Despite marked sex differences in many aspects of human physiology and behaviour, biomedical research continues to be disproportionately biased towards the male sex. For example, women made up only 25% of participants in landmark trials for congestive heart failure and 19.2% of participants for studies in antiretroviral treatment of HIV (Criado-Perez, 2019). Such a skewed evidence base leads to disparities in clinical and non-clinal research applications and it weakens the impact of science-based policies and translational outcomes.

This sex bias or ’sex data gap' – whereby data mainly come from male individuals – has recently received widespread attention (Criado-Perez, 2019), with policy advisers (Buitendijk and Maes, 2015), funders (Lee, 2018; Clayton and Collins, 2014) and publishers (Rippon et al., 2017; Docherty et al., 2019) pushing for better inclusivity in research regarding sex. Embracing these new practices should improve translational outcomes and scientific efficiency, but this will require a two-pronged tactic that both strengthens forces for change and weakens barriers in the field (Karp and Reavey, 2019). The problems that allow sex bias to emerge are multifaceted and closing the data gap will require solutions to be bespoke for each research community.

Here, we explore the sex data gap in the context of human circadian physiology and sleep research. The field focuses on the temporal organization of physiology and behaviour at a daily scale, including rest-activity cycles, diurnal changes in hormone levels and cognitive performance, as well as the non-visual effects of light. First, we describe primary findings on sex differences in circadian physiology and sleep. Next, we discuss the sex data gap in circadian and sleep research based on an analysis of over 150 papers on the non-visual effects of light, and finally we outline recommendations emerging from a virtual workshop on the biomedical implications of sex differences in sleep and circadian physiology (held in June 2020).

While we distinguish between gender identity (how individuals and groups perceive themselves e.g. men, women, non-binary) and sex (the biological attributes that distinguish organisms as female, male or intersex), we note that these terms are often used interchangeably and wrongly in the literature (Tannenbaum et al., 2019). Yet, in biology, sex describes differences in sexual characteristics that go beyond reproductive functions. Furthermore, we acknowledge that there is very little to no research about intersex individuals within circadian physiology and sleep research, constituting an important gap. Addressing this gap may contribute to better granularity and understanding of sex-differentiated biological mechanisms and responses. When reporting on results from the literature, we use the terms used by the researchers in these studies, as we are unable to know whether participants were asked about their sex or their gender.

Human circadian and sleep physiology features well-established sex differences: for instance, circadian timing is phase-advanced (earlier) in female compared to male individuals, as seen in the core body temperature minimum and evening rise in melatonin (Boivin et al., 2016; Cain et al., 2010). Female individuals also have a shorter circadian period of the temperature and melatonin rhythms (Duffy et al., 2011), and larger amplitude of the melatonin rhythm (Cain et al., 2010). Furthermore, sex differences exist in chronotype, the circadian continuum of early (‘larks’) to late (‘owls’) diurnal preference (Chontong et al., 2016), such that more male individuals are late types than females (Fischer et al., 2017; Roenneberg et al., 2004; Fischer et al., 2016; Phillips et al., 2017). With regard to sleep, female individuals have an earlier timing of sleep, longer sleep duration and more slow-wave sleep (Roenneberg et al., 2004; Dijk et al., 1989).

More recently, sleep regularity – the day-to-day consistency in sleep timing and duration – has emerged as an important factor in health (Bei et al., 2016). Irregular sleep is associated with cardiovascular disease (Yoon et al., 2014), inflammation (Okun et al., 2011), metabolic disorders (Patel et al., 2014; Chontong et al., 2016; Spruyt et al., 2011), mental health conditions (Lemola et al., 2013; Vanderlind et al., 2014), and cognitive impairment (McBean and Montgomery-Downs, 2013). The data on sex differences are mixed with reports ranging from no sex differences (Kaufmann et al., 2016; Xu et al., 2018; Minors et al., 1998) to more irregular sleep in female (Mezick et al., 2009; Dillon et al., 2015; Lunsford-Avery et al., 2018) or in male individuals (Roane et al., 2015; Yetish et al., 2018). Chronotype may account for these inconsistencies as ‘owls’ tend to be more irregular sleepers (Duffy et al., 2011; Fischer et al., 2017). Indeed, in revisiting three published datasets (Fischer et al., 2016; Fischer et al., 2020; Keller et al., 2017), more male individuals were found to be irregular sleepers than females when both were a later chronotype.

Finally, while data remain sparse, the adverse health effects of sleep irregularity itself may also differ between the sexes. To the best of our knowledge, only one study examined this question (Roane et al., 2015), finding that variability in sleep duration was significantly associated with weight gain in male but not female students. Overall, despite the far-reaching health implications, sex differences in sleep and circadian physiology remain underresearched.

The sex data gap exists both in in vivo (Curtis et al., 2018) and in vitro research (National Institute for Health and Care Excellence, 2020) and is apparent even in diseases that predominantly affect women (Patel et al., 2014). Critically, the sex gap is not just restricted to inclusion at the experimental design stage: researchers frequently ignore sex as a factor in the analysis, even when males and females are included in the study (Karp and Reavey, 2019).

Apart from vision, light plays a critical role in regulating physiology and behaviour via its influence on the circadian system. These effects are mediated by a multi-component photoreceptor system consisting of rods, cones and intrinsically photosensitive retinal ganglion cells in the eye that transmit information to the circadian clock via the retinohypothalamic tract.

To ascertain whether there is a sex data gap in sleep and circadian physiology research, we focused on the non-visual effects of light on human physiology and behaviour – including how it suppresses melatonin and shifts the circadian system. This topical research area has applicability in lighting standards, regulations and guidelines (CIE, 2019; International WELL Building Institute, 2020), and various efforts are underway to incorporate scientific data from this research area into building recommendations. This highlights a pressing need to understand sex bias in this field.

A preliminary literature search identified 545 papers, which were evaluated against a list of exclusion criteria (see Methods for full details), yielding a total of 180 articles. In this specific analysis, we focused on the reported sex of participants, although in many instances the terms sex and gender were used interchangeably. Each paper was then reviewed by a single reviewer to determine if participant sex and numbers were reported, and where possible, the proportion of female participants was calculated. Of the papers assessed, 14 (7.77%) did not give sufficient information on the sex breakdown of participants for this to be determined. In the remaining 166 articles, females comprised an average of 33.9% of the sample. Seven papers reported studying exclusively female participants, while 56 papers reported studying only males. Figure 1 shows the proportion of female study participants as a function of the publication year, calculated from the per-sex participant sample sizes. We conducted binomial tests to investigate the possibility of deviations from a balanced distribution of sexes, finding a large proportion of studies using only male volunteers. Interestingly, for later years, there were fewer female-only studies (but also fewer studies in total). While this may represent a shift towards more sex-inclusive recruitment, it also means that large parts of the cited literature are based on imbalanced participant samples.

Figure 1

Download asset Open asset

Figure 1—source code 1

R code to produce Figure 1.

https://cdn.elifesciences.org/articles/65419/elife-65419-fig1-code1-v2.zip

Download elife-65419-fig1-code1-v2.zip

Figure 1—source data 1

Excel spreadsheet containing the data underlying Figure 1.

https://cdn.elifesciences.org/articles/65419/elife-65419-fig1-data1-v2.xlsx

Download elife-65419-fig1-data1-v2.xlsx

Next, we examined studies that exclusively involved male or female participants (N = 63). Of these, only eleven (17.5%) provided text to justify this sample choice. For studies with only male participants the justifications included female physiology being subject to confounding factors such as menstruation (N = 3), research into the other sex being unnecessary due to previously published observations (n = 3), the study involving a sex-specific condition (N = 2), not being able to recruit females with a specific genetic polymorphism (N = 1), the study being a case study (N = 1), and the study being conducted in a location (field station) with only male staff (N = 1). We found some evidence that the number of females increased over time, with publication year and proportion of female participants being correlated (r(164)=0.17, P = 0.02895). Interestingly, the total sample size correlates with the fraction of female participants in a given study (r(164)=0.3, P = 0.00008; Spearman’s correlation): larger studies seem to recruit more balanced samples.

In summary, we find a sex data gap in the literature on the non-visual effects of light, which needs to be considered in current efforts to translate research findings in the 'real world'.

In a three-part virtual workshop held in June 2020, the authors of this paper explored practices, barriers, and challenges in designing and executing inclusive research in circadian physiology and sleep research. All materials from the workshop, including the recordings and the programme, are available under the CC-BY license (Fischer and Vidafar, 2020; Karp and Ahluwalia, 2020; White and Lévi, 2020).

The workshop series comprised three 90-minute sessions held a week apart and included invited talks as well as interactive sessions. The workshop was advertised through a range of channels, including Twitter, the UK Clock Club listserv, and the personal networks of the organisers and speakers. A total of 275 participants registered for the entire workshop. Across the three workshops, between 38 and 94 attendees participated in the interactive sessions, with approximately four out of five participants being researchers (82 out of 94 in Workshop 1, 47 out of 60 in Workshop 2 and 31 out of 38 in Workshop 3).

We used the web platform Mentimeter to implement polling amongst participants as well as open-ended questions. Prior to participating in the interactive sessions, attendees were informed that their responses would be used for write-up and published as a peer-reviewed article. Attendees were free to not participate in the interactive sessions. No personal data were collected as part of the interactive Mentimeter sessions. We combined yes/no, ranking and open-ended questions throughout the interactive sessions to vary the response modality. The results discussed below were selected from the results, which can be viewed in full on the Open Science Framework page. The number of responses to individual questions varied somewhat due to dropout during the interactive session as well as a time-limited response window; the total number of responses in the participatory parts are given on the bottom right-hand corner of the Materials document.

Workshop 2: Understanding impact

The second workshop focused on understanding the real-world impact of the participants’ research. In the interactive polling segment following this workshop, participants indicated that their research could mostly influence precision and personalized medicine, occupational timing and shift/rota planning, and guidelines for an indoor ‘circadian’ lighting.

When asked to identify the biggest barriers to addressing sex bias in research, research money or funding and time were the most mentioned factors, followed by guidelines and policies. This indicates a scope for funding agencies to specifically address researchers’ need for funding, as well as an opportunity for institutions, funders, professional bodies, learned societies and journals to develop clear guidance (see Box 1 for an example of a journal implementing a specific policy; and Figure 2). The video recording for Workshop 2 is available here, and the materials related to the participatory part are here.

Box 1

Example journal policy to addressing sex bias.

Amrita Ahluwalia, Editor-in-Chief of British Journal of Pharmacology (BJP)

In 2018, the British Journal of Pharmacology identified the issue of sex bias in pharmacological research as a critical area for attention with respect to the work published in the journal. This came following an internal survey of our published work coupled with recognition of the activities and actions of the National Institutes of Health, in the US, raising the profile of this important issue (National Institutes of Health, 2020). We discovered that in addition to a prevailing reluctance to use female individuals in experimental research, both in vivo and in vitro, there was the unsurprising omission of detail regarding the sex of the source for experimental work involving primary cell culture (Docherty et al., 2019).

To address these issues, we introduced a number of initiatives, including: (1) publishing a themed issue in BJP containing a number of reviews and original articles focused on sex differences in pharmacology; (2) bringing together a collection of articles from all of the journals owned by the BPS in a virtual issue focused on sex; and, most importantly, (3) the elaboration and publication of guidelines for original research published in BJP. The aim of this guidance is to ensure that sex as an experimental variable is no longer ignored in articles published in BJP, but also to provide researchers with the tools to adapt their experimental design to accommodate for sex.

A key aspiration, of course, is that both male and female subjects are used as a default design in the experimental work detailed in all manuscripts submitted to the journal, but we do not mandate this at present. Our hope is that by insisting that these issues are considered within any submitted work, we raise their profile, organically leading to change. Of course, it is the responsibility of those who work with the journal to ensure that change does indeed occur. Indeed, there are many examples where such an advisory approach with other important issues related to transparency and reproducibility appear to have failed (Leung et al., 2018; Avey et al., 2016). Yet our experience in such approaches at BJP – for instance, with our guidelines on design and analysis (Curtis et al., 2018) – gives us strong hope that change will take place. We plan to conduct surveys of published material annually to assess this, and we will publish the outcome of these audits.

Figure 2

Download asset Open asset

Based on the workshop content and discussions, we propose the following guiding principles to address the sex data gap in biomedical research, and to build an evidence base which is better inclusive of sex and gender. The central tenet includes sex and gender analysis as an essential component of research design. The specific recommendations are:

1. Design research to be inclusive and accessible. In many cases, research is designed exclusively by researchers who may not necessarily have sufficient expertise on how to make their study inclusive and accessible. An important step is reaching clarity in recording and reporting participant sex and gender. As an example, one research team reporting the sex of participants may use participant-derived responses on a questionnaire or intake form, and another group may use the sex assigned at birth, based, for instance, on an ID card. While these could give congruent answers, they represent different types of information. Wider engagement with definitions of sex and gender and questions surrounding this topic within a research group or researcher community could lay the groundwork for making research more inclusive and accessible. As a formalised way to ensure inclusivity, we also suggest that research participants be integrated in the research planning process through Patient and Public Involvement (see Box 2) or similar mechanisms.

2. Implement recruitment strategies that lead to a sex-balanced sample. This includes wide advertisement of research studies, and tailoring recruitment strategies by engaging with patients, participants and the general public, for example through Patient and Public Involvement mechanisms (see Box 2). Given fixed resources, recruiting a sex-balanced sample does not simply mean doubling the sampling size, but merely recruiting a sample with 50% female and 50% male participants. A balanced design is recommended to ensure the resulting statistical analysis is robust and that the variance can be decomposed to the factors of interest without confounding these (Collins et al., 2009). While exceptions to this principle may arise from sex-specific research questions, as a general guiding principle there is little to argue against. Furthermore, this will allow sex to be included as a factor in the analysis without compromising sensitivity to a generalizable main effect.

3. Use data visualization to grasp the effect of sex. An informal visualization in the early stages of analyses can be used to ascertain sex difference trends, which can then be followed up with more rigorous statistical testing.

4. Implement statistical analyses that include sex as a factor and/or perform group analyses by sex, where possible. Sex can be included as a factor or a covariate in analyses, or an alternative strategy can be to perform a group analysis by sex. Both require a good understanding of effect sizes and statistical power. Researchers should seek to upskill in statistics to develop advanced analytic strategies.

5. Make participant-level data open and available to facilitate future meta-analytic efforts. This step requires data to be available, which many journals now mandate. The large, combined sample size afforded by the wide availability of data can enable a sex-related effect to be more readily detectable. We also suggest that researchers should include tables reporting the primary data and participant meta-data as supplementary information in articles. A recent analysis of open science practices in circadian rhythms and sleep research journals (Spitschan et al., 2020) has indicated an opportunity to mandate data sharing in journal policies. Journal policies requiring participant-level data sharing could facilitate future analyses incorporating sex.

While none of these actions will suffice on their own, each will contribute to closing the sex data gap. Of course, the research ecosystem not only includes individual researchers but also institutions of varying sizes. We present multiple actions that can be adopted by institutions, funders, as well as professional bodies, learned societies, journals in Figure 2. These actions were developed from an interactive segment of Workshop 3.

To implement a breadth-first search for identifying relevant papers, we employed a pragmatic hybrid strategy, identifying relevant articles through three main sources, as listed in Table 1. We conducted a citation search of three key recent reviews (Brown, 2020; Souman et al., 2018; Lok et al., 2018) on the acute effects of light, producing a total of 88 papers of which 83 were included in the present analysis. We carried out a search for papers specifically discussing the melatonin-suppressive effects of light in SCOPUS (search carried out on 22 October 2019) through the search term “TITLE-ABS-KEY ((light AND melatonin AND suppress*)) AND (LIMIT-TO (DOCTYPE, "ar")) AND (LIMIT-TO (LANGUAGE, "English"))” (search carried out on 22 October 2019). Limiting the analysis to papers with a minimum of 30 citations, we identified 359 further papers (94 of which were included). Finally, relevant systematic reviews were identified in the Cochrane Library through the search terms "(light AND circadian OR sleep OR alertness)", generating 24 results with six relevant for the present analysis. A citation search was again conducted, generating a further 98 papers (of which three were included). Overall, a total of 545 papers were identified and analyzed, as shown in Table 1.

All data generated or analysed during this study are included in the manuscript and supporting files. Source data and source code files have been provided for Figure 1.

1. Report

   1. Arthritis Research UK

   (2017) Patient & Public Involvement: a researcher’s guide

   Arthritis Research UK.

   https://www.versusarthritis.org/media/1376/patient-and-public-involvement-guidance.pdf

   • Google Scholar
2. 1. Avey MT
   2. Moher D
   3. Sullivan KJ
   4. Fergusson D
   5. Griffin G
   6. Grimshaw JM
   7. Hutton B
   8. Lalu MM
   9. Macleod M
   10. Marshall J
   11. Mei SHJ
   12. Rudnicki M
   13. Stewart DJ
   14. Turgeon AF
   15. McIntyre L
   16. Canadian Critical Care Translational Biology Group

   (2016) The devil is in the details: incomplete reporting in preclinical animal research

   PLOS ONE 11:e0166733.

   https://doi.org/10.1371/journal.pone.0166733

   • PubMed
   • Google Scholar
3. 1. Bagley HJ
   2. Short H
   3. Harman NL
   4. Hickey HR
   5. Gamble CL
   6. Woolfall K
   7. Young B
   8. Williamson PR

   (2016) A patient and public involvement (PPI) toolkit for meaningful and flexible involvement in clinical trials - A work in progress

   Research Involvement and Engagement 2:15.

   https://doi.org/10.1186/s40900-016-0029-8

   • PubMed
   • Google Scholar
4. 1. Bei B
   2. Wiley JF
   3. Trinder J
   4. Manber R

   (2016) Beyond the mean: a systematic review on the correlates of daily intraindividual variability of sleep/wake patterns

   Sleep Medicine Reviews 28:108–124.

   https://doi.org/10.1016/j.smrv.2015.06.003

   • PubMed
   • Google Scholar
5. 1. Boivin DB
   2. Shechter A
   3. Boudreau P
   4. Begum EA
   5. Ng Ying-Kin NMK

   (2016) Diurnal and circadian variation of sleep and alertness in men vs. naturally cycling women

   PNAS 113:10980–10985.

   https://doi.org/10.1073/pnas.1524484113

   • PubMed
   • Google Scholar
6. 1. Bressolle F
   2. Joulia JM
   3. Pinguet F
   4. Ychou M
   5. Astre C
   6. Duffour J
   7. Gomeni R

   (1999) Circadian rhythm of 5-fluorouracil population pharmacokinetics in patients with metastatic colorectal cancer

   Cancer Chemotherapy and Pharmacology 44:295–302.

   https://doi.org/10.1007/s002800050980

   • PubMed
   • Google Scholar
7. 1. Brown TM

   (2020) Melanopic illuminance defines the magnitude of human circadian light responses under a wide range of conditions

   Journal of Pineal Research 69:e12655.

   https://doi.org/10.1111/jpi.12655

   • PubMed
   • Google Scholar
8. Conference

   1. Buitendijk S
   2. Maes K

   (2015) Gendered research and innovation: integrating sex and gender analysis into the research process

   League of European Research Universities.

   https://www.leru.org/files/Gendered-Research-and-Innovation-Full-paper.pdf

   • Google Scholar
9. 1. Cain SW
   2. Dennison CF
   3. Zeitzer JM
   4. Guzik AM
   5. Khalsa SBS
   6. Santhi N
   7. Schoen MW
   8. Czeisler CA
   9. Duffy JF

   (2010) Sex differences in phase angle of entrainment and melatonin amplitude in humans

   Journal of Biological Rhythms 25:288–296.

   https://doi.org/10.1177/0748730410374943

   • PubMed
   • Google Scholar
10. 1. Chansky K
    2. Benedetti J
    3. Macdonald JS

    (2005) Differences in toxicity between men and women treated with 5-fluorouracil therapy for colorectal carcinoma

    Cancer 103:1165–1171.

    https://doi.org/10.1002/cncr.20878

    • PubMed
    • Google Scholar
11. 1. Chontong S
    2. Saetung S
    3. Reutrakul S

    (2016) Higher sleep variability is associated with poorer glycaemic control in patients with type 1 diabetes

    Journal of Sleep Research 25:438–444.

    https://doi.org/10.1111/jsr.12393

    • PubMed
    • Google Scholar
12. Conference

    1. CIE

    (2019) CIE position statement on non-visual effects of light: recommending proper light at the proper time

    Human Technology Interaction.

    http://files.cie.co.at/783_CIE%20Statement%20-%20Proper%20Light%20at%20the%20Proper%20Time.pdf

    • Google Scholar
13. 1. Clayton JA
    2. Collins FS

    (2014) Policy: NIH to balance sex in cell and animal studies

    Nature 509:282–283.

    https://doi.org/10.1038/509282a

    • PubMed
    • Google Scholar
14. 1. Collins LM
    2. Dziak JJ
    3. Li R

    (2009) Design of experiments with multiple independent variables: a resource management perspective on complete and reduced factorial designs

    Psychological Methods 14:202–224.

    https://doi.org/10.1037/a0015826

    • PubMed
    • Google Scholar
15. Book

    1. Criado-Perez C

    (2019)

    Invisible women: Data bias in a world designed for men

    New York: Abrams Press.

    • Google Scholar
16. 1. Cristina V
    2. Mahachie J
    3. Mauer M
    4. Buclin T
    5. Van Cutsem E
    6. Roth A
    7. Wagner AD

    (2018) Association of patient sex with chemotherapy-related toxic effects: a retrospective analysis of the PETACC-3 trial conducted by the EORTC gastrointestinal group

    JAMA Oncology 4:1003–1006.

    https://doi.org/10.1001/jamaoncol.2018.1080

    • PubMed
    • Google Scholar
17. 1. Curtis MJ
    2. Alexander S
    3. Cirino G
    4. Docherty JR
    5. George CH
    6. Giembycz MA
    7. Hoyer D
    8. Insel PA
    9. Izzo AA
    10. Ji Y
    11. MacEwan DJ
    12. Sobey CG
    13. Stanford SC
    14. Teixeira MM
    15. Wonnacott S
    16. Ahluwalia A

    (2018) Experimental design and analysis and their reporting II: updated and simplified guidance for authors and peer reviewers

    British Journal of Pharmacology 175:987–993.

    https://doi.org/10.1111/bph.14153

    • PubMed
    • Google Scholar
18. 1. Dallmann R
    2. Okyar A
    3. Lévi F

    (2016) Dosing-time makes the poison: circadian regulation and pharmacotherapy

    Trends in Molecular Medicine 22:430–445.

    https://doi.org/10.1016/j.molmed.2016.03.004

    • PubMed
    • Google Scholar
19. 1. Dennis CL
    2. Dowswell T

    (2013) Interventions (other than pharmacological, psychosocial or psychological) for treating antenatal depression

    The Cochrane Database of Systematic Reviews 2013:CD006795.

    https://doi.org/10.1002/14651858.CD006795.pub3

    • PubMed
    • Google Scholar
20. 1. Dijk DJ
    2. Beersma DG
    3. Bloem GM

    (1989) Sex differences in the sleep EEG of young adults: visual scoring and spectral analysis

    Sleep 12:500–507.

    https://doi.org/10.1093/sleep/12.6.500

    • PubMed
    • Google Scholar
21. 1. Dijk DJ
    2. Duffy JF

    (2020) Novel approaches for assessing circadian rhythmicity in humans: a review

    Journal of Biological Rhythms 35:421–438.

    https://doi.org/10.1177/0748730420940483

    • PubMed
    • Google Scholar
22. 1. Dillon HR
    2. Lichstein KL
    3. Dautovich ND
    4. Taylor DJ
    5. Riedel BW
    6. Bush AJ

    (2015) Variability in self-reported normal sleep across the adult age span

    The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences 70:46–56.

    https://doi.org/10.1093/geronb/gbu035

    • PubMed
    • Google Scholar
23. 1. Docherty JR
    2. Stanford SC
    3. Panattieri RA
    4. Alexander SPH
    5. Cirino G
    6. George CH
    7. Hoyer D
    8. Izzo AA
    9. Ji Y
    10. Lilley E
    11. Sobey CG
    12. Stanley P
    13. Stefanska B
    14. Stephens G
    15. Teixeira M
    16. Ahluwalia A

    (2019) Sex: a change in our guidelines to authors to ensure that this is no longer an ignored experimental variable

    British Journal of Pharmacology 176:4081–4086.

    https://doi.org/10.1111/bph.14761

    • PubMed
    • Google Scholar
24. 1. Duffy JF
    2. Cain SW
    3. Chang AM
    4. Phillips AJK
    5. Münch MY
    6. Gronfier C
    7. Wyatt JK
    8. Dijk DJ
    9. Wright KP
    10. Czeisler CA

    (2011) Sex difference in the near-24-hour intrinsic period of the human circadian timing system

    PNAS 108 Suppl 3:15602–15608.

    https://doi.org/10.1073/pnas.1010666108

    • PubMed
    • Google Scholar
25. 1. Fischer D
    2. Vetter C
    3. Roenneberg T

    (2016) A novel method to visualise and quantify circadian misalignment

    Scientific Reports 6:38601.

    https://doi.org/10.1038/srep38601

    • PubMed
    • Google Scholar
26. 1. Fischer D
    2. Lombardi DA
    3. Marucci-Wellman H
    4. Roenneberg T

    (2017) Chronotypes in the US - Influence of age and sex

    PLOS ONE 12:e0178782.

    https://doi.org/10.1371/journal.pone.0178782

    • PubMed
    • Google Scholar
27. 1. Fischer D
    2. McHill AW
    3. Sano A
    4. Picard RW
    5. Barger LK
    6. Czeisler CA
    7. Klerman EB
    8. Phillips AJK

    (2020) Irregular sleep and event schedules are associated with poorer self-reported well-being in US college students

    Sleep 43:zsz300.

    https://doi.org/10.1093/sleep/zsz300

    • PubMed
    • Google Scholar
28. Website

    1. Fischer D
    2. Vidafar P

    (2020) Sex differences in circadian rhythms and sleep – Understanding difference

    Accessed June 15, 2020.

    https://www.youtube.com/watch?v=6sFAowzymJE
29. 1. Forbes D
    2. Blake CM
    3. Thiessen EJ
    4. Peacock S
    5. Hawranik P

    (2014) Light therapy for improving cognition, activities of daily living, sleep, challenging behaviour, and psychiatric disturbances in dementia

    The Cochrane Database of Systematic Reviews 2014:CD003946.

    https://doi.org/10.1002/14651858.CD003946.pub4

    • PubMed
    • Google Scholar
30. 1. Fustin J-M
    2. Doi M
    3. Yamada H
    4. Komatsu R
    5. Shimba S
    6. Okamura H

    (2012) Rhythmic nucleotide synthesis in the liver: temporal segregation of metabolites

    Cell Reports 1:341–349.

    https://doi.org/10.1016/j.celrep.2012.03.001

    • PubMed
    • Google Scholar
31. 1. Gandhi M
    2. Aweeka F
    3. Greenblatt RM
    4. Blaschke TF

    (2004) Sex differences in pharmacokinetics and pharmacodynamics

    Annual Review of Pharmacology and Toxicology 44:499–523.

    https://doi.org/10.1146/annurev.pharmtox.44.101802.121453

    • PubMed
    • Google Scholar
32. Report

    1. Gompers A

    (2018) Three years in: “sex as a biological variable” policy in practice - and an invitation to collaborate

    Genderscilab.

    https://www.genderscilab.org/blog/three-years-in-sex-as-a-biological-variable-policy-in-practice-and-an-invitation-to-collaborate

    • Google Scholar
33. Website

    1. International WELL Building Institute

    (2020) WELL Building Standard

    Accessed March 23, 2022.

    https://v2.wellcertified.com/en/wellv2/overview
34. 1. Karp NA
    2. Mason J
    3. Beaudet AL
    4. Benjamini Y
    5. Bower L
    6. Braun RE
    7. Brown SDM
    8. Chesler EJ
    9. Dickinson ME
    10. Flenniken AM
    11. Fuchs H
    12. Angelis M de
    13. Gao X
    14. Guo S
    15. Greenaway S
    16. Heller R
    17. Herault Y
    18. Justice MJ
    19. Kurbatova N
    20. Lelliott CJ
    21. Lloyd KCK
    22. Mallon AM
    23. Mank JE
    24. Masuya H
    25. McKerlie C
    26. Meehan TF
    27. Mott RF
    28. Murray SA
    29. Parkinson H
    30. Ramirez-Solis R
    31. Santos L
    32. Seavitt JR
    33. Smedley D
    34. Sorg T
    35. Speak AO
    36. Steel KP
    37. Svenson KL
    38. Wakana S
    39. West D
    40. Wells S
    41. Westerberg H
    42. Yaacoby S
    43. White JK
    44. International Mouse Phenotyping Consortium

    (2017) Prevalence of sexual dimorphism in mammalian phenotypic traits

    Nature Communications 8:15475.

    https://doi.org/10.1038/ncomms15475

    • PubMed
    • Google Scholar
35. 1. Karp NA
    2. Reavey N

    (2019) Sex bias in preclinical research and an exploration of how to change the status quo

    British Journal of Pharmacology 176:4107–4118.

    https://doi.org/10.1111/bph.14539

    • PubMed
    • Google Scholar
36. Website

    1. Karp P
    2. Ahluwalia A

    (2020) Sex differences in circadian rhythms and sleep – Understanding change

    Accessed June 29, 2020.

    https://www.youtube.com/watch?v=JV2NRCrEdUE
37. 1. Kaufmann CN
    2. Gershon A
    3. Eyler LT
    4. Depp CA

    (2016) Clinical significance of mobile health assessed sleep duration and variability in bipolar disorder

    Journal of Psychiatric Research 81:152–159.

    https://doi.org/10.1016/j.jpsychires.2016.07.008

    • PubMed
    • Google Scholar
38. 1. Keller LK
    2. Grünewald B
    3. Vetter C
    4. Roenneberg T
    5. Schulte-Körne G

    (2017) Not later, but longer: sleep, chronotype and light exposure in adolescents with remitted depression compared to healthy controls

    European Child & Adolescent Psychiatry 26:1233–1244.

    https://doi.org/10.1007/s00787-017-0977-z

    • PubMed
    • Google Scholar
39. 1. Kervezee L
    2. Shechter A
    3. Boivin DB

    (2018) Impact of shift work on the circadian timing system and health in women

    Sleep Medicine Clinics 13:295–306.

    https://doi.org/10.1016/j.jsmc.2018.04.003

    • PubMed
    • Google Scholar
40. 1. Lee SK

    (2018) Sex as an important biological variable in biomedical research

    BMB Reports 51:167–173.

    https://doi.org/10.5483/BMBRep.2018.51.4.034

    • PubMed
    • Google Scholar
41. 1. Lemola S
    2. Ledermann T
    3. Friedman EM

    (2013) Variability of sleep duration is related to subjective sleep quality and subjective well-being: an actigraphy study

    PLOS ONE 8:e71292.

    https://doi.org/10.1371/journal.pone.0071292

    • PubMed
    • Google Scholar
42. 1. Leung V
    2. Rousseau-Blass F
    3. Beauchamp G
    4. Pang DSJ

    (2018) ARRIVE has not ARRIVEd: Support for the ARRIVE (Animal Research: Reporting of in vivo Experiments) guidelines does not improve the reporting quality of papers in animal welfare, analgesia or anesthesia

    PLOS ONE 13:e0197882.

    https://doi.org/10.1371/journal.pone.0197882

    • PubMed
    • Google Scholar
43. 1. Lévi F
    2. Focan C
    3. Karaboué A
    4. de la Valette V
    5. Focan-Henrard D
    6. Baron B
    7. Kreutz F
    8. Giacchetti S

    (2007) Implications of circadian clocks for the rhythmic delivery of cancer therapeutics

    Advanced Drug Delivery Reviews 59:1015–1035.

    https://doi.org/10.1016/j.addr.2006.11.001

    • PubMed
    • Google Scholar
44. 1. Lévi F
    2. Okyar A
    3. Dulong S
    4. Innominato PF
    5. Clairambault J

    (2010) Circadian timing in cancer treatments

    Annual Review of Pharmacology and Toxicology 50:377–421.

    https://doi.org/10.1146/annurev.pharmtox.48.113006.094626

    • PubMed
    • Google Scholar
45. 1. Lok R
    2. Smolders K
    3. Beersma DGM
    4. de Kort YAW

    (2018) Light, alertness, and alerting effects of white light: a literature overview

    Journal of Biological Rhythms 33:589–601.

    https://doi.org/10.1177/0748730418796443

    • PubMed
    • Google Scholar
46. 1. Lunsford-Avery JR
    2. Engelhard MM
    3. Navar AM
    4. Kollins SH

    (2018) Validation of the sleep regularity index in older adults and associations with cardiometabolic risk

    Scientific Reports 8:14158.

    https://doi.org/10.1038/s41598-018-32402-5

    • PubMed
    • Google Scholar
47. 1. McBean AL
    2. Montgomery-Downs HE

    (2013) Timing and variability of postpartum sleep in relation to daytime performance

    Physiology & Behavior 122:134–139.

    https://doi.org/10.1016/j.physbeh.2013.09.003

    • PubMed
    • Google Scholar
48. 1. Mezick EJ
    2. Matthews KA
    3. Hall M
    4. Kamarck TW
    5. Buysse DJ
    6. Owens JF
    7. Reis SE

    (2009) Intra-individual variability in sleep duration and fragmentation: associations with stress

    Psychoneuroendocrinology 34:1346–1354.

    https://doi.org/10.1016/j.psyneuen.2009.04.005

    • PubMed
    • Google Scholar
49. 1. Milano G
    2. Etienne MC
    3. Cassuto-Viguier E
    4. Thyss A
    5. Santini J
    6. Frenay M
    7. Renee N
    8. Schneider M
    9. Demard F

    (1992) Influence of sex and age on fluorouracil clearance

    Journal of Clinical Oncology 10:1171–1175.

    https://doi.org/10.1200/JCO.1992.10.7.1171

    • PubMed
    • Google Scholar
50. 1. Minors D
    2. Atkinson G
    3. Bent N
    4. Rabbitt P
    5. Waterhouse J

    (1998) The effects of age upon some aspects of lifestyle and implications for studies on circadian rhythmicity

    Age and Ageing 27:67–72.

    https://doi.org/10.1093/ageing/27.1.67

    • PubMed
    • Google Scholar
51. 1. Montgomery P
    2. Dennis J

    (2002) Bright light therapy for sleep problems in adults aged 60+

    The Cochrane Database of Systematic Reviews 2002:CD003403.

    https://doi.org/10.1002/14651858.CD003403

    • PubMed
    • Google Scholar
52. Website

    1. National Institute for Health and Care Excellence

    (2020) Patient and public involvement policy

    Accessed March 14, 2022.

    https://www.nice.org.uk/about/nice-communities/nice-and-the-public/public-involvement/public-involvement-programme/patient-public-involvement-policy
53. Website

    1. National Institutes of Health

    (2020) NIH Policy on sex as a biological variable

    Accessed March 14, 2022.

    https://orwh.od.nih.gov/sex-gender/nih-policy-sex-biological-variable
54. 1. Okun ML
    2. Reynolds CF
    3. Buysse DJ
    4. Monk TH
    5. Mazumdar S
    6. Begley A
    7. Hall M

    (2011) Sleep variability, health-related practices, and inflammatory markers in a community dwelling sample of older adults

    Psychosomatic Medicine 73:142–150.

    https://doi.org/10.1097/PSY.0b013e3182020d08

    • PubMed
    • Google Scholar
55. 1. Pachito DV
    2. Eckeli AL
    3. Desouky AS
    4. Corbett MA
    5. Partonen T
    6. Rajaratnam SM
    7. Riera R

    (2018) Workplace lighting for improving alertness and mood in daytime workers

    The Cochrane Database of Systematic Reviews 3:CD012243.

    https://doi.org/10.1002/14651858.CD012243.pub2

    • PubMed
    • Google Scholar
56. 1. Patel SR
    2. Hayes AL
    3. Blackwell T
    4. Evans DS
    5. Ancoli-Israel S
    6. Wing YK
    7. Stone KL
    8. Osteoporotic Fractures in Men MrOS
    9. Study of Osteoporotic Fractures SOF Research Groups

    (2014) The association between sleep patterns and obesity in older adults

    International Journal of Obesity 38:1159–1164.

    https://doi.org/10.1038/ijo.2014.13

    • PubMed
    • Google Scholar
57. 1. Petit E
    2. Milano G
    3. Lévi F
    4. Thyss A
    5. Bailleul F
    6. Schneider M

    (1988)

    Circadian rhythm-varying plasma concentration of 5-fluorouracil during a five-day continuous venous infusion at a constant rate in cancer patients

    Cancer Research 48:1676–1679.

    • PubMed
    • Google Scholar
58. 1. Phillips AJK
    2. Clerx WM
    3. O’Brien CS
    4. Sano A
    5. Barger LK
    6. Picard RW
    7. Lockley SW
    8. Klerman EB
    9. Czeisler CA

    (2017) Irregular sleep/wake patterns are associated with poorer academic performance and delayed circadian and sleep/wake timing

    Scientific Reports 7:3216.

    https://doi.org/10.1038/s41598-017-03171-4

    • PubMed
    • Google Scholar
59. 1. Prendergast BJ
    2. Onishi KG
    3. Zucker I

    (2014) Female mice liberated for inclusion in neuroscience and biomedical research

    Neuroscience and Biobehavioral Reviews 40:1–5.

    https://doi.org/10.1016/j.neubiorev.2014.01.001

    • PubMed
    • Google Scholar
60. 1. Rippon G
    2. Jordan-Young R
    3. Kaiser A
    4. Joel D
    5. Fine C

    (2017) Journal of neuroscience research policy on addressing sex as a biological variable: comments, clarifications, and elaborations

    Journal of Neuroscience Research 95:1357–1359.

    https://doi.org/10.1002/jnr.24045

    • PubMed
    • Google Scholar
61. 1. Roane BM
    2. Seifer R
    3. Sharkey KM
    4. Van Reen E
    5. Bond TLY
    6. Raffray T
    7. Carskadon MA

    (2015) What role does sleep play in weight gain in the first semester of university?

    Behavioral Sleep Medicine 13:491–505.

    https://doi.org/10.1080/15402002.2014.940109

    • PubMed
    • Google Scholar
62. 1. Roenneberg T
    2. Kuehnle T
    3. Pramstaller PP
    4. Ricken J
    5. Havel M
    6. Guth A
    7. Merrow M

    (2004) A marker for the end of adolescence

    Current Biology 14:R1038–R1039.

    https://doi.org/10.1016/j.cub.2004.11.039

    • PubMed
    • Google Scholar
63. Report

    1. Safe Work Australia

    (2009) Australian work-related injury experience by sex and age, 2009-2010

    Safe Work Australia.

    https://www.safeworkaustralia.gov.au/system/files/documents/1702/australian_work-related_injury_experience_by_sex_and_age_2009-10.pdf

    • Google Scholar
64. 1. Shuboni-Mulligan DD
    2. Breton G
    3. Smart D
    4. Gilbert M
    5. Armstrong TS

    (2019) Radiation chronotherapy-clinical impact of treatment time-of-day: a systematic review

    Journal of Neuro-Oncology 145:415–427.

    https://doi.org/10.1007/s11060-019-03332-7

    • PubMed
    • Google Scholar
65. 1. Slanger TE
    2. Gross JV
    3. Pinger A
    4. Morfeld P
    5. Bellinger M
    6. Duhme AL
    7. Reichardt Ortega RA
    8. Costa G
    9. Driscoll TR
    10. Foster RG
    11. Fritschi L
    12. Sallinen M
    13. Liira J
    14. Erren TC

    (2016) Person-directed, non-pharmacological interventions for sleepiness at work and sleep disturbances caused by shift work

    The Cochrane Database of Systematic Reviews 2016:CD010641.

    https://doi.org/10.1002/14651858.CD010641.pub2

    • PubMed
    • Google Scholar
66. 1. Smolensky MH
    2. Hermida RC
    3. Ayala DE
    4. Mojón A
    5. Fernández JR

    (2017) Bedtime chronotherapy with conventional hypertension medications to target increased asleep blood pressure results in markedly better chronoprevention of cardiovascular and other risks than customary on-awakening therapy

    Heart Failure Clinics 13:775–792.

    https://doi.org/10.1016/j.hfc.2017.05.011

    • PubMed
    • Google Scholar
67. 1. Souman JL
    2. Tinga AM
    3. Te Pas SF
    4. van Ee R
    5. Vlaskamp BNS

    (2018) Acute alerting effects of light: A systematic literature review

    Behavioural Brain Research 337:228–239.

    https://doi.org/10.1016/j.bbr.2017.09.016

    • PubMed
    • Google Scholar
68. 1. Spitschan M
    2. Schmidt MH
    3. Blume C

    (2020) Transparency and open science principles in reporting guidelines in sleep research and chronobiology journals

    Wellcome Open Research 5:172.

    https://doi.org/10.12688/wellcomeopenres.16111.1

    • PubMed
    • Google Scholar
69. 1. Spruyt K
    2. Molfese DL
    3. Gozal D

    (2011) Sleep duration, sleep regularity, body weight, and metabolic homeostasis in school-aged children

    Pediatrics 127:e345–e352.

    https://doi.org/10.1542/peds.2010-0497

    • PubMed
    • Google Scholar
70. 1. Stein BN
    2. Petrelli NJ
    3. Douglass HO
    4. Driscoll DL
    5. Arcangeli G
    6. Meropol NJ

    (1995) Age and sex are independent predictors of 5-fluorouracil toxicity. Analysis of a large scale phase III trial

    Cancer 75:11–17.

    https://doi.org/10.1002/1097-0142(19950101)75:13.0.co;2-n

    • Google Scholar
71. 1. Tannenbaum C
    2. Ellis RP
    3. Eyssel F
    4. Zou J
    5. Schiebinger L

    (2019) Sex and gender analysis improves science and engineering

    Nature 575:137–146.

    https://doi.org/10.1038/s41586-019-1657-6

    • PubMed
    • Google Scholar
72. 1. Tuunainen A
    2. Kripke DF
    3. Endo T

    (2004) Light therapy for non-seasonal depression

    The Cochrane Database of Systematic Reviews 2004:CD004050.

    https://doi.org/10.1002/14651858.CD004050.pub2

    • PubMed
    • Google Scholar
73. 1. Vanderlind WM
    2. Beevers CG
    3. Sherman SM
    4. Trujillo LT
    5. McGeary JE
    6. Matthews MD
    7. Maddox WT
    8. Schnyer DM

    (2014) Sleep and sadness: exploring the relation among sleep, cognitive control, and depressive symptoms in young adults

    Sleep Medicine 15:144–149.

    https://doi.org/10.1016/j.sleep.2013.10.006

    • PubMed
    • Google Scholar
74. 1. Vidafar P
    2. Gooley JJ
    3. Burns AC
    4. Rajaratnam SMW
    5. Rueger M
    6. Van Reen E
    7. Czeisler CA
    8. Lockley SW
    9. Cain SW

    (2018) Increased vulnerability to attentional failure during acute sleep deprivation in women depends on menstrual phase

    Sleep 41:zsy098.

    https://doi.org/10.1093/sleep/zsy098

    • PubMed
    • Google Scholar
75. Website

    1. White R
    2. Lévi F

    (2020) Sex differences in circadian rhythms and sleep – Understanding impact

    Accessed June 22, 2020.

    https://www.youtube.com/watch?v=mFkHDz63ktU
76. 1. Xu X
    2. Conomos MP
    3. Manor O
    4. Rohwer JE
    5. Magis AT
    6. Lovejoy JC

    (2018) Habitual sleep duration and sleep duration variation are independently associated with body mass index

    International Journal of Obesity 42:794–800.

    https://doi.org/10.1038/ijo.2017.223

    • PubMed
    • Google Scholar
77. 1. Yetish G
    2. Kaplan H
    3. Gurven M

    (2018) Sleep variability and nighttime activity among Tsimane forager-horticulturalists

    American Journal of Physical Anthropology 166:590–600.

    https://doi.org/10.1002/ajpa.23454

    • PubMed
    • Google Scholar
78. 1. Yoon DY
    2. Mansukhani NA
    3. Stubbs VC
    4. Helenowski IB
    5. Woodruff TK
    6. Kibbe MR

    (2014) Sex bias exists in basic science and translational surgical research

    Surgery 156:508–516.

    https://doi.org/10.1016/j.surg.2014.07.001

    • PubMed
    • Google Scholar

Decision letter after peer review:

Thank you for submitting your article "Sex differences and sex bias in circadian rhythms and sleep research" to eLife for consideration as a Feature Article. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Peter Rodgers as the Senior Editor.

The reviewers and editors have discussed the reviews and we have drafted this decision letter to help you prepare a revised submission.

Please also note that eLife does not permit figures in supplementary materials: instead we allow primary figures to have figure supplements (tables, however, can be supplementary materials).

After revision, please note that the eLife Features Editor may also contact you separately about some editorial issues that you will need to address.

Summary:

This article highlights the sex-bias present in chronobiology and sleep medicine research, and it provides recommendations and guidance for sex- and gender-inclusive research practices and policies. To do so, it relies on insight from a workshop conducted to evaluate sex differences in circadian and sleep research. Overall, the authors have provided robust examples from a growing body of evidence which suggests that biological sex influences the human circadian clock and that these sex differences may play a key role in health and disease. However, a number of points need to be addressed to make the article suitable for publication.

Essential revisions:

1. Please shorten the introduction, and reorganize the first two sections "Sex differences are ubiquitous in circadian rhythms and sleep" and "Sex differences in circadian rhythms can impact therapeutic interventions" into one cohesive introduction section which lays the ground for the literature analysis and the workshop discussion. In doing so, please make sure to streamline the text to ensure the reader feels it was written cohesively rather than by different people.

2. Lines 61-198 have a large overlap of topic matter with lines 200-266: please try to reduce this repetition. (For instance, the sub-section "What is the sex data gap?" repeats the conclusions from the previous section).

3. Line 73: Please discuss this point further, as "blockers" are not the only people who may resist change – they are probably a tiny minority of people. Far more people resisting change are probably people who don't know enough about how or why.

4. Line 217 onward: Please discuss why there appears to have been a thorough review of studies involving light impact on human circadian rhythms, but not of other types of studies that are only mentioned briefly (e.g., circadian period or phase)

5. Line 227 onwards: Please clarify whether there was a relationship between the proportion of female participants and the sample size of the study

6. Line 268 onwards: Please provide further information about the workshop participants and how they contributed to the outcomes of the paper.

– For instance, please clarify the methods used to obtain participant reflections/responses (i.e. Did you provide participants with a certain set of questions or surveys to obtain the information described in Workshops 1 and 2? Alternatively, did this information come from discussions which were transcribed and coded?)

– Please indicate whether the participants of the workshop gave consent for their responses/comments to be used in this format? If not, please discuss whether there is a need for a retrospective-IRB approval.

– Please report how many individuals attended each workshop

– For each workshop section, please give all answers, not a selection of them. You can group into categories or ones with similar wording. The number of each question is not needed since there was limited time to respond and participants may not have been able to include all possibilities (so the number of each response does not necessarily reflect the importance of said response).

7. Line 282: Please further discuss the statement that a common misconception in research is that the sample size must be doubled to include both sexes. The authors note that this is a flawed assumption because treatment effects are reported from the entire group. While this is certainly true, the authors should also acknowledge that the issue may be more complex than described, especially because sex differences exist. In the first half of this manuscript, the authors describe important differences that have been identified between sexes on a variety of outcomes (shiftwork tolerance, sleep differences, phase differences, etc.). Doesn't this suggest that pooling of males and females a bad idea, particularly for time-varying differences? For example, on lines 188-192, the authors describe the results of several important studies examining oxaliplatin and irinotecan in male and female patients. The authors note that the optimal timing for females and males were at different times for these treatments. If these reports would have pooled the data for males and females wouldn't that have attenuated the results or obscured the researchers' ability to observe a time-of-day effect at all? In this case, it seems pretty important that the analyses were stratified by sex. If a researcher wishes to examine outcomes by sex, as seems pretty important given the thesis of this manuscript, then one should be powered to conduct separate analyses (Diester CM, Banks ML, Neigh GN, Negus SS. Experimental design and analysis for consideration of sex as a biological variable. Neuropsychopharmacology. 2019;44(13):2159-2162. doi:10.1038/s41386-019-0458-9).

8. Line 343 onwards (Recommendations #2):

– Please discuss further whether 1:1 female:male is the only valid ratio. Would other ratios be ok in some circumstances? When is this ratio more important than other variables (e.g. age)?

– Please acknowledge that recruiting a sample that consists of 50% females and 50% males may increase variability (in some circumstances) compared to studies with 100% of one sex, and therefore change power calculations and/or require larger sample size.

9. Recommendations: Please consider discussing whether one option for closing the gap would be to fund/conduct studies that only include females.

10. Line 315: Please provide a copy of the workshop program booklet or guide as supplemental information, so that "the way [you] constructed these workshops and facilitated discussion around sex could be used as a template for exploring other aspects of difference in future as research progresses". The reviewer was unable to find this information based on the references provided [75-77].

11. Line 433 onwards (Methods): Please report whether one or more authors conducted the bibliometric analyses. If multiple authors contributed, please indicate whether there were any inter-rater reliability tests done to assess accurate coding.

12. Please provide the source data for the bibliometric analysis, which should contain citation data and sex-based coding as presented in Figure 1.

Essential revisions:

1. Please shorten the introduction and reorganize the first two sections "Sex differences are ubiquitous in circadian rhythms and sleep" and "Sex differences in circadian rhythms can impact therapeutic interventions" into one cohesive introduction section which lays the ground for the literature analysis and the workshop discussion. In doing so, please make sure to streamline the text to ensure the reader feels it was written cohesively rather than by different people.

We have now shortened the introduction and combined the two sections, thereby streamlining the flow and harmonising style.

2. Lines 61-198 have a large overlap of topic matter with lines 200-266: please try to reduce this repetition. (For instance, the sub-section "What is the sex data gap?" repeats the conclusions from the previous section).

We have now revised this section and removed repetitions.

3. Line 73: Please discuss this point further, as "blockers" are not the only people who may resist change – they are probably a tiny minority of people. Far more people resisting change are probably people who don't know enough about how or why.

The term blockers was used in reference 7 (Karp et al., 2017) in a change management reflection on the resistance to studying both sexes. So the term blocker doesn’t refer to people but encompasses a myriad of practical, cultural and other issues. We have rewritten the paragraph to improve clarity:

“Sex bias or the 'sex data gap' has recently gained widespread attention (Criado-Perez, 2019), with policy advisers (Buitendijk and Maes, 2015), funders (Lee, 2018; Clayton and Collins, 2014) and publishers (Rippon et al., 2017; Docherty et al., 2019) pushing for the inclusion of both sexes in research. […] This two-pronged approach will unfreeze the status quo and allow new approaches to emerge (Karp and Reavey, 2019).”

4. Line 217 onward: Please discuss why there appears to have been a thorough review of studies involving light impact on human circadian rhythms, but not of other types of studies that are only mentioned briefly (e.g., circadian period or phase)

We use the area of non-visual effects of light as a very tangible example in this article. The area is very topical. We have now clarified this in the text:

“Because there are various efforts underway to incorporate scientific data from this research area into guidelines and recommendations for the provision of lighting in the built environment, there is a pressing need to understand sex bias in this area, which we have prioritised in the analysis reported here.”

5. Line 227 onwards: Please clarify whether there was a relationship between the proportion of female participants and the sample size of the study

We have included this additional analysis. We have added the following sentence:

“Interestingly, we also find that the total sample size correlates with the fraction of female participants in a given study (r(164) = 0.23, p = 0.00275).”

6. Line 268 onwards: Please provide further information about the workshop participants and how they contributed to the outcomes of the paper.

– For instance, please clarify the methods used to obtain participant reflections/responses (i.e. Did you provide participants with a certain set of questions or surveys to obtain the information described in Workshops 1 and 2? Alternatively, did this information come from discussions which were transcribed and coded?)

– Please indicate whether the participants of the workshop gave consent for their responses/comments to be used in this format? If not, please discuss whether there is a need for a retrospective-IRB approval.

– Please report how many individuals attended each workshop

– For each workshop section, please give all answers, not a selection of them. You can group into categories or ones with similar wording. The number of each question is not needed since there was limited time to respond and participants may not have been able to include all possibilities (so the number of each response does not necessarily reflect the importance of said response).

We have expanded upon the description of the workshops in both the workshop sections and have added a methods section describing these aspects. All workshop materials have been made available on the Open Science Framework (https://doi.org/10.17605/OSF.IO/WU2QX). We have limited the discussion of findings to key results from the interactive sessions.

We have added information regarding participant consent in the methods section:

“Prior to participating in the interactive sessions, attendees were informed that their responses will be used for write-up and published as a peer-reviewed article. […] No personal data were collected as part of the interactive Mentimeter sessions.”

We have added information about attendance numbers in the main section:

“A total of 275 individuals registered for the workshop. Across the three workshops, between 38 and 94 attendees participated in the interactive sessions, with approximately 4/5 of all participants being researchers (82/94 in Workshop 1, 47/60 in Workshop 2, 31/38 in Workshop 3). […] Responses to individual questions varied somewhat due to drop-out during the interactive session as well as a time-limited response window; the total number of responses in the participatory parts are given on the bottom right-hand corner of the materials document.”

We have added information on the question methodology as well:

“Throughout the interactive sessions, we combined yes/no, ranking and open-ended questions to vary the response modality. The results discussed below were selected from the results, which can be viewed in full on the Open Science Framework page above.”

7. Line 282: Please further discuss the statement that a common misconception in research is that the sample size must be doubled to include both sexes. The authors note that this is a flawed assumption because treatment effects are reported from the entire group. While this is certainly true, the authors should also acknowledge that the issue may be more complex than described, especially because sex differences exist. In the first half of this manuscript, the authors describe important differences that have been identified between sexes on a variety of outcomes (shiftwork tolerance, sleep differences, phase differences, etc.). Doesn't this suggest that pooling of males and females a bad idea, particularly for time-varying differences? For example, on lines 188-192, the authors describe the results of several important studies examining oxaliplatin and irinotecan in male and female patients. The authors note that the optimal timing for females and males were at different times for these treatments. If these reports would have pooled the data for males and females wouldn't that have attenuated the results or obscured the researchers' ability to observe a time-of-day effect at all? In this case, it seems pretty important that the analyses were stratified by sex. If a researcher wishes to examine outcomes by sex, as seems pretty important given the thesis of this manuscript, then one should be powered to conduct separate analyses (Diester CM, Banks ML, Neigh GN, Negus SS. Experimental design and analysis for consideration of sex as a biological variable. Neuropsychopharmacology. 2019;44(13):2159-2162. doi:10.1038/s41386-019-0458-9).

There are ways in which sex can impact an outcome variable of interest. Firstly, the variables baseline levels can depend on sex. Secondly, the treatment effect can depend on sex. When studying both sexes, a regression analysis will account for the sex differences in baseline levels and this has minimal impact on power. Regression analysis is not pooling the data which is completely inappropriate and can deliver misleading results. Recommending half the animals are male and half are female for a treatment group, is based on the premise that you want to estimate an average effect and then you can see a large interaction if it occurs. If we are interested in these differences, we then need the power to study the interaction which is a different goal.

In typical time course studies, where the time effect is also of interest, a regression analysis would ask at each time point what is the effect of sex, what is the effect of the treatment and does the effect depend on sex.

The section on page 8 has been reworded to give greater clarity and to highlight that pooling is not recommended.

8. Line 343 onwards (Recommendations #2):

– Please discuss further whether 1:1 female:male is the only valid ratio.

Statistically it is good practice to have balanced designs [ref: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2796056/]. ‘In a balanced design the main effects and interactions are orthogonal so that each one is estimated and tested as if it were the only one under consideration, with very little loss of efficiency due to the presence of other factors’. Basically, having a balanced design makes analysis more robust and ensures that factors are independent and the variance can be decomposed in to the individual contributions without confounding. Consequently, this is our recommendation as the default basic design. The following sentence has been added to page 11 “A balanced design is recommended to ensure the resulting statistical analysis is robust and that the variance can be decomposed to the factors of interest without confounding.”

Would other ratios be ok in some circumstances? When is this ratio more important than other variables (e.g. age?)

We have added a paragraph to page 9:

“A common push back is that there are other sources of variation (e.g. age) that should be considered and ask why should sex be the variable that is prioritized. Whenever we conduct an experiment we are simplifying a complex biological world with lots of sources of variation into a testing space and we generalize the findings allowing us to make conclusions. When considering biomedical research, on average the target population will be 50% male and 50% female and we are now understanding that the differences between male and female are significant and stretch beyond hormonal differences. Therefore, as a general rule, including both sexes should be the first variable to be included to significant increase generalizability. However, there may be exceptions, such as the study of sex specific conditions or phenomena, as written into the NIH guidelines.”

– Please acknowledge that recruiting a sample that consists of 50% females and 50% males may increase variability (in some circumstances) compared to studies with 100% of one sex, and therefore change power calculations and/or require larger sample size.

The data is not pooled and hence the sex differences are accounted for. With a regression analysis, when the treatment effect is the same or in the same direction then there is no loss in power to estimating the average effect. We gain then that the conclusions space is larger. When the treatment effect is crossed, then we do lose power on estimating the average effect but we gain in ability to test for an interaction. However, in this situation if you had studied one sex you would have missed a major differences in the understanding of the outcome. We have reworded the section on page 8 to improve the clarity of this discussion.

9. Recommendations: Please consider discussing whether one option for closing the gap would be to fund/conduct studies that only include females.

We do not think that the sex data gap can be closed if data are collected in a sex-selective fashion. If generalizable data are to be collected, then biasing the sample in the opposite direction is not helpful. If the suggestion is to fund data collection including only females post-hoc, then it is not clear what justifies data collection of the first sample with only males. The only way to resolve this is to collect data on males and females from the get-go.

10. Line 315: Please provide a copy of the workshop program booklet or guide as supplemental information, so that "the way [you] constructed these workshops and facilitated discussion around sex could be used as a template for exploring other aspects of difference in future as research progresses". The reviewer was unable to find this information based on the references provided [75-77].

We have now included a copy of the workshop program in the supplemental information on the Open Science Framework page. We have also expanded our description and links in the text:

“In a three-part virtual workshop (June 2020). we explored practices, barriers, and challenges in designing and executing inclusive research in circadian physiology and sleep research. […] Across the three workshops, between 38 and 94 attendees participated in the interactive sessions, with approximately 4/5 of all participants being researchers (82/94 in Workshop 1, 47/60 in Workshop 2, 31/38 in Workshop 3).”

In the workshop descriptions, to make it easier for the reader, we now explicitly write:

“The video recording for Workshop 1 is available at https://osf.io/4mk9g/. The materials related to the participatory part are available at https://osf.io/y3ha2/.”

“The video recording for Workshop 2 is available at https://osf.io/ep6uh/. The materials related to the participatory part are available at https://osf.io/d32qr/.”

“The video recording for Workshop 3 is available at https://osf.io/rtaqu/. The materials related to the participatory part are available at https://osf.io/ntjbv/.”

11. Line 433 onwards (Methods): Please report whether one or more authors conducted the bibliometric analyses. If multiple authors contributed, please indicate whether there were any inter-rater reliability tests done to assess accurate coding.

We have clarified that only one reviewer reviewed the articles.

12. Please provide the source data for the bibliometric analysis, which should contain citation data and sex-based coding as presented in Figure 1.

Source data are included in the supplementary material.

Author details

1. Manuel Spitschan

   Manuel Spitschan is in the Department of Sport and Health Sciences, Technical University of Munich, Munich, Germany, the Translational Sensory and Circadian Neuroscience Research Group, Max Planck Institute for Biological Cybernetics, Tübingen, Germany and the Department of Experimental Psychology, University of Oxford, Oxford, United Kingdom

   Contribution

   Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Visualization, Writing – original draft, Writing – review and editing

   For correspondence

   manuel.spitschan@tum.de

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8572-9268

2. Nayantara Santhi

   Nayantara Santhi is in the Department of Psychology, Northumbria University, Newcastle-upon-Tyne, United Kingdom

   Contribution

   Conceptualization, Writing – original draft, Writing – review and editing

   For correspondence

   nayantara.santhi@northumbria.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4568-1447

3. Amrita Ahluwalia

   Amrita Ahluwalia is at the William Harvey Research Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, United Kingdom

   Contribution

   Authors after the first two authors are listed in alphabetical order, Writing – original draft, Writing – review and editing

   Competing interests

   Editor-in-Chief for the British Journal of Pharmacology

   Additional information

   Authors after the first two authors are listed in alphabetical order

   "This ORCID iD identifies the author of this article:" 0000-0001-7626-6399

4. Dorothee Fischer

   Dorothee Fischer is at the German Aerospace Center, Institute of Aerospace Medicine, Sleep and Human Factors Research, Köln, Germany

   Contribution

   Authors after the first two authors are listed in alphabetical order, Investigation, Methodology, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   Additional information

   Authors after the first two authors are listed in alphabetical order

   "This ORCID iD identifies the author of this article:" 0000-0002-2122-3938

5. Lilian Hunt

   Lilian Hunt is at the Wellcome Trust, London, United Kingdom and in the Equality, Diversity and Inclusion in Science and Health Group, London, United Kingdom

   Contribution

   Authors after the first two authors are listed in alphabetical order, Conceptualization, Resources, Writing – original draft, Writing – review and editing

   Competing interests

   employee of the Wellcome Trust. Her work is unrelated to the Wellcome Trust's Research Enrichment funding mechanisms that funded the workshop

   Additional information

   Authors after the first two authors are listed in alphabetical order

   "This ORCID iD identifies the author of this article:" 0000-0002-0319-7764

6. Natasha A Karp

   Natasha A Karp is in the Data Sciences and Quantitative Biology division, Discovery Science at AstraZeneca R&D, Cambridge, United Kingdom

   Contribution

   Authors after the first two authors are listed in alphabetical order, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   Additional information

   Authors after the first two authors are listed in alphabetical order

   "This ORCID iD identifies the author of this article:" 0000-0002-8404-2907

7. Francis Lévi

   Francis Lévi in at the Warwick Medical School, University of Warwick, Warwick, United Kingdom, the Hepatobiliary Center, Hospital Paul Brousse (AP-HP), Villejuif, France and the UPR Chronotherapy, Cancer and Transplantation, Medical School, Paris-Saclay University, Gif-sur-Yvette, France

   Contribution

   Authors after the first two authors are listed in alphabetical order, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   Additional information

   Authors after the first two authors are listed in alphabetical order

   "This ORCID iD identifies the author of this article:" 0000-0003-1364-7463

8. Inés Pineda-Torra

   Inés Pineda-Torra is in the Centre for Cardiometabolic and Vascular Science, Division of Medicine, University College London, London, United Kingdom

   Contribution

   Authors after the first two authors are listed in alphabetical order, Conceptualization, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   Additional information

   Authors after the first two authors are listed in alphabetical order

   "This ORCID iD identifies the author of this article:" 0000-0002-7349-2208

9. Parisa Vidafar

   Parisa Vidafar is in the Sleep and Circadian Research Laboratory, Department of Psychiatry, University of Michigan, Ann Arbor, United States, and the School of Psychological Sciences and Turner Institute for Brain and Mental Health, Monash University, Clayton, Australia

   Contribution

   Authors after the first two authors are listed in alphabetical order, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   Additional information

   Authors after the first two authors are listed in alphabetical order

   "This ORCID iD identifies the author of this article:" 0000-0002-3990-1047

10. Rhiannon White

    Rhiannon White is in the Department of Experimental Psychology, University of Oxford, Oxford, and the Warwick Medical School, University of Warwick, Warwick, United Kingdom

    Contribution

    Authors after the first two authors are listed in alphabetical order, Investigation, Methodology, Visualization, Writing – original draft, Writing – review and editing

    Competing interests

    No competing interests declared

    Additional information

    Authors after the first two authors are listed in alphabetical order

    "This ORCID iD identifies the author of this article:" 0000-0002-8175-3586

• 1,378

  views

• 198

  downloads

• 13

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.