Scoping ‘sex’ and ‘gender’ in rehabilitation: (mis)representations and effects

Our scoping review employed a 15-year time span, responding to an influx of research and literature that emerged on gender and rehabilitation beginning in the mid-2000s. In order to capture as much of the relevant literature as possible, a broad search strategy inclusive of academic literature was employed. The following three databases were selected due to their perceived relevance to the research question: PubMed, Embase, and CINHAL. We did the initial search in PubMed and made minor modifications to adapt the search strategy to Embase and CINAHL, based on some combination of MeSH categories and other search terms within each database. MeSH categories and search terms included variants of the following gender and sex-related terms: ‘gender,’ ‘sex,’ ‘man,’ ‘woman,’ ‘boy,’ ‘girl,’ ‘male,’ ‘female,’ ‘intersex,’ ‘transgender,’ ‘gender equality,’ ‘gender equity,’ ‘gender identity,’ ‘gender role,’ ‘sexism,’ ‘sex disaggregated,’ ‘sex difference,’ ‘gender bias,’ ‘sex bias,’ ‘sex discrimination.’ Two authors determined the search strategies by reviewing search terms used in existing gender and health systems-related scoping reviews [20, 21] and with the support of two Johns Hopkins University Welch Medical Library informationists. The exact search terms are viewable in a supplemental file.

Each database search combined variants of the above sex and gender-related terms with variants of the following rehabilitation and assistive technology-related terms that were adapted to database-specific MeSH categories: ‘rehabilitation,’ ‘self-help device,’ ‘occupational therapy,’ ‘rehabilitation center,’ ‘physical and rehabilitation medicine,’ ‘psychiatric rehabilitation,’ ‘psychosocial rehabilitation,’ ‘rehabilitation nursing,’ ‘rehabilitation research,’ ‘telerehabilitation,’ ‘physical rehabilitation,’ ‘assistive device,’ and ‘assistive technology.’

Each database search included results in English and from both high-income countries (HICs) and LMICs, operating from an assumption that there would be limited results from LMICs but that an understanding of the scope of literature may help inform the global health research agenda related to gender and rehabilitation. We use the terms HIC and LMIC with the caveat that there is deep sociocultural heterogeneity between and within countries that fall beneath LMIC and HIC categories that limit their usefulness as comparative descriptors. The method yielded 4484 references across all databases.

Because the search strategy yielded a large number of irrelevant references, a set of inclusion criteria was applied to all of the 4484 initial titles and abstracts. Two members of the team reviewed roughly 50 titles and abstracts collaboratively, to ensure consistency, before screening the remaining titles and abstracts separately. References were included after an initial title and abstract screening based on the following criteria: 1) focused on physical rehabilitation and/or post-rehabilitation functioning (including caregiver support and assistive technology access, use, satisfaction, and adherence), and 2) included a clear articulation of sex or gender differences or sex-disaggregated data related to rehabilitation participation and/or outcomes. The second criterion was included after it was found that many articles either controlled for sex/gender or reported sex/gender differences in passing, making it difficult to delineate any clear sex or gender differences. Beyond these inclusion criteria, the abstract and title reviewers intentionally chose not to focus on sex and gender differences in burdens of disease. While disease burden is tied to access, use, adherence, and outcomes, articles which focused on disease burden were not directly about rehabilitation, and as such, did not fit within the scope of this review. This inclusion and exclusion criteria yielded 792 studies from the original 4484 references. After engaging in a second round of reviewing titles and abstracts to determine whether articles included sex and/or gender as a major focus of the article, the reviewers were left with 65 references that underwent a full text review. After a full text review, all 65 references met the inclusion criteria, and rather than further refining our inclusion criteria to narrow down the final articles, we decided that it would be most fruitful to include all 65 articles that reached the step of full text screening in our review to ensure a more systematic exploration of current representations of sex and gender in rehabilitation-related literature. This method yielded 65 publications from the initial 4484 that underwent a full text review, among which all 65 were included in the scoping review (Fig. 1).

Two members of the study team participated in extracting and charting data from the final 65 publications. Study team members divided up the final 65 articles and scheduled frequent check-ins early on during the data extraction process to discuss how articles were being coded, refine the coding scheme, and work through discrepancies until concordance was reached. The following information was recorded from each publication: Author, title, year, journal, type of article, country, type of rehabilitation, methods, participants/target, number of participants/sample size, aims of study/paper, and whether or not the article adequately represented sex and gender based on the schema detailed in Table 2. Articles were also coded for the following rehabilitation-related themes: rehabilitation needs (RN), rehabilitation access and use (RAU), rehabilitation experiences (RE), rehabilitation adherence (RA), assistive technology (AT), health systems (HS), rehabilitation outcomes (RO), post-rehabilitation functioning (PRF), experiences of caregivers (EC), and other (O).

Articles were determined to adequately represent sex and gender if they did all of the following: (1) used the word gender to refer to gender-based social factors and/or sex to refer to sex-based biological factors; (2) presented sex and/or gender as nonbinary, which acknowledges the existence and experiences of people who fall outside of limited female/male and woman/man sex and gender categories; (3) defined the terms sex and/or gender, and (4) used gender-related terms like woman/man to refer to gender-related social factors and/or sex-related terms like female/male to refer to sex-based biological factors. The authors collaboratively developed the criteria for determining adequate representation of sex and gender. The two authors who were involved in data extraction reviewed articles after coding them to determine which criteria they did and did not fulfill. We began by reviewing a few articles together before reviewing the remaining articles independently. We also scheduled frequent check-ins early on in the testing process where we discussed questions that arose and further refined our criteria until we reached concordance. We included defining sex and/or gender due to widespread confusion and misrepresentations in public health literature about the difference between sex and gender, which has harmful implications for scientific understandings of the biological and social determinants of health. These criteria were loosely based on Heidari et al.’s [22] SAGER guidelines and recommendations, as well as on the charting categories of Williams et al.’s [21] scoping review on the integration of sex and gender considerations in health policy making. While the SAGER guidelines and Williams et al.’s (2021) charting categories provide a broad and helpful overview of sex and gender considerations for researchers, our schema (Table 2) provides a more detailed explanation of why certain guidelines are important and how to operationalize them. We recognize the limitations inherent in categorical understandings of sex and gender [4], the frequent inextricability of sex and gender [9, 23], and the necessity of exploring gender as intersectional [24] and relational [4], but sought to develop a preliminary framework/schema that would be implementable among public health researchers exploring sex and/or gender health disparities (Table 2).

In the presentation of results, this scoping review relied as much as possible on the sex and gender terms put forth by researchers, who often used the terms interchangeably and were unclear about their use of particular terms. We use the terminology sex/gender or sex and/or gender to reflect our inability to determine why authors were relying on particular terms and to reflect how they mutually affect and shape health and wellbeing [9, 23]. We do not intend our inclusion of studies or use of researchers’ terms to be seen as an endorsement of the common conflation of sex and gender or marginalization of sexual and gender minorities in public health research; however, without additional information beyond the publications, we were not able to sufficiently decipher authors’ misuse of sex and/or gender terms.

Where possible, we incorporated an intersectional approach into our review of existing literature. Our intersectional approach highlights authors’ mentions of associations between categorical quantitative variables as well as authors’ reflections on the compounding effects of social and structural forces on rehabilitation access, participation, adherence, and outcomes. In theory, an intersectional approach acknowledges that human experiences cannot be reduced to any one single factor or constellation of factors and that the intersectional effects of different social and structural factors on health are not additive but compounding [24]. In practice, however, we did not have the original data to be able to explore complex relations across identities and analytic categories and were therefore mostly limited to authors’ descriptions of associations between categorical quantitative variables. For example, while several literature reviews and articles described women who identify as a racial or ethnic minority as participating less than white women and men in rehabilitation [25‐28], the articles generally did not describe how social factors like gender discrimination and racism might have mutually constituted each other to shape their rehabilitation participation, with the exception of a qualitative article that offered a richer explanation of how social and structural forces like gender discrimination, racism, and social class interact in a compounding way to shape access to pain rehabilitation [28]. This scoping review foregrounds sex and gender as analytic variables but includes other relevant social factors that intersect with sex and gender to inform rehabilitation participation and outcomes as they emerged in gender-focused literature.

The scoping review yielded 65 sources (41 original research articles, nine literature reviews, and 15 miscellaneous references, including editorials and commentaries) that generally described worse rehabilitation access, use, adherence, and outcomes and a higher burden of caregiving among women compared to men (Table 1). However, existing literature has focused disproportionately on a few types of rehabilitation among adults in HIC contexts. For example, of the 65 total articles, 29 (45%) emphasized a sex and/or gender disparity in cardiac rehabilitation participation and outcomes, among which 27 were based in HICs. There is limited research on gender and/or sex differences in rehabilitation participation and outcomes for noncardiac conditions, in LMICs, and among children and adolescents. While 77% of global rehabilitation needs are based in LMICs [2], only seven articles (11%) were based in LMICs [27, 29‐34]. An additional two (3%) global articles focused on both LMICs and HICs [35, 36], with a limited LMIC focus. The literature on assistive technology suggested differences in access and use between men and women depending on the device and country income. All of the articles that explored caregiving as it related to gender and rehabilitation—seven in total—focused on post-stroke caregiving in HICs (n = 7) [37‐43]. Furthermore, only four articles (6%) explored rehabilitation participation and outcomes among children and adolescents [30, 31, 34, 44].

Of the 65 total articles, no articles (n = 0) adequately represented sex and gender based on the schema created (Table 2). Fifty-five articles (85%) were consistent and accurate in their use ‘sex’ and/or ‘gender’ as an analytic category. However, only 14 articles (22%) used gender-related terms to refer to gender-related phenomena and sex-related terms to refer to sex-related phenomena. Only eight articles (12%) provided definitions for the commonly misunderstood terms ‘sex’ and/or ‘gender’ and a mere two articles (3%) presented sex and/or gender as nonbinary. As highlighted in Table 3., five articles (8%) met three out of the four criteria of the schema. Of those five articles, two (3%) did not meet the criteria “defined ‘sex’ and/or ‘gender,’” and three (5%) did not present sex and/or gender as nonbinary. All articles that inaccurately used the terms ‘sex’ and ‘gender’ interchangeably (n = 10) did not meet the remaining three criteria. Only one editorial [82] and one case report [47] acknowledged sexual (e.g., intersex) and gender minorities as sex and gender subcategories, which leads to a dearth of knowledge about how sex and gender interact with social and structural factors to shape rehabilitation participation and outcomes within health systems among sexual and gender minorities (Table 2). Our use of the term ‘sexual minorities’ here is not intended to refer to sexual orientation.

Our review revealed lower rates of rehabilitation participation among women than men, with the exception of women participating more than men in cancer rehabilitation in Denmark [84] and in audiological rehabilitation in Sweden [57]. Sociodemographic, health system, patient, and “other” factors influence women’s rehabilitation access, use, and adherence (Fig. 2). Unfortunately, scoping review articles reflect a widespread problem in health literature where researchers name gender but mostly speak about women, with men appearing mainly as the statistical norm or privileged category [4]. Sociodemographic-level factors that influence this disparity in rehabilitation participation between men and women include: lower education levels [25, 66, 68], living alone [57, 60, 62, 68], being married [26], being unmarried [61, 72], older age [25‐28, 58, 59, 61, 63, 65, 72, 79, 83], lower socioeconomic status [27, 28, 45, 58, 61, 72], and racial or ethnic minority status [25‐28]. Health systems factors like a lack of referral or recommendation from a healthcare provider [26, 27, 29, 35, 47, 58‐62, 64, 65, 77‐81], which is compounded by racial or ethnic minority status [61, 62, 78], socioeconomic status [26, 27, 58, 62, 65, 66, 78], or a combination of racial or ethnic minority and socioeconomic status [62, 65, 78]; inadequate assessment [28, 65]; and inadequate insurance [25, 27, 29, 40, 58, 64, 67, 78, 79] were shown to negatively impact women’s participation in rehabilitation. Just as a lack of referral or recommendation by a health care provider serves as a barrier to rehabilitation access, enthusiastic referral or recommendation serves as a facilitator to access [29, 62, 78]. Referral interventions that are either automatic or that require providers to systematically go through a list of criteria to reduce bias in referral decisions have been successful in reducing a disparity between men and women in access to cardiac rehabilitation [27, 36, 58, 60, 68, 79, 81]. Patient-level factors that adversely affect women’s rehabilitation access, use, and adherence include domestic and familial responsibilities like caregiving [25, 26, 45, 59‐65, 67‐70, 79]; a lack of familial and/or social support [26, 60]; formal employment obligations [29, 45, 60, 63‐65, 70, 71]; physical and psychological comorbidities [25, 26, 60‐65, 67, 68, 72, 77‐79]; exercise limitations and apprehension [25, 26, 28, 59, 62‐65, 67, 70, 77, 79]; and negative beliefs and perceptions related to rehabilitation [26, 64, 68, 70, 79].

We separated sociodemographic-level factors from the patient-level factors to emphasize that even when they are used to describe individuals, they are often the result of larger social and structural forces. While some studies reported patient-level factors like women’s domestic and caregiving responsibilities, a lack of social support, and sociodemographic-level factors such as older age as barriers to rehabilitation access, use, and adherence, other studies reported women’s domestic and caregiving responsibilities [29], the presence of social support [25, 29, 45, 61, 63, 64, 68, 71], and older age [68] as facilitators to their participation in rehabilitation. Other major factors affecting women’s rehabilitation access, use, and adherence, as compared to men, are distance and transportation [27, 29, 60‐65, 67, 78, 79]. Health systems have tried to mitigate a disparity in cardiac rehabilitation participation, with women generally having lower access, use, and adherence than men, through women-only [27, 29, 59, 68‐70, 72] and home-based and hybrid rehabilitation models [27, 59, 62, 67, 69], with mixed results.

While few studies in the scoping review reported on gender and sex differences in functional improvements, among those that did, women tended to have less functional gain during rehabilitation than men and similar improvements in psychosocial measures. That is to say that even when women overcame barriers negatively impacting access, use, and adherence to rehabilitation and successfully completed their programs, women reportedly showed less physical, functional improvements post rehabilitation compared to men participants. Articles that reported women as having less overall functional gain during rehabilitation included studies on stroke rehabilitation in the United States [49]; cardiac rehabilitation in Canada [73]; stroke rehabilitation in Italy [50]; and rehabilitation for nontraumatic spinal cord injury in the United States [51]. Only one study, on audiological rehabilitation in Sweden, reported women as experiencing greater functional gain than men [57]. Although women as a whole experienced less functional gain than men during rehabilitation, there were subareas where women experienced greater functional outcomes than men. In a United States-based study on stroke rehabilitation, for example, women discharged with better functional scores than men in expression and social interaction, eating, and bladder control, while men discharged with higher scores in upper extremity dressing, movement from one surface to another, walking, and stairs than women [49]. In another United States-based study about degenerative spinal disease, women were more independent at discharge than men in self-care and mobility, perhaps because men entered rehabilitation with lower functioning after having waited longer than women to seek care due to widespread masculine social norms that hinder men’s care seeking [51].

Studies exploring gender and/or sex differences in improvements in psychosocial outcomes reported mixed results. A Canadian study reported no sex differences in improvements in anxiety or depression following cardiac rehabilitation [73]. Another Danish study reported men as having more emotional unmet needs than women following cancer rehabilitation [84], while a Canadian study reported women’s many unmet psychosocial needs following cardiac rehabilitation [74]. Several articles highlighted the potential of women-only programs to confer an advantage for psychosocial and other health outcomes compared to mixed-sex cardiac rehabilitation groups [25, 69]. A commentary suggested that existing psychosocial measures may not incorporate gender differences in men’s expressions and experiences of anxiety, depression, or other mental health issues, which limits researchers’ ability to understand men’s post-rehabilitation psychosocial outcomes or unmet psychosocial needs [23].

The scoping review yielded limited literature exploring sex and/or gender differences in post-rehabilitation functioning, which the International Classification of Functioning, Disability, and Health (ICF) model describes as “all body functions, activities and participation” [89]. This scoping review only yielded studies on differences between men and women in return to work rates following rehabilitation, revealing a gap in the literature on people’s post-rehabilitation functioning in other social contexts. One study reported higher return to work rates among women with acquired brain injury in Norway, perhaps because more men than women work in managerial positions that require greater cognitive and communication functioning [52]. Another study described Canadian women who returned to work more quickly than men after rehabilitation for musculoskeletal pain as overburdened by domestic and formal employment responsibilities [45]. The above examples suggest how inequitable gender norms, roles, and relations play an important role in social participation and physical functioning following rehabilitation.

The scoping review revealed gender differences in AT use, which are influenced by gender inequities and norms, with conflicting results in HICs and LMICs. In HIC-based articles, women used mobility devices [44, 86, 87] and vision devices [87] more than men, while men used hearing devices more than women [87, 88]. A Canadian study suggested that men’s lower mobility device usage may be due to the perceived stigma associated with using such devices [44]. The same study highlighted that while women use mobility devices more than men, they are more likely to report unmet needs for electric wheelchairs, walkers, lift devices, grab bars and other aids than men, which suggests that women’s higher usage may not reach their higher need for such devices, or that men may be less likely to report an unmet need for AT [44]. Another United States-based study similarly pointed out that while women use canes more than men, when analyses are adjusted for factors related to disability and health behavior, women are less likely than men to use canes [86].

Unlike in HICs, where women use most forms of AT more than men, men and boys in LMICs use AT more than girls and women [31‐34]. Studies in Ghana [32] and Sierra Leone [34] reported a disparity between women and men in assistive device usage, with significantly lower usage rates among women than men, although population-level rates of assistive device usage were overall very low. The Sierra Leone-based study reported that women who used lower-limb prostheses and orthoses were less satisfied with their devices than men [34]. Researchers in an India-based study about mobile device-based assistive technology for vision impairment, like screen readers, found that while women with vision impairments paid more for mobile devices and had less access to them than men, cellular devices increased their participation in the public sphere [33].

Women’s disproportionate global caregiving burden [18] seemed to hold true for rehabilitation, although the scoping review yielded only seven articles focusing on the gender dimensions of post-rehabilitation caregiving, all focusing on stroke patients and their caregivers in HICs [37‐43]. Existing literature tended to focus less on how gender dynamics shape the experiences of caregivers and more on how sex and gender categories, which were often conflated, influence caregiving measures like quality of life and caregiver burden. Scoping review literature was heteronormative and revealed that women both provide more care to and receive less care from partners who are men, in part because women face more societal pressure to provide care, live longer than men, have adverse health events that require rehabilitation at an older age, and tend to be younger than their partners [38, 39, 41]. Assuming a caregiving role brings significant changes to the structure of daily life [37], including uncertainty, unpredictability, the assumption of new roles and tasks, and social isolation [38, 40], which can lead to caregivers experiencing a lost sense of self [40], depressive symptoms [42], and/or caregiving burden [43]. Scoping review literature highlighted that women stroke caregivers are more likely to provide care to partners with severe disabilities [41], experience more disruptions to their daily activities [37], and have a higher rate of depression than caregivers who are men [42], although men stroke caregivers also experience caregiver burden [40] and a lower quality of life due to their experiences as caregivers [43]. One study from the review suggested that caregiver training prior to patient discharge reduced caregiver burden [43]. None of the articles focused on the effects of receiving care for patient recovery trajectories.