Confidential: for Review Only

BMJ

Confidential: For Review Only

Plenty o f moustaches but not enough women: A cross-sectional study of medical leaders

Journal: BMJ

Manuscript ID: BMJ.2015.028492

Article Type: Christmas

BMJ Journal: BMJ

Date Submitted by the Author: 03-Aug-2015

Complete List of Authors: Wehner, Mackenzie; University of Pensylvania, Nead, Kevin; University of Pensylvania, Linos, Katerina; University of California, Berkeley Law Linos, Eleni; University of California San Francisco, Dermatology

Keywords: gender, women, leadership, moustache, medicine

https://mc.manuscriptcentral.com/bmj Page 1 of 20 BMJ

1 2 3 4 5 6 7 Plenty of moustaches but not enough women: 8 Confidential: For Review Only 9 A cross-sectional study of medical leaders 10 11 12 1 2 13 Mackenzie R. Wehner, MD MPhil , Kevin T. Nead, MD MPhil , Katerina 3 4 14 Linos JD PhD , Eleni Linos, BM BCh DrPH 15 16 both authors contributed equally 17 18 4 19 Corresponding author : 20 Eleni Linos 21 Assistant Professor 22 [email protected] 23 Department of Dermatology, 24 25 University of California San Francisco 26 2340 Sutter St, San Francisco CA 94143, 27 +14155020433 28 United States 29 30 1 31 Resident Physician, Department of Dermatology, Hospital of the University of 32 the Pennsylvania, Philadelphia, Pennsylvania 19104, United States 33 34 2Resident Physician, Department of Radiation Oncology, Hospital of the 35 University of the Pennsylvania, Philadelphia, Pennsylvania 19104, United States 36 37 3 38 Professor of Law, Berkeley Law, University of California Berkeley, CA, , United 39 States 40 41 Key words: Gender, women, leadership, moustache, medicine 42 43 44 Word count: 1520 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 2 of 20

1 2 3 Abstract: 4 5 6 Objectives: To investigate the representation of women compared to men with 7 moustaches in academic medical leadership. 8 Confidential: For Review Only 9 Design: A cross-sectional analysis. 10 11 12 Setting: Academic medical departments in the US. 13 14 Participants: Clinical department leaders (N=1017) at the top 50 National 15 Institutes of Health funded US medical schools. 16 17 18 Main outcome measures: The proportion of women and moustaches across 19 institutions and specialties. The Moustache Index: percent women / percent 20 moustaches. 21 22 Results: Women accounted for 13% of department leaders at the top 50 NIH- 23 24 funded medical schools in the US. Moustachioed men accounted for 19%. The 25 proportion of women department leaders ranged from 0% to 26% across 26 institutions and 0% to 36% across specialties. Only 5 institutions and 5 27 specialties had more than 20% women department leaders. The overall 28 Moustache Index was 0.68 (133 women compared to 195 moustaches). 29 Fourteen of 20 specialties had more moustaches than women (Moustache Index 30 31 < 1). 32 33 Conclusions Men with moustaches outnumber women as leaders of medical 34 departments in the US. We believe that every department and institution should 35 strive for a Moustache Index greater than 1. Known, effective evidence-based 36 37 policies to increase the number of women in leadership positions should be 38 prioritised. 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 3 of 20 BMJ

1 2 3 Introduction 4 5 6 Medicine, a historically male-dominated field, has undergone significant change 7 8 Confidential:in gender representation in recent decades.For WomenReview accounted for Only only 9% of 9 10 11 United States (US) medical students in 1960, but for the past 15 years, almost 12 13 50% of medical students have been women. However, the proportion of women 14 15 in academic medicine remains low and drops with increasing academic rank: 16 17 18 38% of full-time faculty, 21% of full professors and 16% of deans are women. [1 19 20 2] There are not enough women in leadership positions. This is a problem not 21 22 only because of the strong theoretical argument for equality, but also for practical 23 24 25 reasons: in business having more women leaders has been linked with better 26 27 performance[3]. 28 29 30 31 32 Our goal was to investigate the representation of women in academic medical 33 34 leadership. However, given the likely societal habituation to the longstanding lack 35 36 37 of women in leadership positions (i.e. everyone is bored of hearing about this) we 38 39 sought to create a contextual construct to facilitate meaningful discourse. 40 41 Specifically, we compared the proportion of women in leadership positions to 42 43 44 men with the relatively rare but extremely appealing facial adornment: the 45 46 Moustache. Our hypothesis was that fewer women lead academic medical 47 48 departments in the US than men with moustaches. 49 50 51 52 53 Methods 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 4 of 20

1 2 3 Setting and Participants: This is a cross-sectional study of 1,017 leaders of 4 5 6 academic medical departments in the US. We used publicly available data 7 8 Confidential:(http://report.nih.gov/award) to identify For the top Review 50 National Institutes Only of Health 9 10 11 (NIH) funded schools of medicine in the US in 2014. We developed a list of 12 13 clinical medical specialties as defined by the NIH (Anesthesia, Dermatology, 14 15 Emergency Medicine, Family Medicine, General Surgery, Internal Medicine, 16 17 18 Neurology, Neurosurgery, Obstetrics and Gynecology, Ophthalmology, 19 20 Orthopedics, Otolaryngology, Pathology, Pediatrics, Plastic Surgery, Physical 21 22 Medicine and Rehabilitation, Psychiatry, Radiology, Radiation Oncology and 23 24 25 Urology). We then utilized institutional websites to identify the leader (e.g. chair, 26 27 chief, head) of each specialty, which we refer to as department leader regardless 28 29 of individual institutional structure (e.g. department, division, section etc.). For 30 31 32 each department leader we collected their medical specialty, institution, gender, 33 34 and, in men, presence and type of facial hair. Gender and facial hair were 35 36 37 confirmed by photograph. Data were reviewed by three authors (MRW, KTN, EL) 38 39 and collected between May 28th, 2015 and July 19th, 2015. 40 41 42 43 44 Definition of Moustache: Categories of facial hair are shown in Figure 1. We 45 46 defined a moustache as the visible presence of hair on the upper cutaneous lip 47 48 and included both stand-alone moustaches (e.g. Copstash Standard, Pencil, 49 50 51 Handlebar, Dali, Fu Manchu, Supermario) as well as moustaches in combination 52 53 with other facial hair (e.g. Van Dyke, The Balbo, Napoleon III Imperial, The 54 55 Zappa). Department leaders with facial hairstyles that did not include hair on the 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 5 of 20 BMJ

1 2 3 upper lip (e.g. Soul Patch, Mutton Chops, Chin Curtain) were considered not to 4 5 6 have a moustache. 7 8 Confidential: For Review Only 9 10 11 Moustache Index: We examined the proportion of department leaders in 12 13 academic medicine who were women and men with moustaches. We calculated 14 15 a Moustache Index by dividing the percentage who were women by the 16 17 18 percentage who were moustachioed men. These calculations were made across 19 20 institutions and specialties. 21 22 23 24 25 Statistical analysis: Pearson’s correlation coefficients were calculated to 26 27 examine the relationship between the percentage of women and moustachioed 28 29 department leaders across institutions and specialties. We also compared the 30 31 32 percentage of women and moustachioed department leaders across institutions 33 34 using linear regression analysis adjusting for institutional region within the US. 35 36 37 38 39 Results 40 41 There were 1,017 department leaders who met our inclusion criteria and were 42 43 44 included in this study. We found that women accounted for 13% of department 45 46 leaders at the top 50 NIH-funded medical schools in the US. We found that 47 48 moustachioed men accounted for 19% of department leaders at these 49 50 51 institutions. 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 6 of 20

1 2 3 Figure 2a shows the proportion of women department leaders across the 50 4 5 6 institutions studied. The proportion of women department leaders ranged from 7 8 Confidential:0% (n = 3 institutions) to 26%. Only For 5 institutions Review had more than 20% Only women 9 10 11 department leaders. Figure 2b shows the proportion of women department 12 13 leaders by medical specialty. The proportion of women department leaders 14 15 ranged from 0% (n = 1) to 36%. Only 5 specialties had more than 20% women 16 17 18 department leaders: Obstetrics and Gynecology (36%), Pediatrics (31%), 19 20 Dermatology (23%), Family Medicine (21%), and Physical Medicine and 21 22 Rehabilitation (21%). 23 24 25 26 27 Figure 3a shows the proportion of moustachioed department leaders across 28 29 institutions. The proportion of moustachioed department leaders ranged from 0% 30 31 32 (n = 1) to 42%. Twenty institutions had more than 20% moustachioed department 33 34 leaders. Figure 3b shows the proportion of moustachioed department leaders by 35 36 37 medical specialty. The proportion of moustachioed department leaders ranged 38 39 from 2% to 29%. Ten specialties had more than 20% moustachioed department 40 41 leaders, with the greatest moustache density in Psychiatry (29%), Orthopaedics 42 43 44 (29%), and Pathology (28%). Two specialties had fewer than 10% moustaches 45 46 (General Surgery [2%] and Plastic Surgery [4%]). 47 48 49 50 51 The overall Moustache Index of all academic medical departments studied was 52 53 0.68 (133 women compared to 195 moustaches). Figure 4 shows the Moustache 54 55 Index across institutions (a) and specialties (b). Six out of 20 specialties had 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 7 of 20 BMJ

1 2 3 Moustache Indices greater than 1, indicating that there were more women than 4 5 6 moustaches: Pediatrics (1.33), Dermatology (1.50), Physical Medicine and 7 8 Confidential:Rehabilitation (1.75), Obstetrics and For Gynecology Review (1.90), Plastic Surgery Only (2.0) and 9 10 11 General Surgery (2.0). 12 13 14 15 There was a statistically non-significant negative correlation (r = -0.254) between 16 17 18 the percentage of moustaches and women across institutions (p = 0.075). 19 20 Results were similar using linear regression analysis adjusting for region of 21 22 institution (coefficient = -0.192; 95% confidence interval = -0.411 to 0.027;p = 23 24 25 0.084). There was a statistically non-significant positive correlation (r = 0.149) 26 27 between the percentage of mustaches and women across departments (p = 28 29 0.530). 30 31 32 33 34 Discussion 35 36 37 Men with moustaches outnumber women as leaders of medical departments in 38 39 the US. The specialties of Pediatrics, Family Medicine, Obstetrics and 40 41 Gynecology, Physical Medicine and Rehabilitation, and Dermatology had the 42 43 44 highest proportions of women leaders and the highest Moustache Indices. 45 46 General Surgery and Plastic Surgery also had high Moustache Indices, but this 47 48 effect was driven mostly by the absence of moustaches, rather than the number 49 50 51 of women in leadership positions. 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 8 of 20

1 2 3 We believe that every department and institution should strive for a Moustache 4 5 6 Index greater than 1. There are two ways to achieve this goal: either increasing 7 8 Confidential:the number of women in leadership For positions Reviewor by asking men in leadershipOnly 9 10 11 positions to shave their moustaches. In addition to being discriminatory, the latter 12 13 choice could have detrimental effects on workplace satisfaction and emotional 14 15 well-being. Deans are left with one viable option: to hire, retain and promote 16 17 18 more women. 19 20 21 22 Closing the gender gap 23 24 25 Gender discrepancies in leadership are distressingly common in many 26 27 professional fields. Many large employers have taken the first major steps to 28 29 reduce these gaps, by adopting policies against formal discrimination and sexual 30 31 32 harassment, and by introducing family-friendly benefits. These tools, especially 33 34 paid parental leave, have been shown to significantly improve women’s labor 35 36 37 force outcomes[4 5]. However, recent evidence from psychology, sociology and 38 39 economics suggests that two additional strategies may be necessary to fully 40 41 close the gap. One effective, and relatively inexpensive, technique is to define 42 43 44 hiring criteria extensively, in advance of evaluating individual candidates[6]. 45 46 Without clearly defined hiring criteria, evaluators unconsciously redefine what 47 48 they are seeking to match the attributes of the male candidate before them[7]. As 49 50 51 a result, women, and especially mothers, tend to be evaluated more negatively 52 53 than men with the same professional characteristics[8]. 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 9 of 20 BMJ

1 2 3 A second strategy that can significantly improve women’s chances of reaching 4 5 6 leadership positions involves increasing temporal flexibility in job structures[9]. In 7 8 Confidential:many occupations, the ideal worker For is one who Review works long hours eachOnly week, 9 10 11 continuously over many decades. Women suffer significant penalties in status 12 13 and pay for taking even a short time off to care for children[9]. This penalty differs 14 15 dramatically by field: it is lowest in fields like pharmacy, in which organizational 16 17 18 innovations allow workers to easily substitute for one another[10]. In medicine, 19 20 innovations that may have many other benefits, such as creating larger practices 21 22 to enable teamwork, computerized medical records, and shift work, may also 23 24 25 dramatically reduce gender inequality, by reducing the premium for very long 26 27 hours and uninterrupted employment[11]. 28 29 30 31 32 Limitations 33 34 In order to highlight the paucity of women in academic medical leadership, we 35 36 37 wanted to choose an easily identifiable comparator that is unrelated to promotion 38 39 and academic achievement: the moustache. However, facial hair has been 40 41 shown to enhance perceptions of maturity, responsibility, dominance, strength 42 43 44 and self-confidence[12 13]. In addition, men who put on fake beards rate 45 46 themselves as more masculine[14]. If moustaches are linked to workplace 47 48 success, this could bias our Moustache Indices. Misclassification of moustaches 49 50 51 is another potential limitation, and our data are only as accurate as the 52 53 institutional websites that were used for collection. Finally, our sample was 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 10 of 20

1 2 3 limited to clinical departments in NIH-funded medical schools, which may limit its 4 5 6 generalizability. 7 8 Confidential: For Review Only 9 10 11 Conclusion 12 13 We conclude that there are more men with moustaches leading academic 14 15 medical departments than women. Two evidence-based solutions could be 16 17 18 applied to academic medicine to close the gender gap: a) predefining hiring 19 20 criteria and b) innovations that encourage teamwork including job sharing, 21 22 shiftwork and use of electronic medical records. We hope that these solutions will 23 24 25 help increase Moustache Indices across all specialties while maintaining enough 26 27 facial hair in our workplaces. 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 11 of 20 BMJ

1 2 3 4 5 6 What is already known on this topic 7 8 Confidential: For Review Only 9 10 - The number of women in medicine has increased significantly since the 11 1960s, with women now accounting for nearly 50% of US medical 12 students 13 14 - The proportion of women in academic medicine is still low: only 21% of 15 16 full professors are women 17 18 19 20 21 What this study adds 22 23 24 25 - There are fewer women leading medical departments than men with 26 moustaches in the US 27 28 29 Evidence-based policies that increase women in leadership positions 30 - 31 include: 32 33 1) Predefining detailed hiring criteria 34 2) Supporting innovations that allow workers to be interchangeable and 35 encourage working in teams (e.g. shiftwork, jobsharing, electronic 36 37 medical records) 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 12 of 20

1 2 3 Figure 1. Facial hair categories. 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 Image courtesy of Jon Dyer / dyers.org , with permission 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 13 of 20 BMJ

1 2 3 Figure 2: Women department leaders by percent by A) institution and B) 4 5 specialty 6 7 Figure 3: Mustachioed department leaders by percent by A) institution and B) 8 Confidential:specialty For Review Only 9 10 11 Figure 4: Moustache Index (percent women/percent moustachioed men) by A) 12 institution and B) specialty 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 20

1 2 3 Competing interests: MRW, KL, and EL are women. KTN is a man and does not 4 5 currently have a moustache. All authors have completed the Unified Competing 6 Interest form (available on request from the corresponding author) and declare: 7 no support from any organisation for the submitted; no financial relationships with 8 Confidential:any organisations that might have anFor interest Reviewin the submitted work Only in the 9 previous three years, no other relationships or activities that could appear to 10 11 have influenced the submitted work. 12 13 Contributions: MRW, KTN and EL conceived and designed the project and 14 collected data. MRW and KTN analysed the data. MRW, KTN, KL and EL wrote 15 the manuscript. EL is the guarantor. 16 17 18 Acknowledgements: The authors would like to thank Jon Dyer for use of Figure 19 1, Natalie Linos and Elizabeth Linos for content advice and editing. EL thanks the 20 American Academy of Dermatology 2015 Leadership Forum and KL thanks the 21 Hellman family fund. 22 23 24 The Corresponding Author has the right to grant on behalf of all authors and 25 does grant on behalf of all authors, an exclusive on a worldwide basis to the BMJ 26 Publishing Group Ltd to permit this article (if accepted) to be published in BMJ 27 editions and any other BMJPGL products and sublicences such use and exploit 28 all subsidiary rights, as set out in our licence. 29 30 31 The manuscript’s guarantor affirms that the manuscript is an honest, accurate, 32 and transparent account of the study being reported; that no important aspects of 33 the study have been omitted; and that any discrepancies from the study as 34 planned have been explained 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 20 BMJ

1 2 3 References 4 5 6 7 1. AAMC. The State of Women in Academic Medicine: The Pipeline and Pathways to 8 Confidential:Leadership. 2013. https:// www.aamc.org/members/gwims/statistics/ For Review Only. 9 2. Page L. Women Still Face Barriers to Leadership Roles in Academic Medicine. 10 11 2012; (December 2012). 12 https:// www.aamc.org/newsroom/reporter/dec2012/323602/women.html 13 . 14 3. Dezsö CL, Ross DG. Does female representation in top management improve firm 15 performance? A panel data investigation. Strategic Management Journal 16 17 2012; 33 (9):1072-89 18 4. Linos K. Diffusion through democracy. American Journal of Political Science 19 2011; 55 (3):678-95 20 5. Ruhm CJ. The Economic Consequences of Parental Leave Mandates: Lessons from 21 Europe. Quarterly Journal of Economics 1998; :285-317 22 108-1 23 6. Uhlmann E, Cohen GL. Constructed criteria: redefining merit to justify 24 discrimination. Psychol. Sci. 2005; 16 (6):474-80 25 7. Phelan JE, Moss ‐Racusin CA, Rudman LA. Competent yet out in the cold: Shifting 26 criteria for hiring reflect backlash toward agentic women. Psychol. Women Q. 27 28 2008; 32 (4):406-13 29 8. Correll SJ, Benard S, Paik I. Getting a Job: Is There a Motherhood Penalty? 1. 30 American journal of sociology 2007; 112 (5):1297-339 31 9. Goldin C. A grand gender convergence: Its last chapter. The American Economic 32 Review 2014; 104 (4):1091-119 33 34 10. Goldin C, Katz LF. The most egalitarian of all professions: Pharmacy and the 35 evolution of a family-friendly occupation: National Bureau of Economic 36 Research, 2012. 37 11. Briscoe F. Temporal flexibility and careers: The role of large-scale organizations 38 39 for physicians. Ind. Labor Relat. Rev. 2006; 60 (1):88-104 40 12. De Souza AAL, Baumgasten V, Baiao U, et al. Perception of Men's Personal 41 Qualities and Prospect of Employment as a Function of Facial Hair. Psychol. 42 Rep. 2003; 92 (1):201-08 doi: 10.2466/pr0.2003.92.1.201[published Online 43 First: Epub Date]|. 44 45 13. Hellström Å, Tekle J. Person perception through facial photographs: Effects of 46 glasses, hair, and beard on judgments of occupation and personal qualities. 47 Eur. J. Soc. Psychol. 1994; 24 (6):693-705 doi: 48 10.1002/ejsp.2420240606[published Online First: Epub Date]|. 49 14. Wood DR. Self-Perceived Masculinity Between Bearded and Non-Bearded Males. 50 51 Percept. Mot. Skills 1986; 62 (3):769-70 52 15. Reed J, Blunk EM. The influence of facial hair on impression formation. Social 53 Behavior and Personality: an international journal 1990; 18 (1):169-75 54 16. Dawson J, Kersley R, Natella S. The CS Gender 3000: Women in Senior 55 Management: Credit Suisse Research Institute, September 2014. 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 16 of 20

1 2 3 4 5 6 7 8 Confidential: For Review Only 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 250x108mm (300 x 300 DPI) 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 17 of 20 BMJ

1 2 3 4 5 6 7 8 Confidential: For Review Only 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 250x109mm (300 x 300 DPI) 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 18 of 20

1 2 STROBE Statement—checklist of items that should be included in reports of observational studies 3 Item 4 No 5 Location in 6 Recommendation Manuscript 7 Title and abstract 1 (a) Indicate the study’s design with a commonly used term in the 1 8 Confidential:title or the abstract For Review Only 9 (b) Provide in the abstract an informative and balanced summary of 2 10 11 what was done and what was found 12 Introduction 13 Background/rationale 2 Explain the scientific background and rationale for the 3 14 investigation being reported 15 16 Objectives 3 State specific objectives, including any prespecified hypotheses 3 17 Methods 18 Study design 4 Present key elements of study design early in the paper 4 19 20 Setting 5 Describe the setting, locations, and relevant dates, including 4 21 periods of recruitment, exposure, follow-up, and data collection 22 Participants 6 (a) Cohort study —Give the eligibility criteria, and the sources and 4 23 methods of selection of participants. Describe methods of follow- 24 25 up 26 Case-control study —Give the eligibility criteria, and the sources 27 and methods of case ascertainment and control selection. Give the 28 rationale for the choice of cases and controls 29 Cross-sectional study —Give the eligibility criteria, and the sources 30 31 and methods of selection of participants 32 (b) Cohort study —For matched studies, give matching criteria and 33 number of exposed and unexposed 34 Case-control study —For matched studies, give matching criteria 35 and the number of controls per case 36 37 Variables 7 Clearly define all outcomes, exposures, predictors, potential 5 38 confounders, and effect modifiers. Give diagnostic criteria, if 39 applicable 40 Data sources/ 8* For each variable of interest, give sources of data and details of 5 41 42 measurement methods of assessment (measurement). Describe comparability of 43 assessment methods if there is more than one group 44 Bias 9 Describe any efforts to address potential sources of bias 9 45 Study size 10 Explain how the study size was arrived at 4 46 47 Quantitative 11 Explain how quantitative variables were handled in the analyses. If 5 48 variables applicable, describe which groupings were chosen and why 49 Statistical methods 12 (a) Describe all statistical methods, including those used to control 5 50 for confounding 51 (b) Describe any methods used to examine subgroups and 5 52 53 interactions 54 (c) Explain how missing data were addressed N/A 55 (d) Cohort study —If applicable, explain how loss to follow-up was 56 addressed 57 58 Case-control study —If applicable, explain how matching of cases 59 and controls was addressed 60 https://mc.manuscriptcentral.com/bmj1 BMJ Page 20 of 20

1 2 Cross-sectional study —If applicable, describe analytical methods 3 taking account of sampling strategy 4 (e) Describe any sensitivity analyses N/A 5 6 7 Results Location in 8 Confidential: For Review Onlymanuscript 9 Participants 13* (a) Report numbers of individuals at each stage of study—eg numbers 5 10 potentially eligible, examined for eligibility, confirmed eligible, included in 11 12 the study, completing follow-up, and analysed 13 (b) Give reasons for non-participation at each stage 14 (c) Consider use of a flow diagram 15 Descriptive 14* (a) Give characteristics of study participants (eg demographic, clinical, social) 5 16 data and information on exposures and potential confounders 17 18 (b) Indicate number of participants with missing data for each variable of N/A 19 interest 20 (c) Cohort study —Summarise follow-up time (eg, average and total amount) 21 Outcome data 15* Cohort study —Report numbers of outcome events or summary measures over 22 23 time 24 Case-control study—Report numbers in each exposure category, or summary 25 measures of exposure 26 Cross-sectional study—Report numbers of outcome events or summary N/A 27 28 measures 29 Main results 16 (a) Give unadjusted estimates and, if applicable, confounder-adjusted Figures 2-4 30 estimates and their precision (eg, 95% confidence interval). Make clear which Page 6-7 31 confounders were adjusted for and why they were included 32 33 (b) Report category boundaries when continuous variables were categorized N/A 34 (c) If relevant, consider translating estimates of relative risk into absolute risk N/A 35 for a meaningful time period 36 Other analyses 17 Report other analyses done—eg analyses of subgroups and interactions, and 7 37 sensitivity analyses 38 39 Discussion 40 Key results 18 Summarise key results with reference to study objectives 7 41 Limitations 19 Discuss limitations of the study, taking into account sources of potential bias 9 42 43 or imprecision. Discuss both direction and magnitude of any potential bias 44 Interpretation 20 Give a cautious overall interpretation of results considering objectives, 8 45 limitations, multiplicity of analyses, results from similar studies, and other 46 relevant evidence 47 48 Generalisability 21 Discuss the generalisability (external validity) of the study results 9 49 Other information 50 Funding 22 Give the source of funding and the role of the funders for the present study 14 51 and, if applicable, for the original study on which the present article is based 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj2