigue. However, the literature reviewed here also showed the following limitations.

1. There is little methodological consistency in the approach to athlete sleep assessment, with the construct of ‘sleep quality’ poorly operationalised in most studies. Few of the studies reviewed here contextualised the sleep outcomes reported with insomnia symptomatology.

2. In most studies neither the selection nor the interpretation of athlete sleep assessments appears to have been influenced by current theories in sleep research and medicine, and none of the studies offered explanations for athlete sleep disturbances in terms of mechanisms or hypotheses drawn from behavioural sleep medicine.

3. While between-sport differences in sleep outcomes are evident from the review, few studies systematically explore sleep differences between sports and, of those which did, none attempted to explain differences in terms of constructs presumed to drive or support sleep quality.

The programme of research presented in the following chapters, therefore, aimed to both address these limitations and extend our present understanding of sleep quality, and the management of sleep quality, in the context of elite sport performance. Specifically, adopting a sleep assessment methodology for elite athletes informed by current theory and practice in sleep research and medicine, the present research programme set out:

1. To assess within-sport sleep quality differences with a view to exploring theoretical constructs and positions which could explain such differences specifically, and sport-related sleep disturbances more generally;

2. To evaluate the sleep-related impact of training, travelling and competing on aspects of sport performance;

3. To interpret results in the context of current research in sleep medicine; and

4. To assess the use and impact of cognitive-behavioural insomnia symptom management strategies among elite athletes.

Addressing these research aims involved a range of methodological approaches and technologies. In order to provide appropriate context for the research reported here, governance procedures, methods and technologies for the whole programme are fully described in the next chapter.

Chapter 2

2. General methods used in this research programme

2.1 Introduction This chapter describes procedures, constructs and methods used across the present research programme. In presenting this information here the aim is to introduce clarity and avoid repetition. Additional detail, specific to the procedures used, are described within the methods section of the relevant chapters. All the work reported here was approved either by the Internal Review Board of the English Institute of Sport (EIS), or by the Ethics Approvals (Human Participants) Sub-Committee of Loughborough University. In both cases, issues of consent, withdrawal, privacy and confidentiality were observed in line with the declaration of Helsinki 1975 (revised 2013). For each individual study, participants were provided with an approved information sheet detailing the study design and requirements. Potential participants could also ask questions regarding the study before consenting to engage with the research.

2.2 Taxonomy of eliteness The construct of ‘eliteness’ in competitive sport is used throughout this programme. In each study eliteness was operationalised using Swann, Moran and Piggott’s (2015) taxonomy of ‘eliteness’ was employed. This approach utilises the formula shown below in Equation 2.1; terms are defined in Table 2.1. A modified taxonomy of’ eliteness’ was employed in some studies due to insufficient information to complete the full taxonomy, and this shown in Equation 2.2.

Table 2.1: Terms used in Swann, Moran and Piggott’s (2015) equation for a taxonomy of eliteness

Variable/score 1 2 3 4

National level; Involved in talent Regional level; university selected to International level; A. Athletes highest standard of development; 3rd tier level; semi-professional; 4th represent nation; top tier professional performance professional leagues or tier leagues or tours 2nd tier professional leagues or tours tours leagues or tours

Sustained success in major Success at regional, Infrequent success B. Success at the athlete's highest National titles or international, university, semi- at international level success at 2nd/ 3rd tier globally professional, or 3rd/ 4th tier level or top tier recognised

completion Withinsubject comparison C. Experience at the athlete's <2 years 2-5 years 5-8 years 8+ years highest level

Sport ranks outside top 10 Sport ranks 5-10 in Sport ranks top 5 in National sport; D. Competitiveness of sport in in county; small sporting country; small-medium country; medium- large sporting athlete's country nation sporting nation large sporting nation

Occasional Olympic Recent Olympic Regular Olympic sport; World sport with regular sport with frequent Not Olympic sport; world championships limited E. Global competitiveness of sport international major international championships limited to to a few counties; competition; semi- competition; few countries; limited limited international TV

Betweensports comparison global TV audience global TV audience national TV audience audience

Categories are divided into ‘within-participant’ and ‘between-sport’ comparisons (Table 2.1). There are three categories in within-subject comparisons (A, B and C) and two categories in between-sport comparisons (D and E). The classification of participants from the scores derived from Table 2.1 are: 1-4 = semi-elite; 4-8 = competitive elite; 8-12 = successful elite; and 12-16 = world-class elite. The full taxonomy was employed in studies presented in Chapter 5, Chapter 6 and Chapter 7.

Eliteness = [(A + B + C/2)/3] x [(D + E)/2]

Equation 2.1 Taxonomy of eliteness

For the studies reported in Chapter 1, Chapter 3 and Chapter 4 a modification of this taxonomy was employed. In the modified taxonomy two categories from the within-participant comparison were not included (e.g. B and C) and all categories from the between-sport comparisons were included. Due to the fewer categories being included in the modified taxonomy, only two classifications could be judged: >8 ‘= competitive elite’ and <8 = ‘semi-elite’.

Eliteness = A x ((D + E)/2)

Equation 2.2: Modified Taxonomy of eliteness

2.3 Newcastle Ottawa Scale (adapted from cross sectional studies) The Newcastle Ottawa Scale (adapted for cross sectional studies) (NOS) was used to appraise the quality of evidence in the introductory review. Table 2.2 describes the NOS, and all component items and Equation 2.3 shows how study quality scores are calculated. A ‘star rating’ system was used employed to appraise the quality of each study using three broad categories: 1) ‘selection’ of participants (4 items); 2) ‘comparability’ of participants (2 items); and 3) exposure or ‘outcome’ of interest for case–control or cohort studies (2 items). The total range of the NOS is 0-10 stars and categorises evidence quality as low (<5), moderate (5-7), and high (>7). This scale has been employed in systematic

reviews of elite athletes (Ashley et al., 2015) and has established content and inter-rater reliability (Luchini et al., 2017).

Study quality score = “selection” sub-total score + “comparability” sub-total score + “outcome” sub-total score

Equation 2.3 Newcastle Ottawa Scale scoring

Table 2.2 Newcastle Ottawa Scale (adapted for cross sectional studies) component (selection, comparability and outcome) items and ratings Selection Comparability Outcome items Score range Items Score range Items Score range

Control for most Representiveness of Assessment of 0-1 important factor 0-1 0-2 the sample outcome(s) (athletism)

Control for other Statistical test(s) Sample size 0-1 factors (Sex and 0-1 0-1 used age)

Non-respondents 0-1 reported

Description of elite 0-2 athletes

2.4 Psychometric sleep assessments The sleep assessments employed in the current research programme are widely used within sleep research and medicine and have shown to be valid and reliable across a range of populations. The following instruments were used:

2.4.1 Pittsburgh sleep quality index The Pittsburgh Sleep Quality Index (PSQI) is a 19-item self-rating questionnaire designed to assess sleep quality and disturbances over a one-month interval and includes seven equally weighted subscales or ‘components’ each delivering a ‘component score’ on a scale from 0-3. The PSQI components are subjective sleep quality; sleep onset latency; sleep duration; sleep efficiency; sleep disturbances; use of sleep medication; and day-time dysfunction. The sum of the component scores yields a global score, ranging from 0-21. Global scores >5 indicate poor sleep quality. This threshold has been shown to provide sensitive and specific measures of sleep quality and insomnia symptomatology (Buysse et al., 1989). Acceptable measures of internal consistency, reliability and validity for the PSQI have been reported across different sub-populations (See (Mollayeva et al., 2016) and the questionnaire has been previously used in elite athlete populations (see Chapter 1, Table 1.13) with adequate test-retest reliability (Samuels et al., 2016).

2.4.2 Ford Insomnia Response to Stress Test The Ford Insomnia Response to Stress Test (or FIRST) is a 9-item scale which assesses sleep reactivity (i.e. how an individual’s sleep responds when it is challenged or under stress). Scores range from 9-36 with a score of <16 indicating low sleep reactivity, ≥16 indicating high sleep reactivity, and >18 indicating very high sleep reactivity (C. Drake, Richardson, Roehrs, Scofield, & Roth, 2004; Kalmbach, Pillai, et al., 2016; Kalmbach, Vivekpillai, Toddarnedt, & Drake, 2016). The scale has high test-retest reliability (Gelaye et al., 2016; Marques, Allen Gomes, Drake, Roth, & de Azevedo, 2016), temporal stability (Jarrin, Chen, Ivers, & Morin, 2014) and has been validated as a predictive measure of both vulnerability to sleep disturbance (C. L. Drake et al., 2006; C. Drake et al., 2004) and insomnia pathologies (C. Harvey et al., 2014; Kalmbach, Pillai, et al., 2016) in non-clinical

populations. For the purposes of the present research, item 6 “after a bad at work” was replaced with “a bad day at training”.

2.4.3 Dysfunctional Beliefs and Attitudes about Sleep The 16 item Dysfunctional Beliefs and Attitudes about Sleep Scale (DBAS-16) was designed to appraise unhelpful cognitions about sleep (Charles M Morin, Vallières, & Ivers, 2007). By contributing to pre-sleep cognitive arousal, it is assumed that dysfunctional (i.e. erroneous) beliefs about sleep can extend sleep latencies and perpetuate insomnia symptoms. The scale consists of 16 statements clustered around five conceptually derived themes: (a) misconceptions about the causes of insomnia; (b) misattributions or amplifications of the consequences of insomnia; (c) unrealistic sleep expectations; (d) diminished perception of control and predictability of sleep; (e) faulty beliefs about sleep-promoting practices. Participants indicate their level of agreement to each statement on a visual analogue scale, scoring between 1 (‘strongly disagree’) and 10 (‘strongly agree’). An average total score is then reported, with high scores on the DBAS-16 being indicative of pronounced dysfunctional beliefs about sleep. The scale has been shown to have adequate temporal stability (r=0.82) and internal consistency (Chronbach-α= 0.79) (Charles M Morin et al., 2007). The scale has not been used previously in athletes, but has been employed in young adult populations, such as students (Yang, Chou, & Hsiao, 2011). For the purposes of the present research the word “insomnia” was changed to “poor sleep” as, outside of the clinical context, the term was judged to be inappropriate by athletes.

2.4.4 Pre-sleep Arousal Scale The Pre-Sleep Arousal Scale (PSAS) is a 16-item inventory of eight cognitive (e.g. intrusive thoughts) and eight somatic (e.g. sweating) symptoms of arousal experienced at bedtime. Ratings of the degree each symptom is experienced range from 1 (not at all) to 5 (extremely). A total score from 8 to 40 is computed for both subscales with higher scores indicating higher arousal. The scale has broadly shown adequate internal consistency and temporal stability (Jansson- Fröjmark & Norell-Clarke, 2012) . The scale has not been previously used in athletes, but again has been used in young adults (Exelmans & Van den Bulck, 2017). The cognitive symptoms of this scale have been previously shown to be

the mediating factor between stress and poor sleep quality (Robertson, Broomfield, & Espie, 2007).

2.5 Instrumental measures

2.5.1 Heart rate recovery run The Heart Rate Recovery (HRR) run was employed in chapter 6 only. The protocol for the assessment is part of the Great Britain Hockey’s fatigue monitoring programme. Whilst similar protocols have been employed in previous research, the test-retest reliability of this specific protocol is unknown. The protocol and an estimate of the test-retest reliability of the test is provided below.

2.5.1.1 Heart rate recovery run description The test was completed on an indoor running track and players ran at a speed of 2.2 m.s-1 for 5 minutes between 20 m markers. The speed of the run equated to stage 1 of the 30-15 test (Buchheit, Lefebvre, Laursen, & Ahmaidi, 2011) which players were familiar with through player profiling test batteries. All players wore a heart rate monitor (T31, Polar Electro, Kempele, Finland) and each test was supervised by the present author. Immediately following the 5-minute test, players recorded their HR (in beats/minute or bpm) and then were seated for an additional 1 minute and their HR was recorded after that period. The exercising heart rate for the 5-min period (HRex) and the drop in HR between the end of the exercising period and the end of 1 minute of recovery period (HRR) expressed as a percentage (HRR%) were used in analyses.

2.5.1.2 Heart rate recovery run test-retest reliability Test-retest reliability of the HRR run protocol was estimated using 13 hockey players of mixed levels of performance or ‘eliteness’. Six of the players were regional, male hockey players and seven were international, female hockey players. The international players were members of the Great Britain national squad and the regional hockey players were first team players for Henley Hockey Club. The HRR run was conducted on two occasions on the same day. Using the protocol described above one HRR run was performed by all players simultaneously, followed by another performance of the HRR run 10 minutes later. Due to the submaximal intensity of exercise, ten minutes was considered

sufficient for exercising HR to return to resting HR. The test-retest reliability of HRR run variables: HRex, HR at rest (HRrest), HRR, and HRR% was evaluated using calculations of technical error measurement (TEM), coefficient of variation (CV), intraclass correlation coeffcients (ICC) and Bland-Altman plots.

The CV, ICC can be seen in Table 2.3 and Bland-Altman plots can be seen in Figure 2.1. The CV was lowest, and ICC was highest for HRex; and the CV was highest and ICC lowest for HRR%. The Bland-Altman plots revealed HRex had the highest level of agreement (mean = 0 bpm; 95% CI = 13 bpm). However, HRR% had a poorer level of agreement (mean = 3 bpm; 95% CI = 17 bpm).

Table 2.3: Estimates of test-retest reliability of heart rate during exercise (HRex), heart rate during rest (HRrest), heart rate recovery (HRR), and heart rate recovery expressed as a percentage (HRR%) for 13 hockey players.

HRex HRrest HRR HRR% (%) Test-retest reliability (bpm) (bpm) (bpm)

CV (%) 3 5 14 13

ICC 0.84 0.80 0.50 0.63

Notes: CV, co-efficient of variation; ICC, intraclass correlation coefficient; HRex, Heart rate during exercise, HRrest, heart rate 1 minute post-exercise during recovery; HRR, heart rate recovery (the difference between HRex and HRrest)

A 15 B 20 10 10 5

0 0

125 135 145 155 165 70 90 110 130 HRrest2 HRrest2 (bpm)

HRex2 HRex2 (bpm) -5 -10 - - -10 -20

HRex1 HRex1 -15 HRrest1 HRrest1 -30 -20 Average HRrest (bpm) Average HRex (bpm)

C 30 D 15

20 10 5 10

HRR%2 HRR%2 0 -

HRR2 HRR2 (bpm) 0 20 25 30 35 40 (bpm) - -5 30 40 50 60

-10 -10 HRR%1 HRR%1 HRR1 -15 -20 Average HRR% (bpm) Average HRR (bpm)

Figure 2.1: Bland-Altman plots for 13 hockey players: A) heart rate during exercise (HRex); B) during rest immediately following the heart rate recovery (HRrest) run; C) the difference between HRex and HRrest (HRR); and D) the difference between HRex and HRR expressed as a percentage (HRR%)

Considering the findings above, whilst no variable is considered to have excellent test-retest reliability the HRRex was judged to be the most reliable and therefore was adopted as a marker of physiological fatigue in the present research.

2.5.2 Wrist actigraphy Serial measurements of 24 hour sleep-wake patterns were obtained using wrist actigraphy. Two different actigraphic devices were used: 1) the Motionwatch 8 (Camntech, Cambridge, UK); and 2) the Actiwatch 2 (Philips Resperonics, USA). The use of 2 the different devices reflected availability. The protocol employed in the current research, irrespective of the device used, is described below, along with agreement analyses between the two models of actigraphy. The watch specifications can be seen in Table 2.3 and images of the watches can be seen in Figure 2.2.

2.5.2.1 Actigraphy protocol Participants were informed of the device features (i.e. light sensor and accelerometer) and were instructed: to wear the device on their non-dominant wrist for the recording period; and to remove the watch only for training, matches, showering and bathing. Participants were instructed to press the event marker button each night when they started ‘trying to sleep’; and again, each morning when they woke up. Epoch length was set at 1 minute. Following the recording period, activity counts and light intensity data were downloaded and analysed using specialist software. All participants were required to keep a sleep diary (see Appendix 9.2) during recording periods. Combining sleep diary and actigraphy data, the following sleep variables were calculated: time in bed (TIB); sleep onset latency (SOL); total sleep time (TST); and sleep efficiency (SE).

A B

Figure 2.2: Images of: A) the Actiwatch 2 (Philips Resperonics, USA); and B) the Motionwatch 8 (Camntech, Cambridge, UK). Both devices are fitted with a photometer and an event marker.

Table 2.4: Specifications for Actiwatch 2 (AW2) and Motionwatch 8 (MW8) Specification AW2 MW8

Accelerometer Piezo-electric MEMs

Accelerometer range (G) 0.5-2 0.01-8.0

Software ActiWare MotionWare

(Version 6.0.7) (Version 1.0.27)

Sampling rate (Hz) 32 11

Size – L x W x D (mm) 43 x 23 x 10 36 x 28 x 9

Validated threshold Medium (40) Low (20)

Weight (grams) 16 9

Illuminance range (lux) 0-100,000 0-64,000

Notes: MEMs = ; L = length; W = Width; H = Height.

2.5.2.2 Levels of agreement To judge the level of agreement between the two actigraph devices, 10 non- athlete, young adults (5 females) were recruited to take part in a comparative study. Participants were provided with both models of actiwatch and were instructed to wear each watch on the same (non-dominant) wrist for 3 consecutive nights. Participants were asked to press the event markers on both watches simultaneously on settling down to sleep, and on awakening. Overall, then, 30 nights of sleep data were collected. Actigraphy data were scored using each model’s respective software. Thresholds to score sleep were selected according to manufacturer recommendations (e.g. Actiwatch 2 = ‘Medium’ or 40 counts, and Motionwatch 8 = ’low’ or 20 counts). Concordance between the devices was estimated for TST (minutes), SE (%), and SOL (minutes) using calculations of means, intraclass correlation coefficients, and Bland-Altman plots. A paired student’s t-test was conducted to assess for differences in mean values.

The agreement between devices can be seen in Figure 2.3. Overall, agreement between the two devices was good, with a mean bias of 1% and a 4% 95% CI calculated for SE, and 0 h mean bias and 0.3 h 95% CI between the two actiwatch modelss. The level of agreement for SOL was moderate, with a -2 min mean bias and a 6 min 95% CI. There were no significant differences between devices for mean SE (t(12)=1.15; p=0.288) and TST (t(12)=0.39; p=0.99); however there was a significant difference between devices for SOL (t(12)=2.5; p=0.04). Intraclass correlation coefficients revealed strong, moderate, and weak correlations between TST, SE and SOL respectively.

Several limitations should be acknowledged in the study. Firstly, the sample size was small and therefore it is likely that the study was underpowered, and findings should be considered with caution. Secondly, measurements of WASO were not included in the agreement analyses and therefore the level of agreement between models remains unknown. In relation to this, the participants were also not screened for sleep disorders and therefore the sample may not have been homogenous which may have affected agreement scores, with sleep disordered participants likely to report higher levels of WASO. However, SEs reported in the study were above 85% for all participants, and SOLs were low and therefore it is likely that levels of WASO were not generally high (e.g. <30 minutes)

Table 2.5: Agreement of sleep onset latency, sleep efficiency, and total sleep time between Actiwatch 2 and Motionwatch 8 actigraphy devices

AW2 MW8 ICC

a a a a a a b b b SE (%) SOL (min) TST (min) SE (%) SOL (min) TST (min) SE SOL TST

90 (2) 4 (3)* 375 (11) 89 (2) 1 (2) 375 (10) 0.59 0.29 0.99

Notes: SE Sleep efficiency; SOL, sleep onset latency; TST, total sleep time; AW2, Actiwatch 2; MW8 Motionwatch 8; ICC, Intraclass correlation coefficient; a mean (SD) values; b coefficient values; * significantly different to MW8 (P<0.05)

A B 5 4

4 2 3 0 2 0 2 4 6 8 -2

SEMW8 SEMW8 (%) 1 SOLMW8 (min)

- 0 - -4 85 87 89 91 93 95

-1 -6 SEAW2 SEAW2 -2 SOLAW2 SOLAW2 -8 -3 -10 -4 Average SOL (min) Average SE (%)

C 0.3

0.2 0.1

TSTMW8 (h) 0.0 - 4 5 6 7 8 -0.1

-0.2 TSTAW2 -0.3 Average TST (h)

Figure 2.3: Bland-Altman plots for A) sleep efficiency (SE); B) sleep onset latency (SOL); C) total sleep time (TST) between Actiwatch 2 (AW2) and Motionwatch 8 (MW8). Grey line = mean bias; dashed line(s) = 95% confidence intervals

2.6 The Athlete Sleep Management Programme For the intervention and monitoring studies reported in Chapters 6 and 7 a programme of sleep management was designed, employed, and evaluated. The Athlete Sleep Management Programme (ASMP) was derived from materials developed by Morgan et al. (2012) for a self-help cognitive behavioural therapy programme for insomnia symptoms. Cognitive behavioural therapy for insomnia offers a collection of separate interventions which target sleep-related behaviour patterns and habitual styles of thinking. When used singly or in various combinations (see Morgenthaler et al., (2006)), these CBT-I components have shown significant efficacy in reducing the frequency and severity of insomnia symptoms (see van Straten et al., (2017)). In line with this latter research, a recent meta-analysis has shown that interventions targeting cognitive and behavioural self-regulation can also improve sleep quality in adults without a sleep disorder (Murawski, Wade, Plotnikoff, Lubans, & Duncan, 2017). The evidence strongly suggests that the principles of CBT-I could usefully be applied in elite sport to mitigate sleep disturbances before competition (Juliff et al., 2015) or following a night match (Juliff, Peiffer, & Halson, 2017), targeting common mechanisms that disturb sleep during these scenarios (e.g. learning, and cognitive and physiological arousal). More recently, given the high levels of attrition typically attributed to lengthy (i.e. 6-8 sessions) CBT-I programmes (Ellis, Cushing, & Germain, 2015), and the challenge to changing human health-related behaviour (Kelly & Barker, 2016), research has been conducted into the efficacy and feasibility of alternative modes of delivery for CBT-I formats, such as group sessions (Koffel, Koffel, & Gehrman, 2015) internet-based programmes (Ye et al., 2016), self-help booklets (Morgan et al. 2012), mobile phone applications (E. Koffel et al., 2018), and ‘single-shot’ (one-off) CBT-I sessions (Ellis et al., 2015). However, such alternative formats have not previously been explored in elite sport.

The design, content, delivery, and evaluation of the self-help sleep management material used in Chapters 6 and 7 is described below. Table 2.5 provides a summary of each of the ‘workshops’ during which the material was introduced and explained to the athletes. Throughout this description the sleep management programme is referred to as the ASMP.

2.6.1 Programme development The development of the modified CBT-I programme for athletes was influenced by 3 considerations:

1. To fit-in with existing training and competition schedules, the ASMP needed to be delivered in group (team) sessions, allowing opportunities for questions and discussion;

2. The format needed to be rapidly accessible to young athletes for whom sleep management may not be perceived as a key element of their performance;

3. All messages must be practical and actionable within the demands of elite sport participation.

The resulting programme aimed to both provide athletes with cognitive- behavioural strategies to mitigate the impact of sleep challenges and increase levels of athlete confidence in their ability to manage and cope with sleep disturbances.

2.6.2 Programme design The source material for the ASMP was provided by the programme successfully trialled by Morgan et al (2012). Each element of the original content was first considered in relation to the points listed above (paragraph 2.4.1) before being converted into a format suitable text for delivery in a workshop. The resulting ASMP comprised key messages drawn from the host CBT-I programme, and from sleep science. These messages were then translated into four simplified infographics (reproduced in Appendix 9.1.1, 9.1.2, 9.1.3. and 9.1.4) covering the following:

1. Introduction: sleep in elite sport; influence of homeostasis, circadian rhythms and automaticity;

2. Key elements of sleep hygiene and stimulus control;

3. Managing cognitive arousal: simple cognitive strategies

4. Managing and scheduling naps

Each infographic formed the basis of a single component of the ASMP. Components, supported by infographics, could be delivered in separate workshops, or combined in a single workshop. In the studies presented here athletes received two (Chapter 7) and four (Chapter 6) workshops, with one workshop taking place every week.

The content of each component, and the evidence that underpins their inclusion, can be seen in the section below. With less severe symptoms of insomnia being shown as a predictor of attrition to CBT-I programmes (Yeung, Chung, Ho, & Ho, 2015), an abbreviated version of CBT-I (with supporting materials) was seen as a suitable model for providing athletes with tools to manage insomnia symptoms (Edinger & Sampson, 2003).

2.6.3 Programme content

2.6.3.1 Component 1 The first component consisted of an introduction to sleep, focussing mainly on ‘grounding’ the athlete’s understanding of sleep in evidence from sport science and behavioural sleep medicine. Information covered chronotypes, typical sleep durations, the impact of elite sport participation on sleep, sleep cycles, and the (homeostatic, circadian, and psychological) control of sleep onset. The aim was to address any dysfunctional beliefs attitudes and beliefs about sleep and, consequently, reduce unhelpful levels of pre-sleep cognitive arousal. Such an approach has been shown to shown to enhance treatment efficacy and increase participants’ control over sleep (Morin, Blais, & Savard, 2002). In the studies described in Chapters 6 and 7, DBAS responses (from all participants) were used to highlight athlete cognitions and inform the content of that workshop ( Morin et al., 2007). When presenting this component in the workshops athletes were given the opportunity to raise areas of sleep which they would like to explore in more detail, and areas which they believe most impacted their wellbeing and performance.

2.6.3.2 Component 2 The second component focussed on elements of sleep hygiene and stimulus control. Sleep hygiene (SH) refers to a list of behaviors, environmental conditions,

and other sleep-related factors that can be adjusted as a stand-alone treatment or component of multimodal CBT-I treatment (Stepanski & Wyatt, 2003). Although the effect of SH guidelines on sleep outcomes within both the general population (See Irish et al., (2015)) and among athletes (Fullagar et al. 2016; Duffield et al. 2014) has been shown to be inconclusive, poor sleep hygiene has been associated with the perpetuation of insomnia symptoms in athletes (Knufinke, Nieuwenhuys, Geurts, Coenen, & Kompier, 2017). In addition, the inclusion of SH as a practical element in psychoeducational sleep programs, particularly for individuals who do not meet clinical insomnia criteria but exhibit insomnia symptomatology, has been found useful (See Irish et al., (2015)).

Stimulus control, on the other hand, employs operant conditioning to strengthen associations between the bed/bed-room environment and sleep onset; and classical conditioning to extinguish associations between the bed and wakefulness, (Bootzin & Nicassio, 1978). Stimulus control has been shown to reduce the frequency and duration of nocturnal awakenings in individuals with insomnia (Morin & Azrin, 1987) and forms a key component of CBT-I programs (See Morgenthaler et al., (2006)). Given the relative simplicity of stimulus control guidelines, and minimal assistance required to employ stimulus control procedures, this element of CBT-I was considered important for inclusion.

2.6.3.3 Component 3 The third component focussed on pre-sleep cognitive arousal, explaining the role of arousal in delaying sleep onset, and the operation of the ‘vicious cycle of insomnia’ (where arousal results in delayed sleep onset, concerns about which contribute to higher levels of arousal). Two techniques to manage cognitive and physiological arousal during periods of both training and competition were presented: 1) articulatory suppression; and 2) progressive muscle relaxation (PMR).

Articulatory suppression is derived from Baddeley and Hitch’s model of working memory (Baddeley & Hitch, 1974), and Neisser’s (1967) early theory of attention, whereby engaging in a cognitive task, which is low in attentional demand and embedded in inner speech, can interfere with cognitive processes (See (Levey, Aldaz, Watts, & Coyle, 1991)). As a result, engaging in an articulatory task can interrupt and block unwanted pre-sleep cognitive activity (Harvey, 2002) and

may improve problems in relation to sleep onset latency. Although this technique has not been validated as a single treatment component for insomnia (Harvey, 2002) it has been shown effective in reducing sleep latency in case studies (Levey et al., 1991) and has been incorporated into widely used multi-component models of CBT-I (Espie et al., 2012). Again, the simplicity and minimal guidance required to perform the technique provided a suitable rationale to include the component in the current programme.

Progressive muscle relaxation (PMR) is a technique which operates using proprioceptive feedback from skeletal muscles (Jacobson, 1938, 1964) through the volitional contraction and relaxation of specific skeletal muscle groups. Over time, participants learn to exert an element of control over the relaxation of skeletal muscle groups which can induce physiological but also cognitive relaxation. PMR is considered an adjunct therapy to CBT-I and is commonly used in the treatment of insomnia (See Morgenthaler et al., (2006)).The use of skeletal muscles in this technique make it especially relevant for athletes, and this is supported by a recent study whereby PMR was shown to be feasible and accepted by male football players (Sharifah Maimunah & Hashim, 2016).

2.6.3.4 Component 4 Finally, the fourth component focussed on napping and prophylactic sleep extension (or ‘sleep banking’). Although napping is regarded as a maladaptive response to insomnia symptoms, it can also be effective in managing the short- term symptoms of mild sleep deprivation, such as sleepiness and fatigue, in circumstances where bi-phasic sleep patterns do not present a problem. As was clear from the introductory review, napping appears as an embedded practice in elite sport culture. As a result, Component 4 directly considered the optimal timing and duration of naps, and how to plan napping environments. The dangers of excessive napping were also considered. Prophylactic sleep extension was recommended to athletes prior to known periods of disturbance. Sleep “banking” is described as optimising sleep prior to a known period of disturbance. Previous research has shown sleep banking can minimise the impact of sleep deprivation on cognitive function and enhance recovery of cognitive function to baseline levels (Rupp, Wesensten, Bliese, & Balkin, 2009). In a recent study, sleep banking was shown to reduced perception of effort during

cycling following an acute sleep deprivation period (Arnal et al., 2016). Behavioural methodologies to extend sleep, either by day-time or night-time sleep opportunities were presented and discussed.

2.6.4 Workshop materials Following each workshop written support material was made accessible in the form of infographics (See Appendix 9.2.1). All infographics were generated using a specialist online software package (Piktochart, Malaysia), and could be accessed by mobile phones to reinforce the messages presented during the workshops. Following the third Component (which included Progressive Muscle Relaxation), an MP3 recording of PMR instructions from the ‘Sleepful’ online self- help CBT-I programme was made available to download (Sleepful, Version: 2.1.31, Cuttlefish Multimedia Ltd, UK, the programme can be accessed at: https://play.google.com/store/apps/details?id=com.cuttlefish.sleepful&hl=en_B ). Athletes were encouraged to first practice the technique (using the MP3 recording) outside of the bedroom, adopting the technique as a pre-sleep strategy once it had been mastered.

2.6.5 Programme evaluation An online survey (BOS, Bristol University, UK) was presented at the end of both workshop series (described in Chapters 6 and 7) asking athletes to evaluate the content and to what extent they adopted the sleep behaviours recommended. Athletes answered these questions on a 5-point Likert scale ranging from “no, definitely not” to “yes, absolutely”. Questions in regard to overall programme content included:

1. “Did you find the material relevant?”; 2. “Do you feel equipped to deal with sleep disturbances when they arise?”; 3. “Would you recommend this workshop series to a fellow athlete?”; 4. “Have you previously received information about how to manage sleep disturbance during normal training (or) competitions?”; and 5. “Do you think you will use the information provided in the future?”

Two separate closed questions were asked around the adoption of sleep related behaviours recommended These questions were:

1. Sleep promoting activities are activities that are performed before (or at) bed-time to help promote feelings of sleepiness in preparation for sleep, do you regularly perform a sleep promoting activity before bed? Please provide details of what this activity(s) is.

2. Do you use any techniques to help you fall asleep at night (including when you wake-up in the night and cannot fall back asleep)? Please described this technique? And;

3. Do you feel this technique (to help you fall asleep) is effective?

Table 2.6: Overview of sleep management programme content

Component Aims Techniques Support

1 1. To Introduce athletes to the ASMP and provide key messages Cognitive reconstruction Infographic on the regulation of sleep Sleep science education 2. To address attitudes and beliefs held towards sleep by the participant group

2 1. To explain concepts of sleep hygiene and stimulus control, and Sleep hygiene Infographic the role they play in sleep management in sport Stimulus control 2. To suggest behaviours related to these concepts that will help facilitate sleep onset and maintenance

3 1. To introduce pre-sleep arousal in the context of elite sport Articulatory suppression Infographic MP3 of PMR 2. To introduce techniques to offset and control pre-sleep arousal Progressive muscle relaxation instructions

4 1. To discuss sleep duration in the context of sleep quality Napping Infographic

2. To introduce how intently manipulating sleep duration can be Sleep extension employed as a technique to manage sleep

3. To introduce behaviours to facilitate sleep duration manipulation

Chapter 3

3. A comparison of between-sport differences in sleep quality and insomnia symptomatology among elite British athletes

3.1 Introduction As shown in the introductory review (Chapter 1), the prevalence of poor sleep quality and insomnia symptomatology has been reported to be high among elite athletes, and as characterised by regular episodes of sleep dissatisfaction (Rodrigues et al., 2015), unrefreshing sleep (Tuomilehto et al., 2016), long sleep latencies (Schaal et al., 2011; Tuomilehto et al., 2016), and high levels of day-time fatigue (Sargent, Lastella, et al., 2014). Studies reporting global scores for the Pittsburgh Sleep Quality Index (PSQI) have found 43-57% of elite athletes to score above the ‘poor sleeper’ threshold of >5 (Samuels, 2008; Durán et al., 2015; Tsunoda, 2015; Swinbourne et al., 2016), while 22-26% scoring above the more conservative (and more severe) threshold of 8 (Samuels, 2008; Swinbourne et al., 2016). Studies have also reported incidences of more clinical sleep problems, close to insomnia diagnosis, to be as high as 22% in elite athletes (Schaal et al., 2011; Tuomilehto et al., 2016).

It has been estimated that, in a typical sport career, an elite athlete is exposed to over 600 distinct psychosocial, physical and organisational stressors (Arnold & Fletcher, 2012). Again, as shown in the introductory review, sleep stressors commonly arise in training, competition, and international travel. More specifically: the heightened frequency, intensity and volume of training (Schaal et al., 2015; Thornton et al., 2016), particularly during training camps (Pitchford et al., 2016; Thornton et al., 2016); the scheduling of training sessions (Sargent, Halson, et al., 2014; Sargent, Lastella, et al., 2014) and competitions (Fullagar, Skorski, Duffield, Julian, et al., 2016; Sargent & Roach, 2016); pre-competition anxiety (Erlacher et al., 2011; Juliff et al., 2015); and the regular relocation demanded by national (Richmond et al., 2004; Richmond et al., 2007) and international competition (Lastella et al., 2014; Fowler et al., 2015, 2016; Fullagar

et al., 2016b) have all been implicated as factors contributing to sleep disturbance.

However, while the presence of sleep disturbance in elite sport is evidenced by the available literature, the dynamics of such disturbance (in terms of both the risk profiles of athletes and the sleep challenge posed by a sport) is under- researched.

3.1.2 The ‘3-factor’ model of insomnia development The ‘3-factor’ model proposed by Spielman, Caruso and Glovinsky (1987) provides a useful framework for understanding the interplay of individual and situational processes in the aetiology of insomnia. According to this model acute insomnia symptoms become chronic through the interaction of: i) predisposing factors (inherent psychological vulnerability characterised by higher levels of trait anxiety, a susceptibility to cognitive intrusions, and attentional bias); ii) precipitating factors (sleep disturbing physical, psychological or situational events); and iii) perpetuating factors (maladaptive behavioural responses to sleep disturbance which, over time, help to maintain insomnia as a chronic problem). Given the specificity of athlete selection processes (both psychological and physiological) for different sports, and given also the very different training ‘cultures’ which exist across elite sport (in terms of frequency, duration, intensity and scheduling of training), it is reasonable to expect athlete risk profiles (i.e. the degree of ‘predisposition’ in the Spielman and colleagues (Spielman et al., 1987) model, and the challenge to sleep (i.e. the strength of sport-related ‘precipitations’) to vary across different sports. This view is consistent with some limited findings in the sport science and medicine literature.

In a representative sample of elite French athletes, for example, Schaal et al. (2011) reported lifetime prevalence rates of sleep problems, closely related to insomnia diagnosis, to be higher in ‘aesthetic’ (i.e. gymnastics, and synchronised swimming) and ‘contact and combat’ sports than in all other sport types. Between sport differences in actigraphic sleep efficiency (SE), sleep onset latency (SOL) and total sleep time (TST) were also reported by Leeder et al. (2012), with results (see Chapter 1, Table 1.8 and 1.9) most likely reflecting the impact of training schedules. Similarly, Lastella et al. (2015a), using measures of actigraphy-derived sleep outcomes from Australian athletes (SE and SOL) found

that individual sports (cycling, swimming and triathlon) exhibited lower SEs and TSTs, but also longer SOLs and higher wake time after sleep onsets (WASO) when compared with team sports (basketball, rugby and football). However, neither of these latter two studies reported data on the athlete’s experience of sleep quality.

During periods when sleep is expected to be challenged, such as competition, again differences have been observed between individual and team sport athletes. Among elite German athletes Erlacher et al. (2011) reported higher levels of pre-competition sleep difficulties in individual sports when compared with team sports using the Competitive Sport, Sleep and Dreams questionnaire (CSSDQ). Extending these comparisons, athlete age has been identified as a risk factor for sleep disturbances before a major competition (Erlacher et al., 2011; Juliff et al., 2015), though gender and global sleep quality (as assessed by the PSQI), have not (Erlacher et al., 2011; Juliff et al., 2015). Differences in the extent to which age influenced sleep before a major competition was also reported by Juliff, Halson and Peiffer (2015), with individual athletes exhibiting higher levels of disturbance with increasing age, while increasing age among team athletes was associated with improved pre- competition sleep quality.

3.1.3 Between sport differences Cross-sport differences have also been reported in sleep-related coping practices. Sargent et al. (2014) reported a higher incidence of napping among individual compared to team sport athletes, but no differences in nap duration. Similarly, in relation to night time sleep Juliff, Halson and Peiffer (2015) reported a greater number of individual athletes to have a strategy (63%) to help them sleep well the night before a major competition, compared to team sport athletes (41%). Collectively, these studies indicate higher levels of sleep management among individual athletes.

Overall, few studies have reported the prevalence of poor sleep quality and insomnia symptomatology between sport-types among elite athlete populations. The majority of research investigating differences in sleep among sport-types has employed broad categorisations of sports e.g. individual vs. team (Erlacher et al., 2011; Juliff, Halson and Peiffer, 2015; Lastella et al., 2015). It is likely, therefore, that

these broad categories will contain sports that exhibit very different training and competition demands, and very different athlete profiles (age, gender, athleticism, etc.). As a result, other than Schaal et al.’s (2011) study of psychological symptoms in elite French athletes, no study has assessed differences in sleep complaints between categorisations of sport-types, whilst considering sport training and competition demands. In addition to having a practical value in training, coaching and the management of athlete wellbeing, this information could also contribute to the development of theory and understanding in sleep-sport relationships. From the perspective of ‘precipitating factors’ for insomnia symptoms (Spielman et al., 1987), exploring between-sport differences in sleep quality could help to identify those elements of sports performance and culture most likely to impact sleep outcomes. However, from the perspective of ‘predisposing factors’, between-sport comparisons could also inform our understanding of how the selection processes adopted by a given sport might differentially influence the insomnia risk profiles (i.e. the behavioural phenotypes) of elite athletes (in that sport).

3.1.4 Research aims and hypotheses Having identified between-sport differences in sleep quality as an important but under-researched area (Chapter 1) this study set out to explore sleep quality and insomnia symptomatology between sport-types among a representative sample of elite British athletes. The primary aims of the study were:

1. To assess levels of sleep quality and insomnia symptomatology across a broad range of sport-types using a standard sleep quality instrument (the PSQI) supported by information on sleep-related attributions, attitudes and experience; 2. To assess and compare the subjective quality of athlete sleep in relation to the subjectively reported structure of athlete sleep; and identified elite sport risk-factors; 3. To assess and compare levels of sleep management adopted by elite athletes from sport-types, and associations with sleep quality

to test the null hypotheses that:

1. There will be no significant differences in sleep quality (as measured by PSQI) between sport types 2. In common with findings in the general population, cross-sport variations in sleep quality will not be significantly influenced by age and gender. 3. Sleep management adopted by elite athletes will not significantly differ between sport-types, and will not be driven by sleep quality

3.2 Methodology The study was designed as an online survey targeting National Lottery funded British athletes and mounted on the Bristol Online Surveys platform (University of Bristol, UK). Approval was provided by the Ethics Approvals (Human Participants) Sub-Committee of Loughborough University. Over a six-month period (September 2014 – February 2015), athletes were invited to open and complete the survey online, either in a single session or through multiple log-ons. Information about, and electronic links to the survey were distributed through sport science practitioners from the English Institute of Sport or respective national governing bodies. All respondents were required to give consent, and full information on the withdrawal, privacy and confidentiality were provided in the questionnaire’s landing page (see Appendix 9.3). The survey was broken down into four main sections: 1) athlete demographics and training characteristics; 2) sleep quality and insomnia symptomatology; 3) stress induced sleep disturbances; and 4) sleep education and management.

3.2.1 Athlete demographics and training characteristics Demographic information included athlete age, gender, and sport. Sports were categorised into five sport-types: team; endurance; technical; combat; and speed/power sports (See Tables 3.1 and 3.2). No athletes that represented aesthetic sports (e.g. gymnastics and synchronised swimming) took part in the survey (see Table 3.1). Athlete training characteristics consisted of training

volume (e.g. hours per week); training day start and end times; and number of major competitions per year.

3.2.2 Sleep quality and insomnia symptomatology Sleep quality was assessed using the PSQI (see Chapter 2, Para. 2.3.1). Acceptable measures of internal consistency, reliability and validity for the PSQI have been reported across different sub-populations (See (Mollayeva et al., 2016) and the questionnaire has been previously used in elite athlete populations (Chapter 1, Table 1.13) with adequate test-retest reliability (Samuels et al., 2016). Additional closed-ended (i.e. yes or no) and 5-point Likert scale questions in relation to other aspects of insomnia symptomatology which are not captured in the PSQI were included in this section:

1. “Do you usually feel refreshed on awakening?”; 2. “Do you feel you get enough sleep?”; 3. “In the past 12 months do you believe a period of disturbed sleep has led to an episode of poor performance?”; and 4. “In the past 12 months, how often have you experienced general fatigue to the extent that it has resulted in you missing, or performing modified training?”

Table 3.1: Sport categorisations used in the EIS online sleep survey: Sport (n of athletes)

Endurance Team Technical Combat Power

Rowing (n=33) Hockey (n=56) Archery (n=11) Boxing (n=22) Athleticsb (n=16)

Para-rowing (n=8) Rugby sevens (n=31) Para-archery (n=10) Taekwondo (n=11) Para-athleticsb (n=15)

Cycling (n=7) Wheelchair rugby (n=12) Shooting (n=17) Sprint cycling (n=13)

Para-cycling (n=7) Wheelchair basketball (n=24) Para-shooting (n=16) Short track speed skating (n=10)

a Athletics (n=8) Boccia (n=11) Bob skeleton (n=5)

Para-athleticsa (n=6) Disability table tennis (n=10)

Triathlon (n=7)

Para-triathlon (n=10)

Sailing (n=22)

Para-sailing (n=8)

Notes: a e.g. long-middle distance running; b e.g. sprints and throws

3.2.3 Stress induced sleep disturbances Questions detailing risk factors for stress induced sleep disturbances identified in Chapter 1 (see Para. 1.3.5) were included in this section. Using 5-point Likert scales, two questions asked participants to state how strongly they agreed with the following statements: “My sleep is more disturbed the night before a major competition compared to my normal sleep”; and “I sleep worse than usual after I have competed in the evening/ night”. Two more 5-point Likert scale questions were used in relation to travel and training camps: “Following long haul travel, how often is your sleep disturbed?” and “Do you sleep better or worse in a training camp environment?” One 3-point Likert scale question was used to see whether participation in elite sport has degraded sleep quality: “Since reaching your current level of performance how has your sleep changed?”

3.2.4 Sleep management Napping during the day was addressed by two 5-point Likert scale questions on the frequency of napping, and ease of falling asleep: “How often do you nap in the day-time?”; and “How easy do you find it to fall asleep?” Nappers were defined as those who napped ‘sometimes’, ‘often’ or ‘all the time’, while habitual nappers were defined as those who napped ’often’ or ‘all the time’ (Evans, Cook, Cohen, Orne, & Orne, 1977). A single closed-ended question explored athlete sleep education: “have you ever received any form of education on sleep in your sport?”. Finally, a single item addressed personal (night-time) sleep management: “Do you have a personal technique(s) to help you fall asleep when you are unable to do so?”

3.2.5 The sample A total of 412 (183 female) elite British athletes (mean age 26.5, SD = 0.7 y) completed the survey. Athletes were drawn from 25 Olympic and Paralympic (139 athletes) sports (See Table 3.1), representing 66% of all funded sports for the Rio 2016 Olympic and Paralympic cycle. All athletes were at an international level of performance, with 80% having competed at an Olympic or Paralympic Games, Commonwealth games or senior world championships and could be regarded as ‘competitive elite’ according to Swann et al.’s (2015) taxonomy of ‘eliteness’. Athletes’ mean training volume was 18.5 (SD = 7.5) h/week and had competed in 4 (SD = 3) major competitions per year. At the time of completing

the survey 87% (n=359) of athletes were in-season, and 13% were off-season (n=53). Among the participants that were in-season, 75% (of the full sample, n=301) of athletes were in “training”, and 10% (n=40) were in a “competitive” phase; and 5% (n=20) were “injured or had a long term illness”.

3.2.6 Data analysis One-way ANOVAs were employed to assess mean PSQI global scores, and PSQI derived TST, SE and SOL across sport categories. Separate ANCOVAs were employed, using 1) age, gender and 2) age, gender and ability (Paralympic/ Olympic status) as co-variates. Bivariate associations between categorical variables were assessed using Pearson’s chi-square statistic. Relative risk ratios for global scores of >5 and >8 on PSQI were also derived from contingency tables, where between sport (e.g. sport types), within sport (e.g. gender, Paralympic/ Olympic status, age) and elite sport stressors (e.g. sleep disturbance night before competition) were entered as risk factors. A Cronbach alpha coefficient were calculated for the PSQI to assess the metric’s internal consistency reliability in an elite athlete population. Missing data were replaced using the SPSS subroutine for mean imputation. For all statistical inferences, alpha was set at p<0.05. Analyses were performed using SPSS 23 for Windows (Version 23.0, SPSS inc., IBM, Armonk, New York).

3.3 Results Sample characteristics are shown in Table 3.2. Overall, six athletes could not be categorised into a sport-type (2%) due to not providing the name of the sport within which they participate and were therefore excluded from the analyses. Missing data analysis of all other questions revealed 5% incomplete data, and therefore data imputation was minimal across the entire survey. The Cronbach alpha coefficient of the PSQI was 0.7. There was no significant difference

(X2(1)=0.187, p=0.666) in PSQI score distributions during in- (n=359) and off-season phases (n=53). Moreover, among those participants who were ‘in-season’ (n=359), PSQI global score distributions were not significantly different between ‘training’ (n=299), ‘competition’ (n=40) and ‘injured/ illness’ groups (n=20)

(X2(3)=2.8, p=0.416). As a result, participants in all season phases were retained for analyses.

3.3.1 Sleep quality and insomnia symptomatology Global PSQI scores between sport-types can be seen in Figure 3.1. Prevalence and relative risk (RR) ratios of scoring >5 and >8 on PSQI for risk factors listed above are presented in Table 3.3, and TST, SOL, SE and insomnia symptoms between sport-types can be seen in Table 3.5. The global mean PSQI score was 6 (SD=3) and ranged from 1-17. The prevalence of participants (n=412) scoring >5 was 45% (n=186) and those scoring >8 was 16% (n=66). In the previous 12 months, 45% (n=186) of athletes reported that a period of poor sleep had led to an episode of poor athletic performance, however this distribution was similar between sport-types (X2(4)=7.215; p=0.125). Using age and gender as co-variates, there was a significant difference between sport types (F(7,398)=5.35; p<0.001). The addition of ability to the ANCOVA model revealed a significant difference between sport-types (F(6,399)=5.181; p<0.001). However, post-hoc analyses revealed no significant differences (p>0.05).

* 12

PSQI global Score 4

0 Team Sport Endurance TechnicalSkill-based Combat Speed/ power

Figure 3.1: Mean Pittsburgh Sleep Quality (PSQI) Index global scores across sport types.

* significantly. different to speed/ power, p<0.01. Sport type bars and error bars are expressed as means and standard deviations respectively.

3.3.2 Stress-related sleep disturbances Incidence rates of stress related sleep disturbances in response to different stressors within the elite sport environment (e.g. competition, travel and training) can be seen in Table 3.6, while the relative risk of athletes reporting poor sleep in response to each stressor can be seen in Table 3.4. There was a trend (X2(4)=3.687, p=0.055) for athletes who experienced sleep disturbances following long haul travel to score >5 on PSQI, with a relative risk ratio of 1.290 (95% CI = 0.995 to 1.673). PSQI scores in athletes who believe their sleep has deteriorated, “remained unchanged” or “improved” since reaching their current level of performance is shown in Figure 3.2. From the overall sample, 15% (n=62) stated their sleep has ‘deteriorated’ since reaching their current level of performance, with 62% (n=257) of athletes’ sleep “remaining unchanged” and 23% (n=94) stating it has improved. There was no significant difference in distributions of how athletes’ sleep had changed since reaching their current level of performance and sport-type (X2(8)=11.727, p=0.164).

3.3.3 Sleep management The prevalence of the two ‘sleep management’ practices included in the survey (daytime napping and the employment of personal techniques to help falling asleep) are shown in Table 3.6. From the overall sample, 43% (n=176) reported daytime naps “always/often/sometimes” and were categorised as ‘nappers’. Of these, 20% (n=81) were categorised as habitual nappers. There was no significant difference in PSQI global scores (F(1,411)=0.179; p=0.672), SE (F(1,411)=1.765; p=0.185), and TST (F(1,411)=2.61; p=0.107), between habitual nappers and non- nappers (PSQI; 6, SD=3 vs. 6 SD=3; SE; 88, SD=9 vs. 86, SD=10; TST, 7.8, SD=0.9 vs. 7.6, SD=1.1). Proportions of athletes reporting not getting enough sleep were not significantly different between habitual nappers (40%, n=32) and non-nappers

(44%, n=104) (X2(1)=0.48, p=0.49). Among the nappers, 56% (n=99) reported finding falling asleep when they napped ‘very easy’ or ‘easy’, and 43% (n=76) found it ‘difficult’ or ‘neither easy nor difficult’. One athlete reported finding it ‘very difficult’ to fall asleep. There were no differences in distributions of nappers who found falling asleep in the day-time ‘easy’ or ‘very easy’ between sport- types (X2(4)=2.085, p=0.720). From the overall athlete sample (n=412), 66% (n=272)

of athletes reported having not received any form of sleep education within their sport.

Table 3.2: Athlete demographics and training and competition characteristics between sport-types

All Team-sport Endurance Technical Combat Power

(n=412) (n=123) (n=116) (n=75) (n=33) (n=59)

Demographic

Age (y)* 25 (7) 26 (5) 27 (5) 30 (11) 22 (4)a,b,c 23 (6)a,b,c

Age (>27 y, %)# 32 29 39a 43a 12b,c 19c

Gender (female, %)# 44 45 55 33b 33 41

Ability (para-athletes, %)# 34 29 34 63a,b,d,e 0a,b,e 25

Training and competition

Volume (h per week)* 17 (10) 15 (5)b,c,e 21 (8)d 21 (11) 16 (3) 19 (11)

Training start time (hh:mm)* 09:08 (01:50) 09:20 (01:10) 08:30 (01:40)a,c,e 09:50 (02:00) 08:10 (02:50)a,c,e 09:50 (02:40)

Training end time (hh:mm)* 17:40 (02:20) 18:10 (02:40) 17:00 (02:00)a,d 17:30 (02:20) 18:20 (01:40) 17:30 (01:50)

Major competitions (per year)* 4 (3) 3 (3) 5 (2)a,c,d, 4 (2) 3 (2) 4 (3)

Notes: a sig different to team sport; b sig different to endurance; c sig different to technical; d sig different to combat; e sig different to power (all p <0.01). Continuous data are expressed as mean (standard deviation). *One-way ANOVA, #Chi-square test of independence

Table 3.3: Prevalence and relative risk of poor sleep quality between sport-types, ability, gender and age groups

PSQI >5 PSQI >8

Prevalence Relative Prevalence Relative Risk factors 95% CI X2 95% CI X2 (%) Risk (%) Risk

Sport-types

Team 44 0.979 0.635 - 1.485 0.019 11 0.621 0.304 - 1.075 3.067

Endurance 45 1.017 0.746 - 1.386 0.011 14 0.801 0.476 - 1.345 0.72

Technical 52 1.356 0.900 - 2.041 2.135 33a 2.575 1.745 - 4.100 19.6

Combat 49 1.178 0.760 - 1.592 0.240 15 0.926 0.400 - 2.143 0.03

Power 37 0.818 0.576 - 1.160 1.418 12 0.698 0.335 - 1.452 0.98

Demographic

Paralympic 51 1.307 1.003 - 1.545 3.771 27b 2.574 1.661 - 3.988 18.898

Female 44 0.972 0.782 - 1.209 0.066 18 1.144 0.738 - 1.773 0.360

>27 y 49 1.195 0.902 - 1.440 1.585 25c 2.082 1.352 - 3.206 11.24

Notes: PSQI Pittsburgh sleep quality index; a significant difference to all other sports combined into one group (p<0.05); b significant difference to Olympic athletes; c significant difference to <27 y.

Table 3.4 Prevalence and relative risk of poor sleep quality and sport sleep challenges

PSQI >5 PSQI >8

Prevalence Relative Prevalence Relative Risk factors 95% CI X2 95% CI X2 (%) Risk (%) Risk

Elite sport stressorsa

Before comp 64b 1.210 1.027 - 1.425 5.133 20b 1.847 1.127 - 3.027 6.314

After eve comp 42 0.921 0.740 - 1.146 0.553 14 0.738 0.4.68 - 1.163 1.747

After long haul travel 50 1.290 0.995 - 1.673 3.687 18 1.165 0.776 - 1.977 0.804

During a training camp 46 1.085 0.859 - 1.369 0.465 21b 1.355 1.040 - 1.766 4.353

Notes: PSQI Pittsburgh sleep quality index; a represents athletes who experience sleep disturbance in response to risk factors listed; b sig. different to those who do not experience sleep disturbances in a particular elite sport risk factor;

Table 3.5: Sleep quality and insomnia symptoms between sport types

All Team-sport Endurance Technical Combat Power (n=412) (n=123) (n=116) (n=75) (n=33) (n=59)

PSQI Component

TST (h) 7.7 (1.0) 7.6 (0.8) 7.8 (1.0) 7.2 (1.1)e,h 7.6 (1.1) 7.9 (1.0)

SOL (min) 27 (26) 25 (24) 24 (21) 32 (32) 32 (27) 30 (31)

SE (%) 89 (13) 86 (10) 89 (10) 83 (12)e,g 90 (10) 88 (10)

Insomnia symptomatology

SOL >30 min (%)a 18 16 16 20 27 19

Nocturnal/ morning awakenings (%)a 28 23 27 33 52d,h 19

Un-refreshing sleep (%)b 52 54 49 56 63 42

Sleep dissatisfaction (%)c 20 20 16 29 21 17

Fatigue to extent miss training (%)b 16 17 13 20 18 15

Insufficient sleep (%) 44 47 39 52 39 39

Notes: PSQI Pittsburgh sleep quality index; TST Total sleep time; SOL sleep onset latency; SE sleep efficiency; a frequency >3 times per week in the past month; b ‘regular’ occurrence of symptom; c symptom occurrence in the past month; d significantly different to team sport; e significantly different endurance, f significantly different to technical, g significantly different to combat, h significantly different to power (all p< 0.05); Continuous data are expressed as mean (standard deviation).

Table 3.6 Sleep quality, insomnia symptoms and sleep management between sport types

Team-sport Endurance Technical Combat Power All (n=412) (n=123) (n=116) (n=75) (n=33) (n=59)

Stress induced sleep disturbances

Before comp (%) 58 47 65 63 76d 53

After evening comp (%) 45 55 51 11d,e,g,h 36 58

After long haul travel (%) 39 52h 35 38 44 20

During a training camp (%) 41 43 39 45 27 44

Sleep management

Nappers (%) 43 38 53 20d,e,g,h 70d 46

Techniques to fall asleep (%) 31 24 37 28 24 42

Notes: PSQI Pittsburgh sleep quality index; TST, Total sleep time; SOL, sleep onset latency; SE, sleep efficiency; a frequency >3 times per week in the past month; b ‘regular’ occurrence of symptom; c symptom occurrence in the past month; d significantly different to team sport; e significantly different endurance, f significantly different to technical , g significantly different to combat, h significantly different to power (all p< 0.05); Continuous data are expressed as mean (standard deviation)

12 *

4 PSQI globalscores 2

0 Deterioated Remaimed unchanged Improved

Figure 3.2: Pittsburgh sleep quality index (PSQI) global scores between elite athletes who felt their sleep quality has ‘deteriorated’, ‘remained unchanged’ or ‘improved’.

* sig. different to ‘deteriorated’. Category bars and error bars are expressed as standard deviations.

3.4 Discussion The study aimed 1) to assess sleep quality and insomnia symptomatology across a broad range of sport-types; 2) to assess and compare the subjective quality of athlete sleep in relation to sleep patterns, and identified sport risk factors , and 3) to evaluate sleep management strategies adopted by sport-types, and associations to sleep quality. The main findings of the study support a broad hypothesis that differences in levels of sleep quality, insomnia symptomatology, stress induced sleep disturbances and sleep management exist between sport- types, with technical sports reporting a higher prevalence and greater risk of poor sleep quality overall. Sub-populations among the athlete sample who are at greater risk of having poor sleep quality were identified, and these were: Paralympic athletes, and athletes who were >27 y. Moreover, athletes who exhibit stress-related sleep disturbances, specifically before competition and during training camps, were significantly more at risk of having poor sleep quality. However, participation in elite sport does not appear to generally degrade sleep quality.

3.4.1 Between-sport differences in sleep quality (hypothesis 1) The study findings support a conclusion that between-sport differences in sleep quality and insomnia symptoms exist. As a result, the null hypothesis that there will be no significant differences in sleep quality (as measured by PSQI) between sport types can be rejected. Although global PSQI scores were similar between all sport-types after controlling for age and gender, technical sport athletes had a 2.6 greater risk of scoring above the poor sleeper threshold of >8 and these distributions were significant compared to all other sport-types indicating highly disturbed sleep (see Table 3.3). Given the similarities in training volume and that endurance athletes reported significantly earlier training start times (See Table 3.2), the lower TSTs reported in technical sports may not be attributed to differences in training (at least for the training parameters captured in the current study – see Table 3.2). Power sport athletes were significantly younger than technical sport athletes; and there was a sig. difference in Paralympic athletes between technical and endurance sports which may well explain the differences observed. One aspect of training which was not captured in the current study, however, was the energy demands of each sport. As a result, the greater energy expended by endurance and combat athletes during training

and competition may well drive higher SEs and TSTs in these sport types, or vice versa. This claim is supported by Trinder et al. (1985) whereby differences were reported in PSG derived slow wave and non-REM sleep, TSTs and SOLs between endurance and power trained sub-elite athletes. While accepting differences in methodology, this difference in power sport-type composition, therefore, may account for the similarities seen in the current study.

Insomnia symptoms were generally similar across all sport-types (see Table 3.5). However, higher distributions of combat athletes who reported sleep maintenance complaints >3 times per week over a one-month period (52%) were significantly different compared to team sport (23%) and power athletes (19%). In comparison, a nationwide wide study of psychopathology among elite French athletes, by Schaal et al. (2011) reported a significantly different life-time prevalence of sleep problems between sport-types, and reported a particularly higher prevalence among ‘aesthetic’ and ‘contact and combat’ sports compared to all other sport-types. Moreover, in the same study, and like the current study, the prevalence of sleep problems between ‘team sports’ and ‘aiming sports’ (termed technical sports in the current study) were similar. Whilst acknowledging the smaller sub-sample of combat athletes compared to other sport-types in the current study, the higher incidence of sleep maintenance insomnia symptoms observed here (and higher sleep complaints reported in other studies e.g. Schaal et al. (2011)) could be due to the added exposure to other elite sport stressors, that are otherwise absent in other sports, such as weight-regulation. For example, combat athletes (often termed weight- dependent sports) have been reported to practise, often drastic, body mass reduction strategies to ‘meet weight’ for competitive events. The implications of frequent body mass reduction (or energy restriction) practise for athlete health and wellbeing are far reaching, but include insomnia symptoms (see Sundgot- Borgen et al. (2013)).

From the overall sample, the prevalence of stress-induced sleep disturbances was generally high, ranging for 39-58% (see Table 3.6) and were broadly supports findings from Chapter 1 (Para. 1.3.5) stating sleep disturbances are predictable across sport seasons. In the current study, again, combat sports had the highest number of athletes reporting disturbances before competition

(76%), with team sports reporting the least (47%), and these differences were significant. The high levels of sleep disturbance reported by combat sports may be, to an unknown extent, due to the higher levels of anxiety experienced before competition as a result of combat sports being individual sports (Erlacher et al., 2011), but also due to the added demands (and possible affect, see Sundgot- Borgen and Garthe (2011)) of body-mass reduction practises before competition. On the other hand, in technical sports, the levels of physiological arousal induced by the evening physical performance, of say, archery or shooting may be less likely to be at a level enough to induce sleep disturbances.

3.4.2 Risk factors for poor sleep quality (hypothesis 2) Demographic risk factor analyses revealed Paralympic and older (>27 y) athletes were 2.6 and 2.1 times more likely to score >8 on PSQI, respectively (see Table 3.3). However, the proportion of poor sleepers (as defined by both >5 and >8 poor sleeper thresholds) among male and female athletes were similar. As a result, the null hypothesis that cross-sport variations in sleep quality are not significantly influenced by age and gender can be rejected. It is well known, within sleep epidemiology, that insomnia symptomatology is higher among older individuals (Ancoli-Israel, 2009; Ohayon & Sagales, 2010); as a result sports in which athletes compete to an older age e.g. technical sports, are more likely to report high incidences of insomnia. Another common finding is that women tend to report poorer sleep quality than men (Ohayon & Sagales, 2010); however this was not found in the current study and is in contrast to previous work within the sport science and medicine literature. For example, Swinbourne et al. (2016) reported higher PSQI scores in female (mean score=8) compared to male (mean score=6) New Zealand team sport athletes. Moreover, Schaal et al. (2011) reported a greater prevalence of both lifetime (22%) and current (27%) sleep problems (e.g. symptoms <6 months) among females athletes, including greater prevalence of insomnia symptoms, such as long sleep latencies and nocturnal awakenings compared to male athletes. Similarities between gender in the current study could be have been confounded, and therefore explained by, a higher proportion of male participants being Para-athletes compared to female athletes. A higher prevalence of poor sleepers among Paralympic athletes (compared to Olympic athletes), supports the high prevalence of sleep complaints in this sub-population as highlighted in Chapter 1 (see Para. 1.3.4). In

support of this, Duran et al. (2015) reported 70% of Chilean Paralympic athletes to score >7 on the Insomnia Severity Index (indicating sub-clinical insomnia symptoms) and 80% to score ≥5 on PSQI. However, it is recognised that the sleep of Paralympic athletes is under researched in the sport science and medicine literature. The influence of different disabilities as separate risk-factor(s) for poor sleep quality (e.g. visual impairment or spinal cord injury) on sleep quality requires further exploration.

In the current study, for the factors that showed significant differences (or showed a trend) between groups (e.g. disturbed or not disturbed), risk ratios revealed athletes that experienced stress induced sleep disturbances were 1.2 to 1.3 times more likely to score >5, and 1.4 to 1.9 times more likely to score >8 on PSQI. Sleep disturbances the night prior to competition was shown to be a factor that increased the risk of poor sleep quality at >5 and >8 poor sleeper thresholds. This is in contrast to findings in a study by Juliff, Halson and Peiffer (2015) who reported no association between poor sleepers as defined by >5 on PSQI and athletes who experience disturbed sleep prior to competition. Therefore the findings from the current study refutes the claim that “global sleep quality assessments may not display the same efficacy as with the general population, due to situational stressors and events athletes encounter” (Juliff, Halson and Peiffer, 2015, pp.5) and that sleep disturbances around competition cannot be predicted through the PSQI (Juliff et al., 2015). Instead, findings from the current study suggest possible linkages between incidences of stress related sleep disturbances and a global measure of sleep quality.

Despite high levels of both stress related sleep disturbances and insomnia symptomatology (see Table 3.6), however, sleep quality does not appear to be degraded by participation in elite sport, with only 16% of the sample stating their sleep has deteriorated since reaching their current level (e.g. elite) of performance. However, the cross-sectional design of the current study means that whether elite sport degrades sleep quality cannot be confirmed - a point raised in Chapter 1. To adequately address this question, longitudinal research is required across the entire athlete development pathway e.g. when athletes transition from junior through to senior levels of elite performance, and even all the way through to retirement). As almost two thirds of the sample (63%) stated

100

their sleep has remained unchanged, of which a large proportion reported poor sleep quality (see Figure 3.2), it could be argued, therefore, that good sleep quality is not a prerequisite for participation in (or selection for) elite sport. Moreover, similarities in the distributions observed between sport-types here, supports the claim that, for at least for the sport-types listed in this study, participation in elite sport can with-stand high levels of poor sleep quality, with technical sports withstanding higher levels than others. This suggests that sleep quality doesn’t appear to be a distinguishing factor for elite sport participation or success at an elite level of performance. Further research into the sleep quality of successful and unsuccessful elite athletes is needed to further detail elite sport performance-sleep quality relationships.

3.4.3 Sleep management (hypothesis 3) Similar sport-type differences were also observed in the prevalence of sleep management strategies employed, specifically napping in the current study (see Table 3.6). As a result, the third null hypothesis can be rejected. From the entire sample, 43% reported napping, of which 20% (of the entire sample) were categorised as habitual nappers. Using a large sample of South African team sport athletes, Venter et al. (2012) reported 21% to be habitual nappers (e.g. ‘regularly’ or ‘always’ take a nap). Whilst accepting unstandardized definitions of nappers are prevalent in the literature (see Milner and Cote (2009)), Tuomilehto et al. (2016) reported 80% of professional ice hockey players to nap at least twice per week. Although the prevalence of habitual nappers appears to be high in some sports, the overall prevalence of napping reported in the current study (and previous research) appears like high achieving, non-athletic adults. Again, accepting differences in napper definition employed, previous research has reported 68% of Australian students to nap 1-2 times per week (Lovato et al., 2014). Further research is required to assess the prevalence of napping among elite athletes using controlled comparisons to non-athletes, while employing standardised definitions of napping patterns. In the current study, distributions of nappers were significantly lower in technical (20%) compared to endurance, combat and speed/ power sports. Moreover, combat athletes reported the highest prevalence (70%), and this distribution was significantly different compared to team and technical sports. In support of the between sport variability observed here, using actigraphy over a 14 day period,

101

Lastella et al. (2015a) reported higher incidences of day-time napping among individual athletes (15% of 754 days) compared to team sport athletes (11% of 613 days). The different sport-type categorisation criteria and methods to assess napping prevalence employed in the aforementioned study make comparisons to the current study problematic, however it is noteworthy that the individual sport athletes recruited in Lastella et al.’s (2015a) study participated in mainly endurance sports (e.g. race walking, triathlon, cycling and swimming). Making this assumption, then, the distribution of nappers were higher in endurance sports (53%) compared to team sports (38%) in the current study, however these differences were not significant. In the current study, the lower distribution of nappers among technical sports does not appear to be driven by sleep quality or insomnia symptoms (or sleep durations as shown by PSQI TST and distributions in response to the question “do you feel you get enough sleep?”) and therefore naps do not appear to be compensatory. Higher frequencies of napping amongst combat and endurance athletes, to an unknown extent, may reflect training schedule demands, and subsequent opportunity to nap during the daytime, as point recently raised by Sargent et al. (2014). Although opportunities to nap across the day were not captured here, training schedules which exhibited an early morning bias (e.g. combat and endurance athletes), may reflect the differences in napping distributions seen (even though they are not reflected in variations in night-time TST and SEs). In support of this claim, in the current study over half the proportion of nappers reported they found it easy to fall asleep during day, and this was similar across sports. As reasons for deciding to nap were not captured, whether napping patterns are predominantly ‘restorative’ cannot be confirmed. Further research is required to assess the day-time sleep propensity of elite athletes, whilst considering day-time sleepiness, sleep patterns and sleep quality.

Despite a high prevalence of daytime napping among sport-types, the frequency of athletes employing other types of sleep management, such as “personal techniques to help falling asleep” were generally low (31%) and this distribution was similar between sport-types. Moreover, education received on sleep was, again, low across the entire athlete sample (34%). This is dissimilar to previous research, with Venter et al. (2012) reporting 67% of South African athletes to have a “regular sleep routine” and Juliff, Halson and Peiffer (2015)

102

reporting 48% of Australian athletes to have a “strategy” to help them sleep well before a competition. In contrast to the current study, Juliff, Halson and Peiffer (2015) also reported differences between sport-types, with a greater proportion of team sport athletes having no strategy to help them sleep (67%) compared to individual athletes (48%). It was suggested this variation in sleep strategies employed between sport-types, may explain why significantly more individual athletes perceive a poor night of sleep to have no impact on performance compared to team sport athletes (LE Juliff et al., 2015). Moreover, given Tuomilehto et al. (2016) has shown 80% of team sport athletes reported wanting assistance with their sleep provides rationale to investigate sleep management strategies employed by elite athletes and sleep education received on possible linkages to sleep quality and stress induced sleep disturbances, and athletic performance.

3.4.4 Limitations The study had several limitations. Firstly, the study used a cross-sectional design to assess differences in the prevalence of poor sleep quality and insomnia symptoms between sport-types, and therefore cause-and-effect could not be inferred when conducting statistical analyses of relative risk. As a result, bidirectional relationships between stress-induced sleep disturbances (including elite sport participation more broadly – see Figure 3.2) and prevalence of poor sleep quality are plausible. To clarify the direction of this association, a longitudinal study of good and poor sleepers under the experience of stress is required. Secondly, the study did not include a control group, and used a combination of other sport-types as a reference when making comparisons. The addition of a control group would allow for inferences to be made against non- athletic groups of similar age, gender and ability (Paralympic/ Olympic status), and robustly test hypotheses regarding differences in sleep quality and athletism between sport-types. Thirdly, although assessments of sport ‘precipitating’ factors were employed, ‘predispositions’ to stress induced sleep disturbances were not. Through using formal metrics of sleep reactivity or hyperarousal, such as the Ford Insomnia Response to Stress Test (C. Drake et al., 2004; Kalmbach, Vivekpillai, et al., 2016), would allow for associations between ‘predispositions’ of stress- induced sleep disruption and sleep quality to be formally addressed. Finally, the sample did not include athletes from all sport-types. One sport-type which was

103

not represented in the sample was ‘aesthetic sports’ (i.e. gymnastics or synchronised swimming). Given aesthetic sports have been reported to exhibit high levels of insomnia symptoms (see Schaal et al. 2011)), the overall prevalence of poor sleep quality could have been underestimated. However, as aesthetic sports in the UK consist of three Olympic sports (gymnastics, diving, and synchronised swimming) and one Paralympic sport (Paralympic dressage), the number of sports included in total sample can be argued as representative of funded Olympic and Paralympic sports in the UK.

3.5 Conclusion The study supports previous work suggesting high levels of poor sleep quality and insomnia symptomatology in elite athletes (see Chapter 1, Para. 1.3.4). Extending this work, evidence from the current study provides some support to the hypothesis that sport-type related differences exist in sleep quality and insomnia symptomatology, with technical sports reporting lower self-reported TSTs and SEs, and having a higher risk of poor sleep quality (>8 on PSQI). However, due to within group heterogeneity in both age and ability (both, of which, are risk factors for poor sleep quality) this conclusion should be considered with caution. Demographic risk factors for poor sleep quality exist in elite sport and include ability (Paralympic) status, age and experience of stress induced sleep disturbances.

The prevalence of nappers in this population is high, and shows high between sport variability, with technical sports reporting the lowest prevalence. Between-sport variability in napping does not appear to be driven by low levels of sleep quality; nonetheless, a large proportion of athletes appear to be ‘able’ to fall asleep easily during the daytime. Further research is required to investigate daytime napping and daytime sleep tendency in elite athletes. The, overall, low frequency of athletes of who have received sleep education and employ sleep techniques to help falling asleep, along with similar between-sport distributions, demonstrate all athletes could benefit from sleep education and sleep management.

104

Incidences of stress induced sleep disturbances are prevalent in elite sport, but show high between-sport, but also within-sport, variability. Athletes who report stress induced sleep disturbances are associated with poorer sleep quality, yet only a small proportion of athletes reported their sleep had deteriorated since reaching an elite level of performance. This evidence suggests elite athletes across sport-types can with stand high levels of poor sleep quality. Further research is needed to investigate associations between stress induced sleep disturbances and sleep quality in elite athletes, and the formal identification of athletes who are vulnerable to stress induced sleep disturbances.

105

Chapter 4

4. Napping in high performance athletes: sleepiness or sleepability?

This study was published as: Gupta, L; Morgan, K; North, C and Gilchrist, C (2020). ‘Napping in high-performance athletes: sleepiness or sleepability?’ European Journal of Sport Science. 16, pp. 1-25. Doi: 10.1080/17461391.2020.1743765. [E- pub ahead of print]

4.1 Introduction Both the literature review (Chapter 1) and the survey data (Chapter 3) identified daytime napping as a common practice among elite athletes. Among the survey participants, for example, 43% were categorised as ‘nappers’ and of these, 20% were categorised as habitual nappers (see Chapter 3, Para. 3.3.3). These levels of napping appear to be consistent with levels of subjective sleepiness found among athletes. From studies using the Epworth Sleepiness Scale (ESS) (Johns & Hocking, 1997), for example, 28-44% of athletes report symptoms of excessive daytime sleepiness (Silva et al., 2012; Durán et al., 2015; Swinbourne et al., 2016). However, while the survey reported in Chapter 3 (Para. 3.3.3) found between-sport differences in nap frequencies, these differences are not easily attributable to the quantity or quality of night-time sleep. Habitual nappers and non-nappers showed no significant differences in PSQI global scores, with similar proportions “agreeing/disagreeing” with the question “do you feel you get enough sleep?”.

Within the sports science literature there appears to be growing consensus around three aspects of daytime napping: that napping is a common practice in elite sport; that daytime naps can enhance sports performance; and that athlete naps result from (and therefore provide evidence of) sleep debt (see (Fullagar et al., 2015; O’Donnell et al., 2018; Romyn et al., 2018). Support for the first two of these propositions comes directly from research evidence. Studies conducted among national squads, for example, report that 43% of junior (Kölling et al., 2016) and 48% of senior (Sargent, Lastella, et al., 2014) athletes nap during normal training. Regarding benefits, scheduled naps have been reported

106

to significantly improve sprint times in healthy young males (Waterhouse et al., 2007), peak jump velocity in elite female netball players (O’Donnell et al., 2018), and cognitive performance in highly trained male athletes (Petit et al., 2018).

However, the assumption that napping is primarily due to increased homeostatic sleep pressure resulting from inadequate nocturnal sleep (e.g. Davies, Graham and Chow, 2010; Petit et al., 2014; Lastella et al., 2015; O’Donnell et al., 2018; Romyn et al., 2018) is less well supported. In particular, studies evaluating relationships between athlete napping and subjective daytime sleepiness (the cardinal symptom of sleep debt) are lacking. This is an important omission since it is possible that, for some athletes, napping may not be compensatory. Broughton and Dinges (1989), for example, recognised 3 types of napping which differ in their relationship with daytime sleepiness: prophylactic napping (in anticipation of sleep loss); replacement or compensatory napping (in response to sleep loss); and appetitive napping (for convenience or enjoyment). In a factor-analytic study of napping typologies among young adults, Duggan et al. (2018) found that appetitive napping ‘on demand’ was significantly associated with higher quality nocturnal sleep, and was not significantly associated with daytime sleepiness.

This ability to nap on demand has been termed ‘sleepability’ (Harrison & Horne, 1996; Horne, 2010), and has been explored in Multiple Sleep Latency Tests (MSLTs) which use polysomnographic criteria to measure the time taken to fall asleep (sleep latency) when presented with a nap opportunity (Littner et al., 2005). While MSLT scores are responsive to changes in homeostatic sleep pressure (Arand et al., 2005), high sleepability has been identified in healthy young adults who record low MSLT scores (i.e. they tend to fall asleep quickly) independent of daytime sleepiness scores (Harrison & Horne, 1996). Since napping behaviour is a common feature of high-performance sport spanning junior and senior levels, it is possible that athletes may develop a superior ability to sleep on demand during the day. Such a possibility could be tested by comparing the nap behaviour of appropriately matched athletes and non-athletes while controlling for levels of daytime sleepiness. To date, however, attention to athlete nap electrophysiology has not included comparison groups (e.g. Petit et al., (2014);

107

Romyn et al., (2018)), and no study has explored the construct of ‘sleepability’ in athletic populations.

4.1.1 Insomnia and daytime sleepiness Within the sleep medicine literature, it is evident that the symptom of excessive daytime sleepiness (EDS), as opposed to fatigue, is not a typical characteristic of insomnia (Riedel & Lichstein, 2000). MSLT measures, for example, poorly discriminate between those with insomnia and controls (Lichstein et al., 1994). Underpinning this, previous research has also shown that individuals with insomnia exhibit heightened levels of day-time arousal when compared to healthy sleepers (M H Bonnet & Arand, 1997); this has been shown through both physiological and cognitive indices (See Bonnet and Arand (2010)). For example using the MSLT, Bonnet and Arand (2003) demonstrated that a period of partial sleep restriction failed to induce an increase in sleep tendency in individuals with insomnia, but did in healthy sleepers. These findings highlight that heightened basal levels of arousal or hyperarousal can mask the increase in physiological sleep tendency associated with restricted sleep. Moreover, it is also possible that higher levels of trait arousal (or even hyperarousal) could play a role in the successful development of elite athletes since higher levels of alertness (Riemann et al., 2010) could be advantageous for competitive performance. Explorations of daytime sleep tendency in athletes, therefore, could also provide insights into the role of arousal in mediating both performance and sleep management in elite sport.

4.1.2 Sleep Reactivity and ‘first night effects’ Hyperarousal has also been suggested to underpin the construct of sleep reactivity, an inherited predisposition to sleep disruption that manifests as a sleep system that is vulnerable to or ‘reactive’ to stress (C. Drake et al., 2004). Prior investigations have shown that individuals with high sleep reactivity scores, as measured by the Ford Insomnia Response to Stress Test (FIRST), but no current or prior insomnia, show greater sleep disruption in response to ‘sleep system stressors’ like caffeine intake (C. L. Drake et al., 2006), circadian misalignment (Kalmbach et al., 2015), and experimental ‘first night effects’ (C. Drake et al., 2004) Originally described by Agnew, Webb and Williams, (1966) ‘first night effects’ in laboratory- based polysomnographic sleep research are characterised by longer sleep

108

latencies and greater sleep fragmentation on the first attendance at the laboratory in some individuals. Because these “first night” phenomena have proved so robust, novel exposure to the sleep laboratory has been successfully adopted as a model of sleep challenge in both healthy sleepers (Coates, 1981), and those with insomnia (Hauri & Olmstead, 1989). Although predispositions to sleep disturbances have not been formally addressed within the sport science and medicine literature, individual differences in responses to challenged sleep within elite sport environments are evident (see Chapter 1, Para. 1.3.5). During periods of stress, for example, it has been reported that while some athletes demonstrate sleep disturbances (e.g. prolonged sleep latencies, increased awakenings, etc.) others appear much less affected (e.g. Fullagar et al. (2016a)). Exploring links between sleep reactivity measures and responses to sleep challenge among elite athletes, therefore, could help to explain these individual differences.

4.1.3 Research aims and hypotheses Despite the high levels of insomnia-type symptoms (see Chapter 1) and EDS (Silva et al., 2012; Durán et al., 2015; Swinbourne et al., 2016) reported within elite sport, the day-time sleep tendency of elite athletes has not been investigated using standardised MSLT procedures (American Academy of Sleep Medicine, 2005). The present study, therefore, was designed to compare the daytime sleep tendencies of high-performance athletes and non-athlete controls using a single nap opportunity MSLT model involving 2 consecutive attendances at the laboratory: a first (adaptation) trial; and a second (experimental) trial. Adaptation periods are embedded in MSLT protocols (see Littner et al. (2005)) in order to eliminate ‘first night effects’ (atypically disturbed sleep on first experiencing the sleep laboratory (Agnew et al., 1966). Initiating sleep in a novel environment is a common scenario among elite athletes, particularly during competitions (Erlacher et al., 2011; LE Juliff et al., 2015), and training camps (Pitchford et al., 2016). Given this, our adaptation trial served both to familiarize participants with the laboratory environment and simulate the challenge of sleeping in an unfamiliar setting, allowing the study to address the following research questions.

109

1. In: i) novel; and ii) familiar environments, do high performance athletes show greater levels of daytime sleep tendency (as evidenced by shorter sleep latencies) when compared with non-athlete controls?

2. If present, are differences in daytime sleep tendency between athletes and non-athletes primarily mediated by differences in daytime sleepiness?

Broadly, these questions involve the same null hypothesis:

1. sleep tendency measures from a single-trial nap opportunity will show no significant differences between high-performance athletes and non- athlete control participants.

To further explore whether napping behaviour is influenced by factors relating to the degree of athletic eliteness, 2 operationally defined groups of high- performance athletes were included, an elite group, and a sub-elite group.

4.2 Methods

4.2.1 Participant selection Participants were recruited between the 1st and 6th weeks of Semester 2 (to avoid the examination timetable) from the Loughborough University student population using an electronic message-board requesting athlete and fit non- athlete volunteers for a study of night-time sleep quality and daytime sleep tendency. Prior to allocation volunteers completed:

1. The PSQI (Buysse et al., 1989) is fully described in Chapter 2 (Para. 2.4.1). In the present analyses global scores were used as a continuous measure of sleep quality, while scores >5 were used to categorise individuals as ‘poor sleepers’. 2. The Ford Insomnia Response to Stress Test (FIRST) (Drake et al., 2004; Kalmbach et al., 2016) is fully described in Chapter 2 (Para. 2.4.2). In the present analyses, scores <16 were used and were indicative of ‘unreactive’ (i.e. robust) sleepers (Kalmbach et al., 2016) and scores >18 indicated individuals most likely to experience disturbed sleep under circumstances of challenge or stress (Kalmbach et al., 2016)

110

3. The Morningness-eveningness questionnaire (MEQ) has 19-items pertaining to habitual rising and bedtimes, preferred times of physical and mental performance, and subjective alertness after rising and before going to bed (J. A. Horne & Ostberg, 1976). Scores range from 16-86 with chronotype characterized as “evening”, “intermediate” or “morning” types (J. A. Horne & Ostberg, 1976). The questionnaire is valid against rest- activity patterns (Thun et al., 2012) and circadian rhythm of body temperature (Baehr, Revelle, & Eastman, 2000). The psychometric properties of the questionnaire demonstrate adequate internal consistency and reliability (Adan & Natale, 2002). The MEQ has also been used among elite athletes in sleep research (Michele Lastella, Roach, Halson, & Sargent, 2016). Following completion of the questionnaire assessments, Natale and Cicogna’s (2002) classification schema was used to exclude those scoring <30 (“definitely evening types”) or >70 (“definitely morning types”) on the MEQ.

Habitual sleep patterns were then screened over 7 consecutive days using Motionwatch 8 actigraphy (Camntech, Cambridge, UK): Mode 3; epoch length = 30 seconds (See Chapter 2, Para. 2.5.2). Participants were informed of the Motionwatch features (e.g. light sensor) and were instructed to press the event marker button each night when they started ‘trying to sleep’; and again each morning when they woke-up. Alongside the use of the MW, participants were instructed to complete a daily sleep diary on awakening. Captured data was analysed using Motionware software 1.0.27 (Camntech, Cambridge, UK). Using these recordings, mean and standard deviation values for sleep onset latency (SOL) and total sleep time (TST) were calculated for all participants. Those meeting research diagnostic criteria for insomnia symptoms (Edinger et al., 2004) by scoring >5 on the PSQI and recording 7-day averaged actigraphic sleep latencies >30 minutes were also excluded from the study. Selected volunteers were allocated to 1 of 3 groups using the following inclusion criteria. Elite athletes: currently competing at (at least) a national level and judged against Swann, Moran and Piggott's (2015) taxonomy to be ‘semi-elite’ or ‘competitive elite’. Sub-elite athletes: currently competing for the University, but below a national level of competition. Non-athlete controls: not engaged in competitive sport, non-obese and non-smokers/vapors, and meeting the WHO recommended

111

levels of physical activity (i.e. ≥ 30 minutes of moderate to vigorous physical activity performed daily on at least 5 days/week).

4.2.2 Sample size Sample size was estimated using G*Power 3 (Faul, Erdfelder, Lang, & Buchner, 2007), and was based on an effect size of 0.64, calculated from sleep latency data reported in 2 studies in which participants and procedures approximated to the present design: Petit et al’s (2014) study of a single nap opportunity in athletes (mean sleep latency = 7.7 minutes; SD = 3.9 minutes), and Bonnet and Arand’s (1998) study of MSLT performance in physically active young adults (baseline mean MSLT = 13.0 minutes; SD 5 minutes). Setting alpha = 0.05 and beta = 80%, and assuming an effect size of 0.64, a 3-group F-test would require a minimum of 27 participants (9/group) with a critical F-value = 3.4. Present recruitment achieved 10/group. Across the sub-elite and elite athlete groups, 90% of athletes represented endurance-based sports, mainly middle- and long- distance running (n=11), swimming (n=4), and triathlon (n=3); with the remaining representing team-sports (n=2).

4.2.3 Sleep latency recordings Participants reported to the sleep laboratory at 13:30 on two consecutive occasions within one week (separated by a minimum of 3 days) for the adaptation trial and experimental trial respectively. Actigraphic measurements of sleep were obtained from all participants for the night immediately preceding both trials. All pre-trial TSTs were required to fall within 2 standard deviations of that individuals screening average. Trials consisted of a single nap opportunity at 15:00, selected due to the heightened level of sleep tendency associated with this time of day (Dijk, Groeger, Stanley, & Deacon, 2010). The sleep laboratory maintained a regime of calm quietness; trials were conducted in complete darkness and rooms were maintained at 40-50% humidity and a temperature of 20-21°C. Sleep latencies were measured using polysomnography (PSG; Embla Systems, Denver, CO, USA). The montage included: electroencephalography (EEG) - central (C3-A2, C4-A1) and occipital (O1-A2, O2-A1); left and right eye electrooculograms (EOGs); and mental/submental electromyogram (EMG). Once electrodes were attached, and following standard bio-calibration procedures, participants were asked to “lie, assume a comfortable position,

112

keep your eyes closed and try and fall asleep”. According to American Academy of Sleep Medicine (2005) criteria for MSLTs (Arand et al., 2005; Littner et al., 2005), sleep latency was measured as the time from lights out to the first epoch of any stage of sleep (N1, N2, N3, REM) (Silber et al., 2007). Sleep onset was, in turn, defined as the first epoch of greater than 15 sec of cumulative sleep in a 30 sec epoch. The absence of sleep on the nap opportunity was recorded as a sleep latency of 20 min. PSG traces were scored by two independent researchers, and concordance was assessed. Participants who scored ≤8 min were judged to have a ‘high sleep tendency’ according to AASM (2005) criteria.

4.2.4 Sleepiness assessment The Karolinska Sleepiness Scale (KSS) was administered at 14:00, 14:30, and immediately prior to the nap opportunity at 15:00. This 9-point verbally anchored Likert scale measures sleepiness (as subjective sleep tendency) in the 5 minutes preceding the rating. Scale scores range from 1, ‘extremely alert’, through 5 ‘neither alert nor sleepy’, to 9, ‘very sleepy, great effort to stay awake, fighting sleep’ (Akerstedt & Gillberg, 1990). The scale is sensitive to changes in homeostatic sleep pressure (Åkerstedt, Anund, Axelsson, & Kecklund, 2014) and has been used previously among elite athletes (McGuckin, Sinclair, Sealey, & Bowman, 2014).

4.2.5 Statistical analysis Pearson’s correlation coefficient and Cohen’s kappa were used to evaluate inter- and intra-rater reliability for the scoring of sleep latency and sleep onset respectively. One-way analyses of variance were used to compare group mean values. Two measures of sleep tendency were used: i) sleep latency scores (in minutes); and ii) the proportions of those showing ‘high sleep tendency’ (i.e. sleep latency scores ≤8 minutes) within the non-athlete, sub-elite athlete and elite athlete groups. Preliminary ANOVA modelling showed significant interactions between sleep latency scores and both KSS scores and pre-trial (actigraphy) TST scores, violating the assumption of homogeneity of regression slopes for ANCOVA. For the adaptation and experimental trials, therefore, sleep latency scores in the 3 groups were first compared in multiple regression models, with dummy codes designating non-athletes as the reference group and sub-elite and elite athlete groups as covariates (the unadjusted model). This model was

113

then repeated with either 15:00 KSS or pre-trial actigraphic TST scores separately added as covariates. Within all 3 models differences between athlete group and reference means were determined using t-tests. Proportions categorized as ‘high sleep tendency’ within the 3 groups were compared using Fisher’s Exact Test. To assess changes in pre-trial subjective sleepiness, KSS scores were compared in a two-way (group: non-athlete; sub-elite athlete; elite athlete x time: 14:00; 14:30; 15:00) repeated measures ANOVA. All statistics were performed using SPSS 23 for Windows (Version 23.0, SPSS inc., IBM, Armonk, New York).

4.3 Results Participant characteristics are shown in Table 4.1. There were significant differences in training volume (p <0.01) and estimated training energy expenditure (p <0.001) between elite and sub-elite athletes. No significant associations between experimental group membership and the proportions of female participants (FET = 2.34; p = 0.36) or unreactive sleepers (FET = 0.77; p = 0.88) were found. There were no significant between-group differences for age, or the actigraphic sleep variables TST, sleep efficiency (SE), SOL or wake after sleep onset (WASO). Sleep variables for the nights immediately preceding the sleep latency trials are shown in Table 4.2. There were no significant between- group differences in TST, SE, SOL or WASO.

4.3.1 Sleep scoring reliability Inter-rater correlation coefficients for the first and second trial sleep latency scores were 0.83 (p<0.001) and 0.79 (p<0.001) respectively. Intra-rater correlation coefficients for the first and second experiment sleep latency scores were 0.91 (p<0.001) and 0.67, (p<0.001) respectively. The inter-rater Kappa coefficient for the identification of sleep onset within a nap was 0.7 for the adaptation trial and 0.9 for the experimental trial (both p<0.001).

4.3.2 Pre-trial sleepiness Karolinska Sleepiness Scale (KSS) scores prior to each sleep latency trial are shown in Figure 4.1. Two way (group x time) ANOVAs showed a significant main effect of time on subjective sleepiness prior to both the adaptation trial (F(2,54) =

35.51, p<0.001) and the experimental trial (F(2,54) = 25.28; p<0.001), with subjective

114

sleepiness increasing significantly from 14:00 to 15:00 (Figure 4.1). A significant main effect for group was also present for the adaptation trial (F(2,27 )= 4.49, p = 0.021), with paired comparisons showing greater overall sleepiness for sub-elite athletes relative to both non- and elite athletes. The experimental trial showed no main effect of group (F(2,27 )= 0.96, p = 0.34), and neither trial showed a significant group x time interaction effect (adaptation: F(2,27)= 2.23, p = 0.18; experimental: F(2,27 )= 2.28, p = 0.12).

Table 4.1: Participant characteristics

Non- Sub-elite Elite athletes athletes athletes p* (n = 10) (n = 10) (n = 10)

Age (y): Mean( SD) 21.0(1.5) 22.8(4.8) 23.1(3.8) p = 0.63

Female (%) 50% 60% 30% p = 0.53a

Training: mean (SD)

Training volume: h/w NA 9 (4.0) 17 (7.0) p < 0.01

Sport EE: METsb NA 9.8 (1.0) 8.5 (1.0) p = 0.23

Training EE: METs/wb NA 91 (47.0) 156 (91.0) p <0.001

Questionnaire sleep assessments: mean (SD)

FIRST 20.2 (4.9) 17.4 (3.7) 16.7 (4.6) p = 0.20

PSQI 3.9 (1.3) 4.6 (2.1) 5.5 (0.2) p = 0.19

MEQ 51.9 (10.1) 55.5 (12.6) 52.6 (9.1) p = 0.73

7-day actigraphic TST from screening: mean (SD)

TST: hours 7.1 (0.8) 6.6 (0.6) 6.4 (0.8) p = 0.10

SOL: minutes 15.8 (12.0) 9.9 (11.2) 17.0 (16.5) p = 0.46

SE (%) 81.2 (6.4) 82.8 (7.2) 76.9 (5.7) p = 0.12

WASO (min) 77.7 (30.5) 69.5 (34.7) 93.6 (21.9) p = 0.20

Notes: *Probability of F from one-way ANOVA unless stated otherwise; a Pearson Chi Square; b METs for each athlete’s sport derived from (Ainsworth et al., 2011). SD, standard deviation; FIRST, Ford Insomnia Response to Stress Test; PSQI, Pittsburgh Sleep Quality Index; MEQ, Morningness-Eveningness Questionnaire; TST, total sleep time; SOL, sleep latency.

115

Table 4.2: Mean pre-trial sleep patterns (for the nights immediately preceding laboratory visits)

Adaptation trial Experimental trial

Non- Sub-elite Elite p* Non- Sub-elite Elite p* athletes athletes athletes athletes athletes athletes (n=10) (n=10) (n=10) (n=10) (n=10) (n=10)

Sleep patterns

TST (h) 6.8 (0.9) 6.8 (0.9) 6.8 (1.6) 0.99 7.1 (1.4) 6.4 (1.3) 6.5 (0.9) 0.37

SOL (min) 16.2 (19.4) 9.8 (17.4) 14.7 (16.1) 0.70 16.3 (18.5) 5.4 (9.3) 17.3 (20.7) 0.24

WASO (min) 71.9 (29.8) 50.4 (35.9) 77.4 (38.9) 0.21 87.2 (31.9) 63.7 (39.9) 94.7 (32.7) 0.14

SE (%) 81.9 (7.9) 85.0 (8.3) 80.7 (9.8) 0.52 78.9 (7.6) 84.9 (8.7) 77.8 (7.0) 0.11

Notes: values are mean (SD); *p: significance of F for one-way ANOVA; TST: total sleep time; SOL: sleep latency; WASO: wake time after sleep onset; SE: sleep efficiency.

116

8.0 Adaptation Trial

7.0

6.0

score 5.0

4.0

Karolinska sleepinessscale 3.0

2.0

1.0

Experimental Trial 8.0

7.0

6.0

5.0 score 4.0

3.0

Karolinska sleepinessscale 2.0

1.0 14:00 14:30 15:00

Time

Figure 4.1: Mean Karolinska Sleepiness Scale (KSS) scores for non-athletes (○) sub-elite athletes (Δ) and elite athletes (□) prior to: the adaptation trial (above); and the experimental trial (below). Bars indicate 1 standard deviation.

117

4.3.3 Sleep tendency measures For most participants (63%) sleep latency scores declined between the adaptation and experimental trials. To assess the relationship between sleep reactivity and trial performance, Pearson product moment correlations between FIRST and sleep latency scores were calculated across all groups. While FIRST scores correlated significantly (r = 0.45; n = 30; p<0.05) for the adaptation trial, the experimental trial correlation was weaker and nonsignificant (r = 0.33; n = 30; p>0.05).

4.3.3.1 Adaptation Trial Unadjusted mean sleep latency scores for the adaptation trial are shown in Table 4.3 and Figure 4.2. Adjusted trial outcomes can be seen in Table 4.4. For the unadjusted regression analysis the overall model significantly predicted 33% of the variance in adaptation sleep latency scores (F(2,27) = 6.72; p = 0.004), with paired comparisons showing that mean sleep latencies for both the elite (mean = 10.3 min; t=-2.53; p = 0.02) and sub-elite (mean = 8.0 min; t=-3.60; p = 0.001) groups differed significantly from non-athlete control group values (mean = 16.3 min). When 15:00 KSS scores and pre-trial TST scores were separately added to the regression as covariates, the models remained significant overall (KSS adjusted model: F(3,26) = 4.32; p=0.013; TST adjusted model: F(3,26) = 4.823l p=.008) with both elite and sub-elite values differing significantly from the control group control values (see Table 4.4).

Levels of ‘high sleep tendency’ at adaptation showed a significant association with group (Table 4.3; Figure 4.3), with 20% of non-athletes, 80% of sub-elite athletes, and 50% of elite athletes recording a sleep latency of ≤8 minutes (Fisher's Exact Test = 7.02, p<0.05).

118

Table 4.3: Pre-trial sleep assessments and unadjusted trial sleep latency outcomes for adaptation and experimental trials

Adaptation trial Experimental trial

Non- Sub-elite Elite Non- Sub-elite Elite athletes athletes athletes athletes athletes athletes

(n=10) (n=10) (n=10) (n=10) (n=10) (n=10)

Pre-trial sleep assessment: mean (SD)

Mean TST (hours) 6.8 (0.9) 6.8 (0.9) 6.8 (1.6) 0.991 7.1 (1.4) 6.4 (1.3) 6.5 (0.9) 0.371

Mean SOL (minutes) 16.2 (19.4) 9.8 (17.4) 14.7 (16.1) 0.701 16.3 (18.5) 5.4 (9.3) 17.3 (20.7) 0.241

Mean (15:00) KSS scores 5.7 (0.8) 6.6 (0.95) 5.1 (1.3) 0.051 5.5 (1.5) 5.8 (1.4) 5.5 (1.3) 0.861

Unadjusted trial outcomes

Mean sleep latency scores 16.3 (5.0) 8.0 (4.7) 10.4 (5.8) <0.011 13.7 (5.8) 9.1 (4.9) 7.9 (4.8) <0.051 in minutes: mean (SD)

High sleep tendency 2 (13.3%) 8 (53.3%) 5 (33.3%) <0.052 2 (12.5%) 6 (37.5%) 8 (50.0%) <0.052 ratings: n (%)

Sleep onset not observed: 5 (62.5%) 1 (12.5%) 2 (25%) 0.092 4 (80%) 1 (20%) 0 (0%) <0.052 n (%)

Notes: 1significance of F for one-way ANOVA; 2significance of Fisher’s Exact Test; 3Significance of control group v athlete group t-value. SD, standard deviation; TST, total sleep time; SOL, sleep latency; KSS, Karolinska Sleepiness Scale (higher scores = greater subjective sleepiness). *Values are mean differences (95% confidence intervals) between athlete groups and reference (non-athlete) values; negative signs indicate lower sleep latency scores for athlete groups

119

Table 4.4 adjusted (for pre-trial TST and 15:00 KSS) trial sleep latency scores for adaptation and experimental trials

Adaptation trial Experimental trial Sub-elite athletes Elite athletes Sub-elite athletes Elite athletes

(n=10) (n=10) (n=10) (n=10)

Sleep latency scores adjusted for 15:00 KSS -8.3 (-13.3, -3.2) -5.9 (-11.0, -0.9) -4.8 (-9.6, -0.1) -5.9 (-10.6, -1.1) scores* p <0.013 <0.053 <0.053 <0.053 Sleep latency scores adjusted for pre-trial -8.3 (-13.2, -3.4) -5.9 (-10.8, -1.1) -4.5 (-9.5, -0.5) -5.8 (-10.7, -0.8) TST scores* p <0.013 <0.053 0.07 <0.053

Notes: 1significance of F for one-way ANOVA; 2significance of Fisher’s Exact Test; 3Significance of control group v athlete group t-value. SD, standard deviation; TST, total sleep time; SL, sleep latency; KSS, Karolinska Sleepiness Scale (higher scores = greater subjective sleepiness). *Values are mean differences (95% confidence intervals) between athlete groups and reference (non-athlete) values; negative signs indicate lower sleep latency scores for athlete groups.

120

20.0

18.0

16.0

14.0

12.0 10.0

8.0

6.0 4.0 Mean sleep latency scores(min) 2.0

0.0 Non-athlete Sub-elite athlete * Elite athlete *

Figure 4.2: Mean adaptation trial sleep latency scores for a single nap opportunity (n = 10/group). Bars indicate 1 standard deviation.

Between-group ANOVA: (F(2,27) = 6.72, p=0.004)

* significantly different to non-athlete (p<0.05)

121

5 participants) 4

3 High High sleep tendency (number of 2

0 Non-athlete Sub-elite athlete Elite athlete

Figure 4.3: Levels of high sleep tendency (sleep latency scores ≤8 minutes) for the adaptation trial (n = 10/group).

Fisher's Exact Test = 7.02, p<0.03

4.3.3.2 Experimental Trial Mean sleep latency scores for the experimental trial are shown in Table 4.3 and Figure 4.4. Adjusted sleep latency scores can be seen in Table 4.4. For the unadjusted regression the overall model significantly predicted 21% of the variance in experimental sleep latency scores (F(2,27)=3.56; p=0.04), with paired comparisons showing that mean sleep latencies for the elite group (mean = 7.9 min; t=-2.53; p = 0.02) differed significantly from non-athlete control group values (mean = 13.7 min). For the sub-elite group in the unadjusted model, mean sleep latency scores (mean = 9.1 min) were marginally non-significant when compared with control values (t=-2.01; p=0.055). In the adjusted regression (with pre-trial KSS scores added as a covariate), both elite (t = -2.52; p = 0.02) and sub-elite (t=-2.08; p < 0.05) values differed

122

significantly from control values, though the model overall was non-significant (F(3,26) = 2.63; p=0.07).

Levels of ‘high sleep tendency’ for the experimental trial showed a significant gradient across the groups (Table 4.3; Figure 4.5), with 20% of non-athletes, 60% of sub-elite athletes, but 80% of elite athletes recording a sleep latency of ≤8 minutes (Fisher's Exact Test = 7.67, p<0.04).

20.0

18.0

16.0

14.0

12.0

10.0

8.0

6.0

4.0

Mean sleep latency scores Meanscores latency sleep (min) 2.0

0.0 Non-athlete Sub-elite athlete Elite athlete *

Figure 4.4: Mean experimental trial sleep latency scores for a single nap opportunity (n = 10/group). Bars indicate 1 standard deviation.

Between-group ANOVA: (F(2,27) = 3.56, p=0.04)

* significantly different to non-athlete (p<0.05)

123

participants) 4

High High sleep tendency (number of 2

0 Non-athlete Sub-elite athlete Elite athlete

Figure 4.5: Levels of high sleep tendency (sleep latency scores ≤8 minutes) for the experimental trial (n = 10/group).

Fisher's Exact Test = 7.67, p<0.04

4.4 Discussion Participant profiles (Table 4.1), together with the overall pattern of results, justify the selection procedures employed, and are consistent with the assumptions made in designing and executing this study. While participants showed similar sleep quality (PSQI), sleep reactivity (FIRST) and chronotype (MEQ) scores, differences in training volume and energy expenditure are consistent with the criteria used to differentiate the sub-elite athlete and elite athlete groups at recruitment. The high proportion of participants (63%) showing a decline in sleep latency from the adaptation to the experimental trial, and the significant correlation between FIRST scores and adaptation (but not experimental trial) sleep latencies, are both consistent with a generalized ‘first night’ effect. Similarly, the significant main effect of time shown for subjective sleepiness (KSS scores) prior to both the adaptation and experimental trials, with sleepiness increasing steadily and significantly from 14.00, supports the selection of 15.00 as a time offering a realistic nap opportunity to all 3 groups, and emphasizes

124

the sensitivity of the KSS sleepiness measure. Collectively, then, these descriptive findings provide confidence in the ability of the present design to adequately address the study research questions.

4.4.1 Sleep tendency (hypothesis 1) The present results allow for the rejection of the general null hypothesis in relation to both research questions. Relative to non-athlete controls, elite athletes showed significantly shorter sleep latencies and a significantly greater likelihood of achieving the threshold for ‘high sleep tendency’ in the unadjusted comparisons for the adaptation and experimental trials. Furthermore, across both trials adjusted comparisons showed that these significant differences in mean sleep latency were independent of both pre-trial (15:00) sleepiness and total sleep time on the preceding night. It follows, then, that between-group differences in sleep latency were not primarily mediated by differences in daytime sleepiness. A similar, though less consistent pattern of differences was found in the adjusted comparisons for the sub- elite group, which showed shorter sleep latencies independent of pre-trial sleepiness (Table 4.3) in both the adaption and experimental trials, but not independent of TST recorded on the night preceding the experimental trial.

Two broad and interconnected conclusions are supported by these findings. First, since the adjusted analyses of sleep latency scores indicated that athlete-control differences were maintained independent of pre-trial subjective sleepiness (and since sleepiness is the cardinal symptom of increased homeostatic sleep pressure), it is reasonable to infer that in the present study high-performing athletes demonstrated higher “sleepability” as characterized by Harrison and Horne (1996). For the elite group, this conclusion is also supported by the unadjusted analyses, where greater sleep tendency was demonstrated despite elite athletes showing significantly lower KSS scores (in the adaptation trial) or KSS scores equivalent to other groups (in the experimental trial). Second, accepting the explanatory value of the sleepability construct, the present findings suggest that daytime napping per se does not necessarily provide evidence of sleep debt or excessive daytime sleepiness in high performing athletes. The overall pattern of results also suggests a ‘gradient’ in the degree to which athlete napping appears to be disconnected from homeostatic

125

sleep need, with sub-elite athletes showing only partial or intermediate levels of sleepability between the elite- and non-athlete groups.

4.4.2 Sleepability The superior ability shown by elite athletes in the present study to initiate sleep in both novel and familiar environments merits further investigation. One plausible factor which could be influencing these findings concerns the influence and management of hyperarousal. Extensive research supports the view that hyperarousal, and consequent pre-sleep cognitive activity, is a major factor delaying sleep onset and initiating sleep dysfunction in insomnia disorder (Espie et al., 2006; A G Harvey, 2002; Riemann et al., 2010). The sleep latency scores recorded by the athletes in this study may therefore be indicative of either lower constitutional ‘levels’ of hyperarousal, or a superior ability to ‘manage’ (and mitigate) hyperarousal. Whether ‘sleepability’ (and possible arousal management) is a trait favoured by sport selection and training regimes, or is a strategic skill acquired by athletes to manage routine sleep challenges or optimize recovery, cannot be determined by the present results. However, exploring this phenomenon could have useful implications for both understanding and managing sleep in elite sport. While indicative of “sleepability” (Harrison & Horne, 1996) the capacity to initiate sleep ‘on demand’ for so-called “appetitive-naps” (Broughton & Dinges, 1989) has also been linked to higher quality night-time sleep (Duggan et al., 2018). Again, the selection methods used here, which limited variation in sleep quality among participants, do not allow this possibility to be examined using the present data. Nevertheless, it remains an interesting and testable hypothesis that daytime napping among some athletes, instead of indicating sleep inadequacy, might actually be indicative of superior night-time sleep quality.

4.4.3 Limitations The limitations of the present study, which was designed and powered to address only specific questions, should also be recognized. Efforts to reduce intrinsic contributions to pre-trial sleepiness were made at the level of selection by screening for disordered night-time sleep. However, the significantly higher KSS scores for sub-elite athletes at the adaptation trial (Figure 4.1) suggest the influence of additional group-specific sources of sleepiness. It is relevant to note that prior training on the day of each trial was not controlled for. While it is possible that the duration and intensity of uncontrolled pre-trial training could have differentially affected physiological sleep

126

tendency in individuals, it is less clear how such an influence could be uniformly exerted across a single group of 10 athletes participating in a range of sports/events. It should be emphasized, however, that the influence of differential levels of sleep need, as expressed through daytime sleepiness, were controlled in the adjusted analyses. The study inclusion criteria aimed to eliminate individuals with problems with sleep onset, however it cannot be ruled out that sleep disorders may have been present in the selected sample. Although not reflected in levels of daytime sleepiness scores, the selected elite athlete sample reported PSQI scores that crossed the poor sleeper threshold and reported both SEs and WASOs that were below and above clinical thresholds, respectively (e.g. <85% and > 30 min, (Ohayon et al., 2017)). As a result, it is possible that the presence of unknown sleep disorders may have contributed to the higher levels of sleep tendency observed. The identification of REM latencies during naps in addition to PSG assessments of nocturnal sleep may also have helped identify other influencing sleep pathologies (Chervin & Aldrich, 2000; Shelgikar & Chervin, 2013). It is also the case that the present study did not assess, and therefore could not control for, participant’s typical co-sleeping arrangements. Since the MSLT protocol may represent a greater departure from ‘typical’ sleeping arrangements for habitual co-sleepers, this omission should be considered when interpreting the present results. Finally, it should also be acknowledged that the scale of the present study may have contributed to instability in the group mean values and reduced power in the experimental trials. While the relatively modest n-sizes certainly limited the number of covariates which could reasonably be entered in the regression, participant selection procedures resulted in a generally stable pattern of means across the two trials, with no major outliers. In addition, effect sizes (estimated from the unstandardized beta values) showed robust mean differences in unadjusted athlete- control values ranging from 4.6 minutes (experimental trial, sub-elite v controls) to 8.3 minutes (adaptation trial, sub-elite v controls), while Cohen’s f2 for these models (calculated as f2 = r2 / (1 - r2)) adjusted models indicated medium (f2 ≥ 0.15) to large (f2 ≥ 0.35) effect sizes.

4.4.4 Practical application Two practise points are also supported by the present findings. First, given the clear relevance of assessing homeostatic drive in relation to sleep variables, we would recommend that future studies of athlete napping include valid and reliable measurements of daytime sleepiness. And second, given the significant correlations

127

achieved in the present study between FIRST scores and sleep latency scores in the adaptation trial (r = 0.45; n = 30; p<0.05) we would recommend the use of formal metrics to assess sleep reactivity (and predict sleep support need) among athletes sleeping in novel environments.

4.5 Conclusions The present study compared daytime sleep latencies in high performance athletes and non-athlete controls using a single nap opportunity model. In unadjusted comparisons with non-athlete controls, elite athletes showed significantly shorter sleep latencies in both the adaptation and experimental trials. These significant differences were maintained in models controlling for pre-trial KSS scores and pre-trial total sleep time. Sleep latency scores for sub-elite athletes showed similar trends but were more labile. These results are consistent with two conclusions: first, that among elite athletes, napping behaviour can reflect sleepability and may not necessarily result from nocturnal sleep disruption and daytime sleepiness; and second, since superior sleepability was also demonstrated in the adaptation trial, elite athletes show less vulnerability when faced with sleep challenge. However, while not reflected in daytime sleepiness scores, the influence of unknown sleep disorders on sleep tendency cannot be ruled out. Further research employing homogenous groups of ‘good sleepers’ without the presence of sleep disorders are required to confirm the findings from this study. Nonetheless, these conclusions have implications for how athlete sleep is interpreted and managed. While daytime napping may provide evidence of homeostatic sleep pressure and degraded nigh-time sleep in some athletes, it might also provide evidence of superior arousal and sleep management in others. Furthermore, the significant correlations between FIRST scores and sleep latencies in the adaptation trial (coupled with the variability seen among athletes in managing nap opportunities) indicates that measures of sleep reactivity could play a useful role in predicting which athletes are more (or less) vulnerable to sleep challenge. In the light of these findings, the next chapter explores the constructs of sleep reactivity and arousal in the context of ‘real world’ challenges to the sleep of elite athletes.

128

Chapter 5

5. Sleep reactivity during a simulated Olympic tournament in elite British field hockey players

5.1. Introduction In the previous chapter comparisons of sleep latency scores from a single nap opportunity indicated that athletes show superior levels of sleepability both under ‘normal’ (experimental trial) and more challenging (adaptation trial) circumstances. However, the variations in sleep latencies among athletes (Chapter 4, Para. 4.3.3) also indicates that some athletes find daytime napping more difficult to initiate than others. Such variations across athlete populations were also highlighted in the literature review (Chapter 1, Para. 1.3.3) and survey data reported in Chapter 3 (Para. 3.3.3). This chapter examines athlete sleep in the context of real-world competitive sport and explores athlete-specific factors which may help to explain these variations. The broad aim, here, is to identify those athletes whose sleep is most at risk from competitive sport so that support and management strategies can be better targeted.

It was shown in Chapter 3 (see Table 3.4) that athletes who experienced sleep disturbances prior to competition (and in response to other stressors) were at greater risk of scoring >5 and >8 on the PSQI, indicating that sensitivity to acute stress-induced sleep disturbances may be a predisposing factor for generally poor global sleep quality. This finding is supported by a recent study (Dieck, Helbig, Drake, & Backhaus, 2018) which reported that competitive German athletes who scored high of Ford Insomnia Response to Stress Test (FIRST) also scored significantly higher on the PSQI, indicating that a vulnerability for stress‐induced sleep disturbances accompanies poorer sleep quality. This view is complemented by the findings of Juliff, Peiffer and Halson (2017) who reported associations between scores on the Arousal Predisposition Scale, and changes in actigraphy-derived sleep efficiency scores following night match-play in netball players (suggesting that high basal levels of arousal may explain sleep disruption observed following a sleep challenge in athletes). Furthermore, Ehrlenspiel, Erlacher and Ziegler (2016) reported that athletes who had high levels of anxiety at baseline were more likely to report a degradation in sleep quality the night

129

prior to competition. Overall, then, a growing body of research provides evidence for the operation of arousal-related constructs (stress, anxiety, sleep reactivity) which help to explain individual differences in the response to sleep challenge among elite athletes. The implication of this evidence is far reaching, suggesting measurements of arousal or sleep reactivity may predict the degree of sport-related sleep disruption. To date, however, the predictive utility of arousal-type measures has not been fully recognised in sports science. The recently developed Athlete Sleep Screening Questionnaire (ASSQ; Bender et al., (2018)), for example, does not address elements of trait arousal, or sleep reactivity (Chapter 1). Extending the findings obtained so far (and presented in earlier chapters), the present chapter describes the application and interpretation of sleep reactivity measurement (and the adjunctive construct of pre-sleep arousal) in the context of real-world ‘elite competition sleep’.

5.1.1 Elite competition sleep The night before a competition was identified in the literature review (Chapter 1, Para. 1.3.5.1) as the period within an athlete’s season associated with the highest levels of disturbed sleep. For example, in cross-sectional studies levels of athlete sleep disturbance on the night prior to a competition have been reported to range from 59- 71% using the Competition, Sport and Sleep Questionnaire (Erlacher et al., 2011; LE Juliff et al., 2015). Similarly, in the survey reported in Chapter 3 (see Para. 3.3.2), 64% of respondents reported disrupted sleep on the night prior to major competitions (see Table 3.3). General metrics of sleep quality have also shown competition-related sleep disturbances, with 37-78% of elite athletes reporting scores of >5 on the PSQI during periods of elite competition, such as World championships (M.-R. R. G. Silva & Paiva, 2016) and Paralympic games (Rodrigues et al., 2015; A. Silva et al., 2012). Nights following competitive events have also been shown to degrade sleep quality, with impact reported on sleep latencies and sleep efficiency (Fullagar et al., 2016a). While putative mechanisms through which the competition period might impact sleep quality have not been directly researched or manipulated in sports science, the evidence to date strongly implicates a role for pre-sleep arousal.

5.1.2 Pre-sleep arousal Reimann et al. (2010) describes the experience of pre-sleep arousal as “a common pathway resulting from the interplay between arousing and sleep-inducing brain

130

activity, psychological stressors and perpetuating mechanisms including learned sleep preventing associations”. Taking this into consideration, then, two common themes emerge from the findings of Chapter 1 and 3 in relation to the competition period. Firstly, pre-sleep arousal appears to play a mediating role in sleep disturbances both prior to and following competition. Regarding the pre-competition period, Juliff, Halson and Peiffer (2015) reported “thoughts about competition” and “nervousness” as common reasons for experiencing disturbed sleep. Moreover, Ehrlenspiel, Erlacher and Ziegler (2016) showed that athletes reporting high levels of ‘cognitive anxiety’ in the days leading up to a competition were more likely to report impaired sleep quality the night before competition. The same relationship was not seen for somatic anxiety in this study. In relation to the post competition period, however, it appears that the more physiological aspects of arousal tend to predominate. For example, Fullagar et al. (2016a) showed that professional football players reported “pain” and “adrenaline”, and “a strenuous game” were the main factors associated with disturbed sleep following evening competitions. Moreover, O’Donnell et al. (2018) reported heightened levels of cortisol after a match and disrupted sleep in netball players. In the context of findings reported in earlier chapters, the following conclusions can reasonably be drawn from the studies reviewed here. First, pre-sleep arousal manifests as either a cognitive or a physiological precipitating factor for athlete sleep disturbance during competitions and could provide a useful target for intervention in athlete sleep-management programmes. Second, it is likely that the impact of competition-related arousal is greater in ‘predisposed’ individuals (Spielman et al., 1987); And third, sleep reactivity (as described by Drake et al. (2004)) represents a ‘predisposing factor’ in the context of Spielman, Caruso and Glovinsky's (1987) model of insomnia.

5.1.3 Research context, aims and hypotheses Given the arguments presented above, it is reasonable to hypothesise that levels of sleep reactivity, as measured by the FIRST, will significantly predict athlete sleep disturbances in periods of elite competition. This ability to identify ‘sleep vulnerable’ athletes could inform team management strategies and greatly enhance the targeting of supportive sleep protocols during higher-risk periods. This research aim, however, encounters methodological challenges. In the context of elite competitions, the management and protection of ‘performance’ is the highest priority, and research input which requires athlete participation and/or deviations from normal pre-

131

competition practices may be resisted by team managers. Tournament simulations are becoming increasingly adopted as both ‘dress rehearsals’ for real competitions, and opportunities for team managers to make final performance-informed team selections. As such, these simulations closely approximate to the physical and emotional demands of an actual tournament. Simulating competitions have previously been used to assess sleep disturbances in sub-elite athletes in cycling (Lastella et al., 2015b) and football (Fullagar et al., 2016a). However, simulated Olympic (elite) tournaments, which form an important part of an elite athlete’s training, have not previously been used to assess the sleep patterns.

The present study, therefore, was embedded in a simulated field hockey Olympic tournament covering a total of 28 consecutive days (the ‘observation period’) where, following a 7-day baseline period, players competed in 8 matches within 14 days, followed by a 7-day ‘recovery’ period, during which no training or match-play took place. Data collection and analysis was designed to address the following research questions:

1. In a 28-day observation period comprising baseline, competition and recovery weeks, which aspects of athlete sleep are most disturbed during the competition period?

2. In a 28-day observation period comprising baseline, competition and recovery weeks, is sleep reactivity (as indexed by FIRST scores) significantly related to athlete sleep during competition periods?

The null hypotheses associated with these questions are:

1. measurements of sleep quality and quantity will show no significant differences between baseline, competition, and recovery periods; and 2. sleep quality and quantity measurements made in baseline, competition and recovery weeks will show no significant associations with measures of sleep reactivity (as measured by FIRST scores).

132

5.2 Methodology

5.2.1 Participants Twenty male elite field hockey players (mean age =26.5, SD=3.0; mean body mass= 77.6 SD=4.6 kg), volunteered to take part in the study. All athletes were part of the senior Great Britain squad. The sample made up 66% of the entire Great Britain senior men’s squad prior to selection for the Rio 2016 Summer Olympic games. Athletes playing experience at this level of performance (e.g. Great Britain) ranged from 2-14 years. The mean number of Great Britain caps won amongst the participants involved in the study was 45 (range 2-136). Using Swann, Moran and Piggott’s (2015) full taxonomy of ‘eliteness’, participants scored 7 (range 7-8), with 100% of participants being categorised as ‘competitive elite’ athletes.

5.2.2 Study design A schematic of the study design can be seen in Figure 5.1. During October 2015, all participants were observed for a period of 28 days, during which sleep patterns and self-reported sleep and wellbeing was recorded throughout. The 28 day observation period was broken down into three separate sections: 1) normal training (7 days); 2) simulated Olympic tournament (14 days); and 3) recovery (7 days). For the observation period players were based at their homes <10 miles from the training facility (English Institute of Sport, Bisham Abbey National Sports Centre, UK). The normal training week consisted of three weight sessions (10:00-11:00) and two club training sessions (20:00-22:00), which took place away from the training facility within the UK. During the simulated competition period, participants competed in eight games replicating the format (and potential number of games) of an Olympic field hockey tournament (see section 5.4.3). The scheduling of these matches can be seen in Figure 5.1. Of the eight games played three international matches (played at home) were against the Argentinian national squad (who, at that time, were based in the UK), three were training squad matches (played at home) and two were domestic matches (1 played at home and 1 played away). The recovery week was a decentralised training week (i.e. away from the Bisham Abbey training facility) and consisted of recovery and mobility sessions.

During the period of normal training, participants completed a series of online questionnaires relating to sleep quality (see section 5.4.5), sleep reactivity and daytime napping and sleep management. Questionnaires were mounted on the

133

Bristol Online Surveys (Bristol University, UK) platform, and were accessible via participants mobile phones.

5.2.2.1 Olympic tournament format During the Olympic Games, teams are divided into two groups of six nations, playing every team in their group once. Three points are awarded for a victory, one for a draw. The top four teams in each group then qualify for the quarterfinals. From the quarterfinals onwards a knock-out format is imposed with a loss resulting in exit from the tournament. During this time teams will potentially be required to undertake 8 games in 13 days should they reach the finals of the competition. For the Rio 2016 Summer Olympic Games a change to match-play format from previous Olympics games was imposed, with matches divided into 4 x 15 min quarters (previously 2 x 30 min halves).

134

Observation Period Baseline Recovery Week TW1 TW2 Week (BW) (RW) Day 1-7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22-28

Game daysa

11:55 10:25 14:30 12:30

10:30 Game 09:25 14:00-

start time 14:00-18:00

18:00 (hh:mm) Results - - - - - 2-3 2-1b 5-3b club matches; training-squad matches; international matches; a club matches took place on days 1 or 2, and 8 or 9; b Great Britain win

Figure 5.1: A schematic of the study design, and simulated tournament schedule

135

5.2.3 Questionnaire assessment Prior to the observation period, participants completed: 1) the Pittsburgh Sleep Quality Index (PSQI); 2) the Ford Insomnia Response to Stress Test (FIRST); 3) the napping behaviour questionnaire; 4) bespoke questions around sleep management, in general, and 5) a daily self-reported sleep symptom diary/ questionnaire

5.2.3.1 Sleep questionnaires 1. Pittsburgh Sleep Quality Index: The Pittsburgh sleep quality index (PSQI) is fully described in Chapter 2 (Para 2.4.1). In the present analyses global scores were used as a continuous measure of sleep quality, while scores >5 were used to categorise sleep as ‘disturbed’. 2. Ford Insomnia Response to Stress Test: The Ford Insomnia Response to Stress Test (FIRST) is fully described in Chapter 2 (Para. 2.4.2). The test requires respondents to rate the likelihood (Not likely, somewhat likely, Moderately likely, Very Likely) of sleep disturbance associated with 9 scenarios (e.g. “Before an important meeting the next day”). For the purpose of this study, an additional two scenarios were added: “The night before a major match”; and “After losing a match or performing badly” 3. Napping behaviour questionnaire: The napping behaviour questionnaire (NBQ) is a 6-item descriptive instrument (with no overall score) assessing the frequency, duration, and time of day when any naps are taken, and the reason for napping (Lovato et al., 2014). The metric has been used among young adults (Lovato et al., 2014), but not has not previously been used among athletes. 4. Sleep management: One closed-type question on sleep management was employed “When you become aware that you are unable to fall asleep at night (including after you have woken up in the middle of the night), do you use any techniques to help you fall asleep?”

5.2.3.2. Self-reported sleep symptoms Daily assessments of sleep and wellbeing were captured via a bespoke mobile phone application (AER app©, English Institute of Sport, Manchester, UK) each morning during the observation period. Items were scored on a scale of 1 (very low rating) to 10 (very high rating), and included the following:

1. Sleep quality (SQ): “how well did you sleep last night?”;

136

2. Ease to fall asleep (ETFA): “how easy was it to fall asleep last night?”; 3. feelings of ‘freshness; on awakening (FROA): “how refreshed did you feel on awakening this morning? 4. self-reported total sleep time (STST): “how much sleep did you get last night?”; and 5. Mood (MD): “How is your mood today?”

5.2.4 Actigraphy Wrist-watch actigraphy (see Chapter 2, Para. 2.5.2) was employed to assess the sleep patterns of participants using the Actiwatch 2© (AW2; Philips Respironics, Murrysville PA, USA). It was recommended that the Actiwatch was worn across the entire observation period but could be removed for training and matches. Epoch length was set at 1 min. Activity counts and photometer data were downloaded and analysed using ActiWare software (Version 6.0.7, Philips Respironics, Murrysville PA, USA). A ‘medium’ (wake threshold=40) threshold was employed to analyse activity counts. The sleep outcomes selected were total sleep time (TST), sleep efficiency (SE), wake time after sleep onset (WASO), and sleep onset latency (SOL).

5.2.5 Match-play demands Goal keepers were removed from analyses due to the minimal physical demands placed on this position. Match-play running patterns were collected using a 10Hz global positioning system (GPS) (OptimEye S5, Catapult innovation, Melbourne, Australia). During match-play, players wore a fitted harness, within which a GPS device was placed in a pocket situated between the scapula bones on the players’ backs. All players wore a heart rate monitor for each match (Polar T31, Polar Electro OY, Kempele, Finland) and heart rate was recorded (at 5 second intervals) continuously via the GPS unit. Following each match GPS and heart rate data were downloaded and time-motion analysis was performed using OpenField software (Catapult Innovation, Melbourne, Australia). To obtain an indication of the physical demands of tournament match-play, duration of match-play time, the total distance covered (TD), high speed running distance expressed using a maximum aerobic running speed threshold (HSR), and mean speed (S) were reported. Total distance covered has shown to be reliable and valid, and also used to quantify match-play demands in team sports during tournaments, previously (Strauss, Sparks, & Pienaar, 2019). To gauge the players’ physiological responses to each match, average heart

137

rate (recorded over the entire match) and time spent above 90% of a pre-determined maximum heart rate was selected (Bangsbo, 1994). To assess whether match-play demands were similar between tournament weeks, the sum of time and distance variables over both tournament weeks, and weekly average of speed and heart rate during matches were reported.

5.2.6 Statistical analysis All continuous data (e.g. actigraphy variables) are presented as weekly means (SD). Data that was not normally distributed were treated as non-parametric and analysed accordingly. Participants with >50% missing data for each week over the entire observation period were excluded from the analyses (Kölling et al., 2016). Participants were grouped using the FIRST reactive sleeper threshold of >18 (Drake et al. 2014) and independent sample student’s t-test was used to assess between-group (‘reactive sleepers’ v ‘nonreactive sleepers’) differences. To address study research questions, actigraphy and self-reported insomnia symptom variables were then entered into separate two-way (group x time) repeated measures analysis of variance (ANOVA) models and post-hoc tests were employed to assess differences between time points. To further address research question 2, separate linear regression models were employed to assess the proportion of variance shared between FIRST scores and change scores (BL values minus W2 values) in actigraphy variables during the second week of the tournament period (TW2). TW2 was selected due to the higher quality of opposition players competed against in this period (see Figure 5.1). For self-reported sleep symptoms, Pearson’s correlation coefficients were derived between raw values for the same period. To address research question 2, separate linear regression models were employed to assess the proportion of variance shared between FIRST scores and change scores (BL values minus RW values) in actigraphy. A paired sample student’s t-test was used to assess differences in match play demands from the first week of the tournament (TW1) to the next (TW2), and an independent t-test was used to assess differences between groups during both weeks. A level of significance was set at p<0.05. All statistics were performed using SPSS for Windows (Version 23.0, SPSS inc., IBM, Armonk, New York).

138

5.3 Results Missing actigraphy and self-reported sleep symptoms data exceeded 50% for 4 participants; these were therefore omitted from all analyses. Regarding the sleep questionnaires, there were no missing data. Overall, then, 16 athletes were retained for analysis. Dividing the sample at the FIRST score of 18, 8 were rated reactive sleepers (>18) while 8 were rated nonreactive sleepers (0-18). All actigraphy-derived and questionnaire variables were normally distributed, and therefore parametric statistical analyses were conducted. However, all self-reported insomnia symptom data were not normally distributed and therefore were analysed using a related samples Friedman's two-way analysis of variance by ranks and multiple Mann Whitney tests, with Bonferroni correction (adjusted alpha value = 0.02), to assess differences between the reactive and nonreactive sleeper groups for each time point. Spearman’s correlation coefficients were employed to assess relationships between FIRST scores and sleep symptoms during TW2 and RW

5.3.1 Match-play demands Match play demands across the simulated tournament, between TW1 and TW2 can be seen in Table 5.1 (n=16) and match play demands between reactive and nonreactive sleepers can be seen in Table 5.2 (n=8/group). During TW1, 88% players competed in all four matches, and 100% (n=16) competed in at least 3 matches. During TW2, 25% (n=4) competed in all four matches and 81% (n=13) competed in at least 3 matches. Significant differences were observed between TW1 and TW2 in match-play duration, total distance covered (TD), mean speed, the distance covered by high speed running (HSR), and the time spent above 90% of maximum heart rate (HR). However, mean heart rate (HR) during matches did not differ significantly between TW1 and TW2.

Comparing match-play variables in reactive and nonreactive sleepers averaged across both weeks (TW1 and TW2) both match play time and heart rate were significantly higher in the nonreactive group (Table 5.2).

139

Table 5.1: Comparisons of match play demands across the first (TW1) and second (TW2) weeks of a simulated Olympic field hockey tournament

Simulated Olympic tournament Match play demands p TW1 TW2 Matches played 4 (1) 3 (1)* p=0.003 Playing time (min) 234 (44) 185 (48)* p=0.001 Total distance (m) 26,019 (4569) 20,639 (4055)* p=0.004 Mean speed (m/ min) 112 85 (8)* p<0.001 HSR distance (m) 3207 (581) 2232 (553)* p<0.01 HR (bpm) 142 (15) 141 (11) p=0.565 >90% maxHR 55 (30) 14 (7)* p<0.001 Notes: TW1, tournament week 1; TW2, tournament week 2; HSR, high speed running; HR, heart rate; maxHR, maximum HR. *significantly different to TW1

5.3.2 Participant sleep characteristics The mean PSQI score was 5 (SD = 2) with 50% (n=8) scoring >5 and no participant scoring >8. The mean FIRST score was 18 (SD = 3). From the additional sport-specific scenarios added to the FIRST, 19% (n=3) reported disturbed sleep to be “moderately likely” before a major match, however 69% (n=11) reported their sleep is likely/very likely to be disturbed after losing or performing badly. In relation to day-time sleep, 56% (n=9) stated that they napped during the day-time, 78% of this sample (n=7) reported napping 1-2 days per week and 22% (n=2) reported napping 2-3 days per week. Most nappers reported “feeling sleepy during the day” (78%, n=7) as the reason why they napped, while 22% (n=2) reported napping to “recover from training”. Night- time sleep management techniques were adopted by 38% (n=6) of athletes when they were unable to fall asleep.

The sleep patterns and sleep symptoms reported by reactive and nonreactive sleepers are compared in Table 5.3 (n=8/group). Likelihood levels of experiencing disturbed sleep after losing or performing badly are shown in Figure 5.2. At baseline

140

sleep was rated as likely/very likely to be disturbed after “losing or performing badly” among 100% of reactive sleepers (n=8), but only among 38% of nonreactive sleepers

(n=3) (X2(1)=7.27, p=0.007; Figure 5.2).

Positions

Goalkeepers (%) 25 0 NA Defenders (%) 0 38 NA Midfielders (%) 50 38 NA Forwards (%) 25 25 NA Match play demands TW1 n=6 n=8 Games played 4 (2) 3 (1) p=0.342 Total distance (m) 26604 (3578) 25263 (5924) p=0.607 Working time (min) 223 (48) 235 (38) p=0.840 Mean speed (m/min) 110 (10) 114 (7) p=0.381 HSR (m) 3195 (667) 3307 (535) p=0.744 HR (bpm) 135 (13) 149 (9)* p=0.039 HR>90% max (min) 56 (30) 53 (37) p=0.879 TW2 n=6 n=8 Games played 3 (2) 3 (1) p=0.212 Distance (m) 18342 (3963) 22362 (3377) p=0.063 Match play time (min) 162 (30) 207 (28)* p=0.016 Mean speed (m/min) 86 (6) 84 (11) p=0.635 HSR (m) 2259 (624) 2252 (529) p=0.983 HR (bpm) 133 (7) 147 (5)* p=0.001 HR>90% max (min) 17 (6) 10 (8) p=0.133 Notes: TW, training week; HSR, high speed running; max, maximum; HR, heart rate. * significantly different to reactive sleepers

141

Table 5.3: Comparisons of demographic characteristics, sleep patterns, and sleep symptoms between reactive (FIRST score ≤ 18) and nonreactive (FIRST score >18) sleepers prior to a 2-week simulated Olympic field hockey tournament Reactive sleepers Nonreactive sleepers p (n=8) (n=8) Sleep characteristics

Age 25.5 (2.6) 27.9 (2.6 p= 0.088 PSQI 5.0 (5) 4.9 (3) p= 0.883 >5 (%) 63 37 p= 0.317 Day-time nap (%) 44 56 p= 0.614 Sleep strategy (%) 57 43 p= 0.614 Actigraphy SE 82 (7) 79 (6) p=0.391 SOL 14 (16) 13 (10) p=0.942 WASO 82 (42) 80 (38) p=0.928 TST 463 (38) 419 (34) p=0.034* STST 464 (33) 494 (26) p=0.078 Self-reported insomnia symptoms ETFA 6.8 (0.7) 7.6 (0.9) p=0.065 FROA 6.3 (0.9) 7.0 (0.8) p=0.074 SQ 6.8 (0.7) 7.6 (0.9) p=0.053 MD 6.5 (0.5) 7.4 (0.5) p=0.021* Notes: PSQI, Pittsburgh Sleep Quality Index; SE, sleep efficiency; SOL, sleep onset latency; WASO, wake after sleep onset; TST, total sleep time; STST, self-reported total sleep time; ETFA; ease to fall asleep; FROA; freshness on awakening; MD; mood; * significantly different (at p<0.05). (Table 5.3) reactive sleepers reported significantly higher total sleep times (TST), and significantly lower mood (MD).

142

100

40 (within groups) 30

% Likely to experience Likely % to experience disturbed sleep after 'losing a match or 'performed badly 0 Reactive Unreactive

Figure 5.2: Likelihood levels of sleep being disturbed after losing a match or playing badly within groups of reactive and nonreactive sleepers (n=8/ group)

2 Chi square: X (1)=7.27, p=0.007

5.3.3 Sleep symptoms during an Olympic field Hockey tournament

5.3.3.1 Sleep assessments From the analyses of variance of sleep variables across the whole observation period, there was no significant main effects for time on actigraphic measures of: SE

(F(3,42)=1.68, P=0.187); SOL (F(3,42)=0.696, P=0.560); TST (F(3,42)=1.404, P=0.255); or WASO

(F(3,42)=1.48, P=0.235). However, there was a significant main time effect for STST

(F(3,42)=3.886, p=0.015), with pairwise comparisons showing significant higher STST in the recovery week relative to both TW1 and TW2 (Figure 5.3). There were no significant main group effects for all actigraphy variables (SE, F(1,14)=0.54, P=0.475; SOL, F(1,14)=0.01,

0.989; TST, F(1,14)=1.356, P=0.264; WASO, F(1,14)=0.001, P=0.975; STST, F(1,14)=1.084, p=0.315).

There was no significant group x time interaction effect for SE (F(3,42)=0.22, P=0.995),

WASO (F(3,42)=0.489, P=0.692), and STST (F(3,42)=1.265, p=0.299). Changes from baseline values in actigraphically measured TST during the tournament and recovery weeks showed a significant group x time interaction effect (F(3,42)=3.303, P=0.042), with a

143

significant negative gradient being shown for the reactive sleeper group (see Figure 5.4). Analysis of actigraphic SOL change scores showed a similar pattern which failed to achieve significance (F(3,42)=2.665, P=0.060).

Tournament week: Separate linear regression models indicated that FIRST scores did not explain a significant proportion of variance in BW-TW2 change scores for: SE

(r2=0.033; F(1,14)=0.121, p=0.503); SOL (r2=0.109, F(1,14)=1.706, p=0.213); WASO (r2=0.118,

F(1,14)=0.186, p=0.193); or STST (r2=0.107, F(1,14)=1.68, p=0.215). FIRST scores did, however, explain a significant proportion of variance in BW-TW2 change scores for TST values

(r2=0.259, F(1,14)=4.904, p=0.044).

Recovery week: Separate linear regression models indicated that FIRST scores did not explain a significant proportion of variance in BW-TW2 change scores for: SE (r2=0.001,

F(1,14)=0.011, p=0.987); WASO (r2=0.044, F(1,14)=0.644, p=0.436); SOL (r2=0.07, F(1,14)=1.06, p=0.321); TST (r2=0.013, F(1,14)=0.179, p=0.678); or STST (r2=0.035, F(1,14)=0.510, p=0.487).

144

540

520 * * * * 500

480

460

Self Self reported TST (min) 440

420

400

380 BW TW1 TW2 RW Week Figure 5.3: Self-reported TST (STST) values across baseline week (BW), simulated Olympic tournament week 1 (TW1) and week 2 (TW2), and a recovery week (RW) (n=16).

2-way repeated measures ANOVA: F(3,42)=3.886, p=0.015

Post hoc comparisons: * Significantly different to RW

-20

-40

Total sleep time (min) -60 Δ

-80

-100 BW TW1 TW2 RW Week

Figure 5.4: Changes from baseline values in actigraphic total sleep time during the tournament week 1 (TW1) and tournament week 2 (TW2), and recovery week (RW) among reactive (□) and nonreactive (○) sleepers (n=8/ group)

Two-way repeated measures ANOVA (interaction effect): F(3,42)=3.303, p=0.029

145

5.3.3.2 Self-reported insomnia symptoms: Ratings for ETFA, FROA, and SQ across the observation period can be seen in Figures 5.5, 5.6 and 5.7, respectively (n=16). There was a significant time effect for FROA

(X2(3)=14.143, p=0.003), ETFA (X2(3)=11.67, p=0.009, and SQ (X2(3)=11.275, p=0.010).

However, there was no time effect for MD (X2(3)=1.042, p=0.791). There was a significant difference in ETFA between RW and TW1 (See Figure 5.5). Pairwise comparisons revealed there was a significant difference in FROA between TW2 and both BW and RW (See Figure 5.6). There was also a significant difference in SQ between TW1 and RW (See Figure 5.7). Mann Whitney-U tests with Bonferroni correction (adjusted p value=0.02) revealed no significant differences across all time points in all insomnia symptom variables (p>0.02). However, there was a trend for SQ to be different between groups during RW, with reactive sleepers reporting higher SQ values compared to nonreactive sleepers (p=0.028).

Tournament week: There were no significant correlations between FIRST scores and ETFA (R=-0.029, p=0.916), FROA (r=-0.177, p=0.511), MD (r=0.120, p=0.659) and SQ (r=0.122,p=0.653).

Recovery week: There was no significant correlations between FROA (r=0.335, p=0.204), ETFA (R=-0.060, p=0.825), and MD (r=0.132, p=0.627). However, there was a correlation with SQ (r=0.522,p=0.038).

146

10.0

9.0 * * * 8.0

7.0

6.0

Ease fall to asleep (ETFA) 5.0

4.0 BW TW1 TW2 RW Week

Figure 5.6: Ease to fall asleep (ETFA) values across baseline week (BW), tournament week 1 (TW1), tournament week 2 (TW2), and recovery week (RW) (n=16).

Friedman’s Two-way ANOVA by ranks: X2 (3)=11.67, p=0.009

Post hoc comparisons: * significantly different to RW

10.0

9.0

8.0 *

7.0 (FROA) 6.0

5.0 Feelings of freshnessawakening on 4.0 BW TW1 TW2 RW Week Figure 5.5: Feelings of freshness on awakening (FROA) values across baseline week (BW), tournament week 1 (TW1), tournament week 2 (TW2), and recovery week (RW) (n=16).

Friedman’s 2-way repeated ANOVA by ranks (significant main time effect): X2

(3)=14.143, p=0.003

Post hoc comparisons: * significant different to TW2 *

147

10 *

Sleep Sleep quality (SQ)

4 BW TW1 TW2 RW Week

Figure 5.7: Sleep quality (SQ) values across baseline week (BW), tournament week 1 (TW1), tournament week 2 (TW2), and recovery week (RW) (n=16).

Friedman’s 2-way repeated measures ANOVA by ranks: X2(3)=11.275, p=0.010

Post hoc comparisons: * significant difference to RW (p=0.030)

5.4 Discussion The findings illustrated in Figures 5.4, 5.6, 5.6 and 5.7 support assumptions made regarding the impact of simulated competition on sleep quality and insomnia symptoms, with freshness on awakening (FROA), ease to fall asleep (ETFA), sleep quality (SQ) and self-reported total sleep time (STST) showing significant degradation across the tournament period. Despite a higher standard of opposition (e.g. another international side) during the second week of the tournament (Figure 5.1), physical demands were significantly higher during the first week of competition (Table 5.2). This finding could be explained by the ‘style’ of play adopted when playing Argentina when compared to training matches. Irrespective of a higher physical output during tournament week one (TW1), insomnia symptoms did not increase during this period. This finding suggests that physical load isn’t the only driving factor in sleep disturbance

148

during tournaments. Overall, then, it can be concluded that the simulated tournament provided a sleep challenge.

Despite overall differences in physical outputs, between the sleep reactivity groups there were no differences across both weeks of competition (Table 5.2). However, there was a significant difference in cardiovascular responses between groups, with mean heart rate across both weeks being significantly higher among nonreactive compared to reactive sleepers (Table 5.1). This may be accounted for by the distribution of positional demands, and associated fitness levels of each group, with only midfielders and forwards making up the reactive group, following the removal of goal keepers. This is supported by no differences being seen in the total time spent above 90% maximum heart rate between groups (Table 5.2). This finding indicates that levels of physiological strain were similar between groups.

5.4.1 Sleep reactivity (hypothesis 1 and 2) At baseline sleep characteristics were, overall, similar between groups (Table 5.3). However, reactive sleepers had significantly greater baseline TSTs compared to nonreactive sleepers. Additionally, reactive sleepers reported significantly lower ratings of mood (MD) at baseline and a trend for lower ETFA, FROA and SQ values (See Table 5.3). This indicates that baseline assessments of sleep during normal training may not reflect athletes’ true baseline sleep, and training may have presented a level of sleep challenge. This is further supported by most insomnia symptoms being significantly more severe during the tournament when compared to RW, and not BW. Despite no significant differences in either global PSQI scores or the distribution of poor sleepers between groups, the findings here support the notion that, during competitions, reactive sleep is associated with insomnia symptoms among elite athletes.

The evidence presented here allows for both null hypotheses to be rejected. There was a significant negative gradient in total sleep times (TST) among reactive sleepers over the tournament period with values being 30 min less on average. This was supported by FIRST scores explaining a proportion of variance in change scores for TST during TW2. These values did not, however, ‘rebound’ to baseline levels during the recovery week. Whilst in the nonreactive group, TSTs appeared robust throughout the observation period with tournament week TSTs being 5 min higher than baseline on average. Despite no other changes in sleep variables during the tournament period,

149

there was a trend for an interaction effect for SOL, with reactive sleepers having SOLs higher than nonreactive sleepers during TW2, coinciding with a reduction in TST during this period. However, this difference did not reach a level of significance (p>0.05) nor did FIRST scores predict changes in SOL during this period. Even though TSTs were below baseline values during RW, reactive sleepers reported SQ to be 1.3 units higher, on average, compared to nonreactive sleepers during this period. This was shown via a significant moderate, positive correlation with FIRST scores and a trend for SQ to be different between groups. Moreover, SQ appeared to ‘rebound’ more so in the reactive group; sleep quality values were similar between groups during the tournament period and 1.2 units lower than nonreactive group at baseline (see Figure 5.6 and Table 5.3).

Among reactive sleepers, cross-sectional analysis at baseline showed that those “likely/very likely” to experience disturbed sleep reached 100% after losing or performing badly but reached only 37% prior to an important match (See Figure 5.2). A conclusion that can be drawn from these findings is that unreactive sleepers exhibit less sleep disruption to a challenge presented by a simulated Olympic tournament. Despite little evidence for differences in sleep variables or differences in levels of sleep management between the groups at baseline, nonreactive sleepers appeared to preserve sleep duration across the tournament period, more so than reactive sleepers. Interestingly, longer sleep durations at baseline (reported by the sleep reactive group; Table 5.3) were not associated with unreactive sleep. This finding is consistent with the results of Pitchford et al. (2016) who reported that higher actigraphy derived SE values at baseline predicted degradation in SE during a training camp, compared with players showing lower SE values at baseline. In relation to this, it can be concluded that reactive sleepers exhibit greater ‘rebound’ in sleep quality compared to nonreactive sleepers indicating that when a sleep challenge is removed then sleep is restored to ‘normal’ levels of sleep quality. The absence of a TST ‘rebound’ among the nonreactive sleepers also suggests that tournament TST does not reflect a competition-related ‘floor’ effect resulting from temporarily restricted sleep. Had this been the case, then the recovery week would likely have shown a rebound when the physical and mental stresses (and the match schedules) were removed. This conclusion is supported by Lastella et al. (2015b) who reported superior ratings of sleep quality during a recovery week following a simulated cycling grand tour in sub-elite

150

cyclists. Whilst in this study sleep reactivity wasn’t accounted for, the authors reported a rebound in SE but, like the current study, did not show a rebound in TST.

5.4.3 Practical applications These conclusions have research and practical implications. The study extends the concept of sleep reactivity in elite athletes, and highlights the need to explore both concepts of sleep reactivity and resilience in sport-specific scenarios, and also highlights to implication of screening for sleep reactivity in elite athlete groups to identify athletes at risk of stress-induced sleep disturbances, particularly given current metrics utilised did not consider this construct.

5.4.1 Limitations Limitations of the study should be considered when interpreting the results. Whilst attempts were made to simulate an Olympic hockey tournament, it is accepted that some elements of the simulation were missing and therefore the sleep challenge presented, here, may be somewhat underestimated than experienced during an Olympic Games. Firstly, during the Olympic and Paralympic games (and most major international competitions the ‘trio’ of sleep challenges are likely to be present (see Chapter 1). Typically athletes sleep away from home in novel environments, which have been reported to disturb composites of sleep quality, such as SE (Pitchford et al., 2016) due to unfamiliarity (LE Juliff et al., 2015). However, in the current simulation athletes slept in their home sleep environments which were based near to the training facility and as a result sleep disruption due to novel sleep environments were unlikely. Secondly, the psychological stress experienced by athletes during an Olympic Games is likely to be very high. For example, high levels of anxiety are experienced prior to competition particularly during the pre-sleep period has been shown to disturbed sleep (Ehrlenspiel et al., 2016). The psychological toll of competing in a simulated competition, however, would usually be expected to be low, however as the current tournament simulation took place nine months out from the Rio 2016 Summer Olympic Games (October 2015) and during a period of squad selection, psychological stress experienced by players, although not measured here, could be expected to be high. Thirdly, the competition schedule of Olympic field hockey tournaments is highly variable (Rio 2016 Olympic Games match schedule range: 10:00-20:00), with evening matches being common. The physiological arousal induced by competing late in the evening, has been previously reported to cause sleep disturbances in team sport

151

athletes, particularly curtailing TSTs (See Roberts, Teo and Warmington (2019)). However, in the current simulation few evening matches were played and as a result the impact of evening match-play on sleep would have been minimal. Another limitation is the small sample sizes used in each group (e.g. n=8/ group) given the drop out of 25% (n=4) from the original sample due to injury or non-compliance to actigraphy and self-reporting protocols. As a result, the study was likely underpowered and therefore results should be interpreted with caution.

5.5 Summary and conclusion The present study assessed subjective and objective sleep variables among elite athletes across three phases of a simulated Olympic competition: a 1-week baseline period; a 2-week period of elite competition; and a 1-week recovery period. Results can be summarised in relation to the study’s objectives

Which aspects of athlete sleep are most disturbed during the competition period? For the athlete sample as a whole, the lowest subjective ratings of FROA, ETFA, SQ and STST were recorded during the tournament periods, for each variable significant differences were present between one or both of the tournament weeks, and either the baseline or recovery week. While actigraphic measures showed little overall (i.e. main effect) impact of competition period, actigraphic TST showed a significant time x group interaction, with the sleep reactive group showing a significant reduction in TST across both TW1 and TW2.

Is sleep reactivity (as indexed by FIRST scores) significantly related to athlete sleep during competition periods? The division of the sample into reactive and nonreactive sleepers showed some discriminative value. At baseline, for example, reactive sleepers reported significantly higher TST and significantly lower MD. It is also relevant that reactive sleepers were significantly more likely to expect sleep disturbances after “losing or performing badly” in a competition. Subjective sleep quality also showed evidence of greater ‘rebound’ in the sleep reactive group during the recovery period. Finally, regression modelling showed that, overall, FIRST scores significantly explained 26% of the variance in actigraphic TST change scores in the second tournament week (baseline minus TW2).

152

In conclusion the present study provides some evidence that elite competition directly and negatively impacts the athlete’s experience of sleep, and selectively impacts total sleep time in athletes categorised as ‘reactive sleepers’. In general, the sleep reactivity construct may have modest utility in predicting sleep outcomes during elite tournaments.

153

Chapter 6

6. Sleep management in elite team sport athletes using cognitive-behavioural principles: A pilot study

6.4 Introduction The previous Chapter highlighted the sleep challenge experienced by elite athletes during periods of competition and indicated the possible use of the sleep reactivity construct in identifying the most ‘at risk’ athletes. Given the situational nature of these sleep disturbances, and their likely relationship with physical and cognitive arousal mechanisms, timely interventions which adopt evidence-based symptom management strategies developed within behavioural sleep medicine could offer an appropriate response. Such a need is emphasised by the high prevalence of insomnia symptoms shown in the introductory review (Chapter 1, Para. 1.3.4), and by the relative lack of sleep information and advice received by athletes shown in the online survey reported in Chapter 3 (with only 34% of athletes receiving sleep education, and only 33% having a strategy to combat sleep disturbances; see Chapter 3, Para 3.3.3)

6.1.1 Sleep management in elite sport To assess the state of sleep interventions in elite sport prior to the present study, the methodology for the introductory review (Chapter 1, Para 1.2) was repeated with the key-words ‘intervention’, ‘trial’, ‘treatment’ and ‘sleep management’ added to the existing search terms (sleep, insomnia, sleep disorders, sleepiness, napping, performance, training, tapering, competition, recovery, adaptation and travel). The same databases (Web of Knowledge, PubMed and Science Direct) were then searched. Searches were conducted up until April 2017. Studies describing the effects of altitude exposure, sleep deprivation or daytime naps on nocturnal sleep and/or performance were omitted. The search identified 12 intervention studies which targeted a range of sleep outcomes; and can be clustered into three intervention types: ‘sleep extension’ (n = 2, Table 6.1), ‘novel intervention’ (n = 5; Table 6.2) and ‘sleep hygiene and sleep education’ (n = 5, Table 6.3). However, while all the studies listed sleep outcomes as principal dependent variables, many emphasised performance outcomes (rather than improved sleep outcomes) as the main reason for conducting the study. For example, in the study by Mah et al. (2011) monophasic

154

sleep extension, where athletes were encouraged to sleep for at least 10 h per night over 5–7 weeks was shown to lengthen total sleep time (TST), and also improve basketball-specific performance (Mah et al., 2011) among student athletes. In a similar study, collegiate tennis players were encouraged to sleep for 9 h per day (i.e. over 24 hours) with the use of day-time naps. Following 1 week of bi-phasic sleep extension increases in self-reported TST and serving accuracy were observed (Schwartz & Simon, 2015). It is interesting, that in each of these studies few details are provided on how athletes achieved the additional TST over baseline values (beyond simply spending more time in bed).

Research describing protocols to enhance night-time sleep quality using both novel (Table 6.2) and sleep hygiene and sleep education (Table 6.3) focussed intervention in athletes has demonstrated modest results. For example, Abeln et al. (2014) employed a nocturnal “auditory brainwave entrainment” intervention (binaural electronic pulses delivered at 2-8 Hz throughout the sleep period for a total of 8 weeks) and reported significantly enhanced motivational ratings and sleep quality in soccer players. Duffield et al. (2014) assessed the impact of ‘sleep-hygiene recommendations’ on recovery processes in tennis players (specifically, they recommended a moratorium on “electronic stimulants such as television, mobile phones, and computers” 30 minutes before bedtime, and maintaining a bedroom temperature of 19 (SD=2)°C, and a light luminescence of 3 to 8 lux on athlete recovery. It was found that the sleep hygiene regime was associated with significantly improved ratings of sleep quality, and reduced perceptions of joint and muscle soreness. A similar “sleep hygiene strategy” was used by Fullagar et al. (2016a) who, after a late match, required highly trained soccer players to go to their bedrooms (maintained at ~17°C) at 23:45 in dimmed light with the offer of ear plugs and eye-masks. Furthermore, no technological or light stimulation was allowed ~15–30 min prior to bedtime. It is clear, then, that in both the Duffield et al. (2014) and the Fullagar et al. (2016a) studies, “sleep hygiene” mostly reflects a prescriptive bedtime routine coupled with temperature control. From the other studies to include a ‘sleep hygiene’ intervention (Duffield et al., 2014; H. Fullagar, Skorski, Duffield, & Meyer, 2016; O’Donnell & Driller, 2017; Van Ryswyk et al., 2016), only one provided details on sleep hygiene behaviours encouraged. In this study, O’Donnell and Driller (2017) stated “the sleep hygiene education focussed on the following five practical tips: maintaining a regular bed and wake time, ensuring a quiet, cool and dark bedroom environment, avoidance of

155

caffeine and other stimulants prior to sleep, avoidance of light-emitting technology devices in the hours prior to sleep and implementation of relaxation strategies before bed (e.g. progressive muscle relaxation)” (O’Donnell and Driller, 2017, pp. 525). As a result, collectively studies show little standardisation in sleep hygiene education design and delivery, and as a result are difficult critique and replicate.

156

Table 6.1: Sleep extension studies conducted among high-performance athletes

Mean Study n Sleep Sport age (SD)/ Outcomes Treatment effects* design (Gender) intervention Authors Range: y

Mean sleep ↑ TST; Mah et al. Time extension Actigraphic TST; Basketball 11 (M) 18 - 22 (2011) seriesa 110.9 (± 79.7 performance ↑ performance min)

Mean sleep Schwartz and Time Self-reported TST; ESS ↑ TST; ↓ ESS scores; and Tennis 12 (M + F) 20.2: 18-22 extension 108 Richard (2015) seriesa Performance ↑ performance min

Notes: * significant effects at <0.05; a Pre v Post measurements only (no controls); TST, Total Sleep Time; ESS, Epworth sleepiness scale; ↑, increase; ↓, decrease

157

Table 6.2 Novel interventions conducted among elite athletes among high performance athletes

Mean Study n age Sport Sleep intervention Outcomes Treatment effects* Authors design (gender) (SD)/ range: y

binaural beats at 2-8 Subjectively rated sleep ↑ sleep quality and ↑ Controlled trial Soccer 15 (M) 16.3 (1.0) Hz delivered during quality and sports Abeln et al. ‘motivation’. sleep for 8 weeks performance (2014)

Leaflet explaining Harada et al. “…benefits from Sleep diary: (2016) going to bed early, achievement of Shift towards ‘morning- Cohort Soccer 84 (M) 18–22 getting up early and ‘morning-type’ sleep type’ sleep wake patterns taking nutritionally wake pattern rich

30 min whole body Zhao et al. irradiation from red- Serum melatonin levels; ↑ serum melatonin; ↓ PSQI Controlled trial Basketball 20 (F) 18.6 (3.6) (2012) light on 14 and performance scores; and ↑ performance consecutive nights.

Randomised Sleep quality: PSQI, + ↓ PSQI scores, ↑ self- Dekker et al. “eyes open alpha placebo Gymnastics 12 (F) 22 (2.3) self-reported mental reported physical and (2014) power training” control trial and physical shape mental shape

“Restricted ↑ “Freshness”, ↑ “at ease”; ↓ Driller and environmental MDMQ; self-reported “worn-out”, ↓ “exhausted Multi-sport 60 (M + F) NR Agnus (2016) stimulation therapy” muscle soreness items of MDMQ; ↓ + 20 min nap “soreness”

Notes: * significant effects at <0.05; a Pre v Post measurements only (no controls); ↑, increase; ↓, decrease; PSQI, Pittsburgh sleep quality index (decrease in scores = better sleep quality); MSMDQ, multidimensional mood state questionnaire; NR, not reported

158

Table 6.3 Sleep hygiene, psychoeducation and counselling interventions

Mean Study n/ Authors Sport age (SD)/ Sleep intervention Outcomes Treatment effects* design gender Range: y

Duffield et al. Crossover Subjective sleep ↑ Sleep quality rating; and Tennis 8 (M) 20.9 (3.6) Sleep hygiene (2014) trial quality; performance ↓ soreness

Randomised Fullagar et al. “Sleep hygiene ↑ TST in the sleep hygiene cross-over Soccer 20 (M) NR Actigraphy: TST (2016a) strategy” group design

2x1 hour “sleep Van Ryswyk Actigraphy + sleep ↑ subjective TST (by 20 Cohorta Soccer 25 (M) 23.7 (2.0) education sessions” et al. (2016) diary: TST and SE min); and ↑ SE (by 2%) + tailored feedback

“…83% of the athletes Single 2-hour session Tuomilehto Ice Questionnaire at 1- reported having Cohorta 107 (M) 25.4; 17-40 of structured sleep et al. (2016) hockey year follow-up benefitted from the counselling counselling”

Single 1-hour session O’Donnell Acticraphy: TST, SE, “sleep education” + ↑ TST (by 20 min), ↓ “wake and Driller Cohorta Netball 26 (F) 23.0 (6.0) SOL, “wake “sleep hygiene variance” (by 21 min) (2017) variance”b information”

Notes: * significant effects at <0.05; a Pre v Post measurements only (no controls); b “wake variance” = variance in wake-up time relative to mean (O’Donnell and Driller, 2017); NR, not reported; TST, Total Sleep Time; SE, sleep Efficiency.↑, increase, ↓ decrease

159

Three factors limit the ability to synthesise conclusions from the studies shown in Table 6.1, 6.2 and 6.3. First, there is a lack of consistency in the collation of sleep parameters used to demonstrate improvements in sleep quality. Many studies used ‘ad hoc’ measures to quantify sleep quality, while most studies failed to measure sleep quality at all. As a result, it is currently difficult to quantify the holistic effect of these interventions on athlete sleep. Second, few of interventions listed in Table 6.1, 6.2 and 6.3 was directly informed by theory or practice in behavioural sleep medicine. For example, the use of ‘whole body irradiation with red-light’ (Zhao et al., 2012), overnight ‘2-8 Hz binaural beats’ (Abeln et al., 2014), and booklets encouraging ‘morning-type’ sleep patterns and a nutritionally rich breakfast (Harada et al., 2016) were each presented as ‘novel’ interventions (see Table 6.2). And third, there is a clear lack of standardisation in the interventions offered. References to ‘sleep hygiene’ in the studies listed in Table 6.3, for example, include a range of different strategies and practices ‘mostly’ unrelated to protocols developed through clinical research (e.g. (Irish et al., (2015)). In conclusion, Tables 6.1, 6.2 and 6.3 demonstrate that cognitive behavioural approaches to sleep management are poorly developed and under researched in sport science and medicine. Given the wealth of sleep management experience and theory currently deployed in clinical settings, it seems appropriate that the optimal management of athlete sleep health should be informed by existing evidence-based principles and practice.

6.1.2 Cognitive behavioural principles in sleep management Since the mechanisms which mediate episodes of sleep disturbance in athletes (responses to acute stress; the management of pre-sleep arousal) appear to be similar to those involved as precipitating and perpetuating factors in the development of insomnia symptomatology (see Bonnet and Arand (2010)), it is reasonable to suggest that approaches successfully employed in the treatment of insomnia (van Straten et al., 2017) could be effective in helping athletes manage sleep in elite sport. Evidence based cognitive behavioural approaches have been widely researched in sleep medicine among non-athletic populations, with several meta analyses (E. A. Koffel et al., 2015; Okajima & Inoue, 2017; van Straten et al., 2017), and one major review by the Standards of Practice Committee of the American Academy of Sleep Medicine (Morgenthaler et al., 2006) being published. Such approaches include techniques such as sleep restriction therapy (see Miller et al. (2014)), stimulus control (Bootzin, Epstein, & Wood, 1991), sleep hygiene (see Stepanski and Wyatt (2003)) and cognitive

160

reconstruction (Edinger, Wohlgemuth, Radtke, Marsh, & Quillian, 2001). Recent meta- analyses have showed that both individual (van Straten et al., 2017) and group based (Davidson, Dawson, & Krsmanovic, 2017) CBT-I programmes are effective at improving insomnia symptoms with large effects on sleep onset latencies (SOL), sleep efficiencies (SE) and wake time after sleep onset (WASO), with only small improvements in TST among insomnia patients (E. A. Koffel et al., 2015). Relevant to elite athletes, then, such programmes have been shown to be effective in improving sleep outcomes in non-athletic young adults with insomnia (Freeman et al., 2015) and aspects of CBT-I in the form of psychoeducation have been shown to improve sleep quality in college students (Kloss et al., 2016). Whilst it is accepted that both the prevalence and natural history of persistent insomnia in elite sport is under researched (as shown in Chapter 1), similar approaches have been shown to be effective in acute cases of insomnia as a means to prevent further development of the disorder (Boullin, Ellwood, & Ellis, 2016). Moreover, cognitive behavioural approaches have been shown to improve sleep quality in adults without a diagnosable sleep disorder (See Murawski et al. (2017)). Since elite athlete populations as a whole are situationally vulnerable to acute insomnia symptoms, and that for some athletes these acute episodes will interact with a constitutional predisposition to chronic sleep disturbance, then interventions based on cognitive behavioural approaches offer a reasonable, evidence-based investment in athlete sleep health (Kalmbach et al., 2016).

6.1.3 Development of the ASMP The development and content of the Athlete Sleep Management Programme (ASMP) is fully described in Chapter 2 (Para. 2.4). Briefly, the ASMP is a psychoeducational programme which, across 4 separate workshops, explains sleep management strategies based on cognitive-behavioural principles. The 4 workshops cover 1) sleep in elite sport; influence of homeostasis, circadian rhythms and automaticity; 2) key elements of sleep hygiene and stimulus control; 3) managing pre-sleep cognitive arousal; and 4) managing and scheduling naps.

The overall aim of the ASMP was to provide athletes with information and resources which they could apply (as ‘self-help’) when the need arose. The content and organisation of the programme was influenced by 3 specific considerations: 1) to fit- in with existing training and competition schedules; 2) to be readily accessible to young athletes for whom sleep management may not be perceived as a key element

161

of their performance; and 3) to be practical and actionable within the demands of elite sport participation. Given the relative paucity of cognitive-behavioural approaches to sleep management available within the sports science literature, the ASMP was also designed to connect with, and allow future developmental input from, behavioural sleep medicine.

6.1.5. Research aims and hypotheses In order to test the 3 considerations (above) which had influenced the content and organisation of the ASMP, the present study was designed to pilot the programme and 1) evaluate the feasibility, perceived utility and acceptance of the ASMP among participants who have completed the programme; and 2) assess the impact of the ASMP on: sleep attitudes, patterns and quality. To meet these objectives, the study specifically addressed three research questions:

1. Is it practical to deliver the ASMP in the context of a routine Training Period for elite team-sport athletes? 2. Will the delivery of the ASMP provide evidence of a positive impact on sleep quality and insomnia symptoms? 3. Will the content of the ASMP be judged as useful by elite team-sport athletes?

The null hypotheses for the study were, therefore:

1. it is impractical to deliver the psychoeducational sleep management programme in the context of a routine training period for elite team-sport athletes; and 2. The psychoeducational sleep management programme will not be effective in improving sleep quality and insomnia symptoms; and 3. the psychoeducational sleep management programme will be judged to be not useful by athletes

162

6.2 Method

6.2.1. Participants Twenty male elite field hockey players (age = 25.8, SD = 3.1 y; body mass = 83.3, SD = 1.0 kg) were recruited to take part in this study. This sample of elite male hockey players consisted of 50% participants that took part in the study presented in Chapter 5. All athletes were part of the senior Great Britain squad at the time. The sample made up 65% of the entire Great Britain senior men’s squad prior to selection for the Rio 2016 Summer Olympic games. Athletes playing experience at this level of performance (e.g. Great Britain) ranged from 2-11 years. The mean number of Great Britain caps won amongst the participants involved in the study were 35 (range 13-73). Using Swann, Moran and Piggott's (2015) full taxonomy of ‘eliteness’, participants scored 7 (range 7-8), with 100% of participants being categorised as ‘competitive elite’ athletes.

6.2.2 Study Design A schematic of the study design can be seen in Figure 6.1. Following discussions with team managers it was decided that, since the athletes were competing against one another for selection to the squad which will travel to the Rio 2016 Summer Olympic Games, a control trial whereby one group received the sleep management programme and the other group did not would be inappropriate. A pre-post study design was therefore selected for the current study. The study took place over 10 weeks (January–March 2016) during a physical training period and was split into three distinct periods: 1) a baseline period (1 week); 2) the implementation period during which the ASMP was delivered (4 weeks); and 3) a follow up period (1 week and 1 month). A follow up period longer than this was unachievable due to the likely change in squad personnel following a final Olympic squad selection, and the Olympic games itself. Over the full 10-week training period athletes were based at their homes <10 miles from the training facility (English Institute of Sport, Bisham Abbey National Sports Centre, UK). The normal training week consisted of three weight sessions (10:00-11:00) and two club training sessions (20:00-22:00), which took place away from the training facility within the UK.

163

Training Period

Week 1 2 3 4 5 6 7 8 9 10 Base- Follow- Implementation Period Follow-up 2 line Up 1

sleep management; actigraphy; sleep questionnaires; sleep diary

physiological assessment; ? Programme Evaluation

Figure 6.1: Study design for a pre-post-test of the English Institute of Sport - Sleep Management Programme (delivered during the ‘Implementation Period’)

6.2.3 Delivery of the ASMP Following baseline assessments in Week 1 (see Figure 6.1) group-based workshops were conducted in each week of the Implementation Period, with each workshop focussing on a different component of the ASMP (as described in Chapter 2, Para. 2.6.3). All squad members were provided with the option to attend. Workshops were classroom based and lasted approximately 30 minutes. During the first session (Week 2) attendees were asked ‘what do you want get out of the ASMP?’ and ‘what areas of sleep would you like to cover?’ and this provided additional content to cover over the 1-month period. During each subsequent workshop an opportunity was provided for athletes to ask questions regarding any concepts, processes and techniques described throughout the programme. Following completion of the ASMP workshops, athletes completed Follow-up assessments at Week 6 (“Follow-up 1”) together with an online Programme Evaluation questionnaire. At the end of the training period (Week 10) athletes were again asked to complete Follow-up assessments (“Follow-up 2”). For

164

1-week periods during Baseline and Follow-up 1, athletes completed daily sleep diaries and actigraphy data were collected (described below).

6.2.4 Baseline and follow-up Questionnaires The following questionnaires were administered at the baseline and follow-up assessments:

6.2.4.1 Sleep questionnaires 1. Pittsburgh Sleep Quality Index: A description for the Pittsburgh sleep quality index (PSQI) is provided in the general methods section (Chapter 2, Para. 2.4.1) of the thesis. In the present analyses global scores were used as a continuous measure of sleep quality, while scores >5 were used to categorise sleep as ‘disturbed’. 2. Dysfunctional Beliefs about Sleep: A description of the Dysfunctional Beliefs and Attitudes about Sleep Scale (DBAS-16) is provided in the general methods section (Chapter 2, Para. 2.4.3) of the thesis. 3. Pre-sleep Arousal Scale: A description of The Pre-Sleep Arousal Scale (PSAS) is provided in the General Methods section of the thesis (Chapter 2, Para. 2.4.4). Whilst the scale has been previously used to appraise the pre-sleep period for a single night, in the current study a period one week was chosen as the period of assessment e.g. in the past week. 4. The Ford Insomnia Response to Stress Test: A description of the Ford Insomnia Response to Stress Test (FIRST) can be seen in the General methods section of the thesis (See Chapter 2, Para. 2.4.2). A reactive sleeper threshold of >18 was set.

6.2.4.2 Self-reported sleep symptoms Daily assessments of sleep and wellbeing were captured via a bespoke mobile phone application (AER app, English Institute of Sport, Manchester, UK) in the morning prior to each training session. All items were scored as single items on a scale of one (low) to ten (high). Sleep and wellbeing questions included:

1. sleep quality (SQ) “How well did you sleep last night?”; 2. ease to fall asleep (ETFA) “How easy was it to fall asleep last night?”; 3. feelings of ‘freshness’ on awakening (FROA); “how refreshed did you feel on awakening this morning?”, and 4. mood (MD) “How is your mood today?”

165

6.2.4.3 Programme Evaluation The Programme Evaluation questionnaire was made available to each athlete via an online platform (BOS, Bristol University, UK) at the end of the workshop series and is described in Chapter 2 (Para. 2.6.5). To assess the utility of the ASMP, sleep behaviours adopted to help promote feelings of sleepiness prior to sleep onset and pre-sleep de- arousal techniques deployed were assessed. Athletes answered these questions on a 5-point Likert scale ranging from “No, definitely not” to “Yes, absolutely”, and using a closed-question format The questions employed in the evaluation included:

1. “Did you find the material relevant?”; 2. “Do you feel equipped to deal with sleep disturbances when they arise?”; 3. “Would you recommend this workshop series to a fellow athlete?”; 4. “Have you previously received information about how to manage sleep disturbance during normal training (or) competitions?”; And 5. “Do you think you will use the information provided in the future?”

The questions employed to capture sleep related behaviours adopted were: 1. Sleep promoting activities are activities that are performed before (or at) bedtime to help promote feelings of sleepiness in preparation for sleep, do you regularly perform a sleep promoting activity before bed?

2. Do you use any techniques to help you fall asleep at night (including when wake-up in the night and cannot fall back asleep)? Please described this/these technique(s) and;

Do feel this technique (to help you fall asleep) is effective?

6.2.5 Physiological assessment Three physiological outcomes were assessed at Baseline and Follow-up 1: Saliva alpha amylase (sAA); exercising heart rate (HRex); and training load. Each is explained below.

6.2.5.1 Saliva alpha amylase Saliva alpha amylase (sAA) is an enzyme that has been shown to be a surrogate marker of sympathetic nervous system activity, with sAA levels rising in response to

166

physical and psychological stressors. As a result, sAA has been widely used in sport science as a biological marker in relation to physiological stress (Edmonds, Burkett, Leicht, & McKean, 2015), sleep (Chennaoui et al., 2016; Schaal et al., 2015) and sleep during high intensity training (Schaal et al., 2015) and competition (Chennaoui et al., 2016).

6.2.5.2 Saliva collection Prior to collection athletes were asked record whether they had eaten or consumed a caffeinated or sports drink within the last hour. Given sAA is precursor for activation of the central nervous system, ingestion of caffeine around the time of sample collection may confound with the presence of the enzyme in saliva (Laura C. Klein, Bennett, Whetzel, Granger, & Ritter, 2010), particularly in non-caffeine users (Laura Cousino Klein et al., 2014). To control for diurnal variation in sAA samples (Nater, Rohleder, Schlotz, Ehlert, & Kirschbaum, 2007) were collected at 10:00 on the same day each week. Saliva was collected using the passive drool method. Five minutes prior to sample collection, all participants rinsed their mouths with water to remove food debris. Participants were then instructed to be seated, place their heads between their legs and allow saliva to pool in the bottom of their mouths, before pushing the saliva into the sterilised collection vessels. The process was repeated for 2 min or until 2 mL of saliva had been collected. Saliva collections were then centrifuged at 1000 × g for 15 min (K241r refrigerated centrifuge, Centurion®, Chichester, UK), before being aliquoted into tubes for storage at −20 ◦C until analysis.

Saliva samples were assayed using a commercially available sAA-specific kinetic enzyme assay kit (Salimetrics, State College, PA, USA). On the day of analysis, saliva samples were completely thawed at room temperature prior to being diluted 1:200 with commercially supplied diluent. Diluted samples (and controls) were added to a 96-well microtiter plate in duplicate, followed by the addition of pre-heated (37 ◦C) substrate solution. The microtiter plate was then warmed in a pre-heated (37 ◦C) plate shaker (Jitterbug 2, Boekel Scientific, USA) for 1 min, and then transferred to the plate reader for measurement. An initial measurement of optical density at 405 nm was recorded at 1 min, before being placed back into the heated plate shaker. This process was then again repeated at 3 min. sAA activity was determined by subtracting the 1 min reading from the 3 min reading and multiplying by the

167

conversion factor supplied by the manufacturer. All samples were run in duplicate. The inter- and intra-assay coefficients of variation were 11 and 2%, respectively.

6.2.5.3 Heart rate recovery run The Heart rate recovery (HRR) run formed a part of the GB hockey’s ‘physical fatigue’ monitoring battery and was included in this study to capture levels of autonomic physiological arousal. A general description can be seen in the general methods section of the thesis (Chapter 2, Para. 2.5.1). All heart rate recovery runs were completed by all athletes every Monday morning at 0900. During the current study, exercising heart rate (HRex) was employed according to reliability assessments reported in in Chapter 2.

6.2.5.4 Training load As athletes were participating in a period of intensive physical training which has shown to be disruptive to sleep (Pitchford et al., 2016; Schaal et al., 2015; Thornton et al., 2016), it is important that training load was considered in this study. As a result, training load was calculated using a session rating of perceived exertion methodology (sRPE) using an amended RPE scale scaling from 1-10 (C Foster, Daines, Hector, Snyder, & Welsh, 1996). Players entered this information through a mobile application and were encouraged to do so within 30 min after training. The rating was then multiplied by the duration of training to obtain training load (see Carl Foster, Rodriguez-Marroyo, & de Koning, 2017). Total training load for the day was collected and weekly training loads were recorded at Baseline and Follow-up 1).

6.2.6 Actigraphy A general description of the actigraphy protocol can be seen in the General Methods section of the thesis (Chapter 2, Para. 2.5.2). In the current study the Actiwatch 2 (Philips Resperonics, USA) was employed to collect actigraphy data. The sleep outcomes selected were TST, SE, SOL, and WASO.

6.2.7 Statistical analyses All continuous data were presented as means (SD) (e.g. actigraphy data). Continuous data that were not normally distributed were treated as non-parametric accordingly. Participants with >50% missing data for each week over the entire observation period were excluded from the analyses (Kölling et al., 2016). To address the first objective of the study, repeated measures analysis of variance (ANOVA) models were employed

168

to assess differences in sleep questionnaire scores (PSQI, DBAS, and PSAS) scores between Baseline, Follow-up 1, and Follow-up 2. Partial eta squared values were reported for effect sizes. Cochran’s Q test was employed to assess the distribution of poor sleepers as defined by the PSQI cut-off value of >5 across the 3 time points (Baseline, Follow-up 1and Follow-up 2). Furthermore, a paired samples t-test was employed to assess differences in actigraphy (SE, SOL, WASO and TST), sleep sleep symptoms (ETFA, FRSH, SQ,and MD) and day-time arousal data (HRex and sAA) from Baseline to Follow-up 1. Cohen’s d was used for effect sizes. For actigraphy data, insomnia symptoms were defined as SOL >30 min, WASO >30 min and SE ≤85% (Ohayon et al., 2017). Moreover, a short TST threshold of <6 h was set (Hirshkowitz et al., 2015) . The proportion of participants showing these characteristics at Baseline and Follow-up 1 were compared using McNemar’s test. To judge the overall feasibility, acceptability and perceived utility of the ASMP the Programme Evaluation questions were examined using descriptive statistics. All statistics were performed using SPSS for Windows (SPSS inc. Version 23.0,, IBM, Armonk, New York) with alpha set at p<0.05..

6.3 Results

6.3.1 Sample One athlete was omitted from the study due to injury. Adherence to questionnaire completion was good with 100% (n=20) of athletes completing all questionnaires. One athlete experienced an actiwatch malfunction during the recording period. Missing data were accounted for by mean imputation. Adherence to actigraphy protocols was excellent, with no athletes having >50% of missing data. The mean proportion of missing actigraphy data days was 11 (SD = 12)%. Adherence to sleep diary protocols was excellent with no athlete having >50% missing data; the mean proportion of missing data was 21% (SD = 15%). Similarly, missing data for HRR and sAA was 10 (SD = 5)% and 23 (SD = 18)%, respectively. Attendance at group sessions was high, with only two athletes not attending all workshops. All actigraphy, questionnaire and physiological assessment data were normally distributed. However, all sleep diary data were not normally distributed and therefore treated as non-parametric. Overall, 18 athletes were included in the analyses.

169

6.3.2 Training load There was no significant difference in training load between Baseline (mean = 544 training load units; SD = 155) and Follow-up 1 (mean = 592 training load units; SD = 360) measurements: (t(11) = -.790, p = 0.446).

6.3.3 Baseline sleep outcomes Baseline sleep questionnaire and actigraphy scores for the sample are shown in Table 6.4 and 6.5; correlation coefficients between these variables are shown in Table 6.6. From the sample, 67% (n = 12) were close to the poor sleeper threshold (i.e. ≥5). Insomnia symptoms were reported by 28% (n = 5) of athletes in total, with one athlete reporting not being able to fall asleep in <30 min and 17% (n = 3) athletes reporting waking up in the middle of the night or early in the morning more than 3 times per week over the past month. One athlete reported their sleep quality to be ‘poor’ without reporting sleep onset or maintenance symptoms. DBAS-16 scores were the highest for the sleep expectations component of the questionnaire, with 55% (n = 10) scoring >5. Actigraphy data showed that 17% of athletes (n = 3) had SOL ≥30 min, 78% (n = 14) had SE <85%, 100% (n = 118) had WASO >30 min and 11% (n = 2) had TSTs <6 h. PSAS scores revealed higher cognitive component scores than somatic. A small proportion (16%; n = 3) reported receiving advice on sleep management during normal training and (or) competition previously. FIRST scores had moderate correlations with age, PSAS total scores, SE and WASO.

170

Table 6.4: Sleep characteristics of 18 elite field hockey players participating in a 10-week training period.

Age (years) Mean (SD): 25.7 (2.9) Range: 22 - 32

Goal keepers n = 2

Defenders n = 5

Midfield n = 4

Forward n = 7

Questionnaires

PSQI Mean (SD): 4.7 (1.7) Range: 1 – 8

>5 n = 6 (33%)

>8 n = 1 (6%)

FIRST Mean (SD):14.8 (3.7) Range: 9 - 22

>18 n = 4 (22%)

>16 n = 7 (39%)

DBAS Mean (SD): 3 (1) Range: 2 – 5

Consequences Mean (SD): 3 (1) Range: 1 – 6

Worry/ helplessness Mean (SD): 3 (1) Range: 1 – 5

Expectations Mean (SD): 6 (2) Range: 2 – 10

Medication Mean (SD): 1 (1) Range: 0 – 4

PSAS Mean (SD): 26.5 (5.5) Range: 19 – 43

Cognitive Mean (SD):15.4 (4.3) Range: 9 - 24

Somatic Mean (SD): 10.3 (3.2) Range: 8 - 15 Notes: PSQI, Pittsburgh sleep quality index; FIRST, Ford Insomnia Response to Stress Test, DBAS, Dysfunctional Beliefs About Sleep; PSAS, Pre-sleep Arousal Scale

171

Table 6.5 Actigraphy derived sleep patterns of 18 elite field hockey players participating in a 10-week training period. Mean (SD) Range

SE (%) 80.1 (7.0) 59 - 89

SOL (minutes) 17.8 (11.2) 1 - 45

WASO (minutes) 67 (25) 38 - 125

TST (hours) 7.1 (0.9) 5.1 - 8.4

Notes: SE, sleep efficiency; SOL, sleep onset latency; WASO, wake time after sleep onset; TST, total sleep time.

172

Table 6.6: Pearson correlation coefficients between age, sleep questionnaire and actigraphy variables

Age FIRST PSQI DBAS PSAS SOL SE WASO TST

Age - 0.59 -0.183 0.089 0.254 0.31 -0.32 0.12 -0.22

FIRST - - 0.4 0.27 0.67 0.215 -0.56 0.47 -0.44 PSQI - - - 0.286 0.421 -0.12 -0.105 0.3 -0.08 DBAS - - - - 0.081 -0.39 0.13 -0.01 -0.04 PSAS - - - - - 0.103 -0.26 0.16 -0.07 SOL ------0.56 0.049 -0.335

SE ------0.79 0.865 WASO ------0.65 TST ------Notes: Bold type indicates significant at <0.05; PSQI, Pittsburgh sleep quality index; FIRST, Ford Insomnia Response to Stress Test; DBAS, Dysfunctional Beliefs and Attitudes about Sleep; PSAS, Pre-Sleep Arousal Scale; SE, sleep efficiency; SOL, sleep onset latency; WASO, wake time after sleep onset; TST, total sleep time

173

6.3.4 Follow up sleep outcomes

6.3.4.1 Sleep questionnaires Pittsburgh Sleep Quality Index PSQI scores at Baseline, Follow-up 1 and Follow-up 2 are shown in Table 6.7, along with proportions of PSQI ‘poor sleepers’ (PSQI score ≥5) at Baseline, Follow-up 1 and Follow-up 2 shown in Figure 6.2. There was no significant time

effect for PSQI global scores (F(2, 34) = 1.687, p = 0.200, ŋ2 = 0.09). Cochran’s Q test revealed no significant differences in the proportion of poor sleepers (PSQI global

score >5) over time (Baseline = 28%, Follow-up 1 = 17%; Follow-up 2 = 22%; X2(2) = 0.857, p = 0.651). However, using a less conservative cut-off of ≥5 comparisons of distributions

revealed a significant effect over time (X2(2) = 7.09, p = 0.029). Post-hoc comparisons with a Bonferroni correction (adjusted p = 0.02) revealed no significant difference in distributions between Baseline (67% n = 12) and Follow-up 1 (28%, n = 5, p = 0.04) distributions, Baseline and Follow-up 2 (39%, n = 7, p = 0.125) nor Follow-up 1 and Follow-up 2 distributions (p = 0.668). Excluding athletes that scored <4 on PSQI (remaining sample n = 12), there was not a significant time effect for global PSQI

scores (F(2,22)=3.04, p=0.068, ŋ2 = 0.217).

100 90 80 70 60 50 40 30 20

% distribution of poor sleepers poor of distribution % 10 0 Baseline Follow-up 1 Follow-up 2

Figure 6.2: Distribution of poor sleepers (defined by scores of ≥5 on Pittsburgh Sleep Quality Index) at Baseline, and 6 weeks (Follow-up 1) and 10 weeks (Follow-up 2) post baseline (n=18)

Cochran’s Q test (X2(2) = 7.09, p = 0.029)

174

Dysfunctional Beliefs and Attitudes about Sleep DBAS scores at Baseline, Follow-up 1 and Follow-up 2 are shown in Table 6.7 and Figure 6.3. There was a significant time effect for DBAS scores (F(2,34) = 8.821, P = 0.01, ŋ2 = 0.323). Post-hoc comparisons revealed Follow-up 2 scores were significantly lower than Baseline scores (p = 0.01, d = 0.87) and Follow-up 1 scores (p = 0.02, d = 0.83). Separate component analyses revealed a significant time effect for the “worry/helplessness about sleep” component (F(2,34) = 9.949, p <0.0001). Post-hoc comparisons revealed Follow-up 2 scores for this component were significantly lower than Baseline scores (p = 0.01, d = 0.99) and Follow-up 1 scores (p = 0.01, d = 0.90). There were no significant time effects for “sleep expectations” (F(2,34) = 2.678, p = 0.083, = 0.136), “perceived consequences of poor sleep” (F(2,34) = 0.206, P = 0.144) and “medication” (F(2,34) = 0.271, p = 0.081) component scores. There was a significant time effect for the “sport performance concern” question (F(2,34) = 4.04, p = 0.027, ŋ2 = 0.191). However, post-hoc comparisons (corrected p = 0.02) revealed that both Follow-up 1 scores (p = 0.035, d = 0.54) and Follow-up 2 scores (p = 0.031,d = 0.54) were not significantly lower than Baseline scores.

175

4.5

4.0 *

3.5

3.0

2.5

2.0 DBAS scores DBAS 1.5 1.0

0.5

0.0 Baseline Follow-up 1 Follow-up 2

Figure 6.3: Mean Dysfunctional Beliefs and Attitudes about Sleep (DBAS) scores at Baseline, and 6 weeks (Follow-up 1) and 10 weeks (Follow-up 2) post baseline (n=18).

*Significantly different to Baseline (p<0.01).

One-way ANOVA (F(2,22)=4.885; P=0.018; n2=0.306)

Pre-sleep arousal scale Pre-sleep arousal total, cognitive and somatic scores at baseline, Follow-up 1and Follow-up 2 can be seen in Table 6.7. There was no significant time effect for PSAS total (F(2,34) = 0.425, p = 0.657, ŋ2 = 0.024), PSAS cognitive component (F(2,34) = 0.488, p=0.618, ŋ2 = 0.028) nor PSAS somatic component scores

(F(2,34) = 0.390, p=0.680, ŋ2 = 0.022).

176

Table 6.7: Sleep questionnaire scores at Baseline, Follow-up 1, and Follow-up 2 intervention: score values are mean (SD) Baseline Follow-up 1 Follow-up 2 Questionnaires p (n=18) (n=18) (n=18)

PSQI 4.7 (1.7) 4.1 (1.6) 4.1 (1.9) 0.218*

>5 (%) 27% 17% 22% 0.651**

>4 (%) 67% 28% 39% 0.029**

DBAS 3.2 (0.9) 3.0 (1.1) 2.4 (1.2) 0.001*

Expectations 5.9 (2.1) 5.3 (1.8) 4.9 (2.0) 0.083*

Worry/ 3.0 (1.0) 2.7 (1.3) 2.0 (1.4) 0.001* helplessness

Consequences 3.0 (1.3) 2.7 (1.4) 2.3 (1.9) 0.114*

Medication 1.4 (1.3) 1.4 (1.4) 0.7 (0.9) 0.081*

Sport 4.1 (2.9) 2.4 (2.4) 2.8 (2.7) 0.027* performance

PSAS 25.7 (4.1) 24.7 (7.0) 24.3 (7.0) 0.657*

Cognitive 15.4 (4.3) 14.3 (5.2) 14.4 (4.4) 0.619*

Somatic 10.3 (2.3) 10.3 (2.3) 9.8 (2.6) 0.685* Notes: PSQI, Pittsburgh Sleep Quality Index; DBAS, Dysfunctional Beliefs and Attitudes about Sleep; PSAS, Pre-Sleep Arousal Scale. * repeated measures ANOVA; ** Cochran’s Q test (p<0.05)

177

6.3.4.2 Actigraphy Mean values for the actigraphy variables SE, SOL and WASO at Baseline and Follow- up 1 are shown in Table 6.8 and Figures 6.4, 6.5 and 6.6 respectively. Paired samples t- tests revealed that SE values were significantly higher (t(17) -2.41, p = 0.028, d = -0.55), and SOL (t(17) = 3.241, p = 0.005, d = 0.75 ) and WASO (t(17) = 2.491 = 0.023, d = 0.59) values were significantly lower at Follow-up 1 compared to Baseline. However, TST values at Baseline and Follow-up 1 did not differ significantly (t(17) = 0.58, p = 0.578, d = 0.13).

90 * 85

80 75

(%) efficiency Sleep 60

55 50

BaselinePRE FollowPOST1-up 1

Figure 6.4: Sleep efficiency at Baseline and 6 weeks (Follow-up 1) post baseline (n=18).

Paired samples t-test (t 17) -2.41, p = 0.028, d = -0.55)

* significantly different from Baseline (p<0.05)

178

20 * 15

10 Sleep onset latency (min) latency onset Sleep 5

0 PRE POST1 Baseline Follow-up 1

Figure 6.5: Sleep onset latency at Baseline and 6 weeks (Follow-up 1) post baseline (n=18).

Paired samples t-test (t(17) = 3.241, p = 0.005, d = 0.75)

* significantly different from Baseline (p<0.05)

179

100

90 * 80

Waketime ater sleep onset (min) 20

0 PRE POST1 Baseline Follow-up 1

Figure 6.6: Wake time after sleep onset at Baseline and 6 weeks (Follow-up 1) post baseline (n=18).

*Paired samples t-test (t(17) = 2.491,p = 0.023, d = 0.59)

* significantly different from Baseline (p<0.05)

6.3.4.3 Self-reported sleep symptoms Sleep symptom variables are shown in Table 6.8, with MD ratings (“How is your mood today?”) compared for Baseline and Follow-up 1 in Figure 6.7. There were no significant differences between Baseline and Follow-up 1 ratings of FROA (“How refreshed did you feel on awakening this morning?”; Z = -1.604, p = 0.109, d = -0.38) and SQ (“How well did you sleep last night?”; Z = 1.318, p = 0.187, d = 0.33) scores. There was, however, a trend for ratings of ETFA (“How easy was it to fall asleep last night?”) to be higher at Follow-up 1 compared with Baseline (z = 1.897, P = 0.058, d = 0.45). Ratings of MD were significantly lower at POST compared to Baseline (z = -1.903, p = 0.047, d = -0.46).

180

10 9 * 8

7 10) - 6 5

Mood ( Mood 4 3 2 1 0 Baseline Follow-up 1

Figure 6.7: Ratings of mood (MD) for the Baseline and Follow-up 1 weeks.

Wilcoxon test (Z = -2.1, p = 0.039, d = 0.61). Blocks = Median, error bars = maximum and minimum values

*Significantly different to baseline (p<0.05)

6.3.4.4 Physiological assessment Physiological assessment outcomes for HRex and sAA are shown in Table 6.8. There

was no significant difference between baseline and Follow-up 1 values of HRex (t(17) =

0.486, p = 0.633, d=0.11) and sAA (t(17) = -0.112, p = 0.912, d=0.02).

181

Table 6.8: actigraphy, sleep diary and day-time arousal variables pre- post intervention

Baseline Follow-up 1 p d (n=18) (n=18)

Actigraphy

SOLa 18 (11) 10 (5) 0.005§ 0.75

>30 min 17% 0 0.250Ŧ

SEa 80 (7) 82 (5) 0.033§ 0.55

≤85% 78% 61% 0.250Ŧ

WASOa 67 (25) 60 (20) 0.023§ 0.59

>30 min 100% 100% 1.0Ŧ

TSTa 7.1 (0.6) 7.0 (0.8) 0.573§ 0.13

<6 h 11% 11% 1.00Ŧ Sleep symptoms

STSTa 8.1 (0.6) 8.0 (0.4) 0.673҂ 0.12

ETFAb 7.0 (4.0) 7.5 (3.0) 0.058҂ 0.44

FROAb 6.0 (3.0) 6.0 (3.0) 0.109҂ 0.37

SQb 7.0 (4.0) 7.0 (4.0) 0.187҂ 0.31

MDb 7.0 (5.0) 7.0 (4.0) 0.047҂ 0.46 Physiological assessment

HRexa 117 (7) 116 (14) 0.633§ 0.11 sAAa 66.8 (49.0) 68.1 (45.0) 0.912§ 0.03 Notes: SOL, sleep onset latency; SE, sleep efficiency; WASO, wake time after sleep onset; TST, total sleep time; STST, subjective total sleep time; ETFA, ease to fall asleep; FROA, freshness on awakening; SQ, sleep quality; MD, subjective ratings of mood; HRex, exercising heart rate; sAA, saliva alpha amylase; d, Cohen’s d; a mean (SD) values; b median (range) values; § paired samples t- test; Ŧ McNemar’s test; ҂ Wilcoxon signed rank test

182

6.3.5 Programme Evaluation The athletes’ evaluation of the sleep management programme is shown in Figure 6.8. Overall, 89% (n=16) stated that they would use the information presented during the programme in the future, and 100% (n=18) stated they would feel confident using this information in a competition scenario. Following the programme, 22% (n=4) stated that they still feel unequipped to deal with sleep disturbances when they arise.

Maybe 5 No, not at all No, not really Yes, somewhat 4 Yes, definitely

Question number 2

-100 -50 0 50 100 Distribution of responses (%)

1 “Did you find the material presented relevant?”;

2 “Do you think you will use the information provided in future?”;

3 “Do you feel you are unequipped to deal with sleep disturbances when faced?”

4 “Would you recommend this workshop series to a fellow athlete from another sport?”

5 “Would you feel confident in using this information in a competition scenario?”

Figure 6.8: Evaluation of sleep programme

183

6.3.6 Programme utility Differences in the distributions of players adopting sleep promoting activites prior to sleep onset, and adopting pre-sleep de-arousal techniques are seen in Figures 6.9 and 6.10. There was a significant increase (Z=2.8, p=0.005) in players adopting technqiues to fall asleep from baseline (5/18) to follow up 2 (11/18), with all players adopting sleep management (n=11) stating their respective techniques were effective at follow up 2. Five players reported effectiely using elements of stimulus control at follow up 2 (mainly “getting out of bed when unable to sleep, and not returning until sleepy”), and 3 players reported using articularly supression to good effect. The remaining players used “reading” and “breathing exercises”.

20 Yes No * 18

16 14

12 10

8 Number of players of Number 6 4 2

0 BaselineBsaeline Follow up 1 Follow up 2

Figure 6.9 distribution of players performing sleep promoting activities prior to sleep onset (n=18)

Cochron’s Q (X2(2)=7, p=0.03)

*Significant differences between baseline and Follow up 1 (McNemar Post-hoc, Z=2.6, p=0.009)

184

20 Yes No * 18 16

14 12 10

8 6 players of Number 4

2 0 BaselineBsaeline Follow up 1 Follow up 2

Figure 6.10 distribution of players performing pre-sleep de-arousal techniques (n=18)

Cochron’s Q (X2(2)=8.9, p=0.012)

*Significant differences between baseline and Follow up 2 (McNemar Post-hoc, Z=2.8, p=0.005)

6.4 Discussion The present study was designed to assess the feasibility, perceived utility and acceptance of a group-based sleep management programme in elite team sport athletes. Overall, the results from the study allow for positive responses to each of the 3 research questions set out in Para. 6.1.3. The feasibility of the programme is supported by its successful delivery within the context of a routine elite training period. The perceived utility of the programme is reflected in the finding that most athletes stated that the programme material was relevant, that they now felt equipped to manage sleep disruption when it arises, and that they would be confident in using the material in a competition scenario. Evidence that players were adherent can be seen in the sleep behaviours adopted, with greater proportions of players adopting sleep promotion and techniques to help fall asleep. Furthermore, evidence of positive sleep impact was supported by improvements in all actigraphy derived measures of sleep quality and a reduction in dysfunctional beliefs and attitudes towards sleep. Despite no changes in actigraphy derived TST, there were significant reductions in SOL and

185

WASO, however WASO scores at follow up remained above 30 min, and the proportion of players scoring >30 min remained at 100% following intervention. Moreover, there was a modest increase in SE with follow up scores below 85%, and over half of players scoring <85%. Sleep questionnaires revealed a significant medium reduction in total DBAS scores 1-month post, with components of “worry/ helplessness” and “medication” showing a similar reduction. Finally, while there was no significant change in group mean PSQI global scores, there was a significant reduction in the proportion of ‘poor sleepers’ (i.e. those scoring ≥5 on the PSQI).

6.4.1 Cognitive behavioural approaches (hypothesis 1 and 2) A broad conclusion that can be drawn from these results is that cognitive behavioural approaches can be employed in elite sport contexts, and within individuals who report poor sleep quality, without necessarily having a sleep disorder. As a result, the null hypothesis that is impractical to deliver the psychoeducational sleep management programme in the context of a routine training period can be rejected. Whilst athletes with sleep disorders cannot be ruled out, here, scores on PSQI indicate relatively low levels of poor sleep quality overall (e.g. only 28% athletes scoring >5 and two athletes scoring ≥8). Nonetheless, previous research has shown efficacy in cognitive behaviour approaches among groups with poor ‘sleep health’ (see (Murawski et al., 2017), with a medium effect size being reported for PSQI total scores (Hedge’s g = -0.54) in these groups collectively. In studies investigating such approaches in young adults with poor sleep quality, then (e.g. non-athlete students), cognitive behavioural approaches have been shown to be effective in improving sleep quality. Gao et al. (2014) reported Chinese University students with poor sleep quality (as defined by >7 on PSQI) had significant reductions in PSQI scores following 3 x 45 min group-based sessions on consecutive days after 1 month. Using a less conservative cut-off of ≥5, Hershner and O’Brien (2018) have recently reported significant reductions in PSQI scores and increases in levels of sleep awareness following a 20-minute online sleep education programme among college students. Among elite athletes, studies employing cognitive behavioural approaches are fewer, and findings are mixed. In a recent study, Driller et al. (2019) reported medium reduction in PSQI scores following sleep psychoeducation (d = -0.71) in elite cricket athletes. Moreover, in this study very large improvements were observed in actigraphy derived SE (d = 1.38) and SOL (d = -0.81). Interestingly, the magnitude of

186

change observed in the current study for actigraphy derived variables is like that reported by Driller et al. (2019)with elite male team sport athletes. 6.4.2 Sleep beliefs and attitudes While reductions in DBAS scores would be expected following a psychoeducational programme with an emphasis on transferring factual sleep information, it is important to recognise that this improved knowledge and understanding can directly impact sleep quality. Dysfunctional beliefs about sleep have been shown to be a component of models explaining the aetiology of insomnia (see Harvey (2002)), particularly the maintenance of the disorder (Morin et al., 2007). It has been suggested disruptive sleep cognitions are higher in poorer sleepers when compared to good sleepers (Morin et al., 2007). As a result, targeting disruptive sleep cognitions among poor sleepers is a key component of cognitive therapies and has been shown to mediate improvements in sleep quality (Edinger et al., 2001). However, no research exists targeting sleep cognitions among elite athletes. Nonetheless, the DBAS values reported here were similar to those reported in a recent study of college students with poor sleep quality (Jin, Zhou, Peng, Ding, & Yuan, 2018). The present reduction in DBAS scores is in line with previous research showing improvements in DBAS with cognitive behavioural intervention. This finding indicates that the athletes’ cognitions were reduced 1-month post intervention and that the mode of delivery was effective. In the current study there was a significant reduction in the “worry/helplessness” DBAS sub-component. Interestingly, while athletes reported high values in the “sleep expectations” sub-component of DBAS, the reduction observed didn’t reach a statistical level of significance (p = 0.08). Players’ attitudes around concern of sleep impairing sport performance were reduced. Whilst athletes reported no use of medication, their attitude towards using sleeping pills were changed as shown by a reduction in the DBAS score for this component, but this again didn’t reach significance (p=0.081). The delayed significant reduction in DBAS scores (emerging at Follow-up 2) is difficult to explain. However, despite no significant change in DBAS scores at Follow-up 1, change scores (Baseline minus Follow-up 1 scores) in DBAS were significantly related to change scores in SOL (r = -0.486), SE (r = 0.609) and WASO (r = - 0.612). 6.4.3 Sleep programme utility and value (hypothesis 3) Generally, most athletes valued the psychoeducational sleep management programme during a period of training, and therefore the third null hypothesis can be

187

rejected. From the athlete sample, 100% stated they were confident in using the information during competition (see Figure 6.8). This finding highlights the high usability of the programme content, and that translated cognitive-behavioural principles are applicable to elite athletic populations. Although athletes stated they felt confident in using the information, further research is required to see whether athletes adopt sleep related information and cognitive-behavioural techniques during a period of elite competition e.g. when the likelihood of experiencing disrupted sleep is high (see Chapter 1). However, 22% of athletes felt they were unequipped to deal will sleep disturbances. The timing of sleep management programme could provide a possible explanation for this finding. Given the programme was delivered during a period of training, it is possible that some athletes’ sleep was not challenged by the ‘trio of challenges’, and therefore did not employ any sleep management during the study period. This is supported by a significantly greater number of players (61%, n=11) adopting techniques to help fall asleep at follow up 2, but not at follow up 1 (33%, n=6). Few studies have evaluated sleep interventions in this manner within the sport science and medicine literature. However, in a single study by Tuomilehto et al. (2016) following ‘sleep counselling’, which focussed on “basic sleep principles and how to basic sleep principles and how to promote good sleeping by adopting healthy, regular sleep routines” (Tuomilehto et al., 2016, pp. 3 ), 83% (total n = 23) of professional ice hockey players found the counselling was beneficial. Considering the programme evaluation findings in the current study and those in Tuomilehto et al.’s (2016) study, further research is required to see whether cognitive-behavioural sleep management is perceived to support athlete wellbeing and attainment of peak athletic performance.

6.4.4 Limitations The study had several limitations. Firstly, a control group was not employed and therefore the evidence of programme efficacy is limited and weak. The decision was made not to employ a control group due mainly to the timing of the Olympic cycle. The study took place during the year of Rio 2016 Olympic Games, and the sport was going through a process of selection for final squad members for the Olympic games. As a result, to provide some athletes with one intervention and another group with a different intervention, was not considered acceptable by senior coaches and support staff. Nonetheless, the study provides preliminary evidence of efficacy and feasibility for the management programme in elite sport. The limitations of the ‘pre-post’ design

188

are clearly illustrated by the self-reported sleep symptom results, where self-reported ease to fall asleep (ETFA) increased but mood (MD) significantly decreased following the intervention. This reduction in mood may have been be influenced by the demands of physical training over the training period. Athletes taking part in the study were undergoing an intense physical training block, where, although no hockey was being played, other aspects of training were being performed. Efforts were made to control for the influence of training load on sleep outcomes, with no differences (in training load) being observed between observation periods. However, it is possible there may have been an accumulative effect of physical training on wellbeing over the course of the 10-week block, with athletes reporting lower mood at the end of the block of training when compared to the start. In relation to the point above, while the selection of participants was based on criteria within one squad of athletes (i.e. n=20), the eventual sample size was, nonetheless, modest (n=18) and therefore the efficacy of the programme on mean sleep outcomes could be regarded as unstable. However, due to the size of squad within the sport that volunteered to take part in the study a maximum possible sample size would have been 20.

It is interesting to note, that despite most athletes reporting good sleep quality, according to the PSQI poor sleeper threshold of >5 (e.g. 28%), improvements in sleep quality were observed in sleep outcomes. With a less conservative poor sleeper cut- off being employed (e.g. ≥5), moderate improvements in actigraphy, sleep diary and sleep questionnaire outputs were shown. As a result, it can be theorised that larger effects may be seen in a sample of athletes with poorer sleep. To support this, the athlete with who scored 8 on the PSQI at Baseline recorded a score of 6 at Follow-up 2. Moreover, sleep reactivity scores among the squad were also relatively low, with few reporting scores >18. Whether greater improvements would have been seen for more reactive sleepers requires further investigation. The application of the findings of the current study within elite sport should proceed with caution. It should be noted that while the programme was conducted on one squad of male hockey athletes during a time when, arguably, sleep was unaffected by the ‘trio’ of challenges identified in Chapter 1 – training, competition and travel (see Para 1.3.5).

189

6.5 Conclusion The present study provides evidence that a psychoeducational sleep management programme reflecting cognitive behavioural principles, and supported by practical resources, can be successfully delivered to elite team-sport athletes in the context of a routine training period. The results also provide an indication that the programme is valued by athletes, sleep behaviours are adopted and that post-intervention changes in sleep-related variables are consistent with positive sleep outcomes, albeit modest. Since the ASMP was designed to help athletes meet the trio of sleep challenges identified in Chapter 1 (training, travel and competition), the present study offered an important, but limited test of its utility. Additional tests of the ASMP under real-world conditions of travel and competition are necessary to confirm its value as a feasible and effective intervention in elite sport.

190

Chapter 7

7. Delivering group-based psychoeducational sleep management to elite team sport athletes: an open label trial

7.1 Introduction The previous chapter showed that a group-based sleep management programme reflecting cognitive behavioural principles is a feasible intervention when delivered in the context of routine training. The programme was valued by athletes and was effective in reducing both insomnia symptoms and negative attitudes and beliefs towards sleep. While the team athletes who participated in the feasibility study expressed confidence that they would use the information gained from the programme “…in a competition scenario”, the value of the programme during this period is untested. This chapter considers the impact of a modified version of the ASMP delivered in the context of elite competition and international travel to female team sport athletes.

7.1.1 Sleep management during competition The sleep management strategies adopted by athletes during competition has not been documented widely in the sport science and medicine literature. Erlacher et al. (2011) reported that 57% athletes had no strategy to help manage sleep disruption during major competitions in national German athletes. This was replicated by Juliff, Halson and Peiffer (2015) who found that 52% of elite Australian athletes reported having no “sleep strategy” for the nights before major competitions in the lead-up to the London 2012 Olympic Games. Both findings accord with results from the survey reported in Chapter 3, which showed that 69% of elite British athletes have no sleep management strategies. However, Juliff, Halson and Peiffer (2015) also noted between-sport differences in the deployment of sleep management strategies, with 59% of team sport athletes reporting ‘no’ sleep strategy compared to 32% of individual athletes. Erlacher et al. (2011) also reported a tendency for female athletes to be more likely than male athletes to report pre-competition sleep disturbances (though this “tendency” did not achieve significance). Such a ‘tendency’ is in line with results from the introductory review (Chapter 1) which showed that, across the competitive

191

cycle, the prevalence of sleep disturbance was significantly higher among women (Lucidi et al., 2007; Schaal et al., 2011). Collectively, these findings focus attention on the sleep management needs of female team-sport athletes. It is also the case, however, that while competitions have been widely reported to negatively impact athlete sleep, this impact has been mainly demonstrated using within-group pre-post designs (where sleep during competitions is compared with on sleep rest days or during normal training; see Chapter 1; para. 1.3.5.1). Since team-sport athletes are generally selected for competitions from a pool of eligible players, those ‘not’ selected provide a natural control group allowing for between group comparisons.

7.1.2 Sleep and competition performance To date few studies have examined linkages between sleep quality and athletic performance during major competitions, and those that have (Chennaoui et al., 2016; LE Juliff et al., 2015; M.-R. R. G. Silva & Paiva, 2016) provide little evidence of a linear relationship between poor sleep quality and degraded competitive performance. In the study by Juliff, Halson and Peiffer (2015) small proportions (13-17%) of all athletes reported worse competitive performance due to poor sleep following the London 2012 Summer Olympics Games. However, it was also found that a greater proportion of team-sport athletes reported episodes of increased day-time sleepiness on competition days (48%) compared to individual athletes (27%). Complex inter- relationships between competitive performance, sleep quality and expressed daytime sleepiness have been reported elsewhere. Using finishing position as the main outcome, Silva and Paiva (2016) reported that elite gymnasts with poorer sleep quality tended to finish higher in the rankings when compared to good sleepers during a FIG (Fédération Internationale de Gymnastique) competition. Nevertheless, in the same study it’s was also found that high levels of day-time sleepiness (as assessed by the Epworth Sleepiness Scale) were associated with lower rankings. One possible explanation for these findings is that higher levels of arousal could mediate both poorer sleep quality and superior gymnastic performance ‘and’ mitigate the experience of daytime sleepiness (as suggested by the findings of Chapter 4, Para. 4.3.3). To date, however, no study has systematically examined the impact of a sleep management programme during periods of elite competition. The present study, therefore, was designed to further explore sleep-performance relationships in elite athletes and extend the findings of Chapter 6 by assessing the impact of a

192

psychoeducational sleep management programme (a modified ASMP) delivered during a period of elite international competition.

7.1.3 Research questions and hypotheses The research questions of the present study were:

1. How feasible is it to deliver psychoeducational sleep management programme (based on cognitive-behavioural principles) to elite team players preparing for a period of elite competition? 2. What is the impact of a psychoeducational sleep management programme (based on cognitive-behavioural principles) on player sleep, wellbeing, and performance during a period of elite competition?

These research questions were addressed by testing the following null hypotheses:

1. Where competition teams are selected from a larger pool of athletes, measures of sleep quality and sleep quantity will show no significant differences between the selected and unselected players. 2. In comparisons of selected and unselected players, participation in elite international competitions will not significantly degrade sleep quality and quantity among the selected. 3. Among elite team-sport athletes, a psychoeducational sleep management programme (based on cognitive-behavioural principles) will not mitigate the impact on sleep of competing in international competitions. 4. When delivered during a competition preparation period, a psychoeducational sleep management programme will not be valued and utilised by elite team-sport athletes.

7.2 Methods The study was conducted in collaboration with the Great Britain female rugby sevens team during the 2015/2016 World Rugby Sevens Series (WRRS).

7.2.1 Competition format

Typically, the Women WRSS requires international teams to compete across 6 tournaments, in 6 different countries, over a period of 6 months. Specifically, the

193

season consists of 3 sets (called competitions) of two tournaments. Tournaments take place every 5-6 days and competitions take place within ~4 weeks of each other. For the 2015-2016 season, however, only 2 competitions containing two tournaments were held with one competition only containing one tournament. This was due to Rugby Sevens’ debut at the 2016 Summer Olympic Games in Rio de Janeiro at the end of the season. The season championship is determined by points earned in each tournament with first place (cup winners) receiving 20 points, second place (cup runner-up) receiving 18 points and third place (Cup third place play-off winner) receiving 16 points. Teams finishing below these rankings receive 2 points less for each lower ranking position. Each tournament is played over 2-3 days, with up to three matches per day and ~3 h between matches. Matches adhere to modified World Rugby rules on a standard Rugby-15’s size pitch with only seven on-field players on each team, contested over 2 x 7 min halves with a 2 min half-time interval.

7.2.2 Competition samples The three tournaments included in the present study, together with training schedules, are shown in Figures 7.1 and 7.2. These tournaments, which consisted of 60% (3/5) of all tournaments during the WRSS 2015/2016 season, were Atlanta (USA) and Langford (Canada), which took place within one week of each other, followed by Clermont- Ferrand (France) one month later. These were the final two competitions prior to the 2016 Rio de Janeiro Summer Olympic Games.

7.2.3 Participants The Great Britain female rugby sevens team agreed to participate in the study. In the only previous study of rugby seven players during competition, Fowler et al. (2017) reported changes in self-reported sleep quality ratings during four tournaments in Scottish national team male players. The authors in this study also reported high levels of variability in sleep outcomes between individuals. In the present study a sample of 23 British female elite international Rugby Sevens players from a single squad were recruited as participants. The athletes were regarded as ‘world class elite’ according to Swann’s full taxonomy of ‘eliteness’ (Swann et al., 2015), with a median score of 7. On average, players had 5 years (SD=2) of playing experience at an international level with 60% reporting winning a World Rugby Sevens Series (WRSS) as their highest level of achievement.

194

7.2.4 Study design A schematic of the study design is presented in Figure 7.1. Because of the constraints placed upon elite competition squads by coaches and managers, a ‘natural experiment’ approach was adopted simulating a longitudinal ‘Pre-Post Quasi- Experimental’ design (Gilner, Morgan, & Leech, 2009). From the total group of 23 eligible players, squads of 14 and 12 were selected by team management for the first and second WRSS competitions respectively (see Figure 7.1). For the purposes of the present study, selected squads were designated COMP while the remaining athletes, who continued training throughout the competition phases, were designated TRAIN. All players were assessed at Baseline (Week 1) for the seven days prior to the COMP squad departing for Competition 1 (Figures 7.1 and 7.2). COMP players then entered a competition period for two weeks, while TRAIN players continued training at the team’s home base in Surrey, UK. Competition 1 consisted of two tournaments (Figures 7.1 and 7.2), during which players were monitored using actigraphy, and completed self-reported wellbeing diaries. On return from Competition 1, all players were profiled again with questions asked retrospectively regarding how they slept during the competition. During the month following Competition 1, a sleep management intervention was implemented. Immediately after the two-week sleep management intervention was completed, a second squad selection took place before travelling to Competition 2. During this period the monitoring protocol from Competition 1 was repeated (see Figure 7.2). For the purposes of this study weeks 1-3 (including the Baseline, Competition 1, and the return of competing athletes to the Surrey home base) is designated Phase 1, while weeks 4-10 (which included Competition 2) are designated Phase 2. A separate, and additional, ‘programme evaluation questionnaire’ was employed during Phase 2 to obtain player feedback regarding the sleep programme content and impact of the information presented prior to Competition 2.

7.2.5 Baseline and follow up Questionnaires

7.2.5.1 Sleep questionnaires 1 The Pittsburgh Sleep Quality Index: A description of the Pittsburgh Sleep Quality Index (PSQI) can be seen in the general methods section (Chapter 2, Para. 2.4.1). A poor sleeper threshold of >5 was set.

195

2 The Pre-Sleep Arousal Scale: A description of The Pre-Sleep Arousal Scale (PSAS) is provided in the General Methods section of the thesis (Chapter 2, Para. 2.4.4). 3 The Ford Insomnia Response to Stress Test: A description for the Ford Insomnia Response to Stress Test (FIRST) can be seen in the general methods section (Chapter 2, Para. 2.4.2). A reactive sleeper threshold of >18 was set. 4 Dysfunctional Beliefs and Attitudes about Sleep Scale: A description for the Dysfunctional Beliefs and Attitudes about Sleep Scale (DBAS) can be seen in the general methods section (Chapter 2, Para. 2.4.3) of the thesis. An additional item was added for the present study: “I am concerned poor sleep will impact my performance”. All items were scored on a 10-point Likert scale from 1 (Strongly Disagree) to 10 (Strongly Agree). 5 Competition, Sport, Sleep and Dreams Questionnaire: The Competitive Sports, Sleep and Dreams Questionnaire (CSSDQ), previously described by Erlacher and colleagues (2011), is a sport specific questionnaire used to assess the sleep habits and dreams of athletes prior to important competitions and games. The questionnaire is divided into four sections: 1) demographic information; 2) sleep disturbances prior to competition; 3) the subsequent impact on performance; and 4) strategies used to mitigate these difficulties. Given the demographics of the athlete group were already known, section one was omitted from the questionnaire. The subsequent sections aimed to obtain information on athlete sleep habits prior to the competitions that recently took place. If an athlete answered “yes” to having poor sleep during tournaments, they were required to complete a further four closed response questions as follows:

Have you had disrupted or fragmented sleep prior to an important competition or game in the last 12 months?

If yes:

What kinds of problems did you experience with your sleep prior to an important competition or game?

1. Problems falling asleep 2. Waking up early in the morning 3. Waking up at night 4. Unpleasant dreams

196

5. Not feeling refreshed in morning

What reasons were responsible for your sleeping problems prior to an important competition or game?

1. Thoughts about competition 2. Nervousness about competition 3. Not used to surroundings 4. Noises in room or outside

In what manner did the sleeping problems influence your performance during the competition or game?

1. No influence 2. Increased daytime sleepiness 3. Bad mood the following day 4. Worse performance in competition

Which strategies did you use to sleep well in the nights preceding a competition?

1. No Strategy 2. Methods to relax 3. Sleeping pills 4. Reading 5. Watching TV

7.2.5.2 Wellbeing Diary Daily assessments of sleep and wellbeing were captured via a bespoke mobile phone application (AER app, English Institute of Sport, Manchester, UK) in the morning prior to each training session and game. Dissimilar to previous chapters, however, the questions selected for this study were from a list of questions used by the sport daily. All items were ranked on a 1-5 scale (low to high). These questions included:

1 “How well did you sleep last night?” (SQ); 2 “How ready do you feel to perform?” (RTP); and 3 “how stressed do you feel today?” (STRESS).

197

7.2.5.3 Programme evaluation questionnaire A purpose designed 7-item questionnaire was used to evaluate the sleep management programme (described in Chapter 2, Para. 2.6.5). Items were rated on a 5-point scale (1=Maybe; 5 = Definitely not) and included:

1. “Did you find the material presented relevant?” 2. “Do you feel you are unequipped to deal with sleep disturbances when faced?” 3. “Have you previously received information about how to manage your sleep during normal training and (or) competitions?” 4. “Would you recommend this workshop series to an athlete from another sport?” 5. “Did you use this information during the WRSS in Clermont-Ferrand?” 6. “Do you think you’ll use this information in the future?” 7. “Do you feel this information had a positive impact on your wellbeing and (or) performance?”

The questions employed to capture sleep related behaviours adopted, and therefore programme utility, were:

1. Sleep promoting activities are activities that are performed before (or at) bedtime to help promote feelings of sleepiness in preparation for sleep, do you regularly perform a sleep promoting activity before bed?

2. Do you use any techniques to help you fall asleep at night (including when wake-up in the night and cannot fall back asleep)? Please described this technique(s)? And;

3. Do feel this technique to help you fall asleep was effective?

4. A nap is a time during the day when you may doze or sleep regardless when you have planned to or not. With this in mind, did you nap in the lead up to or during the World Series competition(s)?

198

7.2.6 Actigraphy Wristwatch actigraphy was employed to assess the sleep patterns of participants. A general description of the actigraphy protocol can be seen in the General Methods section of the thesis (Chapter 2, Para. 2.5.2). In the current study both the Actiwatch 2 (Philips Resperonics®, USA) was employed to collect actigraphy data. The sleep outcomes selected were total sleep time (TST), sleep efficiency (SE), sleep onset latency (SOL) and wake time after sleep onset (WASO).

7.2.7 Tournament and training running patterns Match-play running patterns were collected using a 10Hz global positioning system (GPS) (OptimEye S5, Catapult innovation, Melbourne, Australia). During match-play, players wore a fitted harness, within which a GPS device was placed in a pocket situated between the scapula bones on the players’ backs. All players wore a heart rate monitor for each match (Polar T31, Polar Electro OY, Kempele, Finland) and heart rate was recorded (5 s intervals) continuously, via the GPS unit. Following each match GPS and heart rate data were downloaded and time-motion analysis was performed using OpenField software (Catapult Innovation, Melbourne, Australia). To obtain an indication of the physical demands of training and tournament match-play, duration of match-play time (min), the total distance covered (TD) and average speed (S) were reported. Total distance covered has shown to be reliable and valid, and used to quantify match-play demands in team sports, previously (see (Cummins, Orr, O’Connor, & West, 2013)). To assess whether match-play demands were similar between training and tournament weeks, mean values were compared.

199

Week 1 (Baseline)

Great Britain female rugby sevens squad (n = 23) agree to participate. All athletes assessed at Baseline

Team selection made for Competition 1

Selected for Competition 1 Not selected for Competition 1 (COMP: n = 14) (TRAIN: n = 9) Phase 1

Weeks 2 & 3 Competition 1: Week 2 - Atlanta Weeks 2 & 3 (USA) & Week 3 - Langford Continued Training at Surrey (Canada)

Weeks 4 & 5 All athletes return to Surrey for training: Follow-up sleep profiles obtained (n = 23)

Weeks 6 & 7 Sleep management intervention period

Week 8 Continued training at Surrey

Team selection review for Competition 2

Selected for Competition 2 Not selected for Competition 2 (COMP: n = 12) (TRAIN: n = 9) Phase 2

Week 9 Week 9 Competition 2: Clermont Continued Training at Surrey (France)

Week 10 All athletes return to Surrey for training: Follow-up sleep profiles

Figure 7.1 Structure of the 10-week study period covering 2 competitions in the 2016 Women's World Rugby Sevens Series (Team GB)’

200

Week 1 2 3 4 5 6 7 8 9 10 Location

Surrey, UK

Atlanta, USA 1

Langford, 1 Canada

Clermont, 2 France

1 WRSS Competition 1 (selected squad only); 2 WRSS Competition 2 (selected squad only);

Rugby activity (all 23 players); Sleep Management Programme; Actigraphy; Sleep Questionnaires;

Wellbeing Diary

Figure 7.2 Outline of the 10 week study protocol

201

7.2.8 The Athlete Sleep Management Programme In order to accommodate the sleep sessions within the Great Britain female rugby sevens training and competition schedule, the sleep management programme employed in this study merged the four-session programme of the ASMP (as described and implemented in Chapter 6) into two 60-minute workshops. These two group- based workshops were provided for all squad members (n=23), with each workshop focussing on different components of the sleep management programme (as described in Chapter 2). During the first session attendees were asked “what do you want get out of the sleep support sessions?” and “what areas of sleep would you like to cover?” and this provided additional content to cover over the 2-week period. During each subsequent workshop an opportunity was provided for players to ask questions regarding any concepts, processes and techniques described throughout the programme.

7.2.9 Statistical analysis All continuous data were presented as means (SD) (e.g. actigraphy data). Continuous data that were not normally distributed were treated as non-parametric accordingly. Participants with >50% missing data for each week for actigraphy and wellbeing diary data were discarded (Kölling et al., 2016). Correlations between sleep measures were assessed using Pearson product moment correlation coefficients.

To address hypothesis 1, COMP and TRAIN baseline values were compared as follows. Mean total scores (and, where appropriate, sub-scale scores) from the PSQI, DBAS, FIRST, and PSAS, and mean baseline period actigraphy scores for TST (minutes), SE (percent), SOL (minutes), and WASO (minutes) were compared using independent t-tests. Proportions scoring >5 and 8 on the PSQI, and >18 on the FIRST were compared using Fisher’s Exact Test (FET). The wellbeing diary outcomes STRESS, RTP and SQ were compared using Wilcoxon Signed Ranks test.

To address hypothesis 2, COMP and TRAIN scores from Competition 1 (see Figure 7.1) were compared as follows. For sleep questionnaire (PSQI, PSAS) and actigraphy (TST, SE, SOL, WASO) scores, one-way ANCOVAs (with baseline values for that measure as the co-variate) were conducted. For wellbeing diary outcomes (STRESS, RTP and SQ) COMP and TRAIN values players were assessed using Wilcoxon Signed Ranks test.

202

Hypotheses 3 was addressed from several different perspectives using only outcome data from the COMP group. To assess the impact of the sleep management programme on COMP group athletes, paired samples t-tests were used to compare Competition 1 and Competition 2 sleep questionnaire (PSQI, PSAScog) and actigraphy (SE, SOL, WASO and TST) derived sleep outcomes. For the same two points of measurement (Competition 1 and Competition 2) McNemar tests were employed to assess the distribution of actigraphy outcomes (SE, TST, WASO, SOL) according to the following dichotomous criteria of insomnia symptomatology: SOL >30 min, WASO >30 and SE ≤85%. (Ohayon et al., 2017). A short TST threshold of <6 h was adopted (Hirshkowitz et al., 2015). A paired samples t-test was also used to compare DBAS scores at Baseline and Competition 2.

Hypothesis 4 was addressed using responses from the programme evaluation questionnaire for the COMP group at Week 10 (see Figure 7.1). To judge the programme’s utility specifically responses to closed-questions (i.e. yes/ no) in regard to utilising 1) pre-sleep arousal techniques, 2) sleep hygiene behaviours and 3) daytime napping during competition were assessed at Competition 1 and Competition 2 using McNemar tests. Responses to the question “did the information presented improve your performance and (or) wellbeing?” were used as an indication of impact on sleep-performance relationships.

Effect sizes were estimated with Cohen’s d statistic using the standard criteria: 0.2-0.5, small effect; 0.5-0.8, medium effect; and >0.8, large effect. Similarly, the judged effect size indicated by correlation coefficients was: 0.10 to 0.30, small; 0.30 to 0.50, medium; and >0.50: large. For all analyses alpha was set at p<0.05. All statistics were performed using SPSS for Windows (Version 23.0, SPSS inc., IBM, Armonk, New York).

7.3 Results 7.3.1 Sample At the Phase 2 team selection review (see Figure 7.1) the COMP group was reduced from 14 to 12 players, all of whom had been members of the Phase 1 COMP group. From the three tournaments Team GB finished 3rd, 1st and 4th respectively. Timing of matches ranged from 11:15 to 19:45. Questionnaire (PSQI, DBAS, PSAS) and

203

actigraphy (TST, SE, SOL, WASO) outcomes all showed normal distribution. Wellbeing diary data was not normally distributed and non-parametric analyses were employed. Adherence to the study’s protocols are reported below. The response rate for completing questionnaires was 89% (8/9) for the TRAIN group, and 100% for the COMP group (14/14); completion rates for the Wellbeing Diary were 100% for both the TRAIN and COMP players. At Phase 1 (Figure 7.1), adherence to actigraphy was good with 100% COMP players providing data, and >50% completion for TRAIN. Within the COMP group adherence to the actigraphy protocol during Competition 1, week 2 (Atlanta) (see Figure 7.1) was 100% (14/14). Adherence to the protocol in Competition 1, week 3 (Langford) was 71% (12/14) providing data for >50% of the competition period. In the TRAIN group, adherence to actigraphy during Competition 1, Week 2 (Atlanta) was 89% (8/9) providing data for >50% of the competition period. However, for Week 3 (Langford) adherence in the TRAIN group was 22% (2/9).

During Phase 2 (Figure 7.1) 100% (12/12) of COMP players and 78% (n = 7/9) of TRAIN players completed wellbeing diaries with >50% of the competition periods. In COMP players, Phase 2 questionnaire response rates were 100% (n =12/12) while only 33% (n = 3/9) of TRAIN players returned completed data. Due to the poor adherence from TRAIN players, only the first tournament week was employed in the baseline- controlled comparison of TRAIN and COMP players to test hypothesis 2, and only COMP players (n=12) were employed in pre-post ASMP comparisons to test hypothesis 3 and 4.

7.3.2 Running patterns During Phase 1, TRAIN players covered an average distance of 8,433 m/week (SD=3844 m; range=3,337-12,225 m). This distance was achieved over an average of 3 rugby sessions/week (SD=1; range=1-4 sessions). For COMP players, the group covered an average of 6,144 m/week (incl. competitions in Atlanta and Langford) and this distance was achieved over an average of 63 min (SD=23 min) of tournament match-play. During Phase 2, players completed a mean distance of 5,455 m (SD=1430 m; range=3228-7936 m). This was achieved over a mean of 64 min tournament match- play (SD=19 min). There was a significant difference in mean speed between Phase 1

(98 m/min; SD=6 m/min) and Phase 2 (85m/min; SD=9 m/ min) phases (t(10)=4.36; p=0.002).

204

7.3.3 Baseline sleep characteristics Baseline sleep characteristics for all players returning assessments (n = 23) are presented in Table 7.1 and 7.2; correlation coefficients calculated for actigraphic and questionnaire sleep variables are shown in Table 7.3. From the entire squad (n = 23), the mean global PSQI score was 6 (SD = 3) with 48% (n = 11) scoring >5 on PSQI scores, and 22% (n = 5) scoring ≥8. The mean FIRST score was 21 (SD = 3) with 82% players scoring >18. The highest DBAS component score was ‘Consequences’ (mean = 6, SD = 2) while the lowest was ‘Medication’ (mean = 2, SD = 3). In response to the statement “I am concerned poor sleep will impact my performance” players average rating was 7 (SD = 3), indicating strong agreement with that item. This rating was similar between the COMP (mean = 8, SD = 2) and TRAIN groups (mean = 6, SD = 4, t(20) = -1.069, p = 0.298). The cognitive component of PSAS was higher (mean = 17, SD = 8) than the somatic component score (mean = 11, SD = 2).

Across the Baseline assessment week actigraphy derived sleep outcomes showed: mean TST values of 6.9 h (SD = 0.9 h): mean SOL values of 11 min (SD = 0.9 min): mean SE values of 81 (SD = 7)%; and mean WASO values of 70 (SD = 32) min. There was a trend for an association between low SE and group membership, with 21% of COMP players, but 88% of TRAIN players reporting an SE of <85% (p =0.054). There were significant positive correlations between PSAS total scores and total FIRST scores (r = 0.45, p<0.05) and global PSQI scores (r = 0.66, p <0.05). DBAS scores showed a significant positive correlation with actigraphy derived SOL (r = 0.46, p <0.05; Table 7.2).

205

Table 7.1 Baseline sleep characteristics for the selected (COMP, n = 14) and non- selected (TRAIN, n = 9) players for a world rugby sevens series competition ALL (n=23) COMP (n=14) TRAIN (n=9) Mean Mean Mean p (SD) Range (SD) Range (SD) Range PSQI 6 (3) 1 - 15 6 (3) 1 – 12 7 (4) 4 - 15 0.65

50% Na 50% Na 50% Na >5 1.00

22% Na 21% Na 22% Na ≥8 1.00

FIRST 21 (3) 15 - 26 20 (3) 15 - 26 23 (3) 19 - 29 0.06

>18 (%) 77% Na 64% Na 100% Na 0.12

DBAS 4 (2) 2 - 8 4 (2) 2 – 8 5 (2) 2 - 8 0.58

Worry/ help 4 (2) 1 - 8 4 (2) 1 – 8 4 (2) 2 - 7 0.75

Expectations 6 (2) 1 - 9 6 (3) 1 – 9 6 (2) 3 - 9 0.81

Medication 2 (3) 0 - 9 2 (3) 0 – 9 3 (3) 0 - 9 0.67

Consequences 5 (3) 1 - 10 6 (3) 1 – 10 5 (2) 1 - 8 0.56

PSAS 28 (9) 19 - 52 30 (10) 19 - 52 26 (9) 19 - 49 0.39

Cognitive 17 (8) 8 - 38 19 (9) 10 - 38 16 (7) 8 - 36 0.43

Somatic 11 (2) 8 - 14 11 (2) 8 -14 11 (2) 8 - 13 0.45 Notes: COMP selected players; TRAIN non-selected players; PSQI Pittsburgh Sleep Quality Index; FIRST Fords Insomnia Response to Stress Test; DBAS dysfunctional beliefs about sleep; TST total sleep time; SOL sleep onset latency; SE sleep efficiency.

206

Table 7.2 Baseline sleep patterns for the entire squad (n = 23), selected players (COMP, n = 14) and non-selected (TRAIN, n = 9) for a world rugby sevens series competition (Continued).

ALL (n=23) COMP (n=14) TRAIN (n=8) Mean Mean Mean p (SD) Range (SD) Range (SD) Range

Actigraphy 6.9 TST (h) (0.5) 5.4 - 7.7 7.0 (0.6) 5.4 - 7.7 6.7 (0.3) 6.2 - 7.0 0.24

<6 h 59% Na 50% Na 75% Na 0.25

SOL (min) 11 (9) 2 - 44 8 (5) 3 – 44 8 (5) 2 - 18 0.26

>30 min 0% Na 0% Na 0% Na 1.00

SE (%) 81 (7) 60 - 91 82 (8) 60 - 91 79 (2) 77 - 84 0.33

<85% 77% Na 64% Na 100% Na 0.05

WASO 70 (32) 10 - 160 66 (34) 33 - 160 78 (29) 10 - 101 0.41

>30 min 91% Na 93% Na 88% Na 0.67 Notes: COMP selected players; TRAIN non-selected players; PSQI Pittsburgh Sleep Quality Index; FIRST Fords Insomnia Response to Stress Test; DBAS dysfunctional beliefs about sleep; TST total sleep time; SOL sleep onset latency; SE sleep efficiency.

7.3.4 Sleep outcomes Competition 1

7.3.4.1 Actigraphy SOL and TSTs for TRAIN and COMP players during competition 1 can be seen in Figures 7.3 and 7.4, respectively. A one-way ANCOVA revealed a significant difference in

SOL between groups (F(1,19) = 12.834, p = 0.002, ŋ2 = 0.13) with higher adjusted means observed in the COMP group compared to TRAIN. Moreover, there was a significant difference in TST between the two groups (F(1,19) = 8.623; p = 0.008, ŋ2 =0.312), with higher mean TST in COMP compared to TRAIN. However, there was no significant difference in SE (F(1,19) = 1.661; P=0.213, ŋ2 = 0.08), with similar mean values between

207

COMP (84, SD =2%) and TRAIN (82 , SD =1%), nor significant differences (F(1,19) = 0.886, p = 0.358, n2 = 0.045) in WASO between COMP (76, SD = 6 min) and TRAIN (82, SD = 8 min).

Table 7.3 Intercorrelations (Pearson product moment correlation coefficients) between sleep variables among a squad of female rugby sevens players (n=23) FIRST PSQI DBAS PSAS TST SOL SE WASO

FIRST - 0.24 -0.38 0.45 0.00 -0.33 0.26 -0.23

PSQI - - 0.2 0.66 -0.23 0.14 -0.15 0.29

DBAS - - - 0.01 -0.14 0.46 -0.3 0.22

PSAS - - - - -0.04 -0.03 -0.04 0.15

TST - - - - - 0.47 0.78 -0.54

SOL ------0.48 0.13

SE ------0.79

Notes: FIRST, Ford Insomnia Response to Stress Test; PSQI, Pittsburgh Sleep Quality Index; DBAS, Dysfunctional Beliefs and Attitudes about Sleep; PSAS, pre-sleep arousal scale; SOL, Actigraphic sleep onset latency; SE, Actigraphic sleep efficiency; WASO, Actigraphic wake time after sleep onset; TST, Actigraphic total sleep time. Bold significant at p<0.05.

208

8 *

Sleep Onset Latency (min) Latency Onset Sleep 2

COMP TRAIN

Figure 7.3 Sleep Onset Latency in competing (COMP; n=14) and training (TRAIN; n=8) players during Competition 1.

One way ANCOVA (F(1,19)=12.834, p=0.002, n2 = 0.13)

*Significantly different to COMP (p<0.05)

8.2 8 7.8

7.6 * 7.4 7.2

(h) Time Sleep Total 6.8

6.6 6.4 COMP TRAIN

Figure 7.4 Total Sleep Time in competing (COMP; n=14) and training (TRAIN; n=8) players during Competition 1.

One way ANCOVA (F(1,19)=8.623; p=0.008, ŋ2=0.312)

*Significantly different to COMP (p<0.05

209

7.3.2.2 Questionnaire assessments Sleep questionnaire scores between TRAIN and COMP players during competition 1 can be seen in Table 7.1

Pittsburgh Sleep Quality Index There was no group effect (F(2,19)=0.103; p=0.752) for PSQI global scores between COMP (7.1, SD = 0.4) and TRAIN (6.8, SD = 0.6) players during Competition 1, and no difference in poor sleeper distributions (FET = 1.87, p = 0.551), with 62% (n = 5) from TRAIN and 71% (n = 10) scoring >5.

Pre-Sleep Arousal Scale There were no significant group effects for PSAS total scores

(F(2,19) = 3.463, p =0.78, ŋ2 = 0.154) (COMP = 30, SD = 2 vs. TRAIN = 35, SD = 2). Moreover, for PSAS components scores there was no significant difference between groups

(COMP = 18, SD = 1 vs. TRAIN = 22, SD = 2) in cognitive (F(2,19)=0.2.974; p=0.101, ŋ2 =

0.135), nor somatic scores (F(2,19) = 1.49, P = 0.237, ŋ2 = 0.073) (COMP = 12, SD = 1 vs. TRAIN = 13, SD = 1).

Competition, Sport, Sleep and Dreams Questionnaire During competition 1, 72% (n = 11) COMP players reported experiencing disturbed sleep compared to 38% (n = 3) TRAIN players, however these differences were not significant (U = 39, p = 0.079). Among COMP players, common reasons for disturbed sleep were problems falling asleep (n = 4), and early morning awakenings (n = 5). These sleep problems were less frequent among TRAIN players, with “unpleasant dreams” (n = 2) being most common.

Wellbeing diary A Man Whitney-U test revealed no significant differences between

COMP and TRAIN in SQ (TRAIN = 3.5 (range = 1) vs. COMP = 3.5 (range = 1); U(22)=70; p=0.365), ENERGY (TRAIN = 4 (range = 3) vs. COMP = 4 (range = 1.5); U(22)=73, p=0.267), and STRESS (2 (range = 4) vs. 2 (range = 2); U(22)=51.5, p=0.764). However, there was a significant difference in RTP (U(22)=87, p=0.035) with COMP (4 (range = 1.5)) reporting higher values compared to TRAIN (3 (range = 2)).

210

7.3.5 Pre-post assessment of the Athlete Sleep Management Programme

7.3.5.1 Questionnaires Pittsburgh sleep quality index PSQI global scores among poor sleepers (identified at Baseline) at Competitions 1 and 2 can be seen in Figure 7.5. There was not a significant difference in PSQI global scores between Competition 1 and Competition

2 (t(11) = 2.126, p = 0.053, d = 0.57). Moreover, there was not a significant difference in the proportion of PSQI-defined poor sleepers at Competition 1 (75%, n=9/12) and competition 2 (17%, n=2, p=0.07). The distribution of players scoring >8 on the PSQI also showed no significant differences between competition 1 (33%, n = 4/12) competition 2 (0%, n=0, p=0.688)., Among players who scored >5 at Baseline (50%, n = 6/12), however, there was a significant reduction in PSQI global scores at Competition 2 (t(5) = 4.339, p = 0.005, d = 1.7).

12.0

10.0

8.0 *

6.0

4.0

score globalPSQI 2.0

0.0 Competition 1 Competition 2

Figure 7.5 Pittsburgh Sleep Quality Index (PSQI) global scores before (Competition 1) and after intervention (Competition 2) in players scoring >5 on PSQI at Baseline (n = 6).

*significantly different to Competition 1 (p<0.05)

Dysfunctional Beliefs and Attitudes about Sleep DBAS component scores for “consequences” can be seen in Figure 7.6, and scores in response to concerns about poor sleep impacting playing performance can be seen in Figure 7.7. There was no

211

significant difference in DBAS total scores between Baseline and Competition 2 (t(11) = 0.690, p = 0.502, d = 0.18). Moreover, there was no significant difference in component scores for “worry/ helplessness” (t(11) = 1.301, p = 0.216, d = 0.34) and “medication” (t(11) = -0.110, p = 0.914, d = -0.03). However, there were significant reductions in component scores indicating heightened concerns for “consequences” (t(11) = 2.249, p = 0.042, d = 0.60), and for the item “concerns about poor sleep impacting playing performance” scores (t(11) = 3.616, p = 0.003, d = 0.97).

9 8 7 * 6

consequences 4 - 3

DBAS DBAS 2 1 0 PRE POST Competition 1 Competition 2

Figure 7.6 Dysfunctional Beliefs and Attitudes about Sleep (DBAS) consequences component before (PRE) and after (POST) the sleep management intervention (n = 12).

Paired samples t-test (t(11) = 2.249, p = 0.042, d = 0.60)

*Significantly different to Competition 1 (p<0.05)

212

10 9 * 8 7 6 5 4

performance 3

2 Concerns about playing about Concerns 1 0 PRE POST Competition 1 Competition 2

Figure 7.7 Rating of agreement to statement “I am concerned poor sleep will impact my playing performance on the pitch” (higher scores = greater concern; n = 12).

Paired samples t-test (t(11) = 3.616, p = 0.003, d = 0.97).

*Significantly different to Competition 1 (p<0.05)

Competition, Sport, Sleep and Dreams Questionnaire

Responses to CSSDQ during competition 1 and 2 can be seen in Table 7.5. There was no significant difference in the distribution of players reporting sleep disturbances during competition 1 (67%, n = 8/12) and 2 (75%, n = 9/12, FET = 1.3; p = 0.50). During Competition 1, ‘other’ reasons for experiencing disturbed sleep reported were “jet lag” and “concern about my injury”, this was reported by two players and one player, respectively. Different ‘other’ reasons were reported during Competition 2 with a “small and uncomfortable mattress” and “nervousness about Olympic selection” being reported by one and two players, respectively.

213

Table 7.4 Responses to Competition, Sport, Sleep, Dreams Questionnaire (CSSDQ) and sleep behaviours questions during world series tournaments in Atlanta, USA and Langford, Canada (Competition 1) and Clermont, France (Competition 2) Competition 1 Competition 2 CSSDQ items (PRE), n = 12 (POST), n = 12 Disturbed sleep 64% 79%

Insomnia symptoms n = 8 n = 9

Problems falling asleep 56% 91%

Waking up in the night 11% 45%

Wakeup too early in the morning 56% 27%

Not feeling refreshed 67% 45%

Reasons for disturbed sleep n =8 n = 9

Thoughts about Competition 67% 72%

56% 72% Nervousness about Competition

Difficulty winding-down in 56% 27% Evening

Frequent use of toilet 56% 27%

Day-time impact n = 8 n = 9

No influence 33% 45%

Bad-mood 56% 27%

Sleepiness 22% 45%

Worse performance 11% 7%

Sleep management n = 12 n = 12

Pre-sleep arousal techniques 42% 67%

Sleep hygiene behaviours 75% 75%

Napping 17% 0%

214

Wellbeing diary SQ and STRESS ratings between competition 1 and competition 2 can be seen in Figure 7.8 and 7.9, respectively. A Wilcoxon signed ranks test revealed no significant difference in STRESS (z = -0.957, p = 0.339); ENERGY (z = -0.137; p = 0.891) SQ (z = -0.453, p = 0.651) and RTP (z=-0.632; p=0.527). between BASE1 and BASE2. However, self-reported STRESS was significantly higher at competition 2 compared to competition 1 (z = -2.201, p = 0.028) and SQ was significantly lower (z = -2.288, p = 0.022). There was not a significant difference in ENERGY (Z = -0.577, p = 0.564) nor RTP (z = -0.322, p = 0.748) between competitions.

* 5.0 4.5

4.0 3.5 3.0

2.5

2.0 Sleep quality quality Sleep 1.5

1.0

0.5 0.0 Competition 1 Competition 2

Figure 7.8 Sleep quality ratings between competition 1 and competition 2

Wilcoxon signed ranked test (z = -2.288, p = 0.022). Blocks = Median, error bars = maximum and minimum values.

*Significantly different to Competition 1 (p<0.05)

215

5.0

4.5 *

4.0 3.5 3.0 2.5 Stress 2.0 1.5 1.0

0.5

0.0 Competition 1 Competition 2

Figure 7.9 Stress ratings between competition 1 and competition 2

Wilcoxon signed ranked test (z = -2.201, p = 0.028). Blocks = Median, error bars = maximum and minimum values.

*Significantly different to Competition 1 (p<0.05)

7.3.5.2 Actigraphy TSTs between competition 1 and 2 can be seen in Figure 7.10. There was no significant difference in SE (t(11) = 0.552, p =0.591, d = 0.15), SOL (t(11) = -1.479, p = 0.163, d = -0.39),

WASO (t(11) = -1.921, p = 0.077, d = -0.51) and TST (t(11) = -1.591, p = 0.136, d = -0.42) between BASE1 and BASE2. McNemar tests revealed no significant differences in distributions in players scoring <6 h TST (p = 0.50), <85% SE (p = 1.00), >30 min WASO (p = 0.250), and <30 min SOL (p = 1.00). During the competition week, there was no significant difference in SE (t(11) = 1.851, p = 0.087, d =0.49), WASO (t(11) = -0.558, p =

0.587, d = -0.15) and SOL (t(11) = -1.449, p = 0.171, d = -0.38) scores. However, there was a significant reduction in TSTs between competition 1 and 2 (t(11)) = 5.543, p <0.001, d = -1.5). Moreover, there was no significant difference in players achieving <6 h TST (p = 1.00), >30 min SOL (p = 1.00), >30 min WASO (p = 1.00) and <80% SE (p = 0.625).

216

8.2

7.8 7.6 * 7.4

7.2 Total sleep time (h) time sleep Total

6.8

6.6 Competition 1 Competition 2 Figure 7.10 total sleep time (TST) between the competition before the intervention (competition 1) and after (competition 2), n = 12.

Paired samples t-test (t(13) = 5.543, p <0.001, d = -1.5).

*Significantly different to Competition 1 (p<0.05)

7.3.6 Sleep programme evaluation Evaluations of the usefulness, relevance and impact of the sleep management programme can be seen in Figure 7.11. From COMP players, 75% (9/12) players stated “yes, definitely” or “yes, somewhat” that they employed the information presented during the world series in Clermont. However, 25% (3/12) players reported feeling unequipped to deal with sleep disturbances when they arise following the sleep management programme. Using the player evaluations of the programme, 50% (6/12) players state “yes, definitely” or “yes, somewhat” that the information provided had a positive impact on their recent performance(s) and wellbeing during the world series tournament in Clermont (France).

7.3.6.1 Programme utility Sleep management adopted by players can be seen in Table 7.5. The number of players reporting the use of pre-sleep de-arousal techniques during competition 1(42%, n = 5/12) and 2 (67%, n = 8/12) were not significantly different (FET = 0.9; p = 0.340). Among the players adopting pre-sleep de-arousal techniques during

217

Competition 2 (n=8), 50% (4/8) players stated the techniques they used were effective at helping them fall asleep. Techniques mainly included “reading” (n=3) and “breathing exercises” (n=2) with one player adopting “articulatory suppression”. The one player that reported “trying to force/ repeat their sleep pattern” during competition 1, did not use that as a technique during competition 2. Sleep hygiene behaviours were adopted by most during both Competition 1 (75%) and 2 (75%), and no players reported napping during competition 2.

Maybe 7 Yes, somewhat

Yes, definately 6 No, not really

5 No, definitely not

Question number Question 3

-100 -50 0 50 100

1: Did you find the material presented relevant? 2: Do you feel you are unequipped to deal with sleep disturbances when faced? 3: Have you previously received information about how to manage your sleep during normal training and (or) competitions? 4: Would you recommend this workshop series to another athlete? 5: Did you use this information during the WRSS in Clermont-Ferrand 6: Do you think you’ll use this information in the future?” 7: Do you feel this information had a positive impact on your wellbeing and (or) performance?”

Figure 7.11 Evaluation of the sleep management programme. n = 12

218

7.4 Discussion The aims of the present study were: to assess the acceptability, feasibility and impact of a psychoeducational sleep management programme (based on cognitive- behavioural principles) among British female rugby sevens players during WRSS competitions; and to assess the extent of sleep challenge experienced by elite team athletes during international competitions. Findings from Phase 1 of the present study confirm that competition-related sleep disruption is broadly characterised by problems falling asleep and waking up early in the morning, with high levels of variability among athletes (see, for example, Fowler et al., 2018). Nevertheless, actigraphy derived TSTs were higher during a tournament when compared to control values, indicating a preservation of sleep duration during competition.

Symptoms of insomnia were generally high across the group, with very high levels of sleep reactivity and poor sleep quality among the squad of players. The sleep programme was associated with modest improvements in PSQI global scores among poor sleepers following a period of competition, with a 100% reduction in players scoring >8 and a 58% reduction in players scoring >5. Moreover, no improvements in actigraphy derived sleep outcomes were observed. As a result, the programme can be not be deemed effective during competition. Many players accepted the sleep programme, with 75% players reporting they used the information presented during Competition 2. And 50% of players stated the information had a positive impact on their wellbeing and (or) their performance during competition. However, the programme’s utility was modest, with few players adopting effective pre-sleep de- arousal techniques, and no players adopting naps following the intervention.

7.4.1 Insomnia symptoms and sleep quality (hypothesis 1) The proportion of players scoring both >5 (50%) and >8 (22%) on the PSQI were within the range for elite athletes reported previously (see Chapter 1 and 3). Nonetheless, it is interesting to note the very high PSQI scores in rugby sevens players, with some players, in the current study, scoring >10. These scores can be judged ‘high’ when compared to other team sports (e.g. Swinbourne et al. (2016)) and other broader sport-types (see Chapter 3, Para. 3.3.1). The findings highlighted a trend in proportions of non-selected players scoring <85% sleep efficiency compared to selected and a trend (p=0.07) for a greater proportion of non-selected players scoring >30 min on WASO (see Table 7.2). As a result, the first null hypothesis can be rejected. Sleep

219

vulnerability was equally high in the rugby sevens players, with 77% players scoring >18 on FIRST. Given the limited use of the FIRST construct within the sport science and medicine literature, comparative judgements cannot be made to other sports. It is reasonable to suggest, however, that the sleep of rugby sevens players is highly challenged by the ‘trio’ of challenges – training scheduling, frequent demands of long-haul travel and international competition across a competitive season.

7.4.2 Competition sleep (hypothesis 2) Constitutes of sleep quality of rugby sevens players during a tournament was shown to be poorer when compared to players who were training at the same time. As a result, the second null hypothesis can be rejected. This was shown by significantly higher SOLs among COMP players compared to TRAIN players in Competition 1 (see Figure 7.2). Consistent with the actigraphic findings, the CSSDQ showed that “problems falling asleep”, “waking up earlier than desired” and “not feeling refreshed on awakening” were the most common insomnia symptoms reported. Furthermore, the CSSDQ also showed that the most common reason players reported for experiencing disturbed sleep were “thoughts about competition”. It is noteworthy, that some players reported “jet lag” as a reason for disturbed sleep in this tournament. This information emphasises the breadth of sleep challenge experienced by international competitors, and the likelihood of individual (and heterogeneous) responses within the squad. Fowler et al. (2017) reported high levels of variability in sleep responses though self-reported ratings between competitions among male rugby sevens players attending two competitions. The longer actigraphy derived SOLs observed in the present study, specifically, contrasts with previous research showing limited change (Romyn et al., 2016; Shearer et al., 2015) or even shorter SOLs (Dunican et al., 2018) around periods of competition. For example, Dunican et al. (2018) reported shorter actigraphy derived SOLs the night before and following a ‘Super Rugby’ match compared to baseline in elite male rugby players, whereas Lalor et al. (2018) reported no differences in actigraphy derived SOL before or after four home Australian Rules Football (AFL) matches in elite AFL players, when compared to habitual sleep patterns. In the current study, despite significant disturbance reported during competition, COMP players attained higher TSTs during the first tournament. This finding is like previous research (Dunican et al., 2018; Richmond et al., 2004; Richmond et al., 2007) whereby TSTs have been shown to increase around periods of competition. In earlier studies by Richmond et al. (2004, 2007) longer TSTs were

220

reported prior to home matches in Australian rugby players. This is supported in a more recent study by Dunican et al. (2018) whereby a trend for longer TSTs in the lead up to a ‘Super Rugby’ Match in Australian male rugby players can be seen. The longer TSTs seen in these groups, and in the current study, may be the result of a perceived sleep need during competition to attain optimal performance. In the current study, players scored a mean of 7 (SD=3) on the item “I need 8 hours sleep to feel refreshed and function well the next day” on DBAS and scored a mean of 8 (SD=3) on the item “I am concerned poor sleep will impair my playing performance”. This indicates, a proportion of players may have an unrealistic expectation around how much sleep they need coupled with concerns around disrupted sleep impairing performance, particularly during a period when sleep is more likely to be disturbed. This finding also demonstrates that players value their sleep during periods of competition, possibly more so than compared to training.

7.4.3 Sleep management (hypothesis 3 and 4) In the current study, the sleep management programme resulted in modest changes in global sleep quality during a period of WRSS competitions, with no change in DBAS, PSAS scores nor sleep quality derived actigraphy outcomes. The sleep programme did, however, result in a significant reduction in TST. Following the programme there was a 58% reduction in the distribution of players scoring >5. Moreover, there was a 100% (n=4) reduction in players reporting PSQI scores >8 following competition 2, however these did not reach a level of significance. Overall, the unresponsiveness of sleep quality metrics means that null hypothesis 3 cannot be rejected. The magnitude of reduction in PSQI scores (and improvement in sleep quality) observed in the current study is like that reported in the pilot study in Chapter 6, where the sleep management programme improved sleep quality among poor sleepers by 50-60% (as defined by a less conservative cut-off of ≥5) during a phase of physical training. However, the finding of limited response in total DBAS scores contrasted with the previous chapter. The latter finding is difficult to explain. Despite DBAS scores being higher on average compared to the previous chapter, a blunted response was observed. The component score of “consequences of sleep disruption” was lower following the programme along with an additional item around “concerns about playing performance”. It may be that differences in players beliefs and attitudes about sleep towards certain components reflect responses to the intervention. Furthermore, in Chapter 6 DBAS showed a decline after 1 month. In the current study, the time course

221

of DBAS scores following the programme was somewhat different and may reflect the responses seen. In the current study, two WRSS competitions separated by 1 month were used as periods of assessment, whereby sleep quality was highly challenged. This finding provides further evidence to support the effectiveness of the programme to improve sleep quality within the context of competition periods and among female team sport athletes. A consideration for this finding is the period of assessment during Competition 2 (and following deployment of the sleep management programme) was shorter (i.e. 1 week in Clermont) compared to Competition 1 (i.e. 2 weeks in Atlanta and Langford), and therefore the higher levels sleep quality reported during Competition 2 may reflect the shorter duration of competition experienced, given the PSQI refers to sleep during the previous month (Buysse et al., 1989). However, self- reported SQ through wellbeing diaries showed poorer sleep quality during Competition 2 indicating a greater sleep challenge. The format in which sleep management support was provided was different between players. Players who scored >8 following Competition 1 received a one-to-one consultation in addition to the group delivery of the sleep management programme (n=4). Following intervention, all these players reported improvements in PSQI scores with three players reporting PSQI score ≤5 and one player scoring 6 following Competition 2. Whilst not confirmed in the current study, it is possible players who received one-to-one sessions in addition to the group sessions had a greater response to the programme that those that received the group sessions only. This is supported by previous research whereby individual sessions have been shown to be more effective when compared to group sessions in primary insomnia populations receiving CBT-I (Yamadera et al., 2013). However, other research in similar populations have reported similarities in the effectiveness of group vs. individual formats (Bastien, Morin, Ouellet, Blais, & Bouchard, 2004) . In regard to more acute cases of insomnia (i.e. < 3 months), Boullin et al. (2016) reported similarities in treatment formats, whereby a “single shot” delivery format resulted similarities in effectiveness in reducing insomnia caseness, as measured by the ISI. However, in this study adherence to CBT-I, as defined by prescribed time in bed, was significantly lower in the group treatment vs. the individual (Boullin et al., 2016). In the context of the current study, differences in the format of CBT-I delivery and insomnia caseness of participants are apparent when compared to previous research. In the current study, a large group (n=22) attended each of the two sleep management sessions which consisted of all athletes (e.g. good sleepers and poor

222

sleepers). Moreover, insomnia caseness of the group was unknown with the PSQI not being a diagnostic tool as defined by DSM-5 criteria (Buysse et al., 1989). On the other hand, In the current study those who received one-to-one support were players with the poorest sleep quality (and more reactive sleep) and therefore could have been more responsive, irrespective of the delivery format.

In the current study, no improvements in actigraphy derived SOL and SE were observed nor were there improvements in self-reported measures of wellbeing between the two competition periods. However, despite the ASMP promoting behaviours to extend sleep (e.g. napping and sleep extension) there was a large reduction in TST during Competition 2. This contrasts with the findings of both the previous chapter (Chapter 6) and previous research (O’Donnell & Driller, 2017; Van Ryswyk et al., 2016). In Chapter 6 of the thesis, improvements in self-reported ease-to- fall-asleep and actigraphy derived SOL were seen following programme delivery. Moreover, following group delivery of sleep hygiene education, O’Donnell and Driller (2017) reported increased TST and reduced wake episode duration among elite female netball players. In the context of competition, Fullagar et al. (2016a) reported a significant increase in TST in a group of semi-professional football players who received sleep hygiene guidance following a simulated night-match compared to when they received no information. However, the authors reported a trend for a decrease in SE and increase in SOL, with no differences in overall perceptual recovery the following morning. The decrease in TST in the current study is difficult to explain, particularly given psychoeducational messaging on sleep extension and napping were delivered. Whether players adopted sleep behaviours to extend sleep is unknown, however given both SE and SOL scores were similar between competitions, it can be assumed time in bed scores were less during Competition 2 which may indicate training/ competition schedules had dictated player sleep schedules. To date, only one study has employed an intervention during a major competition (Rosa et al., 2018). In this study among Brazilian Olympic swimmers at the Rio 2016 Summer Olympic Games, sleep was shown to be delayed and morning cognitive function was subsequently improved up to 7 days following implementation of a light-therapy protocol and sleep hygiene recommendations (Rosa et al., 2018). To provide context to the purpose of the study, a light therapy protocol was employed to delay sleep to accommodate the later scheduling of swimming finals at the 2016 Summer Olympic Games. Whilst the purpose of the sleep management programme designed in the

223

Rosa et al. (2018) study is somewhat different to the programme deployed in the current study, the study, nonetheless, supports the notion that implementation of sleep management during a major competition is feasible and effective. In the current study, while the programme elicited modest outcomes in sleep quality, most of the players however stated they used the information provided during competition 2 Whilst a large proportion of players reported using the information, it is interesting to note only a modest number of players stated they employed pre-sleep arousal management techniques and napping during competition 2, and this was similar to competition 1 distributions. Therefore, null hypothesis 4 that players did not value the psychoeducation sleep management programme can be rejected.

7.4.4 Sleep and athletic performance A broad observation was that 50% (n=6) COMP players reported the sleep management programme had a positive impact on their wellbeing and performance. Notwithstanding differences in sleep phenotypes, it is interesting to note that despite relatively high levels of poor sleep quality and sleep reactivity players selected here were able to attain the highest levels of performance during a WRSS tournament (i.e. winning a tournament). This provides support to the notion that good sleep quality is not a prerequisite for optimal performance, at least in rugby 7s. This is supported by Silva and Paiva (2016) who reported elite gymnasts who finished in top 10 of FIG competition to report lower sleep quality (as scored by PSQI) compared to those who finished outside the top 10. What is more interesting, is that these two groups were defined somewhat by their daytime levels of sleepiness as assessed by the Epworth Sleepiness Scale, suggesting that day-time levels of arousal may play role here. In the context of the current study, whilst day-time levels of sleepiness were not reported, high levels of sleep reactivity were reported across the playing squad with only one player reporting unreactive sleep indicating high levels of hyperarousal. Further research is required to investigate linkages between arousal and performance among elite athletes. It appears team sports could be a candidate for further study given the high levels of sleep reactivity reported, here.

7.4.5 Limitations A possible explanation for similarities between competitions in the level of disturbance and management techniques employed could be the context within which tournaments were set. For example, in the current study assessments of sleep quality

224

were made during two WRSS competitions which were held within different geographical locations. For example, during the first WRSS competition players were required to undertake a long-haul flight to Atlanta, USA (-5 h GMT) followed another long-haul flight (-8 h GMT) to Langford, Canada; however, in the second WRSS competition players were not required to undertake a short-haul flight to Clermont- Ferrand, France (+1 h GMT). As a result, the periods of reference could be argued as somewhat different because of the influence of jet lag or specifically circadian misalignment on sleep quality (PM Fowler et al., 2016). This is supported by some athletes reporting “jet lag” as a reason why their sleep was disturbed during competition 5 days after landing in Atlanta or Langford. Given the tournament in both Atlanta and Langford took place 4-5 days following arrival, it was assumed circadian realignment would have been achieved based on current recommendations (e.g. 3 days for complete circadian alignment) (Waterhouse et al., 2007) , however the finding from the current study suggests this may not be the case for all individuals in the context of competition. Despite Competition 1 appearing to be a greater challenge to sleep quality compared to Competition 2, self-reported ratings of STRESS were reported to be significantly higher and SQ were significantly lower. During this tournament some players reported “nervousness about Olympic selection” as a reason for disturbed sleep due to the competition be the final WRSS before the 2016 Summer Olympics Games. However, higher levels of cognitive arousal during Competition 2 were not confirmed when referring to PSAScog responses. In support of this finding, albeit in a different context Juliff, Peiffer and Halson (2017) reported no change in PSAScog following a night match in state level netball players. Despite higher levels of STRESS and lower self-reported SQ experienced, actigraphy derived sleep quality metrics were maintained between the two competition periods. Whether or not adverse responses in player sleep during Competition 2 would have been more pronounced without the sleep management programme information provided is unknown due to the lack of control group in the current study.

7.5 Conclusion In conclusion, a WRSS competition presents an array of challenges to athlete sleep, with successive WRSS tournaments eliciting similar responses in composites in sleep quality. Responses to WRSS competition is highly variable among a squad of female

225

rugby players exhibiting high levels of poor sleep quality. The sleep management programme was underutilised, ineffective in improving sleep quality among a group of reactive sleepers and did not improve maladaptive beliefs about sleep. However, players with the poorest sleep quality (e.g. >8 on PSQI) showed the most pronounced improvements following intervention. The programme was shown to be accepted by the players, though half the players stated the programme had influence on their wellbeing and (or) performance. The highest performing players (i.e. COMP players) during the WRSS competitions did not report superior sleep quality. Overall, the sleep management programme is an acceptable and feasible programme which can be deployed prior to a period of competition, how the influence on sleep outcomes is minimal. Further research is required into the linkages between sleep reactivity, arousal and performance and the format by which sleep management programmes can be delivered in the context of elite sport.

226

Chapter 8

8 A general discussion of findings

8.1 Introduction As emphasised in Chapter 1 (Para. 1.1.2), the programme of research presented in this thesis set out to explore the construct of sleep quality in the context of elite sport. While findings have been discussed (with reference to specific hypotheses and research questions) in the relevant chapters, the aim of this general discussion will be to: 1) connect findings across chapters; 2) consider the practical implications of this research for elite sport; and 3) signpost an agenda for future research.

8.2 Connecting findings across chapters A summary of the null hypotheses addressed in this thesis can be seen in Tables 8.1 and 8.2, with selected summaries of principal findings from each chapter provided below.

Chapter 1 From a systematic review of the sports science literature it was concluded that:

1. Insomnia symptomatology and poor sleep quality are high among elite athletic populations; and 2. The three main sources of athlete sleep disturbance are training, travel and competition. However, athletes are most vulnerable to insomnia symptoms during periods of competition.

Chapter 3 A between-sport comparison confirmed that the prevalence of insomnia symptoms and poor sleep quality is high among UK Olympic and Paralympic athletes, and found that:

1. Technical sport athletes showed higher levels of sleep disturbance (as indexed by PSQI scores >8; and 2. Poor sleep quality was not significantly related to daytime napping.

227

Chapter 4 This experimental comparison of daytime sleep tendency in elite athletes, sub-elite athletes and non-athletes showed that:

1. Daytime sleep latency profiles provide evidence of “high sleepability” among elite athletes, indicating that daytime napping in elite sport is not necessarily driven by homeostatic sleepiness; and 2. Correlations between sleep latency scores and levels of sleep reactivity for the adaptation trial suggest that the underlying construct of hyperarousal may identify those athletes whose sleep is most influenced by first-night effects.

Chapter 5 Outcomes from a simulated Olympic tournament among elite British field hockey players indicated that: 1. Reactive sleepers (i.e. those with higher FIRST scores) exhibited a greater degradation in sleep duration during the tournament when compared to their unreactive sleeper counterparts; and 2. Self-reported sleep quality showed a greater ‘rebound’ following the tournament simulation in reactive sleepers compared to unreactive sleepers,

Chapter 6 This intervention study assessed the feasibility of delivering a sleep management programme based on cognitive-behavioural principles elite sport team athletes using. The key findings were:

1. The delivery of a sleep management programme was shown to be feasible within the context of a routine elite training period; 2. The utility of the programme was evidenced by most athletes stating that the programme material was relevant, that they felt equipped to manage sleep disruption when it arises, and the adoption of sleep behaviours 3. There was a reduction in DBAS scores, and a significant reduction in the proportion of ‘poor sleepers’ (i.e. those scoring ≥5 on the PSQI) following the management programme.

Chapter 7 Developing the findings of Chapter 6, the sleep management programme was adapted for delivery during a period of international competition to team sport athletes. Again, the emphasis was on cognitive (pre-sleep de-arousal) and behavioural strategies to manage sleep challenge. Key findings were:

228

1. Despite a 100% reduction in players scoring >8 and a 66% reduction in players scoring >5; the programme was deemed ineffective with few positive sleep outcomes being reported following intervention 2. Most players reported that they used the information presented during tournaments, yet overall programme utility was modest with 50% players reporting using ‘effective’ techniques to help fall asleep during competition 3. 50% of players stated the information had a positive impact on their wellbeing and (or) their performance during competition.

229

Table 8.1 null hypotheses presented in Chapters 2-5

Main Sleep Chapter n Null hypotheses √ X constructs addressed

1. The prevalence of poor sleep quality and insomnia Y symptomatology will be low in elite athletes Chapter2 : Sleep quality in 37* 2. Elements of elite sport will not negatively impact on Sleep quality elite sport: a review sleep quality in elite athletes Y

1. There will be no significant differences in sleep Y quality (as measured by PSQI) between sport-types Chapter 3: A comparison of 412 2. Cross-sport variations in sleep quality are not between-sport differences significantly influenced by age, gender and ability Y Sleep quality in sleep quality and (Paralympic athletes) Sleep reactivity insomnia symptomatology 3. Sleep management adopted by elite athletes will among elite British athletes not significantly differ between sport-types, and will Y not be driven by sleep quality

1. Sleep tendency measures from a single-trial nap Chapter 4: Napping in high- 30 opportunity will not show significant differences Daytime sleepiness performance athletes: between high-performance athletes and non- Y Sleep reactivity sleepiness or sleepability? athlete control participants

1. Measurements of sleep quality and quantity will not show significant differences between baseline, Y Chapter 5: Sleep reactivity competition and recovery periods during a simulated Olympic Sleep reactivity 18 2. Sleep quality and quantity measurements made at tournament in elite British Sleep quality baseline, competition and recovery weeks will field hockey players Y show no significant associations with measures of sleep reactivity (as measured by FIRST scores). Notes: √ = Accept null hypothesis; X = Reject null hypothesis; Y, yes; N, No; * studies

230

Table 8.2 null hypotheses presented in Chapters 6-7

Main Sleep constructs Study Null hypotheses √ X n addressed 1. It is impractical to deliver the psychoeducational sleep management programme in the context of a routine Y Chapter 6: Sleep training period management in elite 2. When delivered during a period of training, a Sleep quality team sport athletes 22 psychoeducational sleep management programme will Y Sleep reactivity using cognitive- not be valued by elite male team-sport athletes Daytime sleepiness behavioural principles: 3. A psychoeducational sleep management programme A feasibility study will not improve sleep quality and insomnia symptoms Y

1. Measures of sleep quality and sleep quantity will show no significant differences between the selected and Y unselected players 2. In comparisons of selected and unselected players, Chapter 7: Delivering participation in elite international competitions will not group-based Y significantly degrade sleep quality and quantity among psychoeducational Sleep quality 23 the selected sleep management to Sleep reactivity 3. a psychoeducational sleep management programme elite team sport will not mitigate the impact on sleep quality and athletes: A field study insomnia symptoms during international competitions. Y 4. When delivered during pre-competition training, a psychoeducational sleep management programme will Y not be utilised nor valued by elite team-sport athletes Notes: √ = Accept null hypothesis; X = Reject null hypothessi; Y, yes; N, no

231

Connecting each of these findings is the causal relationship proposed between arousal/hyperarousal and sleep onset in the genesis of insomnia (see, for example, Harvey et al. (2002) and Kalmbach et al. (2018). The most influential source of sleep disturbance identified in the systematic review was competition (Chapter 1, Para. 1.3.5.1), with the literature further identifying nights immediately prior to competitions as characterised by longer sleep latencies (see Chapter 1, Tables 1.14 and 1.15). While little underlying theory is considered in the sports science literature, it is likely that such disturbances are mediated by concerns and anxieties about performance in the pre-competition period. This conclusions places the focus on pre-sleep cognitive activity which, according to Harvey (2002, pp. 886), “…will trigger arousal which, in turn, will trigger selective attention, a process likely to further fuel excessive cognitive activity.” It can be assumed, therefore, that elite athletes are subject to repeated ‘precipitating’ episodes of competitive stress which, through, arousal mechanisms, create sleep loss prior to performance. The resulting sleep loss could help to explain the significant differences in sleep quality found between technical (skill-based) and other athletes, since skill-based performance, which often requires only brief periods of competitive focus, may be less vulnerable to prior sleep restriction (See Thun et al. (2015)).

Because sleep is repeatedly challenged in elite sport, and night-time sleep loss is likely among athletes, the ability to compensate through daytime napping takes on a special importance. The finding in Chapter 3 (Para. 3.3.3), however, that napping was unrelated to night-time sleep quality introduced the possibility that napping in elite sport may, under some circumstances, be less related to homeostatic sleep need, and more related to the construct of ‘sleepability’ as defined by Harrison and Horne (1996). Again, the construct of arousal is relevant, since an ability to nap ‘on demand’ requires also an ability to effectively manage the arousal mechanisms which promote (or successfully de-arouse from) daytime wakefulness

These assumptions were confirmed in Chapter 4, which also supplied evidence for the influence of arousal mechanisms in predicting first night effects in elite athletes. Correlations between sleep latency and FIRST scores (which are presumed to reflect a propensity for hyperarousal; see Kalmbach et al. (2018))

232

for the adaptation trial suggest that sleep reactivity may identify those athletes whose sleep is most influenced by first-night effects.

Two arousal-related aspects of athlete sleep management were then addressed. In Chapter 5 the utility of the FIRST measurements was tested in a simulated Olympic tournament, where it was found that FIRST scores predicted both levels of sleep disturbance, and levels of rebound. And finally, given the pervasive influence of arousal mechanisms on athlete sleep quality, Chapters 6 and 7 focussed on the delivery of sleep management programmes which, importantly, emphasised strategies to combat and manage pre-sleep arousal. That the programme reduced maladaptive beliefs about sleep and increased the utitlity of cognitive-behavioural techniques in the feasibility trial (Chapter 6) offers evidence for the successful transfer of ‘de-arousal’ strategies, since DBAS measures faulty expectations and perceptions about sleep which “… are instrumental in producing emotional distress, heightening arousal, and in feeding on the vicious cycle of insomnia” ((Morin, Vallières & Ivers, 2007; pp. 1548). Despite unresponsive sleep quality outcomes, the acceptance of the sleep management programme (Chapter 7) suggests targeted management is feasible among athletes during competition, however the timing, content and modality of psychoeducation delivery may require further modification.

It is a reasonable overall conclusion from this programme of research that the systematic assessment and control of arousal-related mechanisms can improve the ability of sport management to predict and protect sleep outcomes among elite athletes.

8.2. Practical application of findings While it is generally accepted that sleep is regarded as important for health, wellbeing and performance by athletes (Rachel Elizabeth Venter, 2014) coaches (Fallon, 2007) and practitioners (Fallon, 2007) alike, the level at which sports engage with ‘sleep science’ support is heavily influenced by other factors. Typically, any form of sleep implementation (e.g. measurement and/or management), is logistically in direct competition for time with sport training and competition schedules, and other more established areas of sport science and medicine need (i.e. nutrition, physiotherapy etc.). As a result, ‘sleep science’ continues to occupy a low priority on wider health, wellbeing and performance

233

agendas across high performance sport. This research programme has attempted to address these challenges, and the practical application of findings from the research programme are discussed below.

8.2.1 Measurement Sleep quality Chapters 1 and 3, generally highlighted high levels of poor sleep quality and insomnia symptoms among a representative sample of Olympic and Paralympic athletes (as indicated by PSQI scores), and therefore measuring sleep quality ‘formally’ is highly relevant in this population. Within sport science and medicine research, however, sleep quality definitions vary widely. In a recent meta-analysis, a total of 30 parameters to assess sleep quality in elite sport research were identified (See Claudino et al., 2019). To date most studies have employed instrumental measures of sleep quality, with less emphasis on athlete sleep experience and symptoms. Though instrumental measures provide valuable contributions to overall athlete sleep quality appraisals (as shown in Chapters 4, 5, 6 and 7), the omission of self-evaluations of sleep quality requires attention. As such, self-report appraisals form an important component of sleep quality definitions (Allison et al., 2008) and insomnia diagnostics (American Academy of Sleep Medicine, 2014; American Psychiatric Association, 2013).

The development of psychometric instruments has allowed self-appraisals of sleep quality to be formally addressed, while allowing subjective perceptions to be quantified, and therefore standardised. The PSQI has been shown to be widely used instrument and has shown to be a valid metric of sleep quality in both clinical and non-clinical populations (see (Mollayeva et al., 2016)). Within elite sport, the PSQI is a common tool employed in both research and applied practise (see Chapter 1, Table 1.13). However, despite widespread use and empirical support, the metric has come under scrutiny regarding its validity in elite athletic populations (See (Bender et al., 2018; Samuels et al., 2016). It is important to highlight limitations to PSQI evaluations made in recent sport science and medicine research (see (Samuels et al., 2016)). Firstly, the PSQI was designed to provide a reliable, standardised, valid measure of ‘sleep quality’ (Buysse et al., 1989) and was not designed to capture clinical caseness of multiple sleep disorders. Secondly, it has been reported that sleep quality has no reference or gold standard (Shelgikar & Chervin, 2013) and, as a result, low concordance

234

rates are to be expected as differences of opinion between physicians/ therapists as likely to be high (Mollayeva et al., 2016). Finally, it has been suggested that different populations are likely to have highly variable characteristics in terms of exposure to environmental/ societal stressors, medical pathology, and general sleep attitudes and beliefs (Mollayeva et al., 2016). Therefore, it is plausible that the PSQI would function differently among elite athletes.

In contrast to previous sport science research (i.e. Samuels et al. (2016)), the current thesis has highlighted the value and feasibility of employing standardised measurements of ‘global’ sleep quality (e.g. the PSQI), with clear practical advantages emerging. Firstly, in Chapter 1 and 3 the instrument’s total scores and subsequent “poor sleeper” thresholds (e.g. >5 and >8) provided ‘global’ sub-groups which enabled cross-population comparisons (see Chapter 1, Para. 1.4.2). In the field, then, the instrument allows for both categorisation sleep quality sub-groups which can guide time and resource around targeted management (as seen in Chapter 6), but also allows for interrogation of specific insomnia symptoms through component scores, and individual items. In addition to this latter point, the option within the instrument to add additional sleep problems in a delimit manner is hugely advantageous for a practitioner or trainer, as it allows athletes to answer completely with personalised answers in their own words about their sleep problems. Overall, the use of the PSQI does not require a high level of expertise to use in the field by sport science practitioners; and therefore, can be regarded as a useable metric. As a result, the PSQI provides a simple, cost-effective mode of assessment to capture athlete self-appraisals of sleep quality.

Sleep reactivity Assessments of sleep reactivity are limited in the sport science and medicine literature. The absence of this construct from sport science and medicine research agendas and practise requires attention, and again, highlights the need for the adoption and translation of sleep medicine derived constructs. Like the PSQI, the FIRST is widely used across clinical and non-clinical populations (Kalmbach et al., 2018). The instrument provides a simple tool to identify athletes who are vulnerable to stress-induced sleep disruption (C. L. Drake, Pillai, & Roth, 2014). Evidence for the relevance of sleep reactivity

235

instruments in sleep measurement batteries comes from the current research programme, and previous research:

1 The identification of reactive sleepers can occur in the absence of insomnia symptoms or poor sleep quality (Ellis et al., 2018) 2 Elite sport fundamentally challenges (or ‘stresses) athlete sleep through training, travel and competition (see Chapter 1 and 3); 3 Levels of sleep reactivity appear high among sub-groups of elite athletes (Chapter 7), 4 High levels of sleep reactivity are predictive of sleep disruption in response to a sleep challenge (see Chapter 4 and 5), but also (from a long-term health perspective) insomnia disorder development (Kalmbach et al., 2016). 5 Newly developed sleep assessment metrics in sport science and medicine (i.e. the ASSQ; Samuels et al. (2016)) do not include sleep reactivity constructs (Chapter 1)

Evidence for the utility of the FIRST instrument was provided throughout the research programme. Chapter 4 (Para. 4.3.3) highlighted that high levels of sleep reactivity (as measured by FIRST) were predictive of lower ‘sleep ability’ in a novel environment (e.g. first visit to a laboratory), but not following adaptation. Given many athletes are required to sleep within novel environments in the lead up to competitions (LE Juliff et al., 2015) or on training camps (Pitchford et al., 2016), measurements of sleep reactivity allow for the identification of athletes who are ‘likely’ to experience disruption, or who are vulnerable under a sleep challenge (e.g. night before/after a competition, training camps and/or long-haul travel) who may otherwise exhibit good sleep quality. Findings from Chapter 5 support this notion further with unreactive sleepers being able to protect sleep duration during a simulated tournament. From a pragmatic viewpoint, then, identification of resilient or robust sleepers can assist practitioners in focussing attention and resources on those more ‘vulnerable’. The construct of sleep reactivity may help explain differences observed in responses to sleep challenges in elite sport, and further aid the prediction of sleep disruption across a competitive season (see Chapter 1). As a result, while accepting sleep reactivity remains a novel

236

construct within elite sport, the FIRST provides a promising, relevant, and simple tool to assess sleep vulnerabilities among athletes.

Daytime sleepiness The measurement of sleepiness has received limited attention in the sport science and medicine literature. The limited number of studies conducted to date have employed Instruments that measure ‘general’ levels of sleepiness, such as the Epworth Sleepiness Scale (ESS) (Johns & Hocking, 1997). While these studies broadly show high levels of general day-time sleepiness, few studies have employed physiological measures of sleepiness, such as the MSLT. In Chapter 4, an MSLT model was adopted to assess physiological sleep tendency alongside state assessments of sleepiness via the KSS. From a research perspective, the SLT employed in Chapter 4 provides a protocol by which daytime napping and associated sleep constructs (e.g. sleepiness and sleepability) can be investigated further. The full MSLT protocol (i.e. overnight PSG following by 5 naps; (Arand et al., 2005) is unlikely to translate into elite sport due to time constraints of daily training, and other sport commitments.

Other modes of assessment have been used to indicate daytime sleepiness. Despite little empirical evidence, the observation of napping behaviour in athletes has been suggested to indicate inadequate night-time sleep, or specifically daytime sleepiness. However, this was shown in the current research programme to be problematic. Chapter 3 (Para. 3.3.3) highlighted that napping behaviours of elite athletes are not necessarily driven by poor sleep quality, and that levels of self-reported sleep tendency are varied among habitual nappers. Furthermore, Chapter 4 (Para. 4.3.3) provided evidence that high levels of physiological sleep tendency in elite athletes (relative to controls) are independent of differences in daytime sleepiness or pre-test total sleep time. As a result, researchers are recommended to use valid assessments of state daytime sleepiness when assessing athlete nap behaviours. In the field, evidence of frequent napping in elite athletic squads should not be assumed to indicate high levels of daytime sleepiness, and evidence from sources other than observed sleep tendency should be sought. Moreover, the influence of other sleep disorders not characterised by sleep onset difficulties e.g. obstructive sleep apnoea and rest leg syndrome should be considered, when interpreting daytime sleep tendency in elite athletes.

237

Operationalisation of sleep measurement Standard metrics of sleep quality, and other supporting constructs, were used in numerous contexts through the research programme and form pragmatic guidance for practitioners around when and why such metrics could be used within a sport. These are listed below:

1. To describe sleep phenotypes across a squad of athletes as part of ‘profiling’ or ‘screening’ 2. To monitor competitive-season variations in sleep quality 3. To identify ‘at-risk’ insomnia symptom sub-groups of athletes 4. To assess sleep support need 5. To assess the efficacy of targeted sleep management interventions

The thesis highlighted the importance and relevance of employing sleep measurement in elite sport using standardised metrics. Regarding sleep management, interventions in elite sport are generally delivered at a squad level (i.e. a group-based delivery) without an understanding of individual sleep phenotypes. Through understanding elite athlete sleep quality and insomnia symptoms, and sport sleep challenges across a season, interventions can be tailored and may, therefore, become more accepted and consequently effective.

8.2.2 Sleep management The thesis demonstrated that programmes of sleep management based on cognitive behavioural principles are feasible and accepted among elite team sport athletes (Chapters 6 and 7), however when delivered during competition periods appear ineffective and underutilised (Chapter 7). Overall, while tailoring information to individual sleep profiles is warranted, there is merit in providing modified CBT-I messages to athletes with poor sleep quality, without necessarily meeting diagnostic criteria for insomnia disorder. However, challenges in the design and implementation of psychoeducational sleep management exist and were highlighted in Chapters 6 and 7.

Mode of delivery Among sports that request sleep support for their athletes, delivery is more likely to be group-based (e.g. including both good and poor sleepers) and brief (i.e. one-shot) in practice. In the current thesis, frequency of delivery and format were dictated by time constraints and a perception of

238

suitability from sport’s management. Chapter 6 highlights these points. In the study, due to the timing of delivery in the Olympic cycle and upcoming squad selection for the Olympic Games, it was perceived that all athletes should receive the sleep management programme to ensure that some athletes were not at an advantage. This requirement created a challenge in delivering ‘targeted’ content. Providing elite athletes with information around sleep requires careful consideration. For example, as observed in Chapters 6 and 7, over emphasising the benefits of sleep to competitive performance (particularly when delivery is near a competition period) may create performance related anxiety in individuals exhibiting insomnia symptoms. However, among athletes without insomnia symptoms (and an absence of sleep related worry) yet exhibit other aspects of poor sleep health may benefit from this approach. Despite a high proportion of athletes being shown to exhibit high levels of insomnia symptoms, the challenge in programme design described here highlights the importance of measuring sleep within squads of athletes. This is so intervention content can be adapted appropriately to sub-groups or individuals, if appropriate. Furthermore, the timing of delivery of a sleep programme requires consideration. In Chapter 6 implementation of the ASMP during training was utilised and was effective in improving sleep outcome; however, Chapter 7 showed, when delivered during a period of competition, both utility and efficacy are low. It is plausible, therefore, that psychoeducation delivery during competition is not feasible and more suited to training periods where sleep behaviours can be rehearsed, similar to other sport science intervention (e.g. nutrition). Moreover, an extended follow up of sleep measurement was not attainable due to the Olympic Games. In practice, follow-up measurements typically do not take place once a sleep programme has been administered, however it is imperative, particularly when delivering a group-based format, that follow-up assessments take place. The research presented here has provided a range of assessments and an evaluation questionnaire which can be administered with ease and without specialist support. These considerations emphasise the importance of planning a sleep programme into a competitive season, and again highlight the importance of sleep measurement in order to monitor acceptability, utility and efficacy.

239

Content The content delivered as a programme was shown to be effective in improving sleep quality during training and other constructs such as sleep cognitions, but not during a period of competition in team sport athletes. However, which components were most feasible and effective/ ineffective remains unknown. What is highlighted in this thesis, however, is that understanding the sport-specific sleep challenges likely to be faced within a competitive season, and the mechanisms by which sleep is disrupted during each challenge, is paramount in designing sleep programme content.

Another important practical point to consider, widely across elite sporting systems is that currently ‘sleep science’ does not sit neatly within a specific discipline such as nutrition or psychology, nor is it a functioning discipline. Sleep science and medicine, as practiced within elite sport, currently includes individuals with expertise and disciplinary backgrounds in exercise physiology, physiotherapy, nutrition, medicine and psychology - all providing support around sleep in their respective sports. Across these different Olympic and Paralympic sports, which specialist provides sleep support to athletes is anecdotally highly varied, and largely based around which practitioners the sport employs as part of their support team (e.g. one physiotherapist and a nutritionist). As a result, there is a need from a system viewpoint to standardise sleep management interventions in sport and make sleep science messaging more accessible to minimise ‘discipline bias’ when working with elite athletes e.g. nutritionists providing nutrition advice to support athlete sleep. In support of this, Miles et al. (2019) stated a “Lack of resources and knowledge are the major barriers to the more frequent implementation [of sleep measurement and management]” in team sport coaches and support staff (Miles et al. (2019); pp.3). Although not specifically explored in the current programme of research, ‘discipline specific advice’ (i.e. nutrition) is invaluable in sleep programme design. To support this the point, throughout the thesis sleep quality has been shown to be a complex construct that is influenced by a multitude of factors (e.g. Chapter 1 and 7). As a result, while understanding the science of sleep, specifically the underpinning psychophysiology of sleep quality, is essential, the additional input from different members of sports’ multidisciplinary teams will help accelerate problem solving around sleep support, but also provide vital context within which an athlete’s sleep complaint may sit. Arguably, sleep should be high up practitioners’

240

agendas (from any discipline), however for sports to be capable of providing an appropriate level of first-line support, improved education and access to messaging and resources on sleep science is required.

Operationalisation of sleep management The practical messages can be summarised into considerations when planning a sleep management programme for practitioners:

1. Understand sleep challenges faced by a sport, and the mechanisms by which they are challenged 2. Carefully plan sleep programme delivery in the context of a competitive season 3. Ensure sleep management programmes incorporate periods of sleep measurement 4. Tailor sleep management according to sleep phenotype sub-groups or individuals 5. Consider a delivery format (e.g. mode, frequency and length) which is accessible to athlete cohorts

8.3. Future research directions Given the investigations into sleep quality in the context of elite sport is sparse, inevitably outcomes from each chapter lead to further research questions and hypotheses around sleep constructs discussed here.

Insomnia symptoms A conceptual diagram of Spielman, Caruso and Glovinsky's (1987) model of insomnia development in the context of elite sport is shown in Figure 8.1. While the incidence of poor sleep quality among athletes was shown to be high throughout the research programme, evidence for insomnia disorder, particularly in terms of severity and chronicity, could not be confirmed. Research into the aetiology of insomnia disorder in elite athletes is limited and warrants further investigation. Studies summarised in Chapter 1 (see Table 1.12) highlighted that caseness of insomnia close to diagnosis was found to be as high as 22% (Schaal et al., 2011). While such investigations will provide a snapshot of insomnia disorder on an epidemiological scale, the cross-sectional study design does not, however, provide insights into the dynamics of insomnia development over athletic careers. Therefore, in order to fully address the question “does elite

241

sport degrade sleep quality?” data are required on sleep quality at different stages of athletic careers. Investigations of athletes at the end of their careers, or even after their active sporting careers (i.e. transitioning out of sport) would provide insight into the consequence of elite sport participation in terms of insomnia caseness (e.g. when elite sports ‘precipitations’ have been removed). As a result, the prevalence of specific symptoms of poor sleep quality and insomnia using standardised metrics has not been conducted in cohorts of former athletes. Further research into the sleep of former athletes would provide insight into: 1) the role of sleep for career longevity and; 2) the impact of sleep quality-elite sport interactions beyond athletic careers (see Figure 8.1).

ATHLETE SLEEP

Predisposition Quality, reactivity, day-time sleepiness, timing personality

ATHLETE ATHLETE LONG CAREER TERM HEALTH LONGEVITY Health, wellbeing, Health, wellbeing, quality of life, performance insomnia

SPORT SLEEP CHALLENGE(S)

Precipitation Training, comp, travel

Figure 8.1 a conceptual diagram showing the application of Spielman’s (Spielman et al., 1987) model of insomnia development in elite sport. Black arrows = evidenced in the current research programme, Grey arrows = hypothesised link

242

Athletic performance In a similar manner, the programme of research explored sleep quality (and related constructs) among elite athletes. However, the sleep phenotypes of successful and unsuccessful athletes were not captured. Whilst the evidence generated from the thesis seems to suggest that good sleep quality is not a determinant of elite sport participation (see Chapter 3, Para 3.4.3), it remains unknown whether successful athletes have superior (or more resilient) sleep than their unsuccessful counterparts. A cross-sectional design of Olympic medallists and World champions (and multiple Olympic champions) would provide a snapshot of the extent of sleep complaints in this sub-population. Moreover, investigating the sleep of athletes that have endured multiple Olympic cycles (and Olympic/ Paralympic Games) and enjoyed long, illustrious careers would provide additional insight into the interaction between sleep and athlete career longevity (and continued success). In regard to answering the question “does sleep play a role in achieving continuous, elite sporting success?” attention is required further down the athletic development pathway. What this means is that by investigating sleep quality among younger individuals (i.e. before they embark on athletic careers) it would be possible to gain insight into the role of sleep in the identification and/or selection for elite sporting programmes. The development of athlete sleep phenotypes through adolescence will likely be influenced by sport participation, and vice versa.

Some support for this point comes from the literature. Lastella et al. (2016) reported that elite Australian athletes with morning chronotypes were more likely to participate in sports with early morning training schedules. While the cross- sectional study design used here does not allow inference of causality, it is interesting that a sleep phenotype-training schedule interaction was observed. The question remains, however, whether participation in sports that demand early morning performances (e.g. swimming, rowing, triathlon etc.) shape the sleep phenotypes (e.g. chronotype, sleep reactivity/resilience etc.) of adolescent athletes, allowing them to athletically progress, or whether the emerging sleep phenotypes of adolescent athletes causes athletes to gravitate to a particularly sport-type meaning that they are more capable to athletically progress. Inevitably, other factors will play a role in the participation of sport-types by individuals (i.e. influence of family and friends, accessibility to facilities etc.) (Greyson, Kelly, Peyrebrune, & Furniss, 2010). Nonetheless, it is plausible the role

243

of sleep characteristics will influence participation and subsequent performance in sports that adopt early morning training schedules i.e. swimming (Gudmundsdottir, 2020), particularly given added scheduling demands, such as school commitments (Greyson et al., 2010). Longitudinal research into the development and influence of sleep phenotypes on selection for and participation in elite sport participation requires further attention.

Sleep management content One aspect of the programme which was explored in the thesis was sleep management programme content (see Chapter 2, Para. 2.6). In the thesis the content employed was translated from a clinically valid programme of CBT-I (Morgan et al., 2012). The modified programme included components on sleep hygiene, mild-sleep restriction, stimulus control, arousal management, and optimal sleep scheduling (e.g. napping). While these components are the same as those included in more traditional CBT-I programmes (van Straten et al., 2017), it remains unknown whether all components are equally valuable in the context of elite sport, given the frequent demand for high levels of day-time performance. For example, sleep restriction has been shown to be potent regarding reducing insomnia symptom severity and chronicity (see Maurer, Espie and Kyle (2018)). However, in the early stages (e.g. up to three months) of treatment high levels of day-time sleepiness and impairments in cognitive function have been observed in insomnia patients (Kyle et al., 2014). Moreover, partial sleep deprivation in elite athletes has shown impairments in cognitive function, wellbeing and physical performance (see Fullagar et al. (2015)) albeit not shown as part of a sleep restriction protocol. In combination, these findings would suggest that some forms of insomnia management may be inappropriate in a competitive sporting context, where performance is the paramount consideration. Moreover, whether an adapted version of sleep restriction is required for elite sport e.g. supplemented with daytime napping and (or) use of caffeine to manage day-time sleepiness symptoms, around performances requires further investigation. Overall, components of CBT-I require review and investigation in the context of elite sport regarding their feasibility and efficacy with athletes exhibiting insomnia symptoms.

244

The research programme adopted a group delivery format of cognitive behavioural principles of sleep management over 2-4 sessions. As stated earlier (Para. 8.3), the frequency of delivery, and length of sessions were dictated by sport availability, and competition/ training scheduling. As a result, traditional formats of CBT-I delivery do not apply and are required to be modified. Previous research has shown the positive impact of “one-shot” delivery of cognitive behavioural therapy for insomnia in different populations (Boullin et al., 2016; Ellis et al., 2015). Whilst these sessions have been shown to be beneficial among individuals with acute insomnia, whether a similar ‘brief’ format of delivery is equally effective in athletes with poor sleep quality and insomnia symptoms is unexplored but seems plausible. In the current study, group-based programme of psychoeducation was attended by most athletes in Chapters 6 and 7. Although this format was found to be feasible in regards to implementing a sleep management programme, the optimal frequency of delivery was not explored. Therefore, further research into minimal frequency of group-based sessions are required. Moreover, a group-based format of delivery is not feasible across all Olympic and Paralympic sports in the UK. As a result, research into the use of more novel delivery formats requires further exploration. The use of online programmes has shown to be effective in increasing accessibility and adherence to CBT-I in insomnia groups (Ye et al., 2016) with promising outcomes in insomnia severity. As a result, the use of online programmes of sleep management may be a useful format for elite athletes, particularly given the conflicting demands for time in athlete schedules.

8.4 Summary The aim of the thesis was to explore the construct of sleep quality in the context of elite sport performance. It was acknowledged at the beginning of this programme that a disconnect existed between the importance placed upon sleep by athletes (Rachel Elizabeth Venter, 2014) and coaches (Fallon, 2007) and the current evidence-base by which principles of sleep profiling and management are conceptually developed, researched and ultimately practised. As a result, while it is recognised as essential that sport science and medicine continue to develop its own sleep science research agendas, it is equally important that existing metrics, assumptions and theories derived from other, larger bodies of evidence in relevant areas of sleep medicine are

245

accepted, considered and translated. In relation to this, it is important that future research continues to connect, rather than separate, the behavioural and physiological sleep phenomenology of elite athletes and their non-athletic counterparts to accelerate the evolution of complex theoretical and conceptual sleep constructs, and the development of subsequent protocols of sleep management delivery and applied practice.

246

Chapter 9

9. Appendices

This section provides additional material, which is referred to throughout the thesis, and include:

Appendix A – sleep infographics (deployed in Chapters 6 and 7);

Appendix B – a sleep diary (used throughout alongside actigraphy); and

Appendix C – a copy of the consent form completed online by athletes for chapter 3 (and on other landing pages where sleep questionnaires were deployed online – Chapter 5, 6 and 7).

247

9.1 Appendix A - ASMP infographics

9.1.1 Workshop one

248

9.1.2 Workshop two

249

9.1.3 Workshop three

250

9.1.4 Workshop four

251

9.2 Appendix B - Sleep diary

DAILY SLEEP DIARY

252

9.3 Appendix C - Sport Sleep Survey landing page

253

254

255

Chapter 10

10. References

Abeln, V., Kleinert, J., Strüder, H. K., & Schneider, S. (2014). Brainwave entrainment for better sleep and post-sleep state of young elite soccer players - A pilot study. European Journal of Sport Science. https://doi.org/10.1080/17461391.2013.819384

Adan, A., & Natale, V. (2002). Gender differences in morningness-eveningness preference. Chronobiology International, 19(4), 709–720. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12182498

Agnew, H. W., Webb, W. B., & Williams, R. L. (1966). The first night effect: an EEG study of sleep. Psychophysiology, 2(3), 263–266. https://doi.org/10.1111/j.1469-8986.1966.tb02650.x

Ahrberg, K., Dresler, M., Niedermaier, S., Steiger, A., & Genzel, L. (2012). The interaction between sleep quality and academic performance. Journal of Psychiatric Research, 46(12), 1618–1622. https://doi.org/10.1016/j.jpsychires.2012.09.008

Ainsworth, B. E., Haskell, W. L., Herrmann, S. D., Meckes, N., Bassett, D. R., Tudor- Locke, C., … Leon, A. S. (2011, August). 2011 compendium of physical activities: A second update of codes and MET values. Medicine and Science in Sports and Exercise, Vol. 43, pp. 1575–1581. https://doi.org/10.1249/MSS.0b013e31821ece12

Åkerstedt, T., Anund, A., Axelsson, J., & Kecklund, G. (2014). Subjective sleepiness is a sensitive indicator of insufficient sleep and impaired waking function. Journal of Sleep Research, 23(3), 242–254. https://doi.org/10.1111/jsr.12158

Akerstedt, T., & Gillberg, M. (1990). Subjective and objective sleepiness in the active individual. The International Journal of Neuroscience, 52(1–2), 29–37. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/2265922

256

Allen, M., Greenlees, I., & Jones, M. (2013). Personality in sport: a comprehensive review. International Review of Sport and Exercise Psychology, 6(1), 184– 208. https://doi.org/10.1080/1750984X.2013.769614

American Academy of Sleep Medicine. (2005). International Classification of Sleep Disorders II: Diagnostic and Coding Manual. In Diagnostic and Coding Manual. Retrieved from http://www.amazon.com/dp/0965722023%5Cnhttp://www.esst.org/adds/I CSD.pdf

American Academy of Sleep Medicine. (2014). International Classification of Sleep Disorders: Diagnostic and Coding Manual. 3nd edition. In Diagnostic Coding Manual. https://doi.org/10.1111/febs.12678

American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5). In Diagnostic and Statistical Manual of Mental Disorders 4th edition TR. https://doi.org/10.1176/appi.books.9780890425596.744053

Ancoli-Israel, S. (2009). Sleep and its disorders in aging populations. Sleep Medicine, 10(SUPPL. 1), S7–S11. https://doi.org/10.1016/j.sleep.2009.07.004

Arand, D., Bonnet, M., Hurwitz, T., Mitler, M., Rosa, R., & Sangal, R. B. (2005). The clinical use of the MSLT and MWT. Sleep, 28(1), 123–144.

Arnal, P. J., Lapole, T., Erblang, M., Guillard, M., Bourrilhon, C., Léger, D., … Millet, G. Y. (2016). Sleep Extension before Sleep Loss: Effects on Performance and Neuromuscular Function. Medicine and Science in Sports and Exercise, 48(8), 1595–1603. https://doi.org/10.1249/MSS.0000000000000925

Arnold, R., & Fletcher, D. (2012). A Research Synthesis and Taxonomic Classification of the Organizational Stressors Encountered by Sport Performers. Journal of Sport & Exercise Psychology, 34(3), 397–429. https://doi.org/10.1123/jsep.34.3.397

Ashley, P., Di Iorio, A., Cole, E., Tanday, A., & Needleman, I. (2015, January). Oral health of elite athletes and association with performance: A systematic review. British Journal of Sports Medicine, Vol. 49, pp. 14–19.

257

https://doi.org/10.1136/bjsports-2014-093617

Baddeley, A. D., & Hitch, G. (1974). Working memory. The Psychology of Learning and Motivation Advances in Research and Theory. https://doi.org/10.1016/S0079-7421(08)60452-1

Baehr, E. K., Revelle, W., & Eastman, C. I. (2000). Individual differences in the phase and amplitude of the human circadian temperature rhythm: with an emphasis on morningness-eveningness. Journal of Sleep Research, 9(2), 117–127. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10849238

Bangsbo, J. (1994). The physiology of soccer - With special reference to intense intermittent exercise. Acta Physiologica Scandinavica, Supplement, 151(619), 1–155.

Barclay, N. L., & Ellis, J. G. (2013). Sleep-related attentional bias in poor versus good sleepers is independent of affective valence. Journal of Sleep Research, 22(4), 414–421. https://doi.org/10.1111/jsr.12035

Bastien, C. H. (2011). Insomnia: Neurophysiological and neuropsychological approaches. Neuropsychology Review, 21(1), 22–40. https://doi.org/10.1007/s11065-011-9160-3

Bastien, C. H., Morin, C. M., Ouellet, M.-C., Blais, F. C., & Bouchard, S. (2004). Cognitive-Behavioral Therapy for Insomnia: Comparison of Individual Therapy, Group Therapy, and Telephone Consultations. Journal of Consulting and Clinical Psychology, 72(4), 653–659. https://doi.org/10.1037/0022-006X.72.4.653

Bender, A. M., Lawson, D., Werthner, P., & Samuels, C. H. (2018). The Clinical Validation of the Athlete Sleep Screening Questionnaire: an Instrument to Identify Athletes that Need Further Sleep Assessment. Sports Medicine - Open, 4(1), 23. https://doi.org/10.1186/s40798-018-0140-5

Bleyer, F. T. de S., Barbosa, D. G., Andrade, R. D., Teixeira, C. S., & Felden, É. P. G. (2015). Sleep and musculoskeletal complaints among elite athletes of Santa Catarina. Revista Dor, 16(2), 102–108. https://doi.org/10.5935/1806- 0013.20150020

258

Bonnet, M H, & Arand, D. L. (1997). Hyperarousal and insomnia. Sleep Medicine Reviews, 1(2), 97–108. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15310517

Bonnet, M H, & Arand, D. L. (1998). Sleepiness as measured by modified multiple sleep latency testing varies as a function of preceding activity. Sleep, 21(5), 477–483. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9703587

Bonnet, Michael H, & Arand, D. L. (2003). Situational insomnia: consistency, predictors, and outcomes. Sleep, 26(8), 1029–1036. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14746386

Bonnet, Michael H, & Arand, D. L. (2010). Hyperarousal and insomnia: state of the science. Sleep Medicine Reviews, 14(1), 9–15. https://doi.org/10.1016/j.smrv.2009.05.002

Bootzin, R. R., Epstein, D., & Wood, J. M. (1991). Stimulus Control Instructions. In Case Studies in Insomnia (pp. 19–28). https://doi.org/10.1007/978-1-4757- 9586-8_2

Bootzin RR, & Nicassio PM. (1978). Behavioural treatment for insomnia. In E. RM & M. Miller (Eds.), Progress in behaviour modification (pp. 1–45). New York: Academic Press.

Boullin, P., Ellwood, C., & Ellis, J. (2016). Group vs. Individual Treatment for Acute Insomnia: A Pilot Study Evaluating a “One-Shot” Treatment Strategy. Brain Sciences, 7(1), 1. https://doi.org/10.3390/brainsci7010001

Broughton, R. J., & Dinges, D. F. (1989). Napping: a ubiquitous enigma. In R. . Broughton & D. F. Dinges (Eds.), Sleep and Alertness: Chronobiological, Behavioural, and Medical Aspects of Napping (pp. 1–7). NY: Raven Press.

Buchheit, M., Lefebvre, B., Laursen, P. B., & Ahmaidi, S. (2011). Reliability, Usefulness, and Validity of the 30–15 Intermittent Ice Test in Young Elite Ice Hockey Players. Journal of Strength and Conditioning Research, 25(5), 1457–1464. https://doi.org/10.1519/JSC.0b013e3181d686b7

Buysse, D. J., Reynolds, C. F., Monk, T. H., Berman, S. R., & Kupfer, D. J. (1989).

259

The Pittsburgh sleep quality index: A new instrument for psychiatric practice and research. Psychiatry Research, 28(2), 193–213. https://doi.org/10.1016/0165-1781(89)90047-4

Chennaoui, M., Arnal, P. J., Sauvet, F., & Léger, D. (2015). Sleep and exercise: A reciprocal issue? Sleep Medicine Reviews, 20, 59–72. https://doi.org/10.1016/j.smrv.2014.06.008

Chennaoui, M., Bougard, C., Drogou, C., Langrume, C., Miller, C., Gomez- Merino, D., & Vergnoux, F. (2016). Stress biomarkers, mood states, and sleep during a major competition: “Success” and “failure” athlete’s profile of high-level swimmers. Frontiers in Physiology, 7, 94. https://doi.org/10.3389/fphys.2016.00094

Chervin, R. D., & Aldrich, M. S. (2000). Sleep onset REM periods during multiple sleep latency tests in patients evaluated for sleep apnea. American Journal of Respiratory and Critical Care Medicine, 161(2 I), 426–431. https://doi.org/10.1164/ajrccm.161.2.9905071

Claudino, J. G., Gabbet, T. J., De Sá Souza, H., Simim, M., Fowler, P., De Alcantara Borba, D., … Nassis, G. P. (2019, January 1). Which parameters to use for sleep quality monitoring in team sport athletes? A systematic review and meta-analysis. BMJ Open Sport and Exercise Medicine, Vol. 5, p. bmjsem-2018-000475. https://doi.org/10.1136/bmjsem-2018-000475

Coates. (1981). First Night Effects in Good Sleepers and Sleep-Maintenance Insomniacs When Recorded at Home. SLEEP, 4(3), 293–298. Retrieved from https://www.ncbi.nlm.nih.gov/labs/articles/7302460/

Cummins, C., Orr, R., O’Connor, H., & West, C. (2013). Global Positioning Systems (GPS) and Microtechnology Sensors in Team Sports: A Systematic Review. Sports Medicine, 43(10), 1025–1042. https://doi.org/10.1007/s40279-013- 0069-2

Cummiskey, J., Natsis, K., & Papathanasiou, E. (2013). Sleep and athletic performance. European Journal of Sports Medicine, 1(1), 13–22. Retrieved from http://www.eujsm.com/index.php/EUJSM/article/view/53

260

Davenne, D. (2009). Sleep of athletes – problems and possible solutions. Biological Rhythm Research, 40(1), 45–52. https://doi.org/10.1080/09291010802067023

Davidson, J. R., Dawson, S., & Krsmanovic, A. (2017). Effectiveness of Group Cognitive Behavioral Therapy for Insomnia (CBT-I) in a Primary Care Setting. Behavioral Sleep Medicine, 1–13. https://doi.org/10.1080/15402002.2017.1318753

Davies, D. J., Graham, K. S., & Chow, C. M. (2010). The effect of prior endurance training on nap sleep patterns. International Journal of Sports Physiology and Performance, 5(1), 87–97. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20308699

Dekker, M. K. J., Van den Berg, B. R., Denissen, A. J. M., Sitskoorn, M. M., & van Boxtel, G. J. M. (2014). Feasibility of eyes open alpha power training for mental enhancement in elite gymnasts. Journal of Sports Sciences, 32(16), 1550–1560. https://doi.org/10.1080/02640414.2014.906044

Dickinson, R., & Hanrahan, S. (2009). An investigation of subjective sleep and fatigue measures for use with elite athletes. Journal of Clinical Sport, 3(3), 244–266. Retrieved from https://www.researchgate.net/profile/Stephanie_Hanrahan/publication/43 523337_An_investigation_of_subjective_sleep_and_fatigue_measures_for_u se_with_elite_athletes/links/545a07d20cf2cf5164840f72.pdf

Dieck, A., Helbig, S., Drake, C. L., & Backhaus, J. (2018). Validation of the German version of the Ford Insomnia Response to Stress Test. Journal of Sleep Research, 27(3), e12621. https://doi.org/10.1111/jsr.12621

Dijk, D.-J., Groeger, J. A., Stanley, N., & Deacon, S. (2010). Age-related reduction in daytime sleep propensity and nocturnal slow wave sleep. Sleep, 33(2), 211–223. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20175405

Drake, C. L., Jefferson, C., Roehrs, T., Roth, T., Ohayon, M. M., Leger, D., … Glovinsky, P. B. (2006). Stress-related sleep disturbance and polysomnographic response to caffeine. Sleep Medicine, 7(7), 567–572.

261

https://doi.org/10.1016/j.sleep.2006.03.019

Drake, C. L., Pillai, V., & Roth, T. (2014). Stress and Sleep Reactivity: A Prospective Investigation of the Stress-Diathesis Model of Insomnia. Sleep, 37(8), 1295– 1304. https://doi.org/10.5665/sleep.3916

Drake, C., Richardson, G., Roehrs, T., Scofield, H., & Roth, T. (2004). Vulnerability to stress-related sleep disturbance and hyperarousal. Sleep, 27(2), 285–291. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15124724

Driller, M. W., & Argus, C. K. (2016). Flotation restricted environmental stimulation therapy and napping on mood state and muscle soreness in elite athletes: A novel recovery strategy? Performance Enhancement & Health, 5(2), 60– 65. https://doi.org/10.1016/j.peh.2016.08.002

Driller, M. W., Lastella, M., & Sharp, A. P. (2019). Individualized sleep education improves subjective and objective sleep indices in elite cricket athletes: A pilot study. Journal of Sports Sciences, 37(17), 2121–2125. https://doi.org/10.1080/02640414.2019.1616900

Driver, H. S., & Taylor, S. R. (2000). Exercise and sleep. Sleep Medicine Reviews, 4(4), 387–402. https://doi.org/10.1053/smrv.2000.0110

Driver, S., & Taylor, S. R. (1996). Sleep disturbances and exercise. Sports Medicine, 21(1), 1–6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/8771281

Duffield, R., Murphy, A., Kellett, A., & Reid, M. (2014). Recovery from repeated on-court tennis sessions: Combining cold-water immersion, compression, and sleep interventions. International Journal of Sports Physiology and Performance. https://doi.org/10.1123/IJSPP.2012-0359

Duggan, K. A., McDevitt, E. A., Whitehurst, L. N., & Mednick, S. C. (2018). To Nap, Perchance to DREAM: A Factor Analysis of College Students’ Self-Reported Reasons for Napping. Behavioral Sleep Medicine, 16(2), 135–153. https://doi.org/10.1080/15402002.2016.1178115

Dunican, I. C., Higgins, C. C., Jones, M. J., Clarke, M. W., Murray, K., Dawson, B., … Eastwood, P. R. (2018). Caffeine use in a Super Rugby game and its

262

relationship to post-game sleep. European Journal of Sport Science, 18(4), 513–523. https://doi.org/10.1080/17461391.2018.1433238

Durán Agüero, S., Arroyo Jofre, P., Varas Standen, C., Herrera-Valenzuela, T., Moya Cantillana, C., Pereira Robledo, R., & Valdés-Badilla, P. (2015). Sleep quality excessive daytime sleepiness and insomnia in Chilean paralympic athletes. Nutricion Hospitalaria, 32(6), 2832–2837. https://doi.org/10.3305/nh.2015.32.6.9893

Edinger, J. D., Bonnet, M. H., Bootzin, R. R., Doghramji, K., Dorsey, C. M., Espie, C. A., … Stepanski, E. J. (2004, December 15). Derivation of research diagnostic criteria for insomnia: Report of an American Academy of Sleep Medicine work group. Sleep, Vol. 27, pp. 1567–1596. https://doi.org/10.1093/sleep/27.8.1567

Edinger, J. D., Fins, A. I., Glenn, D. M., Sullivan, R. J., Bastian, L. A., Marsh, G. R., … Vasilas, D. (2000). Insomnia and the eye of the beholder: Are there clinical markers of objective sleep disturbances among adults with and without insomnia complaints? Journal of Consulting and Clinical Psychology, 68(4), 586–593. https://doi.org/10.1037/0022-006X.68.4.586

Edinger, J. D., & Sampson, W. S. (2003). A primary care “friendly” cognitive behavioral insomnia therapy. Sleep, 26(2), 177–182. https://doi.org/10.1093/sleep/26.2.177

Edinger, J. D., Ulmer, C. S., & Means, M. K. (2013). Sensitivity and specificity of polysomnographic criteria for defining insomnia. Journal of Clinical Sleep Medicine, 9(5), 481–491. https://doi.org/10.5664/jcsm.2672

Edinger, J. D., Wohlgemuth, W. K., Radtke, R. A., Marsh, G. R., & Quillian, R. E. (2001). Does Cognitive-Behavioral Insomnia Therapy Alter Dysfunctional Beliefs About Sleep? Sleep, 24(5), 591–599. https://doi.org/10.1093/sleep/24.5.591

Edmonds, R., Burkett, B., Leicht, A., & McKean, M. (2015). Effect of Chronic Training on Heart Rate Variability, Salivary IgA and Salivary Alpha-Amylase in Elite Swimmers with a Disability. PLOS ONE, 10(6), e0127749. https://doi.org/10.1371/journal.pone.0127749

263

Ehrlenspiel, F., Erlacher, D., & Ziegler, M. (2016). Changes in subjective sleep quality before a competition and their relation to competitive anxiety. Behavioral Sleep Medicine, 9, 1–14. https://doi.org/10.1080/15402002.2016.1253012

Elbayoumy, A., & Elbayoumy, A. (2015). Relationship of Sleep Quality and Copetitve Anxiety in a Sample of Swimming Athletes. The Swedish Journal of Scientific Research, 2(11), 2001–2011.

Ellis, J. G., Allen, S., Perlis, M., Grandner, M., Gardani, M., & Espie, C. (2018). Sleep Reactivity and Psychological Health in Normal Sleepers: the moderating role of stress. https://doi.org/10.31234/OSF.IO/NXZ42

Ellis, J. G., Cushing, T., & Germain, A. (2015). Treating Acute Insomnia: A Randomized Controlled Trial of a “Single-Shot” of Cognitive Behavioral Therapy for Insomnia. SLEEP, 38(6). https://doi.org/10.5665/sleep.4752

Erlacher, D., Ehrlenspiel, F., Adegbesan, O., & El-Din, H. G. (2011). Sleep habits in German athletes before important competitions or games. Journal of Sports Sciences, 29(8), 859–866. https://doi.org/10.1080/02640414.2011.565782

Espie, C. A., Broomfield, N. M., MacMahon, K. M. A., Macphee, L. M., & Taylor, L. M. (2006). The attention-intention-effort pathway in the development of psychophysiologic insomnia: a theoretical review. Sleep Medicine Reviews, 10(4), 215–245. https://doi.org/10.1016/j.smrv.2006.03.002

Espie, C. A., Kyle, S. D., Williams, C., Ong, J. C., Douglas, N. J., Hames, P., & Brown, J. S. L. (2012). A Randomized, Placebo-Controlled Trial of Online Cognitive Behavioral Therapy for Chronic Insomnia Disorder Delivered via an Automated Media-Rich Web Application. Sleep, 35(6), 769–781. https://doi.org/10.5665/sleep.1872

Evans, F. J., Cook, M. R., Cohen, H. D., Orne, E. C., & Orne, M. T. (1977). Appetitive and replacement naps: EEG and behavior. Science, 197(4304), 687–689. https://doi.org/10.1126/science.17922

Exelmans, L., & Van den Bulck, J. (2017). Binge Viewing, Sleep, and the Role of

264

Pre-Sleep Arousal. Journal of Clinical Sleep Medicine : JCSM : Official Publication of the American Academy of Sleep Medicine, 13(8), 1001– 1008. https://doi.org/10.5664/jcsm.6704

Fallon, K. E. (2007). Blood tests in tired elite athletes: Expectations of athletes, coaches and sport science/sports medicine staff. British Journal of Sports Medicine, 41(1), 41–44. https://doi.org/10.1136/bjsm.2006.030999

Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191. https://doi.org/10.3758/BF03193146

Ferrara, M., & De Gennaro, L. (2001, April). How much sleep do we need? Sleep Medicine Reviews, Vol. 5, pp. 155–179. https://doi.org/10.1053/smrv.2000.0138

Fietze, I., Strauch, J., Holzhausen, M., Glos, M., Theobald, C., Lehnkering, H., & Penzel, T. (2009). Sleep quality in professional ballet dancers. Chronobiology International, 26(6), 1249–1262. https://doi.org/10.1080/07420520903221319

Foster, C, Daines, E., Hector, L., Snyder, A. C., & Welsh, R. (1996). Athletic performance in relation to training load. Wisconsin Medical Journal, 95(6), 370–374. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/8693756

Foster, Carl, Rodriguez-Marroyo, J. A., & de Koning, J. J. (2017). Monitoring Training Loads: The Past, the Present, and the Future. International Journal of Sports Physiology and Performance, 12(Suppl 2), S22–S28. https://doi.org/10.1123/ijspp.2016-0388

Fowler, P. M., Murray, A., Farooq, A., Lumley, N., & Taylor, L. (2017). Subjective And Objective Responses To Two Rugby 7’s World Series Competitions. Journal of Strength and Conditioning Research, 1. https://doi.org/10.1519/JSC.0000000000002276

Fowler, P, Duffield, R., Howle, K., Waterson, A., & Vaile, J. (2015). Effects of northbound long-haul international air travel on sleep quantity and

265

subjective jet lag and wellness in professional Australian soccer players. International Journal of Sports Physiology and Performance, 10(5), 648–654. https://doi.org/10.1123/ijspp.2014-0490

Fowler, Peter, Duffield, R., & Vaile, J. (2014). Effects of domestic air travel on technical and tactical performance and recovery in soccer. International Journal of Sports Physiology and Performance, 9(3), 378–386. https://doi.org/10.1123/IJSPP.2013-0484

Fowler, PM, Duffield, R., Lu, D., Hickmans, J., & Scott, T. (2016). Effects of long- haul transmeridian travel on subjective jet-lag and self-reported sleep and upper respiratory symptoms in professional rugby league players. International Journal of Sports Physiology and Performance, 11(7), 876–884. https://doi.org/10.1123/ijspp.2015-0542

Freeman, D., Sheaves, B., Goodwin, G. M., Yu, L.-M., Harrison, P. J., Emsley, R., … Espie, C. A. (2015). Effects of cognitive behavioural therapy for insomnia on the mental health of university students: study protocol for a randomized controlled trial. Trials, 16(1), 236. https://doi.org/10.1186/s13063-015-0756-4

Fullagar, H. H. K., Duffield, R., Skorski, S., White, D., Bloomfield, J., Kölling, S., & Meyer, T. (2016). Sleep, travel, and recovery responses of national footballers during and after long-haul international air travel. International Journal of Sports Physiology and Performance, 11(1), 86–95. https://doi.org/10.1123/ijspp.2015-0012

Fullagar, H., Skorski, S., Duffield, R., Hammes, D., Coutts, A. J., & Meyer, T. (2015). Sleep and athletic performance: The effects of sleep loss on exercise performance, and physiological and cognitive responses to exercise. Sports Medicine, Vol. 45, pp. 161–186. https://doi.org/10.1007/s40279-014- 0260-0

Fullagar, H., Skorski, S., Duffield, R., Julian, R., Bartlett, J., & Meyer, T. (2016). Impaired sleep and recovery after night matches in elite football players. Journal of Sports Sciences, 34(14), 1333–1339. https://doi.org/10.1080/02640414.2015.1135249

Fullagar, H., Skorski, S., Duffield, R., & Meyer, T. (2016). The effect of an acute

266

sleep hygiene strategy following a late-night soccer match on recovery of players. Chronobiology International, 0528(March), 1–16. https://doi.org/10.3109/07420528.2016.1149190

Gao, R., Lv, Y., Li, X., Zhou, K., Jin, X., Dang, S., & Li, N. (2014). Effects of comprehensive sleep management on sleep quality in university students in mainland China. Sleep and Biological Rhythms, 12(3), 194–202. https://doi.org/10.1111/sbr.12063

Gelaye, B., Zhong, Q.-Y., Barrios, Y. V, Redline, S., Drake, C. L., & Williams, M. A. (2016). Psychometric Evaluation of the Ford Insomnia Response to Stress Test (FIRST) in Early Pregnancy. Journal of Clinical Sleep Medicine : JCSM : Official Publication of the American Academy of Sleep Medicine, 12(4), 579–587. https://doi.org/10.5664/jcsm.5696

Gilner, J., Morgan, G., & Leech, N. (2009). Research methods in applied setting: An integrated approach to design and analysis (2nd ed.). NY: Routledge/ Taylor & Francis Group.

Greyson, I., Kelly, S., Peyrebrune, M., & Furniss, B. (2010). Interpreting and Implementing the Long Term Athlete Development Model: English Swimming Coaches’ Views on the (Swimming) LTAD in Practice. International Journal of Sports Science & Coaching, 5(3), 403–406. https://doi.org/10.1260/1747-9541.5.3.403

Gudmundsdottir, S. L. (2020). Training schedule and sleep in adolescent swimmers. Pediatric Exercise Science, 32(1), 16–22. https://doi.org/10.1123/pes.2019-0067

Gupta, L., Morgan, K., & Gilchrist, S. (2017). Does elite sport degrade sleep quality? A systematic review. Sports Medicine, 47(7), 1317–1333. https://doi.org/10.1007/s40279-016-0650-6

Halson, S. (2008). Nutrition, sleep and recovery. European Journal of Sport Science, 8(2), 119–126. https://doi.org/10.1080/17461390801954794

Halson, S. (2014a). Monitoring training load to understand fatigue in athletes. Sports Medicine, 44(Suppl 2), S139-47. https://doi.org/10.1007/s40279-014-

267

0253-z

Halson, S. (2014b). Sleep in elite athletes and nutritional interventions to enhance sleep. Sports Medicine, 44(S1), 13–23. https://doi.org/10.1007/s40279-014-0147-0

Harada, T., Wada, K., Tsuji, F., Krejci, M., Kawada, T., Noji, T., … Takeuchi, H. (2016). Intervention study using a leaflet entitled “three benefits of ‘“go to bed early! get up early! and intake nutritionally rich breakfast!”’ a message for athletes” to improve the soccer performance of university soccer team. Sleep and Biological Rhythms, 14(1), S65–S74. https://doi.org/10.1007/s41105-015-0035-5

Harrison, Y., & Horne, J. A. (1996). “High sleepability without sleepiness”. The ability to fall asleep rapidly without other signs of sleepiness. Neurophysiologie Clinique, 26(1), 15–20. https://doi.org/10.1016/0987- 7053(96)81530-9

Harvey, A G. (2002). A cognitive model of insomnia. Behaviour Research and Therapy, 40(8), 869–893. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12186352

Harvey, Allison G., Stinson, K., Whitaker, K. L., Moskovitz, D., & Virk, H. (2008). The subjective meaning of sleep quality: A comparison of individuals with and without insomnia. Sleep, 31(3), 383–393. https://doi.org/10.1093/sleep/31.3.383

Harvey, C., Gehrman, P., Espie, C., Spielman, A., Caruso, L., Glovinsky, P., … Stein, M. B. (2014). Who is predisposed to insomnia: a review of familial aggregation, stress-reactivity, personality and coping style. Sleep Medicine Reviews, 18(3), 237–247. https://doi.org/10.1016/j.smrv.2013.11.004

Hauri, P. J., & Olmstead, E. M. (1989). Reverse first night effect in insomnia. Sleep, 12(2), 97–105. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/2711096

Hausswirth, C., Louis, J., Aubry, A., Bonnet, G., Duffield, R., & LE Meur, Y. (2014). Evidence of disturbed sleep and increased illness in overreached

268

endurance athletes. Medicine and Science in Sports and Exercise, 46(5), 1036–1045. https://doi.org/10.1249/MSS.0000000000000177

Hershner, S., & O’Brien, L. M. (2018). The impact of a randomized sleep education intervention for college students. Journal of Clinical Sleep Medicine, 14(3), 337–347. https://doi.org/10.5664/jcsm.6974

Hirshkowitz, M., Whiton, K., Albert, S. M., Alessi, C., Bruni, O., DonCarlos, L., … Adams Hillard, P. J. (2015). National sleep foundation’s sleep time duration recommendations: Methodology and results summary. Sleep Health, 1(1), 40–43. https://doi.org/10.1016/j.sleh.2014.12.010

Horne, J. (2010). Sleepiness as a need for sleep: When is enough, enough? Neuroscience & Biobehavioral Reviews, 34(1), 108–118. https://doi.org/10.1016/j.neubiorev.2009.07.009

Horne, J. A., & Ostberg, O. (1976). A self-assessment questionnaire to determine morningness-eveningness in human circadian rhythms. International Journal of Chronobiology, 4(2), 97–110. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1027738

Irish, L., Kline, C., Gunn, H., Buysse, D., & Hall, M. (2015). The role of sleep hygiene in promoting public health: A review of empirical evidence. Sleep Medicine Reviews, 22, 23–36. https://doi.org/10.1016/j.smrv.2014.10.001

Jacobson, E. (1938). Progressive relaxation (2nd ed.). Chicago: University Press.

Jacobson, E. (1964). Anxiety and tension control: a physiologic approach. Philidelphia: Lippincott.

Jansson-Fröjmark, M., Bermås, M., & Kjellén, A. (2013). Attentional bias in insomnia: The dot-probe task with pictorial stimuli depicting daytime fatigue/malaise. Cognitive Therapy and Research, 37(3), 534–546. https://doi.org/10.1007/s10608-012-9486-z

Jansson-Fröjmark, M., & Norell-Clarke, A. (2012). Psychometric properties of the Pre-Sleep Arousal Scale in a large community sample. Journal of Psychosomatic Research, 72(2), 103–110. https://doi.org/10.1016/j.jpsychores.2011.10.005

269

Jarrin, D. C., Chen, I. Y., Ivers, H., & Morin, C. M. (2014). The role of vulnerability in stress-related insomnia, social support and coping styles on incidence and persistence of insomnia. Journal of Sleep Research, 23(6), 681–688. https://doi.org/10.1111/jsr.12172

Jin, L., Zhou, J., Peng, H., Ding, S., & Yuan, H. (2018). Investigation on dysfunctional beliefs and attitudes about sleep in chinese college students. Neuropsychiatric Disease and Treatment, 14, 1425–1432. https://doi.org/10.2147/NDT.S155722

Johns, M., & Hocking, B. (1997). Daytime sleepiness and sleep habits of Australian workers. Sleep, 20(10), 844–849. https://doi.org/10.1093/sleep/20.10.844

Juliff, Laura, Peiffer, J. J., & Halson, S. (2017). Night Games: Physiological, Neuroendocrine and Psychometric Mechanisms to Explain Poor Sleep. International Journal of Sports Physiology and Performance, 1–22. https://doi.org/10.1123/ijspp.2016-0809

Juliff, LE, Halson, L., & Peiffer, J. (2015). Understanding sleep disturbance in athletes prior to important competitions. Journal of Science and Medicine in Sport, 18(1), 13–18. https://doi.org/10.1016/j.jsams.2014.02.007

Kahn, M., Sheppes, G., & Sadeh, A. (2013). Sleep and emotions: bidirectional links and underlying mechanisms. International Journal of Psychophysiology, 89(2), 218–228. https://doi.org/10.1016/j.ijpsycho.2013.05.010

Kalmbach, D. A., Cuamatzi-Castelan, A. S., Tonnu, C. V., Tran, K. M., Anderson, J. R., Roth, T., & Drake, C. L. (2018). Hyperarousal and sleep reactivity in insomnia: Current insights. Nature and Science of Sleep, Vol. 10, pp. 193– 201. https://doi.org/10.2147/NSS.S138823

Kalmbach, D. A., Pillai, V., Cheng, P., Arnedt, J. T., Drake, C. L., McMenamin, T. M., … al., et. (2015). Shift work disorder, depression, and anxiety in the transition to rotating shifts: the role of sleep reactivity. Sleep Medicine, 16(12), 1532–1538. https://doi.org/10.1016/j.sleep.2015.09.007

270

Kalmbach, Pillai, V., Arnedt, J. T., Anderson, J. R., Drake, C. L., Jarrin, D. ., … Al., E. (2016). Sleep system sensitization: evidence for changing roles of etiological factors in insomnia. Sleep Medicine, 21, 63–69. https://doi.org/10.1016/j.sleep.2016.02.005

Kalmbach, Vivekpillai, J., Toddarnedt, C., & Drake, C. . (2016). Identifying at-risk individuals for insomnia using the Ford Insomnia Response to Stress Test. SLEEP, 39(2). https://doi.org/10.5665/sleep.5462

Kelly, M. P., & Barker, M. (2016). Why is changing health-related behaviour so difficult? Public Health, 136, 109–116. https://doi.org/10.1016/J.PUHE.2016.03.030

Killer, S. C., Svendsen, I. S., Jeukendrup, A. E., & Gleeson, M. (2015). Evidence of disturbed sleep and mood state in well-trained athletes during short-term intensified training with and without a high carbohydrate nutritional intervention. Journal of Sports Sciences, 35(14), 1402–1410. https://doi.org/10.1080/02640414.2015.1085589

Klein, Laura C., Bennett, J. M., Whetzel, C. A., Granger, D. A., & Ritter, F. E. (2010). Caffeine and stress alter salivary α-amylase activity in young men. Human Psychopharmacology: Clinical and Experimental, 25(5), 359–367. https://doi.org/10.1002/hup.1126

Klein, Laura Cousino, Whetzel, C. A., Bennett, J. M., Ritter, F. E., Nater, U. M., & Schoelles, M. (2014). Caffeine administration does not alter salivary α- amylase activity in young male daily caffeine consumers. BMC Research Notes, 7, 30. https://doi.org/10.1186/1756-0500-7-30

Kloss, J. D., Nash, C. O., Walsh, C. M., Culnan, E., Horsey, S., & Sexton-Radek, K. (2016). A “Sleep 101” Program for College Students Improves Sleep Hygiene Knowledge and Reduces Maladaptive Beliefs about Sleep. Behavioral Medicine, 42(1), 48–56. https://doi.org/10.1080/08964289.2014.969186

Knufinke, M., Nieuwenhuys, A., Geurts, S. A. E., Coenen, A. M. L., & Kompier, M. A. J. (2017). Self-reported sleep quantity, quality and sleep hygiene in elite athletes. Journal of Sleep Research. https://doi.org/10.1111/jsr.12509

271

Koffel, E. A., Koffel, J. B., & Gehrman, P. R. (2015). A meta-analysis of group cognitive behavioral therapy for insomnia. Sleep Medicine Reviews, 19, 6– 16. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/24931811

Koffel, E., Kuhn, E., Petsoulis, N., Erbes, C. R., Anders, S., Hoffman, J. E., … Polusny, M. A. (2018). A randomized controlled pilot study of CBT-I Coach: Feasibility, acceptability, and potential impact of a mobile phone application for patients in cognitive behavioral therapy for insomnia. Health Informatics Journal, 24(1), 3–13. https://doi.org/10.1177/1460458216656472

Kölling, S., Steinacker, J. M., Endler, S., Ferrauti, A., Meyer, T., & Kellmann, M. (2016). The longer the better: Sleep-wake patterns during preparation of the World Rowing Junior Championships. Chronobiology International, 33(1), 73–84. https://doi.org/10.3109/07420528.2015.1118384

Kubitz, K. A., Landers, D. M., Petruzzello, S. J., & Han, M. (1996). The effects of acute and chronic exercise on sleep. A meta-analytic review. Sports Medicine, 21(4), 277–291. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/8726346

Kyle, S. D., Miller, C. B., Rogers, Z., Siriwardena, A. N., MacMahon, K. M., & Espie, C. A. (2014). Sleep Restriction Therapy for Insomnia is Associated with Reduced Objective Total Sleep Time, Increased Daytime Somnolence, and Objectively Impaired Vigilance: Implications for the Clinical Management of Insomnia Disorder. Sleep, 37(2), 229–237. https://doi.org/10.5665/sleep.3386

Lalor, B. J., Halson, S. L., Tran, J., Kemp, J. G., & Cormack, S. J. (2018). No Compromise of Competition Sleep Compared With Habitual Sleep in Elite Australian Footballers. International Journal of Sports Physiology and Performance, 13(1), 29–36. https://doi.org/10.1123/ijspp.2016-0776

Lastella, M, Roach, G. D., Halson, S. L., Martin, D. T., West, N. P., & Sargent, C. (2015). The impact of a simulated grand tour on sleep, mood, and wellbeing of competitive cyclists. The Journal of Sports Medicine and Physical Fitness, 55(12), 1555–1564. Retrieved from

272

http://www.ncbi.nlm.nih.gov/pubmed/25286890

Lastella, Michele, Roach, G. D., Halson, S. L., Gore, C. J., Garvican-Lewis, L. A., & Sargent, C. (2014). The effects of transmeridian travel and altitude on sleep: Preparation for football competition. Journal of Sports Science and Medicine, Vol. 13, pp. 718–720.

Lastella, Michele, Roach, G. D., Halson, S. L., & Sargent, C. (2015). Sleep/wake behaviours of elite athletes from individual and team sports. European Journal of Sport Science, 15(2), 94–100. https://doi.org/10.1080/17461391.2014.932016

Lastella, Michele, Roach, G. D., Halson, S. L., & Sargent, C. (2016). The chronotype of elite athletes. Journal of Human Kinetics, 54(1), 219–225. https://doi.org/10.1515/hukin-2016-0049

Leeder, J., Glaister, M., Pizzoferro, K., Dawson, J., & Pedlar, C. (2012). Sleep duration and quality in elite athletes measured using wristwatch actigraphy. Journal of Sports Sciences, 30(6), 541–545. https://doi.org/10.1080/02640414.2012.660188

Leger, D., Metlaine, A., & Choudat, D. (2005). Insomnia and sleep disruption: relevance for athletic performance. Clinics in Sports Medicine, 24(2), 269– 285. https://doi.org/10.1016/j.csm.2004.12.011

Levey, A. B., Aldaz, J. A., Watts, F. N., & Coyle, K. (1991). Articulatory suppression and the treatment of insomnia. Behaviour Research and Therapy, 29(1), 85–89. https://doi.org/10.1016/S0005-7967(09)80010-7

Lichstein, K. L., Wilson, N. M., Noe, S. L., Aguillard, R. N., & Bellur, S. N. (1994). Daytime sleepiness in insomnia: behavioral, biological and subjective indices. Sleep, 17(8), 693–702. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7701180

Lima, P. F., Medeiros, A. L. D., & Araujo, J. F. (2002). Sleep-wake pattern of medical students: Early versus late class starting time. Brazilian Journal of Medical and Biological Research, 35(11), 1373–1377. https://doi.org/10.1590/S0100-879X2002001100016

273

Littner, M. R., Kushida, C., Wise, M., Davila, D. G., Morgenthaler, T., Lee-Chiong, T., … Standards of Practice Committee of the American Academy of Sleep Medicine. (2005). Practice parameters for clinical use of the multiple sleep latency test and the maintenance of wakefulness test. Sleep, 28(1), 113– 121. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15700727

Lovato, N., Lack, L., Wright, H., Brooks, A., Lack, L., Debarnot, U., … Gallasch, J. (2014). The Napping Behaviour of Australian University Students. PLoS ONE, 9(11), e113666. https://doi.org/10.1371/journal.pone.0113666

Luchini, C., Stubbs, B., Solmi, M., & Veronese, N. (2017). Assessing the quality of studies in meta-analyses: Advantages and limitations of the Newcastle Ottawa Scale. World Journal of Meta-Analysis, 5(4), 80. https://doi.org/10.13105/wjma.v5.i4.80

Lucidi, F., Lombardo, C., Russo, P. M., Devoto, A., & Violani, C. (2007). Sleep complaints in Italian Olympic and recreational athletes. Journal of Clinical Sport Psychology, 1(2), 121–129. https://doi.org/10.1123/jcsp.1.2.121

Lund, H. G., Reider, B. D., Whiting, A. B., & Prichard, J. R. (2010). Sleep Patterns and Predictors of Disturbed Sleep in a Large Population of College Students. Journal of Adolescent Health, 46(2), 124–132. https://doi.org/10.1016/j.jadohealth.2009.06.016

Mah, C. D., Mah, K. E., Kezirian, E. J., & Dement, W. C. (2011). The Effects of Sleep Extension on the Athletic Performance of Collegiate Basketball Players. Sleep, 34(7), 943–950. https://doi.org/10.5665/SLEEP.1132

Marques, D. R., Allen Gomes, A., Drake, C. L., Roth, T., & de Azevedo, M. H. P. (2016). Assessing Stress-Induced Sleep Reactivity in College Students: The European Portuguese Version of the Ford Insomnia Response to Stress Test (FIRST). Behavioral Sleep Medicine, 1–12. https://doi.org/10.1080/15402002.2016.1210151

Maurer, L. F., Espie, C. A., & Kyle, S. D. (2018, December 1). How does sleep restriction therapy for insomnia work? A systematic review of mechanistic evidence and the introduction of the Triple-R model. Sleep Medicine Reviews, Vol. 42, pp. 127–138. https://doi.org/10.1016/j.smrv.2018.07.005

274

McCloughan, L. J., Hanrahan, S. J., Anderson, R., & Halson, S. R. (2016). Psychological recovery: Progressive muscle relaxation (PMR), anxiety, and sleep in dancers. Performance Enhancement and Health, 4(1–2), 12–17. https://doi.org/10.1016/j.peh.2015.11.002

McGuckin, T. A., Sinclair, W. H., Sealey, R. M., & Bowman, P. (2014). The effects of air travel on performance measures of elite Australian rugby league players. European Journal of Sport Science, 14(sup1), S116–S122. https://doi.org/10.1080/17461391.2011.654270

Miles, K. H., Clark, B., Fowler, P. M., Miller, J., & Pumpa, K. L. (2019). Sleep practices implemented by team sport coaches and sports science support staff: A potential avenue to improve athlete sleep? Journal of Science and Medicine in Sport, 22(7), 748–752. https://doi.org/10.1016/j.jsams.2019.01.008

Miller, C. B., Espie, C. A., Epstein, D. R., Friedman, L., Morin, C. M., Pigeon, W. R., … Kyle, S. D. (2014, October). The evidence base of sleep restriction therapy for treating insomnia disorder. Sleep Medicine Reviews, Vol. 18, pp. 415–424. https://doi.org/10.1016/j.smrv.2014.01.006

Milner, C. E., & Cote, K. A. (2009). Benefits of napping in healthy adults: impact of nap length, time of day, age, and experience with napping. Journal of Sleep Research, 18(2), 272–281. https://doi.org/10.1111/j.1365- 2869.2008.00718.x

Mollayeva, T., Thurairajah, P., Burton, K., Mollayeva, S., Shapiro, C. M., & Colantonio, A. (2016). The Pittsburgh sleep quality index as a screening tool for sleep dysfunction in clinical and non-clinical samples: A systematic review and meta-analysis. Sleep Medicine Reviews, 25, 52–73. https://doi.org/10.1016/j.smrv.2015.01.009

Montserrat Sánchez-Ortuño, M., Edinger, J. D., Means, M. K., & Almirall, D. (2010). Home is where sleep is: An ecological approach to test the validity of actigraphy for the assessment of insomnia. Journal of Clinical Sleep Medicine, 6(1), 21–29. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20191934

275

Morgan, K., Gregory, P., Tomeny, M., David, B., & Gascoigne, C. (2012). Self- help treatment for insomnia symptoms associated with chronic conditions in older adults: A randomized controlled trial. Journal of the American Geriatrics Society, 60(10), 1803–1810. https://doi.org/10.1111/j.1532- 5415.2012.04175.x

Morgenthaler, T., Kramer, M., Alessi, C., Friedman, L., Boehlecke, B., Brown, T., … Swick, T. (2006). Practice Parameters for the Psychological and Behavioral Treatment of Insomnia: An Update. An American Academy of Sleep Medicine Report. Sleep, 29(11), 1415–1419. https://doi.org/10.1093/sleep/29.11.1415

Morin, C. M., Blais, F., & Savard, J. (2002). Are changes in beliefs and attitudes about sleep related to sleep improvements in the treatment of insomnia? Behaviour Research and Therapy, 40(7), 741–752. https://doi.org/10.1016/S0005-7967(01)00055-9

Morin, C M, & Azrin, N. H. (1987). Stimulus control and imagery training in treating sleep-maintenance insomnia. Journal of Consulting and Clinical Psychology, 55(2), 260–262. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/3571683

Morin, Charles M, Vallières, A., & Ivers, H. (2007). Dysfunctional beliefs and attitudes about sleep (DBAS): validation of a brief version (DBAS-16). Sleep, 30(11), 1547–1554. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/18041487

Murawski, B., Wade, L., Plotnikoff, R. C., Lubans, D. R., & Duncan, M. J. (2017). A systematic review and meta-analysis of cognitive and behavioral interventions to improve sleep health in adults without sleep disorders. Sleep Medicine Reviews. https://doi.org/10.1016/J.SMRV.2017.12.003

Mysliwiec, V., McGraw, L., Pierce, R., Smith, P., Trapp, B., & Roth, B. J. (2013). Sleep Disorders and Associated Medical Comorbidities in Active Duty Military Personnel. Sleep, 36(2), 167–174. https://doi.org/10.5665/sleep.2364

Natale, V., & Cicogna, P. (2002). Morningness-eveningness dimension: is it really a continuum? Personality and Individual Differences, 32(5), 809–816.

276

https://doi.org/10.1016/S0191-8869(01)00085-X

Nater, U. M., Rohleder, N., Schlotz, W., Ehlert, U., & Kirschbaum, C. (2007). Determinants of the diurnal course of salivary alpha-amylase. Psychoneuroendocrinology, 32(4), 392–401. https://doi.org/10.1016/j.psyneuen.2007.02.007

Neisser, U. (1967). Cogntive psychology. New York: Appleton-Century-Croft.

Netzer, N. C., Kristo, D., Steinle, H., Lehmann, M., & Strohl, K. P. (2001). REM sleep and catecholamine excretion: A study in elite athletes. European Journal of Applied Physiology, 84(6), 521–526. https://doi.org/10.1007/s004210100383

O’Donnell, S., Beaven, C. M., & Driller, M. (2018a). The Influence of Match-Day Napping in Elite Female Netball Athletes. International Journal of Sports Physiology and Performance, 13(9), 1143–1148. https://doi.org/10.1123/ijspp.2017-0793

O’Donnell, S., Beaven, C. M., & Driller, M. W. (2018b). From pillow to podium: a review on understanding sleep for elite athletes. Nature and Science of Sleep, 10, 243–253. https://doi.org/10.2147/NSS.S158598

O’Donnell, S., Bird, S., Jacobson, G., & Driller, M. (2018). Sleep and stress hormone responses to training and competition in elite female athletes. European Journal of Sport Science, 18(5), 611–618. https://doi.org/10.1080/17461391.2018.1439535

O’Donnell, S., & Driller, M. W. (2017). Sleep-hygiene Education improves Sleep Indices in Elite Female Athletes. International Journal of Exercise Science, 10(4), 522–530. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/28674597

Ohayon, M. M. (2002). Epidemiology of insomnia: what we know and what we still need to learn. Sleep Medicine Reviews, 6(2), 97–111. https://doi.org/10.1053/smrv.2002.0186

Ohayon, M. M., & Sagales, T. (2010). Prevalence of insomnia and sleep characteristics in the general population of Spain. Sleep Medicine, 11(10),

277

1010–1018. https://doi.org/10.1016/j.sleep.2010.02.018

Ohayon, M., Wickwire, E. M., Hirshkowitz, M., Albert, S. M., Avidan, A., Daly, F. J., … Vitiello, M. V. (2017). National Sleep Foundation’s sleep quality recommendations: first report. Sleep Health, 3(1), 6–19. https://doi.org/10.1016/j.sleh.2016.11.006

Okajima, I., & Inoue, Y. (2017). Efficacy of cognitive behavioral therapy for comorbid insomnia: a meta-analysis. Sleep and Biological Rhythms, 1–15. https://doi.org/10.1007/s41105-017-0124-8

Petit, E., Bourdin, H., Tio, G., Yenil, O., Haffen, E., & Mougin, F. (2018). Effects of a 20-Min Nap Post Normal and Jet Lag Conditions on P300 Components in Athletes. International Journal of Sports Medicine, 39(07), 508–516. https://doi.org/10.1055/a-0599-0888

Petit, E., Mougin, F., Bourdin, H., Tio, G., & Haffen, E. (2014). A 20-min nap in athletes changes subsequent sleep architecture but does not alter physical performances after normal sleep or 5-h phase-advance conditions. European Journal of Applied Physiology, 114(2), 305–315. https://doi.org/10.1007/s00421-013-2776-7

Pires, M. L. N., Benedito-Silva, A. A., Mello, M. T., Del Giglio, S., Pompeia, C., & Tufik, S. (2007). Sleep habits and complaints of adults in the city of São Paulo, Brazil, in 1987 and 1995. Brazilian Journal of Medical and Biological Research, 40(11), 1505–1515. https://doi.org/10.1590/S0100- 879X2006005000170

Pitchford, N. W., Robertson, S. J., Sargent, C., Cordy, Fj., Bishop, D. J., & Bartlett, J. D. (2016). A Change in Training Environment Alters Sleep Quality But Not Quantity in Elite Australian Rules Football Players. International Journal of Sports Physiology and Performance, 1–18. https://doi.org/10.1123/ijspp.2016-0009

Rechtschaffen, A., & Kales, A. (1968). A manual of standardised terminology, techniques, and scoring system for sleep stages of human subjects. Los Angelos: UCLA Brain Information Service .

278

Reid, V. L., Gleeson, M., Williams, N., & Clancy, R. L. (2004). Clinical investigation of athletes with persistent fatigue and/or recurrent infections. British Journal of Sports Medicine, 38(1), 42–55. https://doi.org/10.1136/bjsm.2002.002634

Richmond, L., Dawson, B., Hillman, D., & Eastwood, P. (2004). The effect of interstate travel on sleep patterns of elite Australian Rules footballers. Journal of Science and Medicine in Sport, 7(2), 186–196. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15362314

Richmond, L. K., Dawson, B., Stewart, G., Cormack, S., Hillman, D. R., & Eastwood, P. R. (2007). The effect of interstate travel on the sleep patterns and performance of elite Australian Rules footballers. Journal of Science and Medicine in Sport, 10(4), 252–258. https://doi.org/10.1016/j.jsams.2007.03.002

Riedel, B. W., & Lichstein, K. L. (2000). Insomnia and daytime functioning. Sleep Medicine Reviews, 4(3), 277–298. https://doi.org/10.1053/smrv.1999.0074

Riemann, D., Spiegelhalder, K., Feige, B., Voderholzer, U., Berger, M., Perlis, M., & Nissen, C. (2010). The hyperarousal model of insomnia: A review of the concept and its evidence. Sleep Medicine Reviews, 14(1), 19–31. https://doi.org/10.1016/j.smrv.2009.04.002

Roberts, S. S. H., Teo, W. P., & Warmington, S. A. (2019, April 1). Effects of training and competition on the sleep of elite athletes: A systematic review and meta-analysis. British Journal of Sports Medicine, Vol. 53, pp. 513–522. https://doi.org/10.1136/bjsports-2018-099322

Robertson, J., Broomfield, N., & Espie, C. (2007). Prospective comparison of subjective arousal during the pre-sleep period in primary sleep-onset insomnia and normal sleepers. Journal of Sleep Research, 16(2), 230–238. https://doi.org/10.1111/j.1365-2869.2007.00579.x

Robey, E., Dawson, B., Halson, S., Gregson, W., Goodman, C., & Eastwood, P. (2014). Sleep quantity and quality in elite youth soccer players: A pilot study. European Journal of Sport Science, 14(5), 410–417. https://doi.org/10.1080/17461391.2013.843024

279

Rodrigues, D. F., Silva, A., Rosa, J. P. P., Ruiz, F. S., Veríssimo, A. W., Winckler, C., … De Mello, M. T. (2015). Sleep quality and psychobiological aspects of Brazilian Paralympic athletes in the London 2012 pre-Paralympics period. Motriz. Revista de Educacao Fisica, 21(2), 168–176. https://doi.org/10.1590/S1980-65742015000200007

Romyn, G., Lastella, M., Miller, D. J., Versey, N. G., Roach, G. D., & Sargent, C. (2018). Daytime naps can be used to supplement night-time sleep in athletes. Chronobiology International, 35(6), 865–868. https://doi.org/10.1080/07420528.2018.1466795

Romyn, G., Robey, E., Dimmock, J. A., Halson, S. L., & Peeling, P. (2016). Sleep, anxiety and electronic device use by athletes in the training and competition environments. European Journal of Sport Science, 16(3), 301– 308. https://doi.org/10.1080/17461391.2015.1023221

Rosa, J. P. P., Silva, A., Rodrigues, D. F., Simim, M. A., Narciso, F. V., Tufik, S., … Mello, M. T. de. (2018). Effect of bright light therapy on delayed sleep/wake cycle and reaction time of athletes participating in the Rio 2016 Olympic Games. Https://Doi.Org/10.1080/07420528.2018.1459660. https://doi.org/10.1080/07420528.2018.1459660

Rupp, T. L., Wesensten, N. J., Bliese, P. D., & Balkin, T. J. (2009). Banking sleep: realization of benefits during subsequent sleep restriction and recovery. Sleep, 32(3), 311–321. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19294951

Samuels, C. (2008). Sleep, Recovery, and Performance: The New Frontier in High-Performance Athletics. Neurologic Clinics, 26(1), 169–180. https://doi.org/10.1016/j.ncl.2007.11.012

Samuels, C., James, L., Lawson, D., & Meeuwisse, W. (2016). The Athlete Sleep Screening Questionnaire: A new tool for assessing and managing sleep in elite athletes. British Journal of Sports Medicine, 50(7), 418–422. https://doi.org/10.1136/bjsports-2014-094332

Sargent, C., Halson, S., & Roach, G. D. (2014). Sleep or swim? Early-morning training severely restricts the amount of sleep obtained by elite swimmers.

280

European Journal of Sport Science, 14(Supp.1), S310-15. https://doi.org/10.1080/17461391.2012.696711

Sargent, C., Lastella, M., Halson, S. L., & Roach, G. D. (2016). The validity of activity monitors for measuring sleep in elite athletes. Journal of Science and Medicine in Sport, 19(10), 848–853. https://doi.org/10.1016/j.jsams.2015.12.007

Sargent, C., Lastella, M., Halson, S., & Roach, G. (2014). The impact of training schedules on the sleep and fatigue of elite athletes. Chronobiology International, 31(10), 1160–1168. https://doi.org/10.3109/07420528.2014.957306

Sargent, C., & Roach, G. D. (2016). Sleep duration is reduced in elite athletes following night-time competition. Chronobiology International, 33(6), 667– 670. https://doi.org/10.3109/07420528.2016.1167715

Sateia, M. J., Nowell, P. D., Morin, C., Sateia, M., Doghramji, K., Hauri, P., … Morin, C. (1993). Insomnia. Lancet, 364(9449), 1959–1973. https://doi.org/10.1016/S0140-6736(04)17480-1

Schaal, K., Le Meur, Y., Louis, J., Filliard, J. R., Hellard, P., Casazza, G., & Hausswirth, C. (2015). Whole-body cryostimulation limits overreaching in elite synchronized swimmers. Medicine and Science in Sports and Exercise, 47(7), 1416–1425. https://doi.org/10.1249/MSS.0000000000000546

Schaal, K., Tafflet, M., Nassif, H., Thibault, V., Pichard, C., Alcotte, M., … Toussaint, J. F. (2011). Psychological balance in high level athletes: Gender- Based differences and sport-specific patterns. PLoS ONE, 6(5), e19007. https://doi.org/10.1371/journal.pone.0019007

Schwartz, J., & Simon, R. D. (2015). Sleep extension improves serving accuracy: A study with college varsity tennis players. Physiology and Behavior, 151, 541–544. https://doi.org/10.1016/j.physbeh.2015.08.035

Shapiro, C. (1981). Sleep and the athlete. British Journal of Sports Medicine, 15, 51–55. Retrieved from http://bjsm.bmj.com/content/15/1/51.abstract

Sharifah Maimunah, S. M. P., & Hashim, H. A. (2016). Differential Effects of 7 and

281

16 Groups of Muscle Relaxation Training Following Repeated Submaximal Intensity Exercise in Young Football Players. Perceptual and Motor Skills, 122(1), 227–237. https://doi.org/10.1177/0031512515625383

Shearer, D. A., Jones, R. M., Kilduff, L. P., & Cook, C. J. (2015). Effects of competition on the sleep patterns of elite rugby union players. European Journal of Sport Science, 15(8), 681–686. https://doi.org/10.1080/17461391.2015.1053419

Shelgikar, A. V., & Chervin, R. (2013, February). Approach to and evaluation of sleep disorders. CONTINUUM Lifelong Learning in Neurology, Vol. 19, pp. 32– 49. https://doi.org/10.1212/01.CON.0000427214.00092.0f

Silber, M. H., Ancoli-Israel, S., Bonnet, M. H., Chokroverty, S., Grigg-Damberger, M. M., Hirshkowitz, M., … Iber, C. (2007, March 15). The visual scoring of sleep in adults. Journal of Clinical Sleep Medicine, Vol. 3, pp. 121–131. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17557422

Silva, A., Queiroz, S. S., Winckler, C., Vital, R., Sousa, R. A., Fagundes, V., … De Mello, M. T. (2012). Sleep quality evaluation, chronotype, sleepiness and anxiety of Paralympic Brazilian athletes: Beijing 2008 paralympic games. British Journal of Sports Medicine, 46(2), 150–154. https://doi.org/10.1136/bjsm.2010.077016

Silva, M.-R. R. G., & Paiva, T. (2016). Poor precompetitive sleep habits, nutrients’ deficiencies, inappropriate body composition and athletic performance in elite gymnasts. European Journal of Sport Science, 16(6), 726–735. https://doi.org/10.1080/17461391.2015.1103316

Soldatos, C. R., Dikeos, D. G., & Paparrigopoulos, T. J. (2000). Athens Insomnia Scale: Validation of an instrument based on ICD-10 criteria. Journal of Psychosomatic Research, 48(6), 555–560. https://doi.org/10.1016/S0022- 3999(00)00095-7

Spielman, A. J., Caruso, L. S., & Glovinsky, P. B. (1987). A behavioral perspective on insomnia treatment. The Psychiatric Clinics of North America, 10(4), 541– 553. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/3332317

282

Stepanski, E. J., & Wyatt, J. K. (2003). Use of sleep hygiene in the treatment of insomnia. Sleep Medicine Reviews, 7(3), 215–225. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12927121

Strauss, A., Sparks, M., & Pienaar, C. (2019). The use of GPS analysis to quantify the internal and external match demands of semi-elite level female soccer players during a tournament. Journal of Sports Science and Medicine, 18(1), 73–81.

Sundgot-Borgen, J., & Garthe, I. (2011). Elite athletes in aesthetic and Olympic weight-class sports and the challenge of body weight and body compositions. Journal of Sports Sciences, 29(sup1), S101–S114. https://doi.org/10.1080/02640414.2011.565783

Sundgot-Borgen, J., Meyer, N. L., Lohman, T. G., Ackland, T. R., Maughan, R. J., Stewart, A. D., & Müller, W. (2013). How to minimise the health risks to athletes who compete in weight-sensitive sports review and position statement on behalf of the Ad Hoc Research Working Group on Body Composition, Health and Performance, under the auspices of the IOC Medical Commission. British Journal of Sports Medicine, 47(16), 1012–1022. https://doi.org/10.1136/bjsports-2013-092966

Swann, C., Moran, A., & Piggott, D. (2015). Defining elite athletes: Issues in the study of expert performance in sport psychology. Psychology of Sport and Exercise, 16, 3–14. https://doi.org/10.1016/j.psychsport.2014.07.004

Swinbourne, R., Gill, N., Vaile, J., & Smart, D. (2016). Prevalence of poor sleep quality, sleepiness and obstructive sleep apnoea risk factors in athletes. European Journal of Sport Science, 16(7), 850–858. https://doi.org/10.1080/17461391.2015.1120781

Taylor, L., Chrismas, B., Dascombe, B., Chamari, K., & Fowler, P. (2016). Sleep medication and athletic performance-The evidence for practitioners and future research directions. Frontiers in Physiology, 7(83). https://doi.org/10.3389/fphys.2016.00083

Taylor, S. R., Rogers, G. G., & Driver, H. S. (1997). Effects of training volume on sleep, psychological, and selected physiological profiles of elite female

283

swimmers. Medicine and Science in Sports and Exercise, 29(5), 688–693. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9140908

Thornton, H. R., Duthie, G. M., Pitchford, N. W., Delaney, J. A., Benton, D. T., & Dascombe, B. J. (2016). Effects of a Two-week High Intensity Training Camp on Sleep Activity of Professional Rugby League Athletes. International Journal of Sports Physiology and Performance, 1–19. https://doi.org/10.1123/ijspp.2016-0414

Thun, E., Bjorvatn, B., Flo, E., Harris, A., & Pallesen, S. (2015). Sleep, circadian rhythms, and athletic performance. Sleep Medicine Reviews, 23, 1–9. https://doi.org/10.1016/j.smrv.2014.11.003

Thun, E., Bjorvatn, B., Osland, T., Martin Steen, V., Sivertsen, B., Johansen, T., … Pallesen, S. (2012). An Actigraphic Validation Study of Seven Morningness- Eveningness Inventories. European Psychologist, 17(3), 222–230. https://doi.org/10.1027/1016-9040/a000097

Tsunoda, K. (2015). Sleep status in male wheelchair basketball players on a Japanese national team. Journal of Sleep Disorders & Therapy, 4(4), 8–11. https://doi.org/10.4172/2167-0277.1000210

Tuomilehto, H., Vuorinen, V.-P., Penttilä, E., Kivimäki, M., Vuorenmaa, M., Venojärvi, M., … Pihlajamäki, J. (2016). Sleep of professional athletes: Underexploited potential to improve health and performance. Journal of Sports Sciences, 1–7. https://doi.org/10.1080/02640414.2016.1184300 van de Laar, M., Verbeek, I., Pevernagie, D., Aldenkamp, A., Overeem, S., Ancoli-Israel, S., … Rosenblate, R. (2010). The role of personality traits in insomnia. Sleep Medicine Reviews, 14(1), 61–68. https://doi.org/10.1016/j.smrv.2009.07.007

Van Ryswyk, E., Weeks, R., Bandick, L., O’Keefe, M., Vakulin, A., Catcheside, P., … Antic, N. A. (2016). A novel sleep optimisation programme to improve athletes’ well-being and performance. European Journal of Sport Science, 1–8. https://doi.org/10.1080/17461391.2016.1221470 van Straten, A., van der Zweerde, T., Kleiboer, A., Cuijpers, P., Morin, C. M., &

284

Lancee, J. (2017). Cognitive and behavioral therapies in the treatment of insomnia: A meta-analysis. Sleep Medicine Reviews. https://doi.org/10.1016/j.smrv.2017.02.001

Venter, R. (2012). Role of sleep in performance and recovery of athletes: a review article. South African Journal for Research in Sport, Physical Education and Recreation, 34(1), 167–184.

Venter, Rachel E, Potgieter, J. R., & Barnard, J. G. (2010). The use of recovery modalities by elite South african team athletes. South African Journal for Research in Sport, Physical Education & Recreation, 32(133), 45. https://doi.org/10.4314/sajrs.v32i1.54106

Venter, Rachel Elizabeth. (2014). Perceptions of team athletes on the importance of recovery modalities. European Journal of Sport Science. https://doi.org/10.1080/17461391.2011.643924

Violani, C., Devoto, A., Lucidi, F., Lombardo, C., & Russo, P. M. (2004). Validity of a Short Insomnia Questionnaire: The SDQ. Brain Research Bulletin, 63(5), 415–421. https://doi.org/10.1016/j.brainresbull.2003.06.002

Waterhouse, J, Atkinson, G., Edwards, B., & Reilly, T. (2007). The role of a short post-lunch nap in improving cognitive, motor, and sprint performance in participants with partial sleep deprivation. Journal of Sports Sciences, 25(14), 1557–1566. https://doi.org/10.1080/02640410701244983

Waterhouse, Jim, Reilly, T., Atkinson, G., & Edwards, B. (2007). Jet lag: trends and coping strategies. The Lancet, 369(9567), 1117–1129. https://doi.org/10.1016/S0140-6736(07)60529-7

Wells, G., Shea, B., O’Connell, D., Peterson, J., Welch, V., Losos, M., & Tugwell, P. (2016). The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Retrieved April 15, 2020, from Ottawa Hospital Research Institute website: http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp

Wong, W. S., & Fielding, R. (2011). Prevalence of insomnia among Chinese adults in Hong Kong: A population-based study. Journal of Sleep Research,

285

20(1), 117–126. https://doi.org/10.1111/j.1365-2869.2010.00822.x

Yamadera, W., Sato, M., Harada, D., Iwashita, M., Aoki, R., Obuchi, K., … Nakayama, K. (2013). Comparisons of short-term efficacy between individual and group cognitive behavioral therapy for primary insomnia. Sleep and Biological Rhythms, 11(3), 176–184. https://doi.org/10.1111/sbr.12019

Yang, C.-M., Chou, C. P.-W., & Hsiao, F.-C. (2011). The Association of Dysfunctional Beliefs About Sleep With Vulnerability to Stress-Related Sleep Disturbance in Young Adults. Behavioral Sleep Medicine, 9(2), 86–91. https://doi.org/10.1080/15402002.2011.557990

Ye, L., Hutton Johnson, S., Keane, K., Manasia, M., & Gregas, M. (2015). Napping in college students and its relationship with nighttime sleep. Journal of American College Health, 63(2), 88–97. https://doi.org/10.1080/07448481.2014.983926

Ye, Y.-Y., Chen, N.-K., Chen, J., Liu, J., Lin, L., Liu, Y.-Z., … Jiang, X.-J. (2016). Internet-based cognitive-behavioural therapy for insomnia (ICBT-i): a meta- analysis of randomised controlled trials. BMJ Open, 6(11), e010707. https://doi.org/10.1136/bmjopen-2015-010707

Ye, Y. Y., Chen, N. K., Chen, J., Liu, J., Lin, L., Liu, Y. Z., … Jiang, X. J. (2016). Internet-based cognitive-behavioural therapy for insomnia (ICBT-i): A meta- analysis of randomised controlled trials. BMJ Open, 6(11), e010707. https://doi.org/10.1136/bmjopen-2015-010707

Yeung, W.-F., Chung, K.-F., Ho, F. Y.-Y., & Ho, L.-M. (2015). Predictors of dropout from internet-based self-help cognitive behavioral therapy for insomnia. Behaviour Research and Therapy, 73, 19–24. https://doi.org/10.1016/j.brat.2015.07.008

Youngstedt, S. D., O’Connor, P. J., & Dishman, R. K. (1997). The effects of acute exercise on sleep: a quantitative synthesis. Sleep, 20(3), 203–214. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9178916

Zhao, J., Tian, Y., Nie, J., Xu, J., & Liu, D. (2012). Red light and the sleep quality

286

and endurance performance of chinese female basketball players. Journal of Athletic Training, 47(6), 673–678. https://doi.org/10.4085/1062- 6050-47.6.08

287