The incorporation of digital devices and wearables for personal health data has the potential to reduce barriers to health access (1). Perceived barriers resulting in delayed care include discrimination and access (i.e., ability to make appointments, transportation to care location, limited care hours) (2). Digital devices, as inanimate objects, do not discriminate while increasing patient engagement by facilitating the sharing of data with providers (3). Additionally, device sensors can be connected to care providers around the clock to provide real-time personalized medicine in a remote format (4), eliminating the need to make appointments or travel to care locations. While digital devices hold promise, we acknowledge that cost (5) and greater adoption remain barriers to full integration into health care (6).

An important component of health is cardiorespiratory fitness. Cardiorespiratory fitness reflects the ability of the circulatory and respiratory systems to provide skeletal muscle with oxygen necessary for energy production during exercise and physical activity (7). Determining cardiovascular fitness is important because moderate to high levels are associated with a reduced risk of all-cause mortality (8), regardless of adiposity, age, alcohol intake, ethnicity, and smoking status (9). Cardiorespiratory fitness is traditionally measured by laboratory-based metabolic analysis systems and the collection of expired gasses during incrementally increasing intensity exercise tests (10). Advances in technology have resulted in the development of valid wearable portable metabolic analysis systems (11) which allows sport and exercise scientists to obtain measurements outside the laboratory.

Additional demographic variables commonly reported in the sport and exercise science literature include sex and gender. Sex refers to biological factors (i.e., genetic, hormonal, and or anatomical traits) (12) while gender is a social construct with roles and activities that are learned through a socialization process (13). For this reason, it is recommended that data on sex and gender be collected in a two-step process, (1) sex assigned at birth, and (2) current gender (14). Acknowledging diversity in gender is important, and includes people who are non-binary, transgender, and gender fluid across many indigenous cultures (i.e., Two Spirit in North America, Hijra in India, fa'afafine in Samoa) (15). There is a need for greater inclusion in sport and exercise science research, particularly when considering sex and gender diverse people (16). A study of over 800 published papers on exercise science-related topics reported that three investigations collected participant sex or gender with options other than the female-male binary (17). Allowing participants to identify their sex and gender in inclusive terms is important because misgendering (when an individual is described using terminology that is inconsistent with their gender identification) can affect statistical outcomes and interpretation of data in typical sample sizes employed in sport and exercise science research (18). It is possible that a barrier to inclusion for sport and exercise scientists exists in the software and equipment most regularly used.

Metabolic analysis systems are a common piece of equipment in exercise physiology laboratories and provide data for many exercise investigations (19). The software connected to these devices require the input of sex data, and to our knowledge, is limited to the binary options of female or male. Because metabolic analysis systems and the derived data should not be affected by the sex selected in the software, it was hypothesized that there would be no differences in metabolic or pulmonary data between identical exercise bouts conducted by the same individual when categorized as a female or male. Thus, the purpose of this investigation was to test the hypothesis by having the same participants perform identical self-paced walking bouts and running bouts under the female and male categories in the software.

All data met the assumptions for normality. No differences between the female and male conditions were observed for any metabolic or pulmonary variables during self-paced walking (see Table 1). The shared variance was between 54% and 96% (see Table 1).

Similar to walking, no differences between the female and male conditions were observed for any metabolic or pulmonary variables during self-paced running (see Table 2). The shared variance was between 67% and 97% (see Table 2).

According to the randomization schema, 14 participants (n = 7 female, n = 7 male) began the study under the condition that aligned with their self-identified sex (i.e., female participant randomized into the female condition first), while 6 participants (n = 3 female, n = 3 male) began the study under the condition that was not aligned with their self-identified sex (i.e., female participant randomized into the male condition first). No differences for any variable were observed when the aligned with sex condition was compared to the condition not aligned with participants' sex for walking (see Table 3) or running (see Table 4).

The present study did not aim to evaluate sex differences, so we did not test for them. However, to align with the Sex and Gender Equity in Research (SAGER) guidelines (26) and support future meta-analyses, we present disaggregated metabolic and pulmonary data in Table 5. Table 6 presents disaggregated metabolic and pulmonary data when participants’ self-identified sex aligned with the randomized condition, and when it did not align.

We hypothesized that a portable metabolic analysis system measuring metabolic and pulmonary variables would not be affected by participant sex entered into the software environment, and as such, no differences in the variables would be observed between the female or male sex categories available in the software during self-paced walking or self-paced running. Our hypothesis was supported, as oxygen uptake (expressed in absolute and relative terms) carbon dioxide output, respiratory exchange ratio, pulmonary ventilation, respiratory rate, and accumulated kilocalories were unaffected by the sex data entered into the device software.

These results provide evidence that, while differences among females and males may be present for metabolic and pulmonary variables during exercise (20, 27), the COSMED K5 and accompanying Omnia software designed to measure such variables do not utilize the input of demographic sex data to measure the metabolic or pulmonary variables investigated in the present study. We believe that any deviation between conditions is due to the usual physiological variation that occurs between bouts of exercise (28, 29). However, it must be pointed out in the present study that workload in terms of walking and running speed was maintained from one bout to the next. Because the device software limits the sex input to two options (female and male), a potential consequence is an unintentional barrier, engineered into the software, to the inclusion of sex and gender diverse people. We suggest a similar unintentional barrier may exist with much of the equipment found in exercise physiology (i.e., electrocardiogram machines, electromyography systems, blood lactate analyzers, spirometers) and biomechanics (motion capture systems, force platforms, isokinetic dynamometers) laboratories and the accompanying software running these devices. While testing similar to what was conducted in the current investigation is likely unnecessary for all equipment, an opportunity exists, particularly with respect to sex and gender input options, for device software to be updated to be more inclusive of people who are sex and gender diverse. These findings may have regulatory implications, as they suggest that a more critical evaluation of sex-based algorithms in medical devices and digital health tools is warranted. Future guidelines should prioritize evidence-based inclusion of demographic variables.

Belief in sex differences in energy expenditure and resting metabolic rate is well-entrenched dogma (30). The Harris and Benedict equation for estimating basal metabolic rate was published in 1918 and assumed a difference between females and males of up to 7% (31). The Mifflin-St Joer equation for estimating resting metabolic rate was published in 1990 and assumed a sex difference of 166 kcal (32). With respect to physical activity, sex differences in energy expenditure have been reported in individuals completing a three-month wilderness expedition; however, it should be noted that measures were estimated from actigraphy (33). Additionally, males have been reported to expend significantly greater energy expenditure than females while playing golf, however kcal was derived as estimates via heart rate devices rather than indirect calorimetry (34). The energy estimations for physical activity and exercise derived from actigraphy and wearable device estimations are likely influenced by the same assumptions embedded within most resting metabolic rate equations. While it is outside the scope of this investigation to explore underlying assumptions, we can state that the portable metabolic analysis system utilized in the present study was not influenced by the same limitation, as no differences were observed between sex input as an independent variable. Wearable and other digital device manufacturers whose devices measure or estimate energy expenditure may wish to reconsider certain assumptions and make their software more inclusive. Additionally, many fitness wearables require users to input their sex to create a profile, but if the measured outcome is step count, our research suggests the input is not needed.

A knowledge gap in the sport and exercise science literature among females and males has been documented because males make up the majority of participants tested, at around 65% (17, 35, 36). The consequence of this knowledge gap is a lack of appropriate scientific recommendations for females in areas such as training, recovery, and rehabilitation (36). We propose that a much wider chasm exists for sex and gender diverse people, such as people who are intersex, transgender, and/or non-binary. Because of this, a leading professional organization in the field, the American College of Sports Medicine, is only able to provide a single paragraph in recommendation to gender diverse people among the over 500-page text (37). While there is a disparity of reporting sex and gender diverse people in the sport and exercise science literature (17), it is possible that a true lack of inclusion may not be present. Similar to what we report in the present study, a barrier may be present at the software- and equipment-level running through the study design, which results in data on sex and gender diverse people not being collected or ultimately reported in the literature. We have suggested that investigators consider how sex and gender data are collected (18). If researchers rely on obtaining these data from their connected devices and software, having insufficient options for sex and gender maintains the present knowledge gap and exclusivity of the sport and exercise science literature.

This study is not without limitations. The authors acknowledge that while we propose an opportunity for greater inclusivity in sport and exercise science research exists, the current study only tested participants who identified as cisgender females or males. While we did not exclude gender diverse individuals from participating, our findings cannot directly speak to the experiences of those outside the binary framework. A potential limitation may be in our recruitment and enrollment methodology, which could have reduced the likelihood of participation by people who identify as sex or gender diverse. To address this, future studies could employ more targeted recruitment of gender diverse individuals by partnering with LGBTQ + community groups or gender clinics.

In conclusion, we report that an individual's sex, as designated in the Omnia software and associated COSMED K5 portable metabolic analysis system, does not affect measures of relative and absolute oxygen uptake, carbon dioxide production, ventilation, respiratory exchange ratio, respiratory rate, or energy expenditure. From a practical standpoint, researchers who inadvertently enter a participants sex incorrectly into the Omnia software can be reassured the mistake will have no adverse effects on the outcome variables. We propose that limiting the input of sex options to the female-male binary is a barrier to researchers’ ability to include sex and gender diverse people. Manufacturers of digital devices and equipment whose measurements or estimations are not affected by sex could remove this barrier by updating software with inclusive options, potentially enabling the reporting of disaggregated data as recommended by the SAGER guidelines (26). These actions could help to address the knowledge gap present for people who are sex and gender diverse in sport and exercise science research. Until digital devices allow for more inclusive sex and gender options, researchers are encouraged to be intentional in their approach for collecting sex and gender using a two-step process [sex, acknowledging more than binary options (i.e., intersex); and gender, allowing for an open-response option].