Does global health governance walk the talk? Gender representation in World Health Assemblies, 1948–2021

Background While an estimated 70%–75% of the health workforce are women, this is not reflected in the leadership roles of most health organisations—including global decision-making bodies such as the World Health Assembly (WHA). Methods We analysed gender representation in WHA delegations of Member States, Associate Members and Observers (country/territory), using data from 10 944 WHA delegations and 75 815 delegation members over 1948–2021. Delegates’ information was extracted from WHO documentation. Likely gender was inferred based on prefixes, pronouns and other gendered language. A gender-to-name algorithm was used as a last resort (4.6%). Time series of 5-year rolling averages of the percentage of women across WHO region, income group and delegate roles are presented. We estimated (%) change ±SE of inferred women delegation members at the WHA per year, and estimated years±SE until gender parity from 2010 to 2019 across regions, income groups, delegate roles and countries. Correlations with these measures were assessed with countries’ gender inequality index and two Worldwide Governance indicators. Results While upwards trends could be observed in the percentage of women delegates over the past 74 years, men remained over-represented in most WHA delegations. Over 1948–2021, 82.9% of delegations were composed of a majority of men, and no WHA had more than 30% of women Chief Delegates (ranging from 0% to 30%). Wide variation in trends over time could be observed across different geographical regions, income groups and countries. Some countries may take over 100 years to reach gender parity in their WHA delegations, if current estimated trends continue. Conclusion Despite commitments to gender equality in leadership, women remain gravely under-represented in global health governance. An intersectional approach to representation in global health governance, which prioritises equity in participation beyond gender, can enable transformative policymaking that fosters transparent, accountable and just health systems.


INTRODUCTION
Global and national health leadership continues to be dominated by men. While an estimated 70% of health workers are women, this percentage is not reflected in higherwage healthcare occupations, nor the leadership roles of most international and national health organisations. 1 This has continued to be illustrated during the COVID-19 pandemic; a study on gender representation in national COVID-19 task forces, revealed that only 3.5% of 115 identified COVID-19 decision-making and expert task forces had gender parity, and 85.2% were majority men. 2 Additionally, while the economic contribution of women in global health is valued at US$3 trillion annually, half of women's contribution is in the form of unpaid care work. The pay gap between men and women in healthcare remains around 28%. 1 Women from low-income and middle-income countries are particularly under-represented in global health governance, holding less than 5% of senior leadership roles. 3 Only 20% of global health institutions exhibit gender parity on their board of directors, and a mere 25% show gender parity at the senior management level. The current situation in global health leadership is reflected in broader national government positions. Over the last 5 years (as of April 2022), only 14% of countries had a woman as head of government, 21% as head of state and only 42% as Ministers of Health (own analysis).
Collective global efforts have led to greater advocacy for policies that have aimed to increase the status of women and girls-including their participation in political processes and governance. Yet, the overall BMJ Global Health under-representation of half of the world's population in positions of leadership is still displayed in global decisionmaking bodies such as the World Health Assembly (WHA). 4 The WHA is the central decision-making body of the WHO, the lead normative and technical actor within the global health sector, where priorities and agendas are set for the global community by delegations representing each Member State. In 2017 and 2018 WHA delegations reached a peak at 30% of Chief Delegates (head of delegation) being inferred as women. Yet in years to follow, progress has stalled, with merely 24% of delegations headed by a woman in 2019, 22% in 2020 and 24% in 2021 (own analysis).
Global health actors are increasingly aware of the unaddressed lack of gender diversity within global health governance. In 2020, the Gender Equal Health and Care Workforce Initiative was launched by WHO, the French government and Women in Global Health. 5 6 As part of its commitments during the Generation Equality Forum in 2021, the WHO pledged to 'promote and encourage gender parity in WHA delegations, WHO panels and advisory groups'. 5 While progress to increase women's representation in positions of leadership has been made across global health governance since the inception of the WHA in 1948, further action by Member States is likely needed to achieve gender equitable representation of WHA delegations. Here, we present a full analysis of gender representation of Member State, Associate Member and Observer (country/territory) WHA delegations over the past 74 years . This longitudinal, descriptive analysis serves to identify patterns of progress and/or stagnation across regions which can help orient priorities for action.

BMJ Global Health
lists of delegates (see online supplemental table 2). In case a gendered prefix was not available (eg, Dr, Prof.), gendered pronouns (eg, she/he/they) or other gendered language (eg, 'husband') used in WHO documentation or publicly available online documentation (eg, government websites, online biographies) was used to infer likely gender (eg, woman, man, non-binary) by 14 authors. If the full name was available or found through online searches, but no gendered prefixes, pronouns or other gendered language could be retrieved, a genderto-name algorithm (https://genderize.io/) was used based on historical databases combining first name and country (n=3274, 4.6% of the inferred genders of delegation members). This tool has been applied and checked for robustness in multiple previous studies. [8][9][10] The algorithm's inferred gender was only accepted when the probabilistic certainty score was ≥0.50. If a likely gender could not be inferred after this approach, it was classified as 'unknown' (n=4383, 5.8% of total delegation members). Due to the inability of gender-to-name algorithms to identify people outside the gender binary and their reduced quality for inferring gender for non-Western names, this option functioned as a last resort. Gender was not inferred based on gender expression/ presentation (phenotype) in images/photos from delegation members, due to the subjectivity of this method. As few people were inferred to be non-binary (n=2, 0.003% of delegation members), they were included as a gender minority in the categorisation of 'women' for the purpose of this analysis.

Data cleaning and coding
We assigned current Member States to their corresponding WHO region group (2022) (Africa, Americas, Eastern Mediterranean, Europe, South-East Asia, Western Pacific), 11  to participate in selecting their government, freedom of expression, freedom of association and free media 15 16 ') and Government Effectiveness indicator ('a reflection of the perceived public services quality, civil service quality and degree of independence from political pressure, policy formulation and implementation quality, and the credibility of government's commitment to policies 15 16 ').
Estimates of governance performance on these indicators ranges from 2.5 (strong) to −2.5 (weak). 15 16 Countries that changed their name (but not geographical boundaries) over the past 74 years have been re-coded to their current (2022) name (eg, Swaziland to Eswatini, Burma to Myanmar) to enable longitudinal analysis. Countries that have changed their geographical boundaries and/or geopolitical context have not been re-coded (eg, Yugoslavia, Ruanda-Urundi). Online supplemental table 3 displays the re-coded and not re-coded (former) countries, territories and political parties, including their relevant geopolitical contexts. To include these (former) countries, territories and political parties in regional longitudinal analyses, we grouped them under the geographical 'UN', 'WHO' and 'WB' regions that they would theoretically fall in based on their geographical location (eg, Yugoslavia was categorised as 'Europe', 'Eastern Europe' and 'Europe and Central Asia' for the 'WHO', 'UN' and 'WB' regions, respectively) as seen in online supplemental table 4. The WB income group and GII were not extended, as this would have required longitudinal data reflecting the countries change in income group and GII over 1948-2021.

Data analysis and visualisation
We present backward 5-year rolling averages for the percentage (%) of women across WHO, UN, WB region and income groups and across delegates' roles (eg, Chief Delegate) to generate time series over 1948-2021. Inferred gender composition of each delegation was further categorised into majority women (>55% women), gender parity (45%-55%) and majority men (>55% men)-this was presented as total number of delegations and percentage of delegations with majority women, gender parity and majority men over time .
Binomial 95% CIs were calculated for the proportions of interest (per cent of women). To estimate the per cent of change in women per year ±SE, we first aggregated the data over intervals of 10 years (backwards), and calculated the overall per cent of women over these intervals. This was then used to fit a linear regression model and to estimate the number of years until gender parity with 2010-2019 as baseline. The (i) estimated proportion ±95% CI of inferred women delegation members at the WHA in 2019, (ii) estimated change (%)±SE of inferred women delegation members at the WHA per year and (iii) estimated years±SE until gender parity from 2010 to 2019-were presented by WHO region, income group, WHA function and country. The p values for trend (β) were adjusted using the false discovery rate. The former (i, ii and iii) were presented separately BMJ Global Health for countries that were Member States in 2019 with an adjusted p value for trend <0.01, 0.01>p value<0.05 and p value>0.05. This (i, ii and iii) was separately presented for the three Observers in 2019 (Order of Malta, Holy See and Palestine).
Selecting only the countries with an adjusted p value for trend (β) of <0.05-i, ii and iii were plotted at a country level against the GII (2019), the Voice and Accountability Worldwide Governance Indicator (2019) and the Government Effectiveness Worldwide Governance Indicator (2019). Linear regression models were fitted and the Pearson's Correlation Coefficient was calculated. To assess whether there is a difference in the distribution of i, ii and iii between countries who have had a woman HoS, HoG or MoH in the past 5 years (2017-2022) and those who have not-the non-parametric Wilcoxon signedrank test was used, and distributions were presented using boxplot violin plots.
Missing values were excluded from all analyses. All statistical analyses and data visualisations were conducted in Stata V.16 and R V.4.0.5 (R Foundation, Vienna, Austria, www.r-project.org). For data visualisation, the tidyverse, dplyr, pals, and ggplot packages were used.

Ethical considerations
All data used for this study were not restricted nor sensitive, nor did they require permission to access or collate. Data were publicly available and accessible, eliminating the need for additional ethical approval.

Research team
The research team was composed of an internationally diverse group of researchers from a wide variety of sociocultural backgrounds and languages (Arabic, Bengali, Chinese (Mandarin), Dutch, English, Farsi, French, German, Hausa, Kazakh, Nepali, Polish, Russian, Spanish, Swahili, Urdu, Yoruba) which allowed the team to include non-English/non-Western sources and perspectives.

RESULTS
A total of 75 815 delegates, representing 10 944 delegations of 228 unique (including former) countries, territories and political parties of Member States, Associate Members and Observers were included over 1948-2021. Online supplemental table 6 exhibits summary characteristics of all collected data.
Overall, upward trends could be observed in the percentage of inferred women delegates over 1948-2021 across different WHO regions and income groups (figure 1). The Americas and Europe have seen gender parity achieved in their delegations within the last decade. Simultaneously, the Eastern Mediterranean Region, despite significant progress over time, has women representing just 25% of the WHA delegations-while the African region has had stagnant representation of women of around 25%-30% over the last 20 years (figure 1A). Similar trends can be observed when using different regional groupings, such as the UN and WB regional groupings (online supplemental figure 1). Based on current trends in the per cent increase of women's representation per year, some WHO regions will take at least several decades before reaching gender parity of WHA delegation members across their region (figure 2, top). When assessing women's representation across WB income groups, a widening gap of women's representation between high-income and low-income countries in recent decades can be observed. Representation of women in low-income country delegations has been stagnant around 20%-25% since the 1990s, whereas women's  Women have been less represented in higher-powered delegation roles (here considered to be Chief Delegate and Deputy Chief Delegate) in all WHO regions over 1948-2021. Less than 30% of Chief Delegates and Deputy Chief Delegates were inferred to be women in 2021, ranging from 0% to 30% (see online supplemental figure  3). At the current rate of change, it is estimated to take over 40 years to achieve gender parity in the role of Chief Delegate across all WHA delegations (figure 2, middle). In contrast, women's participation in WHA delegations is higher in the roles of an Adviser or Alternate, with over 55% of delegation advisers being women in 2021 (online supplemental figure 3A).  figure 6C, 7C). This may suggest correlations between women's representation and governance performance on these indicators.
Lastly, when comparing the distribution of i, ii and iii, no significant differences in the distributions were observed between countries with a woman HoS or HoG and countries without a woman HoS or HoG in the past 5 years (2017-2022). However, statistically significant differences could be observed for the distributions of i, ii and iii between countries that had a woman MoH in the past 5 years-with a higher per cent of women being part of 2019 delegations in countries who had a woman MoH (p value=0.004), a higher per cent change in women delegation members per year (p value=0.003) and a lower number of years until parity (p value=0.009) compared with countries who did not have a woman MoH.

DISCUSSION
While highlighting the progress of women's representation in WHA delegations over time, this study presents a timely expose on the prevailing levels of gender inequality and exclusion of gender-diverse voices in global health leadership and decision-making bodies. Our quantitative data from 10 994 delegations and 75 815 delegation members spanning 74 years (1948-2021), illustrates that men remain over-represented in most WHA delegations of Member States, Associate Members and Observers to date. From 1948 to 2021, 82.9% of delegations were represented by a majority of men and no WHA had more than 30% of women Chief Delegates in

BMJ Global Health
the 74-year period. Wide variation in trends over time could be observed across different geographical regions, income groups and countries. This is likely the result of different prolonged and multifaceted context-specific social, cultural and institutional factors that inhibit meaningful equitable participation within different countries. Based on our estimated trends, some countries may take over 100 years to reach gender parity (45%-55% women) in their WHA delegations. The lack of representation at this lead global health governance platform is in stark contrast to the health workforce where women constitute over 70%. This unjust disparity in representation fuels real world inequities experienced by women globally.
Unsurprisingly, a higher GII, which indicated higher disparities between women and men in a country, seems to be correlated with a lower proportion of women delegation members in 2019, lower per cent change per year and more years until gender parity. Simultaneously, correlations with the aggregate WB's Voice & Accountability and Governance Effectiveness indicators seem to have the opposite direction. Countries that are perceived to have a strong performance on the Voice & Accountability indicator (citizen's participation in selecting their government, freedom of expression, association and free media) as well as countries that have a perceived strong performance on the Governance Effectiveness indicator (quality of public and civil services, quality of policy formulation and implementation and credibility of government's commitments to such policies) are estimated to have a higher proportion of women delegation members, higher per cent change per year and fewer years until gender parity. These indicators may serve as a proxy for other factors related to the political and socioeconomic context, history and culture of a country that influence its societal norms and structures which may enable higher gender equality and/or participatory governance. 17 Importantly, while useful for broad cross-country comparisons and trends over time, these broad composite represent complex phenomena which cannot be used to elucidate direct or clear associations and therefore to inform specific action for governance reforms. 15 Hence interpretation of these correlations should be approached with caution.
The progress seen today may partially be attributed to a culmination of decades of advocacy, focused on gender equality in international governance. Established in 1946, UN Commission on the Status of Women was the first global intergovernmental body within the UN entirely dedicated for advocating gender equality and BMJ Global Health the empowerment of women. 18 Nearly 50 years later, the 1995 Beijing Platform for Action was adopted at the World Conference on Women in Beijing, highlighting 'Women in power and decision-making,' as one of the 12 critical areas where urgent action was demanded to ensure greater equality for women and girls. 19 In the Beijing Platform for Action, 189 country governments committed to having women in 30% of their decisionmaking roles and the proportion of women in countries' governing bodies nearly doubled since. 20 Subsequent years continued to witness a growth of women in leadership roles across global health governance, including Dr Gro Harlem Brundtland, who served as the first woman in the role of WHO Director-General in 1998. 21 These  BMJ Global Health changes emerging from global governance platforms and international declarations/commitments, are accompanied by overall societal shifts surrounding gender equity such as changing perspectives around gender roles, identities and expectations-enabling improvements in women's participation in governance and leadership across the world. 22 Policy implications: responsibilities of the WHO Through its leadership and normative authority, the WHO holds a central role in promoting gender equity in global health leadership. Global health governance benefits from the inclusion of a variety of perspectives in order to inform more comprehensive and transformative health systems programmes and policies. 2 23 24 Diverse teams (gender, ethnicity, etc) tend to be higher performing, are more innovative and can contribute to inclusion and equality in wider communities. 25 Rolemodelling diversity in WHO staff (not only in gender, but also across other socio-demographic factors through an intersectional lens) encourages other global health organisations and governments to follow suit, and use the rich dividends of diverse expertise, experiences and perspectives in global health.
WHO has signalled its commitment to promoting gender equity in WHA delegations, 5 and could consider more active strategies to ensure this is achieved. This may include supporting Member States to develop leadership programmes or the implementation of gender diversity quotas within delegations. However, it is important to recognise that while gender quotas can be an important method to establish standards for representation, they do not directly correlate to an influence in decisionmaking. Furthermore, an increased number of women at the table does always or necessarily equate to more gender-diverse, inclusive and improved decision-making. For example, women may still not have as many opportunities to speak, access to power and quotas do not take into consideration the formal and informal mechanisms through which gender inequity in leadership occurs. 26 Policies on inclusive leadership should consider more than representation in numbers, but consider the entire enabling environment for the inclusion of diverse voices and perspectives using an intersectional approach to global health decision-making and policy.
The WHO could further commit to monitoring progress on delegation representation over time through the collection of data that is disaggregated by gender and other social identities that may affect the participation of under-represented people, perspectives and expertise in global health leadership. 27 Tracking this information is an important factor for accountability-while presenting this information to Member States may also serve as powerful impetus for meaningful change. When collecting data on WHA participants, WHO should acknowledge that gender is not binary and provide appropriate options during registration, including the ability to self-identify as gender non-binary or -conforming.

Policy implications: responsibilities of Member States
The responsibility for elevating this continuing imbalance in gender participation in WHA delegations lies with each individual Member State and should be supported by the WHO and the international community. While it is imperative to increase women's participation in country delegations, women and gender minorities of diverse backgrounds and origin, should also be meaningfully included in leadership positions within government and international organisations. This would reflect their existing roles, work, expertise and contributions to the global health field, and will further inclusive engagement in conversations related to their own health and wellbeing. Advocates and academics alike have suggested a spectrum of interventions relating to environmental, institutional and individual factors to encourage and empower women's continued involvement and leadership in global health roles. 28 These structural and systematic interventions such as leadership grants, formal policies to safeguard women in the workplace and peertraining and mentorship opportunities could facilitate the meaningful participation in decision-making and leadership roles. 28 29 Individual Member States play an important role in ensuring fair and equitable representation in global health governance, including in their WHA delegations. Member States with a current commitment to gender equity should remain dedicated to their individual country targets, as outlined by their recent and relevant commitments. 30 31 Member States without a current commitment to gender equity in governance, could make explicit public commitments and adopt strategies, policies and practices to enable equitable participation in global health governance. As factors influencing equitable participation in global health governance will differ across settings, this will require active commitments to identify, assess and respond to the prolonged and multifaceted social, cultural and institutional factors that inhibit meaningful equitable participation in global health governance across different country contexts. A recent systematic review on leadership in health identified that cultural change and leadership commitment across five emergent categories were of particular importance to facilitate meaningful equitable participation in leadership: organisational processes, training and development, awareness and engagement, mentoring and networking, and organisational support tools. 32 However, it is important to note that, the inclusion of more women does not explicitly assure the full spectrum of gender transformative policies nor can it be assumed that women are always gender-inclusive advocates. 26 33 Beyond gender parity in representation, it is imperative to recognise that women are not a homogenous group and differences in class, income, race, religion, ability and nationality must also be considered in the development and implementation of global health policies. 26 27 Tacking such an intersectional approach to global health can be used to address not only representation in global BMJ Global Health health multilateral systems, but the systemic inequalities and power hierarchies that influence power in global health decision-making. 2 24 26 27 34 Strengths and limitations Our study has several strengths. The scope of this analysis provides the first comprehensive large-scale longitudinal quantitative assessment of delegation's gender representation since the WHO's inception in 1948. The data disaggregation enabled further descriptive evaluation of trends between countries, regions, income groups and delegate roles. Furthermore, the data generated provides a strong foundation for further gender equitable data collection and in-depth analysis the authors and/or WHO may want to commit to monitor progress in WHA participation over time.
However, our study also has several limitations. First, while our analysis enables the assessment of gender representation over time, it does not allow us to assess the influence delegates have on WHA decision-making processes-nor allow us to directly assess whether increasing gender diversity in WHA delegations may produce more equitable gender-transformative global health policies and agreements at the WHA. Arguably, many decisions on behalf of nations may have already been agreed on before the WHA actually convenes, limiting the influence of representative delegation members on formal decision-making processes.
Second, inferring likely gender was largely limited to binary definitions of gender, as authors were dependent on the prefixes and other gendered language used in WHO/online documentation and a binary gender-to-name algorithm, instead of delegates selfidentification. As a result, some inferred genders may have misrepresented the gender identity of delegates. These limitations further point to the need for better data collection around gender and sex in order to promote transparency and accountability in genderinclusive governance (eg, options for delegates to selfidentify their gender in the WHA registration process).

CONCLUSION
Despite some progress in recent decades, women continue to be under-represented in global health leadership and decision-making at the highest level. The ongoing under-representation of women has implications for not only gender equality but also for global health systems worldwide, from the global to local level. Prioritising equitable intersectional approaches, which prioritise equity of various forms beyond gender, and inclusive representation in decision-making enables transformative policy-making that fosters transparent, accountable, functional and just health systems. Urgent action is required by the global health community, with particular attention to regions and Member States (countries) where progress has been stagnant in the past 74 years. Twitter Kim Robin van Daalen @DaalenKim, Maisoon Chowdhury @machowdh, Sara Dada @dadasara3, Salma El-Gamal @Salma234sal, Laura Jung @lauraejung, Charlotte Anne O'Leary @charlieaoleary, Henry Charles Ashworth @HenryCAshworth, Anna Socha @annamsocha, Dolapo Olaniyan @olaniyan_ dolapo, Fajembola Temilade Azeezat @DrPrish, Toyyib Abdulkareem @toyyibdimeji, Roopa Dhatt @RoopaDhatt and Dheepa Rajan @dheepa_rajan Acknowledgements The authors thank Richard Arthur John Dear (Data scientist and PhD Candidate in Neuroscience, University of Cambridge) for his expert advice on data visualisation in R. We also thank Dr Amit Prasad (Data, Analytics and Delivery for Impact Division WHO HQ) for reviewing the methodology, and thank Dr Magda Robalo and Dr Ann Keeling (Women in Global Health) for their feedback on the manuscript.
Contributors KRvD conceived the presented idea with support from MC, SD, PK and LJ. KRvD, MC, SD, PK, SE-G, GK, LJ, RO, CAO'L, HCA, AS, DO, FTA, SA and TA contributed to the data collection and validation. KRvD, MC, SD, PK and SE-G contributed to the data cleaning. KRvD conducted all data visualisation and analysis in R and Stata. KRvD, MC, SE-G, CAO'L, SD and PK wrote the first draft of the manuscript and supplementary materials. DR and RD provided expert advice and critical feedback during all project phases. All authors contributed intellectually to the final manuscript. MC, SD, PK and SE-G share second authorship. KRvD and DR were responsible for the overall content.
Funding This research did not receive specific funding. KRvD receives funding by the Gates Cambridge Scholarship (OPP1144) for her PhD research. Publication fees were paid by the Bill & Melinda Gates Foundations.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Patient consent for publication Not applicable.
Ethics approval Not applicable.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data ara available in a public, open access repository. The data to support the findings of this study are available upon request from the corresponding author, and will soon be made publicly available on an online repository.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible