Confidential: for Review Only Q Fever: Superstition As a Coping Mechanism? a Randomised Controlled Non-Inferiority Trial

BMJ

Confidential: For Review Only Q fever: Superstition as a coping mechanism? A randomised controlled non-inferiority trial

Journal: BMJ

Manuscript ID BMJ-2019-051945

Article Type: Christmas research

BMJ Journal: BMJ

Date Submitted by the 31-Jul-2019 Author:

Complete List of Authors: Brookfield, Charlotte; Lancashire Teaching Hospitals NHS Foundation Trust, Department of microbiology Phillips, Patrick; University of California San Francisco, Division of Pulmonary and Critical Care Medicine Shorten, Robert; Lancashire Teaching Hospitals NHS Foundation Trust, Department of Microbiology; University College London, Centre for Clinical Microbiology

Keywords: Christmas, Superstition, Quiet, Medical Microbiology

https://mc.manuscriptcentral.com/bmj Page 1 of 19 BMJ

1 2 3 Title Q fever: Superstition as a coping mechanism? A randomised 4 controlled non-inferiority trial 5 6 7 Authors Names CR Brookfield (0000-0002-6493-2574), PPJ Phillips, RJ Shorten 8 Address and titles CR Brookfield Registrar in Medical Microbiology, Department of 9 10 Microbiology, Lancashire Teaching Hospitals NHS Foundation Trust, 11 Sharoe Green Lane, Fulwood, Preston PR2 9HT 12 Confidential: For Review Only 13 PPJ Phillips Assistant Professor, Centre for Tuberculosis, 14 Department of Medicine, Division of Pulmonary & Critical Care 15 Medicine, Zuckerberg San Francisco General Hospital, 1001 Potrero 16 Avenue, Building 5, Room 5K1, San Francisco, CA 94110 17 18 RJ Shorten Clinical Scientist, Department of Microbiology, 19 Lancashire Teaching Hospitals NHS Foundation Trust, Sharoe Green 20 21 Lane, Fulwood, Preston PR2 9HT 22 & 23 Centre for Clinical Microbiology, University College London, Royal 24 Free Campus, Rowland Hill Street, London NW3 2QG 25 26 Corresponding author Correspondence to: [email protected] 27 28 29 30 ABSTRACT 31 32 Objectives: The primary aim of this study was to evaluate the superstition that audible utterance of 33 34 the word “quiet” increases clinical workload. 35 36 37 Design: This was a prospective randomised, controlled, non-inferiority study. 38 39 Setting: The study was conducted within the microbiology department at a large teaching hospital 40 41 in Lancashire, UK. 42 43 Participants: Workload was measured each day over a 61 day period. 44 45 46 Interventions: Each day was randomly allocated, to either saying “Today will be a quiet day” 47 (intervention) (n=29) or refraining from saying the word “quiet” (control) (n=32). 48 49 50 Main outcome measures: The primary outcome was a composite end point of number of clinical 51 52 consultations “calls”, clinically significant results “alerts”, and validated results processed by the duty 53 microbiologist during a 24-hour period. Secondary outcomes include the individual components of 54 55 the primary outcome. An absolute increase of 30 clinical episodes was considered as a reasonable 56 57 margin of non-inferiority. We calculated that a total sample size of 60 days would demonstrate non- 58 inferiority with 80% power. 59 60

1 https://mc.manuscriptcentral.com/bmj BMJ Page 2 of 19

1 2 3 Results: There were a mean of 139.0 clinical episodes on control days compared to 144.9 on days 4 5 when the experimental intervention was uttered, a difference of 5.9, 95% confidence interval (-12.9, 6 7 24.7). The upper bound was less than our pre-specified margin of 30, providing evidence for non- 8 inferiority. There was no evidence of a difference in workload between interventions with any of the 9 10 four components, whether considering unadjusted or adjusted analyses, or looking at the sub- 11 12 groups ofConfidential: week days or weekends. For Review Only 13 14 Conclusions: Our study refutes the long-held superstition that audible utterance of the word “quiet” 15 16 impacts on clinical workload and should not be avoided. In the era of significant staff shortages and 17 18 increased work-related stress, doctors should look to other methods to increase resilience and 19 protect their wellbeing and mental health. 20 21 22 Trial registration: Not required. Registered with trust research department (SE-259) 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

2 https://mc.manuscriptcentral.com/bmj Page 3 of 19 BMJ

1 2 3 INTRODUCTION 4 5 6 It’s oh so quiet (shhh), 7 8 It’s oh so still (shhh), 9 10 11 You’re all alone (shhh) and so peaceful until……1 12 Confidential: For Review Only 13 …someone wishes you a quiet shift and then the whole thing is a total disaster. 14 15 16 In the modern era of evidence-based medicine, superstitions abound. One long held superstition is 17 18 that utterance of the word “quiet” will negatively affect one’s shift with increased workload and 19 complexity of cases.2 As Dr Adam Kay says in his book ‘This is going to hurt: Secret Diaries of a Junior 20 21 Doctor’, “Say the Q word to a doctor and you’re all but performing an incantation, summoning the

22 3 23 sickest patients in the world to your hospital” . Avoidance of the word “quiet” has therefore become 24 common practice among health care professionals in the NHS and has also been described in the US 25 26 4 and Japan 5. 27 28 The evidence to support the harmful effect of the word “quiet” is mainly anecdotal with only one 29 30 other UK study concluding that the word “quiet” did impact on on-call shifts in an orthopaedic 31 32 department .6 We therefore determined to conduct a randomized controlled non-inferiority trial to 33 evaluate whether audible utterance of the word “quiet” increases clinical workload within a busy 34 35 microbiology department at a teaching hospital in the North West of England. A secondary aim of 36 37 the trial was to answer another mystery of the medical world - what does the Medical Microbiologist 38 actually do? 39 40 41 42 43 44 METHODS 45 46 Design 47 48 49 We conducted a randomized non-inferiority trial to evaluate the hypothesis that audible utterance 50 of the word “quiet” increases clinical workload, as measured by the number of clinical consultation 51 52 “calls”, clinically significant results, hereon in referred to as “alerts”, and validated results processed 53 54 daily by the duty medical microbiology team. 55 56 The null hypothesis for the study was that audible utterance of the word “quiet” increases clinical 57 58 workload by no more than an average of 30 clinical episodes per day. 59 60 Ethics and trial registration

3 https://mc.manuscriptcentral.com/bmj BMJ Page 4 of 19

1 2 3 Following MRC / NHS guidelines, full ethics approval was not required, and the study was registered 4 5 with the Trust’s research department as service improvement since data collected will be used to 6 7 enhance our clinical service. Furthermore, since the subjects of the trial were healthcare providers 8 and not patients, we followed guidance from the International Committee of Medical Journal Editors 9 10 stating that trial registration was not required.7 The trial protocol was developed following SPIRIT 11 8 12 guidelinesConfidential: and finalized prior to study start. For Review Only 13 14 Study setting 15 16 17 The microbiology department of Lancashire Teaching Hospitals is based at Royal Preston Hospital 18 and additionally provides services to Chorley and South Ribble Hospital. It provides district general 19 20 hospital services to 370,000 people in Preston and Chorley, and specialist care to 1.5 million people 21 22 across Lancashire and South Cumbria, including oncology, major trauma, disablement services, 23 neurology and neurosurgery, renal, and vascular. The laboratory team comprises over 60 whole time 24 25 equivalent (WTE) laboratory and support staff and 5.8 WTE Microbiology Consultants and trainees. 26 27 The service received over 530,000 specimens during for the financial year 2018/19. Additionally, the 28 department runs a home and clinic based Outpatient Parenteral Antimicrobial Therapy (OPAT) 29 30 service which treated over 300 individual patients in 2018. On any given day 1-2 members of the 31 32 team is assigned to complete the duty work included in this study. 33 34 Interventions and randomization 35 36 37 The intervention was either saying “Today will be a quiet day” out loud (experimental intervention) 38 or refraining from saying the word “quiet” (control intervention) between the hours of 8:00 and 39 40 9:00am at the beginning of each day. This was carried out within the microbiology department by a 41 42 member of the duty clinical team for that day. Each day over a 61 day period was randomly allocated 43 44 to the experimental or control intervention using a list prepared prior to study start by the trial 45 statistician (PPJP) using the big stick procedure 9 to minimise imbalance between arms during the 46 47 trial. Randomisation was not stratified and was implemented using sequentially numbered opaque 48 49 envelopes that contained that day’s allocation. The clinical team was not blinded to the intervention 50 since the outcomes are objective and not likely to be affected by the knowledge of the intervention. 51 52 53 Outcome measures 54 55 The primary outcome was a composite end point of number of calls, alerts, and validated results 56 57 processed by the duty medical microbiology team per day from 9:00am until 8:59am the next day. 58 59 These markers represent clinically relevant objective measurements of workload within the 60 department. Secondary outcomes include the individual components of the composite primary

4 https://mc.manuscriptcentral.com/bmj Page 5 of 19 BMJ

1 2 3 outcome with calls being further divided in to 9am-5pm and 5pm-9am on week days. The data was 4 5 gathered retrospectively from laboratory computer systems, except for telephone consultations and 6 7 alerts received out of hours, which were recorded prospectively by the team. No patient or 8 healthcare worker identifiable information was recorded. 9 10 11 Sample size 12 Confidential: For Review Only 13 Using data collected over a period of 30 consecutive days in January 2019, we expected a mean of 14 15 156 and standard deviation of 41 for the composite primary outcome. Based on discussions with 16 17 clinical colleagues, we considered an absolute increase of 30 as a reasonable margin of non- 18 inferiority reflecting only a small increase in workload. Using a one-sided 2.5% level of significance 19 20 and assuming no difference in the mean primary outcome measure between arms, we calculated 21 22 that a total sample size of 60 days would be adequate to demonstrate non-inferiority with 80% 23 power. 24 25 26 Due to uncertainty in the sample size assumptions, we planned to undertake a blinded sample size 27 28 re-estimation at the half-way point. This involved calculation of the aggregate standard deviation 29 (pooled across arms). The sample size re-estimation was undertaken in line with the study protocol. 30 31 The calculated standard deviation was slightly less than anticipated indicating slightly higher power 32 33 than expected so the trial continued to completion without amendment of the sample size. Since 34 the analysis was blinded to treatment allocation (pooled across arms), there was no impact on 35 36 overall type I error or bias in treatment effect estimates. 37 38 Statistical analysis 39 40 41 Day was considered as the unit of analysis; all consecutive days were included in the analysis. The 42 43 difference in mean of the primary outcome measure between the intervention and control groups 44 was calculated with 95% confidence interval. If the upper bound of this 95% confidence interval was 45 46 less than the margin of non-inferiority of 30, then non-inferiority was to be declared. The mean 47 48 difference and 95% confidence intervals were also calculated for each component of the composite 49 outcome. The analysis was repeated in the sub groups of weekdays (Monday-Friday) and weekends 50 51 (Saturday and Sunday) to identify whether the treatment effect differs in these two subgroups. Bank 52 53 holidays occurring during the study were also included as weekends. In post hoc analyses, we also 54 evaluated whether workload increased on days with a full moon or when a Friday coincided with the 55 56 13th day of the month. 57 58 59 RESULTS 60

5 https://mc.manuscriptcentral.com/bmj BMJ Page 6 of 19

1 2 3 The trial was run from 1st May to 30th June 2019, a period of 61 days. This included 41 week days, 2 4 5 full moons (18th May and 17th June) and no Friday 13ths (Figure 1, Table 1). Considering the primary 6 7 composite outcome (Figure 2), there were a mean of 139.0 clinical episodes on Control days (n=32) 8 compared to 144.9 on days when the experimental intervention was uttered (n=29), a difference of 9 10 5.9 episodes, 95% confidence interval (-12.9, 24.7). The upper bound was less than our pre-specified 11 12 margin ofConfidential: on-inferiority of 30, therefore For providing Review evidence for non-inferiority. Only Although the 13 workload was greater on week days than weekends (a mean increase of 34.6 episodes), the 14 15 differences between interventions was consistent in the subgroups of week days or weekends only 16 17 with no evidence for an interaction between treatment and type of day (p=0.870). The result was 18 also consistent when adjusting for type of day (weekend or week day), -4.8 95% CI (-22.8, 13.1). The 19 20 greatest contribution to workload was results validation with a mean if 97.0 on control days and 96.2 21 22 on days when the experimental intervention was uttered (Figure 3A). There was no evidence of a 23 difference in workload between interventions with any of the four components, whether 24 25 considering unadjusted or adjusted analyses, or looking at the sub-groups of week days or weekends 26 27 (Figures 3B-D). Workload during the study period is summarised by component in Figure 4. 28 29 The mean overall workload was 137.5 on days with a full moon compared to 142.0 on days without a 30 31 full moon, a mean difference of 4.5 95% CI (-49.5, 58.5). 32 33 DISCUSSION 34 35 Our study demonstrates that utterance of the word “quiet” has no impact on clinical workload, and 36 37 this result holds during week days or weekends. Sub group analysis demonstrated no individual 38 39 element of the combined workload was impacted by the intervention. Use of the word quiet should 40 not be avoided and perhaps even encouraged since the sentiment in wishing a colleague a quiet shift 41 42 remains true. 43 44 45 A previous study conducted within an UK orthopaedic department found that saying the word 46 “quiet” out loud did impact on on-call workload and suggested appointment of a ‘Q word specialist 47 48 manager’ as well as considering nationwide public initiatives to reduce use of the word.6 Their study 49 50 only included referrals during night shifts, did not have any description of sample size justification 51 suggesting the convenience sample might have introduced bias in the results, and was not reported 52 53 following CONSORT guidelines making it difficult to judge whether there were other important 54 55 sources of bias. In contrast, our study followed a pre-specified protocol following SPIRIT guidelines, 56 we have reported our results in line with CONSORT guidelines and we conducted the study over 61 57 58 days and looked at several different components of workload. We also included both day time and 59 60 on call sessions. A second similar study was performed in an emergency department in Japan.5 The

6 https://mc.manuscriptcentral.com/bmj Page 7 of 19 BMJ

1 2 3 authors also found no detrimental effect on workload of the utterance of the word “quiet”. 4 5 However, some of the primary and secondary outcome measures were subjective, although 6 7 attempts were made to control for this. Our findings add further weight to these conclusions given 8 the robust statistical powering of the study, combined with the measurement of objective data. 9 10 11 There are several limitations of our study. Firstly, we could not control for the use of the word 12 Confidential: For Review Only “quiet” within other departments of the hospital. Secondly, we did not incorporate ward rounds in 13 14 to our data collection as these are non-standardised and difficult to measure but may have impacted 15 16 on number of clinical enquiries on any given day. Thirdly, we could not control for other confounding 17 18 factors such as black cats, cracked mirrors or lone magpies, and too few days with full moons were 19 included to be evaluated. Fourthly, there was a chance imbalance in treatment allocation between 20 21 weekdays and weekends (Table 1). Randomisation was not stratified and chance imbalances can 22 23 occur. This, however, did not seem to impact our results as adjusted and unadjusted analyses were 24 consistent. Our study was a single centre study in one speciality and therefore the results may not 25 26 be generalisable to other populations. 27 28 Clearly this study is a little tongue and cheek but it highlights an important issue. The Health 29 30 Foundation and Kings Fund 2018 report on the healthcare workforce in England highlighted that 31 32 “There are significant staff shortages across the NHS. There are over 100,000 vacancies across NHS 33 trusts (1 in 11 posts). In addition, the staff that are in post are under increasing stress: the latest NHS 34 35 staff survey showed that 38% of staff had felt unwell during the previous 12 months due to work- 36 37 related stress”10 and A Royal College of Physicians report in 2016 notes that the NHS is underfunded, 38 39 underdoctered and overstretched resulting in falling morale, falling productivity and falling patient 40 experience.11 In the face of such immovable obstacles is it any wonder that staff hope that luck falls 41 42 on their side? 43 44 Our study confirms, what is probably already known, that such superstitions will not ease the heavy 45 46 workload that faces us all. As demonstrated by this study, a single microbiologist can encounter over 47 48 140 clinical episodes in a single shift. Health care professionals need to be resilient and mindful to 49 care for their own wellbeing as well as those around them. For any doctor that is struggling to cope 50 51 there are a number of national professional support resources available such as the BMA Wellbeing 52 53 Support Services Peer Support Counselling (https://www.bma.org.uk/advice/work-life-support/your- 54 wellbeing/counselling-and-peer-support), DocHealth, a confidential psychotherapeutic consultation 55 56 service for doctors, (https://www.dochealth.org.uk/), and the NHS Practitioner Health Programme 57 58 (www.php.nhs.uk). More general resources include the mental health charity MIND 59 60

7 https://mc.manuscriptcentral.com/bmj BMJ Page 8 of 19

1 2 3 (https://www.mind.org.uk/) as well as the Samaritans who can be contacted 24 hours a day on 116 4 5 123. 6 7 In summary saying the word “quiet” out loud does not impact on clinical workload and its use should 8 9 not be avoided. Medical Microbiologists belong to a small speciality that faces large challenges, not 10 11 least trying to slow the increase of antimicrobial resistance. They are there to support you with all 12 Confidential: For Review Only aspects of infection management from individual complex cases and travel associated infections to 13 14 stewardship or infection control advice. 15 16 17 Areas for further research include “can placing horse shoes outside patient isolation rooms prevent 18 the transmission of resistant organisms?”, “would having a rabbit’s foot in theatre reduce surgical 19 20 site infections”, and “does lucky heather have any antiseptic properties?”. 21 22 23 Word count: 2184 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

8 https://mc.manuscriptcentral.com/bmj Page 9 of 19 BMJ

1 2 3 Table 1. Description of study days. 4 5 Control intervention Experimental Total 6 7 intervention 8 9 Total randomised 32 (52%) 29 (48%) 61 10 11 Week days 17 (41%) 24 (58%) 41 12 Saturday,Confidential: Sunday, or 15 (75%) For5 (25%) Review Only20 13 14 Bank Holiday Monday 15 16 Full moons 2 (100%) 0 (0%) 2 17 Friday 13th 0 0 0 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

9 https://mc.manuscriptcentral.com/bmj BMJ Page 10 of 19

1 2 3 ACKNOWLEDGEMENTS 4 5 6 The authors would like to acknowledge all the contributions from the team of Medical Microbiogists, 7 Healthcare Scientists and pathology IT team at Lancashire Teaching Hospitals NHS Foundation Trust. 8 9 10 11 12 Confidential: For Review Only STATEMENTS 13 14 15 The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of 16 17 all authors, a worldwide licence to the Publishers and its licensees in perpetuity, in all forms, formats 18 and media (whether known now or created in the future), to i) publish, reproduce, distribute, display 19 20 and store the Contribution, ii) translate the Contribution into other languages, create adaptations, 21 22 reprints, include within collections and create summaries, extracts and/or, abstracts of the 23 Contribution, iii) create any other derivative work(s) based on the Contribution, iv) to exploit all 24 25 subsidiary rights in the Contribution, v) the inclusion of electronic links from the Contribution to 26 27 third party material where-ever it may be located; and, vi) licence any third party to do any or all of 28 the above 29 30 31 All authors have completed the Unified Competing Interest form (available on request from the 32 33 corresponding author) and declare: no support from any organisation for the submitted work, no 34 financial relationships with any organisations that might have an interest in the submitted work in 35 36 the previous three years, no other relationships or activities that could appear to have influenced 37 38 the submitted work.” 39 40 The lead author, Charlotte Brookfield, (the manuscript’s guarantor) affirms that the manuscript is an 41 42 honest, accurate, and transparent account of the study being reported; that no important aspects of 43 44 the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, 45 registered) have been explained. Charlotte Brookfield and Robert Shorten devised the study, 46 47 supervised the opening of the pre-sealed allocation envelopes and collected the data. Patrick Phillips 48 49 contributed to the study design, provided randomisation and power calculations and conducted 50 statistical analysis. All three authors contributed to the literature review and writing the protocol 51 52 and manuscript. 53 54 No authors have any competing interests to declare. No funding was secured for this study. All work 55 56 was conducted by the authors without payment. 57 58 59 No patients were recruited for this study. No identifiable patient data was collected or analysed. 60

10 https://mc.manuscriptcentral.com/bmj Page 11 of 19 BMJ

1 2 3 REFERENCES 4 5 6 1. Björk, It’s Oh So Quiet 1995 7 2. Warriner D. “The Q word” BMJ 2009;338:b1286 8 9 3. Kay A. This is going to hurt: Secret Diaries of a junior doctor. Picador 2017 10 11 4. Vinson DR. Superstitions in Medicine: Bad luck or Bad logic? Annals of Emergency Medicine. 12 Confidential: For Review Only 1998. 31(5):650-65. 13 14 5. Kuriyama A, Umakoshi N, Fujinaga J, et al. Impact of Attending Physicians' Comments on 15 16 Residents' Workloads in the Emergency Department: Results from Two J(^o^)PAN 17 18 Randomized Controlled Trials. PLoS One. 2016. 11(12):e0167480. 19 6. Lamb JN, Howard AJ, Marciniak J et al. Does the word “quiet” really make things busier. RCS 20 21 Bulletin. 2017. 99(4):133-136. 22 23 7. icmje.org [Internet]. International Committee of Medical Journal Editors; Clinical Trials 24 Registration [cited July 2019]. Available from: http://www.icmje.org/about- 25 26 icmje/faqs/clinical-trials-registration/ 27 28 8. Chan AW, Tetzlaff JM, Gotzsche PC et al. SPIRIT 2013 explanation and elaboration: guidance 29 30 for protocols of clinical trials. BMJ. 2013;346:e7586. 31 32 33 9. Soares JF & Wu CFJ. Some restricted randomization rules in sequential designs. 34 35 Commun. Stat. Theory Methods. 1983. 12:2017-2034 36 37 38 10. The Heath Foundation, Kings Fund and Nuffield Health: The healthcare workforce in 39 England: Make or Break? November 2018. Available at: 40 41 https://www.kingsfund.org.uk/publications/health-care-workforce-england 42 43 11. Royal College of Physicians Underfunded. Underdoctored. OverStretched. The NHS in 2016. 44 September 2016. Available at: https://www.rcplondon.ac.uk/guidelines- 45 46 policy/underfunded-underdoctored-overstretched-nhs-2016 47 48 49 50 51 52 53 54 55 56 57 58 59 60

11 https://mc.manuscriptcentral.com/bmj BMJ Page 12 of 19

1 2 3 Figure 1. CONSORT flow chart. Note the unit of randomization is calendar day. 4 5 6 7 8 9 Enrollment 10 Assessed for eligibility (n= 61) 11 12 Confidential: For Review Only 13 Excluded (n= 0) 14

15 16 17 18 19 20 Randomized (n= 61) 21 22 23 24 25

26 Allocation Allocated to intervention (n= 29) 27 Allocated to intervention (n= 32)  Received allocated intervention (n= 29) 28  Received allocated intervention (n= 32) 29  Did not receive allocated intervention (give  Did not receive allocated intervention (give 30 reasons) (n=0) reasons) (n= 0) 31 32 33 34 35 Follow-Up 36 Lost to follow-up (give reasons) (n= 0) Lost to follow-up (give reasons) (n= 0) 37 38 Discontinued intervention (give reasons) (n= 0) Discontinued intervention (give reasons) (n= 0) 39 40 41 42 43 Analysis 44 Analysed (n=29) Analysed (n=32) 45  Excluded from analysis (give reasons) (n= 0)  Excluded from analysis (give reasons) (n= 0) 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 13 of 19 BMJ

1 2 3 Figure 4. Summary of workload during the study period. Weekends (including bank holidays) are 4 5 indicated by grey shading; days with a full moon are indicated with a yellow . 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 14 of 19

1 2 3 Figure 2. Summary of mean overall workload (primary outcome) by treatment group, Control (C) and 4 5 Experimental (E). The primary analysis was unadjusted, results of the mean difference adjusted for 6 7 type of day (weekend or week day) are also shown. 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj Page 15 of 19 BMJ

1 2 3 Figure 3. Summary of components of workload by treatment group, Control (C) and Experimental (E) 4 5 6 7 8 Confidential: For Review Only 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 2

43 https://mc.manuscriptcentral.com/bmj 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 16 of 19

1 2 3 Q fever protocol 4 5 6 1. Title: Q-Fever: A randomised non-inferiority study assessing the impact of the word Quiet on 7 clinical workload within a microbiology department at a busy teaching hospital. 8 9 2. Trial registration: not applicable 10 11 3. Protocol version 1.4 12 Confidential: For Review Only 4. Funding: No funding resourced 13 14 5. Roles and responsibilities: 15 16 a. Dr Charlotte Brookfield ST4 Medical Microbiology Lancashire Teaching Hospitals 17 18 Foundation Trust- study design, data collection and analysis 19 b. Dr Rob Shorten Clinical Scientist Lancashire Teaching Hospitals Foundation Trust- 20 21 study design, data collection and analysis 22 23 c. Assistant Professor Patrick Phillips, UCSF Centre for Tuberculosis- study design and 24 statistics, data analysis 25 26 27 Introduction 28 29 6. Background and rationale: A long held superstition amongst many health professionals is 30 31 that an individual or team can be condemned to a busy shift by the out loud utterance of the 32 word “Quiet”. The deliberate, or mistaken, release of this word can lead to disharmony 33 34 amongst colleagues and maladaptive coping strategies to relieve anxiety. However, the 35 36 evidence base for the impact of the word Quiet is lacking. In the era of pressured resources 37 and resilient working this trial aims to investigate the impact of the word Quiet on clinical 38 39 workload in order to inform future practice. 40 41 7. Null hypothesis: Audible utterance of the word ‘Quiet’ increases clinical workload, as 42 measured by results alerts, results validation, and clinical queries, by no more than an 43 44 average of 30 per day. 45 46 8. Study design: This is a non-inferiority randomised control trial. The unit of randomisation is 47 a 24-hour day following the application of the experimental intervention (or control) at 48 49 8:00am. 50 51 52 Methods 53 54 9. Study setting: Department of Microbiology, Lancashire Teaching Hospitals Foundation Trust 55 56 10. Eligibility criteria: Each day over a 60 day period will be randomised in to a study or control 57 arm. 58 59 60

https://mc.manuscriptcentral.com/bmj Page 17 of 19 BMJ

1 2 3 11. Interventions: The intervention is saying either saying “Today will be a quiet day” out loud 4 5 (experimental intervention) or refraining from saying the word “Quiet” (control 6 7 intervention) at between 8:00 – 9:00am.This will take place within the microbiology 8 department by a member of the duty clinical team for that day. 9 10 12. Outcome: The primary outcome will be a composite end point of number of clinical 11 12 consultationsConfidential: “calls” and clinically significantFor resultsReview “alerts” and validatedOnly results processed 13 by the duty medical microbiology team per day from 9:00am until 9:00am the next day. 14 15 These markers represent clinically relevant objective measurements of workload within our 16 17 department. Secondary outcomes include the individual components of the composite 18 primary outcome. 19 20 13. Participant timeline: Data will be collected only during each day on which the intervention is 21 22 applied. 23 14. Sample size: 24 25 26 Based on data collected over a period of 30 consecutive days in January 2019, we expect a mean 27 28 of 156 and standard deviation of 41 in the composite primary outcome. Based on discussions 29 with clinical colleagues, we consider an absolute increase of 30 as a reasonable margin of non- 30 31 inferiority reflecting only a small increase in workload. Using a one-sided 2.5% level of 32 33 significance and assuming no difference in the mean primary outcome measure between arms a 34 total sample size of 60 days is adequate with this margin to demonstrate non-inferiority with 35 36 80% power. We will therefore run this clinical trial over a period of 60 consecutive days, applying 37 38 the experimental intervention or control at the beginning of each day. 39 40 Total Margin of Expected Expected 41 number non- daily standard Power 42 of days inferiority mean deviation 43 44 60 20 156 41 46% 45 60 25 156 41 64% 46 60 30 156 41 80% 47 60 35 156 41 90% 48 49 60 40 156 41 96% 50 51 52 After approximately half of the days have been completed, we will undertake a blinded sample size 53 54 re-estimation to ensure that our assumptions are reasonable and our power is adequate. We will 55 56 analyse the data, reviewing the aggregate data across arms calculating the pooled mean and 57 standard deviation. The sample size may be adjusted at this point to achieve adequate power. Since 58 59 60

https://mc.manuscriptcentral.com/bmj BMJ Page 18 of 19

1 2 3 the analysis is blinded to treatment allocation (pooled across arms), there is no impact on overall 4 5 type I error. 6 7 15. Recruitment: In order to achieve the required sample size, the trial will be run over a period 8 9 of 60 consecutive days (including weekends). 10 11 16. Randomisation: 12 Confidential: For Review Only a. The randomisation list will be prepared prior to study start by the trial statistician 13 14 using the big stick procedure to minimise imbalance between arms during the trial. 15 16 Randomisation will not be stratified. 17 18 b. Randomisation will be implemented using opaque sequentially numbered envelopes 19 containing that day’s allocation. At between 8:00 and 9:00am, the day’s envelope 20 21 will be opened and the contents will indicate whether or not to utter the word 22 23 “Quiet”. 24 c. Daily allocation to be selected by a member of the clinical team for that day.. The 25 26 individual selecting the allocation will read the selected piece of paper for that day 27 28 and will either day: 29 i) Say nothing (Control) 30 31 ii) Say “Today will be a quiet day” (Intervention) 32 33 17. Blinding: There will be no blinding of the clinical team to the intervention since the 34 outcomes are objective and not likely to be affected by the knowledge of the intervention. 35 36 18. Data collection: numbers of clinical calls, alerts within work hours and validated results can 37 38 be gathered from laboratory computer systems. For out of hours work individual on call 39 members of the team have agreed to manually record their calls and alerts. 40 41 19. Data management: No patient identifiable information will be recorded. Data will be 42 43 recorded on to an Excel spreadsheet by CB and RS. 44 20. Statistical methods: The difference in mean of the primary outcome measure between the 45 46 two groups will be calculated with 95% confidence interval. All days will be included and 47 48 analysed in the group to which they were randomised (irrespective of whether the word 49 “Quiet” was uttered). The upper bound of the 95% confidence interval of the difference 50 51 (intervention – control) will be compared to the margin of non-inferiority. If the upper 52 53 bound is less than the margin, then non-inferiority will be declared. Mean difference and 54 95% confidence intervals will also be calculated for each component of the composite 55 56 outcome. The analysis will be repeated in the sub groups of weekdays (Monday-Friday) and 57 58 weekends (Saturday and Sunday) to identify whether the treatment effect differs in these 59 two subgroups. 60

https://mc.manuscriptcentral.com/bmj Page 19 of 19 BMJ

1 2 3 21. Data monitoring committee: not required 4 5 22. Harms: adverse events not anticipated- no patient recruitment 6 7 23. Auditing: not required 8 9 Ethics 10 11 12 24. EthicsConfidential: approval: not required For Review Only 13 25. Protocol amendments: to be disseminated between CB, RS and PP 14 15 26. Consent: not required 16 17 27. Confidentiality: No personal data to be recorded 18 28. Declaration of interests: None known or anticipated 19 20 29. Access to data: Limited to CB, RS and PP. 21 22 30. Post trial care: not applicable 23 31. Dissemination policy: Intention to submit trial results for publication in a peer-reviewed 24 25 journal. No intention to use professional writers. 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

https://mc.manuscriptcentral.com/bmj