Tension-type headache (TTH) is the most common primary headache disorder with an estimated global prevalence of 26%. ^{Reference Stovner, Hagen, Linde and Steiner1} Notably, differences in the epidemiology across demographic factors, including sex, race and ethnicity, have been reported. In a telephone study of 13,345 participants from Maryland, USA, there was a significantly higher prevalence of TTH in females than males by 15.9% and in Black or African Americans than White participants by 17.3%. ^{Reference Schwartz, Stewart, Simon and Lipton2} However, more studies are needed to elucidate the etiology underlying demographic differences in TTH. Within neurological disease clinical trials, there have been enrollment disparities in the demographic composition of participants compared to global disease burdens. For instance, there has been underrepresentation of females in stroke clinical trials ^{Reference Strong, Pudar and Thrift3} and of racial and ethnic minority groups in Alzheimer’s disease. ^{Reference Raman, Quiroz and Langford4} Overall, persistent underrepresentation of minority groups in clinical research endorses inequitable therapy evaluation and compromises study generalizability. ^{Reference Charleston5} Clinical trial enrollment proportions that reflect global disease burden can improve our understanding of how TTH affects various populations and enable targeted therapy development. We investigated sex, racial and ethnic representational disparities in TTH clinical trials.

The study methodology was adapted from that of a previous study. ^{Reference Steinberg, Turner and Weeks6} “Tension headache” was searched on ClinicalTrials.gov from inception (02/2000) to 01/2023. Study workflow is summarized in Figure 1. Randomized controlled trials with participants with TTH, at least double-masking, and publicly accessible results were included. Data collection was conducted in duplicate by two independent reviewers. Trial (start year, study location, phase and masking) and demographic (sex, race and ethnicity per US Census Bureau definitions) data were collected from results posted on ClinicalTrials.gov and/or online publications. The Global Burden of Disease (GBD) database was searched for sex-based TTH disease burdens. We sorted the database by start year, country and age group (containing two standard deviations of the mean trial participant age) to obtain the female disease burden proportion for each trial. We calculated female disease burden by dividing the number of female cases by total TTH cases. Across trials, we calculated the absolute number and relative proportions of participants for each demographic group. For females, we calculated participation-to-prevalence ratios (PPRs) and 95% confidence intervals (CIs) with PPR values of 0.8–1.2 representing equitable enrollment. ^{Reference Poon, Khanijow and Umarjee7} Across trial characteristics, we conducted Kruskal–Wallis tests with post-hoc analyses using Mann–Whitney U test with Bonferroni continuity correction. A p-value<0.05 was considered statistically significant. Analysis was completed in RStudio 2022.02.0 + 443.

Figure 1. Study methods and selection of tension-type headache clinical trials for inclusion. ^aReasons for study exclusion at initial screening included studies that were non-interventional, non-randomized, and/or of the wrong condition (e.g., migraine headache). ^bReasons for study exclusion following full record evaluation included studies that had no results or publications available (n = 8), single demographic enrollment (n = 1) and open-label or single-blinded (n = 33).

Study characteristics and enrollment are summarized in Table 1. Of 72 records identified, 13 randomized controlled trials were included (Supplementary Table 1). Trials were started between 2006 and 2020 (median: 2011; interquartile range: 7 years). There were 0 phase I, 2 (15.4%) phase II, 4 (30.8%) phase III, 1 (7.7%) phase IV) and 6 (46.2%) medical device trials (phase not applicable). By blinding status, 10 trials (76.9%) were double-, 1 (7.7%) triple- and 2 (15.4%) quadruple-blinded. For both study and sponsor sites, 3 trials (23.1%) were in the USA and 10 (76.9%) international/non-US. By intervention, 6 trials (46.2%) were drug-related and 7 (53.8%) were device- or procedure-related.

Table 1. Study characteristics and enrollment of 13 tension-type headache clinical trials by sex, race and ethnicity

p-values < 0.05 are in bold.

^aRace reported as unknown (e.g., participant declined to indicate which race).

^bNo phase I studies were eligible for inclusion.

^cUnable to calculate p-value due to insufficient statistical levels.

^dUnable to calculate 95% CI due to insufficient number of observations. AIAN, American Indian or Alaskan Native; NHOPI, Native Hawaiian or Other Pacific Islander; PPR, participation-to-prevalence ratio; CI, confidence interval; NA, not applicable; NR, not reported.

Of 1337 participants across trials, females comprised 69.56% (n = 930) and were overrepresented relative to their disease burden (PPR:1.43; 95% CI:1.27–1.58), which was appreciated across all trial characteristics. Females were overrepresented in device/procedure studies compared to drug studies (p = 0.03). Female enrollment was higher when the sponsor originated internationally rather than the USA (p = 0.04). Three and two trials reported race and ethnicity data, respectively. By absolute enrollment, White participants comprised the majority (n = 252; 61.17%), while Black or African American (n = 71; 17.23%) participants comprised the minority. Participants of Hispanic ethnicity comprised the minority (n = 74; 21.39%). Asian, American Indian or Alaskan Native (AIAN), and Native Hawaiian or Other Pacific Islander (NHOPI) participants were not represented (all n = 0). Enrollment proportions for each race and ethnicity did not vary significantly across trial characteristics.

Across 13 TTH clinical trials, there was female overrepresentation when trials were equally weighted, which was seen across all trial characteristics. This suggests that TTH clinical trials does not mirror the sex distribution of TTH in the larger population, thus decreasing generalizability and external validity of trial findings. While it is promising that there is more than adequate representation of females, who are traditionally underrepresented in research, male underrepresentation may result in an incomplete understanding of TTH in males. For instance, there may be a lack of understanding of sociocultural factors (e.g., gender role expectations) preventing males from reporting pain. Notably, females are significantly more likely to report headache-related conditions and greater associated pain and to seek medical care for headaches than males. ^{Reference Celentano, Linet and Stewart8} Moreover, male underrepresentation results in treatments tailored to females which may be ineffective for males. For instance, females with TTH are hypothesized to benefit from therapies centered around nociceptive mechanisms and psychosocial stressors, while males benefit from addressing sleep quality and depressive symptoms. ^{Reference Fuensalida-Novo, Jimenez-Antona, Benito-Gonzalez, Cigaran-Mendez, Paras-Bravo and Fernandez-De-Las-Penas9} To improve male enrollment, we call for TTH clinical trials to evaluate barriers to participation (e.g., male stigma regarding pain leading to reluctance to seek care), implement policies to encourage equitable sex representation and revise enrollment criteria to promote male participation in pain reporting. Furthermore, we recommend TTH trials to enroll all gender identities (e.g., transgender and non-binary) and include enrollment criteria and data collection that encompasses the broader gender spectrum.

While we were unable to evaluate PPRs for race and ethnicity due to GBD’s lack of racial and ethnic reporting, Black or African American (n = 71; 17.23%) and Hispanic (n = 74; 21.39%) participants comprised the minority of TTH clinical trial participants in the racial and ethnic categories, respectively. Moreover, there was an absence of Asian, AIAN and NHOPI enrollment in TTH trials. Notably, we may be underestimating the disparity between group enrollment proportions as racial and ethnic reporting is not mandated on ClinicalTrials.gov. Despite limited literature on TTH prevalence across racial and ethnic groups, a 2015 survey found that the prevalence of migraine or severe headaches is highest in NHOPI and AIAN and lowest in Asians. ^{Reference Burch, Rizzoli and Loder10} This is striking as we found no NHOPI and AIAN representation in TTH clinical trials. Inequitable racial and ethnic representation can also affect treatment, which can vary across demographics. For instance, Black or African American individuals are more likely to cancel follow-up appointments and less likely to attend headache treatment than White individuals. ^{Reference Charleston5} To promote diverse enrollment, we call for TTH clinical trials to promote leadership representation from racial and ethnic minority groups, mandate racial and ethnic data reporting, and address participation barriers (e.g., mistrust in healthcare and research, socioeconomic considerations). Ultimately, disparities in racial and ethnic enrollment in TTH trials impede our understanding of how TTH presents in diverse populations and the development of and access to effective therapeutics.

The use of GBD longitudinal consecutive data (1990–2019) stratified by year, region and age groups allowed us to estimate disease burdens personalized for each trial setting. Moreover, GBD’s consistent methodology increased our study reliability and reduced biases. Our use of PPRs to evaluate equitable clinical trial sex representation accounted for TTH prevalence, which improved trial sample representativeness and generalizability. However, our study was limited by the number of TTH trials and participants, which restricted statistical testing for several trial characteristics. Additionally, TTH are often mild, and patients do not seek clinical care or only visit primary care settings, which often are not involved in clinical trial participation. Thus, these trials may not fully reflect the true prevalence of TTH. Another limitation is the lack of racial and ethnic data in the GBD database. While this limits the assessment of adequate enrollment, previous studies have reported higher prevalence of TTH in White individuals. ^{Reference Schwartz, Stewart, Simon and Lipton2} Moreover, ClinicalTrials.gov did not include other racial and ethnic groups beyond those designated by the US Census Bureau. Finally, due to the lack of gender-specific data, our findings were confined to the scope of binary sex options. This does not account for the non-binary spectrum of gender, which deserves further study.

Overall, TTH clinical trial enrollment overrepresents females, which inadequately reflects the global disease, and underrepresents racial and ethnic minority groups by absolute enrollment. Male underrepresentation can perpetuate the social attitudes and stereotypes of masculinity that initially prevent men from seeking medical care, which can further worsen health outcomes. ^{Reference Lancet11} Moreover, racial and ethnic underrepresentation hinders our understanding of how different populations respond to various interventions and restricts equitable access to effective medical therapies, which continues the cycle of health disparities. Moreover, it undermines participant trust in clinical research and healthcare and deters participation among certain populations. ¹² These disparities highlight the importance of enrolling a representative participant demographic. We recommend increased diversity within leadership, trial policies that promote diverse enrollment and elimination of barriers to participation. Moreover, future studies should include and address the population of individuals with TTH who do not seek clinical care or only present to primary care settings to capture the true TTH prevalence. Future investigations could include other clinical trial databases, racial and ethnic groups, gender, and intersectionality to capture the full spectrum of diversity. Ultimately, study enrollment that adequately reflects the disease burden improves our understanding of how TTH in different populations and aids the development of effective tailored treatments.