The Role of the Social Determinants of Health on Engagement in Physical Activity or Exercise among Adults Living with HIV: A Scoping Review

Physical activity (PA) and exercise are an effective rehabilitation strategy to improve health outcomes among people living with HIV (PLWH). However, engagement in exercise among PLWH can vary. Our aim was to characterize the literature on the role of social determinants of health (SDOH) on engagement in PA or exercise among adults living with HIV. We conducted a scoping review using the Arksey and O’Malley Framework. We searched databases between 1996 and 2021. We included articles that examined PA or exercise among adults with HIV and addressed at least one SDOH from the Public Health Agency of Canada Framework. We extracted data from included articles onto a data extraction charting form, and collated results using content analytical techniques. Of the 11,060 citations, we included 41 articles, with 35 studies involving primary data collection 23 (66%) quantitative, 8 (23%) qualitative, and four (11%) mixed methods. Of the 14,835 participants, 6398 (43%) were women. Gender (n = 24 articles), social support (n = 15), and income and social status (n = 14) were the most commonly reported SDOH in the literature with the majority of studies addressing only one SDOH. Future research should consider the intersection between multiple SDOH to better understand their combined impact on engagement in PA or exercise among PLWH.


Introduction
In Canada, approximately 62,050 people were living with HIV (PLWH) in 2018, with an incidence of 2242 new cases [1]. HIV prevalence has been increasing in Canada, due to a combination of increased incidence and lengthened life expectancy due to the advent of antiretroviral therapy (ART) [2,3]. As individuals are living longer, HIV is considered a chronic condition [4] characterized by health consequences related to HIV, aging and other health conditions, such as mental health issues [5], cardiovascular disease [6], opportunistic infections [7], and musculoskeletal disorders [8]. These health-related consequences can be defined as disability, including physical, cognitive, mental and emotional symptoms and impairments, which add challenges to carrying out daily activities, social inclusion, and worrying about future health and uncertainty [2,[9][10][11].
Int. J. Environ. Res. Public Health 2022, 19, 13528 3 of 37 Step 1: Identifying the Research Question We set out to answer the following research question: what is the nature and extent of evidence related to the relationship of SDOH on engagement in PA or exercise among adults living with HIV? We defined nature of evidence as the type of literature on the relationships between SDOH and in engagement of PA or exercise among adults living with HIV (e.g., study location, study design, relationship of SDOH and engagement in PA or exercise among adults living with HIV). We defined extent of evidence as the amount of literature on the relationships between SDOH and engagement in PA or exercise and their role with adults living with HIV (e.g., number of publications).

Step 2: Identifying Relevant Studies
We searched Medline, CINAHL, Embase, PsycInfo, SPORTDiscus and Web of Science to identify research studies pertaining to SDOH, PA or exercise and HIV. We used a two concept Medline search on PA or exercise and HIV, using both a combination of keywords and MeSH headings and validated search filter with modifications [28]. Supplementary search terms were added to capture research relevant to the PHAC framework on childhood experiences, healthy behaviors, biology and genetic endowment. We adapted the final Medline search for syntax and subject headings for the other databases. See Supplementary File S2 for the Medline search strategy.
We included articles that addressed (i) PA or exercise, (ii) adults living with HIV (18 years and older) and (iii) addressed at least one SDOH as conceptualized by the PHAC SDOH Framework. We included articles published from 1996 and onwards and were published in English language. We limited our search to 1996 onwards, as this is when Highly Active Antiretroviral Therapy (HAART) emerged resulting in prolonged survival of PLWH [29].

Step 3: Study selection
We imported citations (titles and abstracts) yielded from the search into EndNote referencing software Version X9 [30] (Clarivate: Philadelphia, PA, USA) and then into DistillerSR software Version 2.35 [31] (DistillerSR Inc., Ottawa, ON, Canada) to facilitate the process for screening citations and abstracts for study selection. We reviewed studies for inclusion in three stages, title review, abstract review, and full-text review.
Title review: Two reviewers (FS, NM) independently screened each title for the ability to address HIV, PA or exercise and SDOH by answering the following question: "Is the focus of the article to examine, explore, measure (or refer to) the associations of SDOH with PA and/or exercise?" with "Yes", "No" and "Unsure" options. If any two of three components (HIV, SDOH, and PA or exercise) were included in the title, reviewers responded "Yes". The article was excluded if both reviewers responded "No" to the above-mentioned question. If a reviewer responded "Unsure", the article was included in the subsequent stage for abstract review.
Abstract review: The same two reviewers (FS, NM) piloted the study selection criteria with five initial articles by independently screening five abstracts and answering the following questions, based on the inclusion criteria: process. Two reviewers (FS and NM) independently completed the abstract review. If both reviewers responded "Yes" to all of four inclusion questions, these articles were included in the full article review. If both reviewers responded "No" to any of the four inclusion questions, then the article was excluded. If any of the responses were "Unsure" by at least one of the reviewers for an article, the article was included in the subsequent stage of full-text review.
Full-text review: Two reviewers (FS and NM) independently reviewed full text of articles to determine inclusion. We piloted the process with 10 articles involving the four questions above for inclusion. All authors met to discuss the findings from the full text review. If any uncertainty arose pertaining to article inclusion, a third reviewer (KKO or SR) determined the final inclusion. Two authors (FS and KKO) contacted authors if we could not retrieve a full-text article or if data specific to PLWH were not included in the article. We refined our study selection criteria to include articles with aims or objectives that directly related to examining the relationship of SDOH on PA or exercise among PLWH (opposed to articles that did not a priori consider SDOH in their aims or objectives, but may have discussed SDOH as an emergent finding).

Step 4: Charting the Data
We extracted data from included studies for the following variables: title, author(s), year of publication, study location, city and country of lead author, study purpose, study objectives, study design, type of article (e.g., primary data collection, review article, etc.), data collection method, study participant characteristics (including number and proportion of women), types of outcomes measured, intervention (if applicable), conceptual framework used (if applicable), SDOH reported in the article, authors' results and conclusions, authors' results and conclusion specific to PA and SDOH, our reviewer interpretations, and additional notes. Two reviewers (FS and NM) extracted the data from included studies using a data charting form on Microsoft Excel [32] (Microsoft, Redmond, WA, USA). See Supplementary File S3 for the operational definitions of concepts extracted from the included articles.
We piloted the data extraction form in conducted two rounds. Two reviewers (FS and NM) independently extracted data from five included articles using the data extraction form to ensure consistency, followed by an additional five articles. Following each round of the pilot, the entire team met to refine the data charting form. Two reviewers extracted data from the 10 pilot articles using the data extraction form, which was reviewed by the team to ensure comprehensiveness and consistency in the data extraction process.
Step 5: Collating, summarizing and reporting the results We described characteristics of included articles using frequencies and percent for categorical variables and content analytical techniques for text data (e.g., author's results and conclusions related to PA and SDOH) [25,[33][34][35][36][37][38]. The research team met four times throughout this step to discuss our synthesis and reporting of results.

Results
We yielded 15,836 citations from the search strategy from CINHAL (n = 2690), Embase (n = 3463), Medline (n = 3423), Web of Sciences (n = 4641), PsycInfo (n = 1326), and Sport-Discus (n = 293) databases. After removing 4754 duplicates, this resulted in 11,060 citations. After screening title and abstracts, we retained 89 articles for full-text review. Of the 89 fulltext articles reviewed, 41 articles were included in the review. See Figure 1 Table 1 describes the characteristics of the included articles. Among the 41 included articles, 35 (85%) were research studies involving primary data collection [21, one (2%) was a secondary analysis [68], three (7%) were systematic reviews [69][70][71], and two (5%) were narrative reviews [22,72] (Figure 1).  Table 1 describes the characteristics of the included articles. Among the 41 included articles, 35 (85%) were research studies involving primary data collection [21, one (2%) was a secondary analysis [68], three (7%) were systematic reviews [69][70][71], and two (5%) were narrative reviews [22,72] (Figure 1).  The encouragement participants received from their families and community members helped them adhere to the exercises, and, they encouraged their children, partners, and neighbors to start exercising. Financial constraints limited access to institutional care and contributed to food scarcity, which affected full participation in the home-based rehabilitation intervention. An inhibitor to exercise was HIV stigma and, in some cases, additional discrimination associated with living with disability.     Quotations from the article: "the need for social support or peer support, but essentially trustful and confidential support, from a good friend"; "In the disadvantaged community, for example, women generally are not seen running in the street for exercise and health reasons. This situation might be due to a lack of safety and security in such areas, but also due to social-cultural norms and attitudes about women in the Black African community"; "Transport problems were mentioned, because it was a challenge to come to campus in holidays due to lack of funding ("The school is closed and my aunt is not giving me money for transport and in that way I can't be present at the gym")."    Cost was a barrier to initiating and maintaining exercise among participants living with HIV. "Well, first of all, it is the cost. They have many different fee structures which they won't advertise or let you know of, until they let you walking there and get you into a high pressure sales man." Motivating factors to initiate exercise was location/availability. "if it's not going to be convenient, I'm not going to do it." Social support was an important factor for motivating and continuing exercise.  Sedentary jobs prevented participants accumulating adequate steps, "The week was challenging in that I was working shifts and I have to sit on a chair the whole shift". When participants were busy at work, they were less likely to follow their program, "Tired after I was doing the house work and working in the shop the whole weekend". The state of the weather was frequently voiced as a barrier and complaints ranged from the weather being too hot, cold or raining a lot, "I am fine the weather is the problem". A social environmental barrier was the incidences of domestic abuse and crime that influenced participants' lives.An important facilitator to PA was the support and motivation received from friends and family.

Semi structured interviews
Exercise readiness described as a dynamic spectrum ranging from not thinking about exercise, to routinely engaging in daily exercise.
Participants described the importance of social support as facilitating readiness to exercise. Several participants indicated that having someone to exercise with would improve their willingness to engage in exercise. Some described how an HIV-specific exercise program would facilitate their readiness by creating a safe and inclusive environment, eliminating the challenges associated with disclosure.
When describing the conditions that influenced readiness to engage in exercise, most participants expressed the importance of accessibility. For some, a perceived lack of financial accessibility created obstacles to engagement and hindered their readiness to exercise. Higher educational level was associated with higher physical activity levels in 6/7 studies. Gender differences were inconsistently reported, i.e., while six studies indicated men engaged in more physical activity than women, another reported the opposite, while eight other studies showed no difference between genders. While one study reported a higher physical activity levels in the non-white population, another reported lower levels and four studies reported no associations. Having a manual labor versus non-manual labor job and a higher annual income were, all in one study significantly associated with a higher physical activity level while only one of two studies found that having a job was associated with more physical activity. Two of three studies (67%) reported on social support as a potential positive correlate to PA.  week (p = 0.48). No statistically significant differences in the frequency of exercise between men and women participants living with HIV. Removing low-intensity walking significantly decreased the average amount and number of bouts of exercise per week for men and women (all, p < 0.01). These findings indicate that women living with HIV may have access to more exercise resources than men or that they are more likely to take advantage of resources, resulting in higher intensity, more balanced exercise patterns.   [34,35,37,38,40,42,44,45,[48][49][50][51][52][53][54]57,59,[61][62][63][64][65][66], eight (23%) qualitative [21,36,39,41,47,55,56,58], and four (11%) mixed method studies [43,46,60,67]. Among the 23 quantitative studies, 18 (78%) studies were cross-sectional [34,37,38,40,42,44,45,[48][49][50][52][53][54]59,[62][63][64]66], two (9%) cohort studies [35,61], two (9%) intervention studies [51,57], and one (4%) was a randomized controlled trial (RCT) [65].

Characteristics of Participants
In this section we describe characteristics among studies involving primary data collection. Two studies included the same study population within the same study, hence we report characteristics of the sample from one article only [38,63] resulting in a total of 34 research studies in this review. See Table 1 for characteristics of participants in included studies.
Among the 14,835 participants in the 34 research studies, sample sizes ranged between 8 and 8104 participants. The study population were PLWH in all but one study, where 30 out of 59 participants (51%) were HIV positive and the remainder were HIV negative [67]. Among the 34 studies, four (12%) included women only [37,43,51,59] and one (3%) study included men only [55]. Of the total sample of participants in the 34 studies (n = 14,835), 57% (8437) were men and 43% (6398) were women.

Evidence Related to Relationship of Social Determinants of Health with Physical Activity and Exercise
Nine out of 12 SDOH were addressed across the 41 included articles. The number of SDOH addressed in each article ranged from one [34,37,39,40,[43][44][45]49,50,53,56,[64][65][66] to eight [70] ( Table 2). The most common SDOH addressed across included articles was gender (n = 24; 47%), followed by social support and coping skills (n = 15; 36%), income and social status (n = 14; 34%), education and literacy (n = 9; 22%), employment and working conditions (n = 9; 22%), physical environment (n = 9; 22%), healthy behaviors (n = 9; 22%), race/racism (n = 6; 15%), and culture (n = 5; 12%). (Table 2; Figure 3). Three SDOHs (childhood experiences, access to healthcare, and biology and genetic endowment) were not addressed in any of the included articles. None of the articles utilized a pre-determined SDOH framework to report their findings.  Given the heterogeneity of the study aims, study design and SDOH addressed across the articles, we describe results of the individual articles for each of the SDOH below.    Given the heterogeneity of the study aims, study design and SDOH addressed across the articles, we describe results of the individual articles for each of the SDOH below.

Gender
Gender was reported among 24 (58%) of the 41 articles with mixed results [34,35,37,42,[44][45][46]48,50,[52][53][54]57,59,61,62,[64][65][66][67][68][69][70][71]. Several cross-sectional studies reported that on average, women living with HIV engaged in lower levels of PA than men living with HIV. For example, Chisati et al. (2020) reported 51% of women participants had low PA levels compared to 22% of men with HIV with less than half of women engaged in high intensity PA compared to men (16.7% vs. 37%) [34]. One cross-sectional study collected PA data by using the Global Physical Activity Questionnaire (GPAQ) and found men living with HIV engaged in more PA (overall PA score on GPAQ 480.3 min/week) than women (overall PA score 269.1 min/week) in work-related, transport-related, and leisure-related PA, but especially in work-related PA (men: 279.3 min/week, women: 125.9 min/week) [52]. Additionally, another study reported women had higher odds of not adhering to PA recommendations compared with men (Odds Ratio (OR) 1.62; 95% confidence interval (CI): 1.01 to 2.57) [38]. Muronya et al. (2011) reported greater physical inactivity among women than men (28% vs. 25%); however, gender was not statistically associated with PA [53]. A cross-sectional study found that women living with HIV engaged in less PA than men, but only during middle adulthood (36-50 years) (women: 2.4 h of exercise/ week, men: 4.5 h/week p < 0.05) [64]. Two cohort studies reported similar findings, where men were more physically active than women [35,61]. One RCT aimed to evaluate the 3-and 6-month effects of an intervention (System CHANGE) on PA and dietary quality in PLWH at high risk of developing cardiovascular diseases (CVD) and reported women were consistently associated with engaging in less PA compared with men [65]. Wright et al. (2020) found that men engaged in more intentional PA, but women engaged in more PA related to activities of daily living, indicating that the definition of and type of PA was correlated with gender differences [67]. One study found a low level of PA among women from low SES, with limited time spent in vigorous activities among groups of women from Rio Di Janeiro (3 min/day on average) and Maputo (5 min/day on average), with daily light PA higher among women living in Rio (111 min/day) compared with Maputo (84 min/day). However, data were not compared to PA among men [37]. Dang et al. (2018) was the only study to find that women gender was weakly associated with higher PA levels as measured by the International Physical Activity Questionnaire (IPAQ) (correlation coefficient 0.25; 95% CI: 0.11, 0.39) [42]. Several studies found no significant associations between gender and PA level [44,50,57]. Two systematic reviews could not conclude whether men or women engaged in more PA due to inconsistent findings among included studies [70,71]. Another systematic review and meta-analysis investigated the prevalence and predictors of treatment dropout in PA interventions in PLWH and reported a lower proportion of men participants (ß = 1.15, SE = 0.49, z = 2.0, p = 0.048) moderated the higher dropout rates [69].

Social Support and Coping Skills
Social support and coping skills, and PA among PLWH were reported in 15 (37%) of the 41 articles [21,22,36,38,40,41,47,49,51,55,[58][59][60]63,70]. Fourteen of the 15 studies found that social support was an important facilitator of PA among PLWH. Gray et al. (2019) conducted a qualitative study with active and inactive PLWH in which they found a lack of social support to be a barrier in engaging in PA [47]. Another qualitative study conducted in Canada reported social influence, such as encouragement from friends and health care providers, as a facilitator to PA [58]. A narrative review and a qualitative study reported that having an exercise partner or encouragement from staff members at the community facilities, such as churches or community centres increased motivation and adherence to PA [21,22]. Similarly, another qualitative study reported that encouragement received from families and community members helped participants adhere to exercise, and, in many cases, encouraged their children, partners, and neighbors to start exercising [41]. Authors reported that support from family members was a facilitator for exercise. For instance, a cross-sectional study measuring benefits and barriers to PA among PLWH in the United States showed that not receiving encouragement from a spouse, significant other, or family member was a perceived barrier to engagement in PA [59].
The social setting where PLWH were engaging in PA was also important. Mabweazara et al. (2018) found that community/group exercise settings helped PLWH stay motivated within a cost-effective environment that facilitated social-bond formation [22]. This was especially useful among PLWH since stigma often led to lower levels of social support. Other studies found social support to be an important facilitator to engagement in PA, including Neff et al. (2019) who reported "social support was an important factor for motivating and continuing exercise" [55]; Ley et al. (2015) reported "the need for social support or peer support, but essentially trustful and confidential support, from a good friend" [51]; and Roos et al. (2015) reported "an important facilitator was the support and motivation received from friends and family through participating in sessions and walking with participants" [60]. In summary, social support including encouragement from friends, family members, staff at the gym, and peers in group exercise all appeared to be important facilitators to engaging in PA among PLWH.

Income and Social Status
Fourteen (34%) of the 41 articles examined relationships between income and social status on PA among PLWH [21,22,36,39,41,47,51,55,58,59,62,67,70,72]. All 14 articles reported that financial constraints or costs were a barrier to participation in PA among PLWH, within different contexts and geographic locations. For example, a qualitative study among older exerciser and non-exerciser groups reported both groups identified costs as a barrier to exercise [36]. Similarly, two qualitative studies among older adults living with HIV found cost was a barrier to PA [55,58]. A mixed method study found expense as a barrier to exercise in both HIV positive and negative groups [67]. Two of the studies focused on women from disadvantaged urban settings in South Africa [51] and from the southern United States [59] and reported similar results. Examples of financial barriers were costs of a gym membership [39] and transportation to PA facilities [51]. Financial insecurity also led to food scarcity, which was a barrier to exercise among PLWH in Kwazulu-Natal South Africa [41]. A systematic review incorporated one study in their review that highlighted the association between higher income and higher PA [70].

Education and Literacy
Thirteen (32%) of the 41 articles reported on education and literacy, and PA or exercise among PLWH with varying results [35,38,42,46,48,52,54,[61][62][63]68,70,72]. For example, authors of a cohort study reported PA levels did not correlate with years of education (r = 0.09, p = 0.61) [35]. Mabweazara et al. (2018) reported education was not a significant predictor (p = 0.057) of overall PA among PLWH [22], and Dang et al. (2018) found that high school education did not differ regarding PA compared to PLWH with lower levels of education (OR 1.52, 95% CI: 0.94, 2.46) [42]. Another study reported no significant difference in PA stages of change related to education level [63]. Hsieh et al. (2014) found higher levels of PA were significantly associated with lower education (OR: 0.50, 95% CI: 0.27, 0.91) among Chinese individuals living with HIV [48]. Alternatively, some studies found the association between lower levels of education and lower PA level among PLWH. For instance, a study reported participants who never attended school were less likely to engage in physical exercise compared to those who had secondary or higher education levels (Adjusted Odds Ratio (AOR): 0.22; 95% CI: 0.08, 0.55; p = 0.001) [54]. Similarly, another study found low education (up to 4 years of study, prevalence ratio 1.71) was associated with physical inactivity [62]. A narrative review conducted in South Africa highlighted physical inactivity and obesity were strongly related to lower education level [72]. Nevertheless, a few studies found an association between higher education and higher levels of PA. Mabweazara et al. (2021) showed that educational attainment (β = 0.127; p < 0.001) significantly predicted total moderate-to-vigorous PA among PLWH of low SES in Cape Town, South Africa [68]. Another study reported, participants with higher education reported more free-time PA than those with lower education among PLWH from the Swiss HIV Cohort Study [61]. A mixed methods study examining patients' preferences for HIV treatment in Colombia found higher educated patients identified PA as the most important HIV treatment attribute, whereas low educated patients valued accessibility to clinic over PA [46].

Employment and Working Conditions
Nine (22%) of the 41 articles reported on the relationship between employment and working conditions and PA among PLWH [38,42,48,52,60,63,[68][69][70]. Results from these individual studies suggest that employment was associated with engagement in PA among PLWH, with higher levels of PA and exercise among those in non-sedentary jobs. A crosssectional study showed that higher levels of PA was significantly associated with a higher likelihood of manual labor versus non-manual labor occupation (p = 0.002) [48]. Another cross-sectional study reported unemployed PLWH had a 2.81 (95% CI: 2.00, 3.94) higher odds of not adhering with PA recommendations compared with employed PLWH [38]. A cross-sectional study found that blue-collar workers or farmers (OR 2.24; 95% CI: 1.27, 3.95) were more likely to have a higher International Physical Activity Questionnaire (IPAQ) score and were classified as physically active compared with white-collar workers [42]. Similarly, a RCT found that sedentary jobs prevented PLWH from achieving adequate step counts, as well as following their PA programme outside of work due to fatigue and busy schedule among PLWH in South Africa [60]. A cross-sectional study reported employment was a significant predictor of overall PA [22]. Conversely, a cross-sectional study found no difference in employment status between the different stages of change in regard to PA [63]. One meta-analysis revealed dropout rates in PA interventions in PLWH were not moderated by employment status (β = −1.81, 95% CI, −5.69, 2.06; p = 0.36) [69].

Physical Environments
Thirteen (32%) of the 41 articles reported on the relationship between physical environment and PA or exercise among PLWH [21,36,47,51,55,56,[58][59][60]67,[70][71][72]. Physical environment often was described as either a barrier or facilitator to participating in PA. A qualitative analysis described how physical environmental factors such as gym culture, feeling unsafe in their neighborhood, or lack of affordable housing were barriers to PA among PLWH [36]. Nguyen et al. (2017) also found in their qualitative inquiry that perceived safety or culture in the gym environment was a barrier, specifically pertaining to exercising in "non-positive" spaces due to fear of stigma [56]. Another qualitative study reported HIV-positive exercise programs were a facilitator to PA, by creating a safe and inclusive environment [21]. A mixed methods intervention study reported a lack of safety and security in disadvantaged areas as a barrier to PA among South African women living with HIV [51]. Another mixed methods study found safety concerns was a common barrier to PA among PLWH in Uganda [67]. In their narrative review, Ley and Barrio (2012) reported that limited space and desolate living conditions failed to offer suitable environments to engage in PA in South Africa [72]. Two qualitative and one mixed method studies found that climate constraints were a barrier to engaging in PA [47,58,60]. Lastly, a qualitative study among older PLWH reported that convenience of location to an exercise facility was a motivating factor to initiate exercise [55]. While the SDOH "physical environments" can refer to many different concepts, common themes were gym culture, safety, and climate.

Healthy Behaviors
Nine (22%) of the 41 articles reported on healthy/unhealthy behaviors in relation to PA among PLWH that largely concluded smoking, and substance use did not have an influence on engagement in PA [36,38,48,54,[61][62][63]69,70]. We did not include "physical activity" as a healthy behavior since it was our outcome of interest. Hsieh et al. (2014) found that smoking (OR: 1.49, 95% CI: 0.81, 2.77), and alcohol use (OR: 1.56, 95% CI: 0.75, 3.21) were not significantly related with PA [48]. Similar findings were reported in a cross-sectional study where no association was found between smoking (p = 0.06) and alcohol use (p = 0.15) and adherence to PA [38]. One qualitative study reported drug use as a barrier to PA and exercise [36]. Only one systematic review examined drop out rates among PLWH engaged in PA or exercise and reported smoking status did not moderate dropouts rate (β = −1.78, 95% CI, −3.67, 0.11, p = 0.07) [69].

Culture
Five (12%) of the 41 articles reported on culture with PA or exercise among PLWH [41,43,51,71,72]. A narrative review and one mixed methods intervention study examined the cultural barriers to exercise among women in the Black African community [51,72]. "In the disadvantaged community, for example, women generally are not seen running in the street . . . due to social-cultural norms and attitudes . . . " [51]. A mixed method study also discussed cultural perceptions of exercise as a barrier for PLWH engaging in PA: "many participants were apprehensive about exercise in terms of their cultural practices and were not sure if this was something they would really do" [43]. Finally, a qualitative study examined stigma as a cultural factor as a barrier to engaging in PA [41].

Race/Racism
Six (15%) of the 41 articles reported on race and racism among PLWH with mixed results [35,57,62,69,70,72]. Two studies showed no race difference in PA. For instance, a longitudinal study in the United States found PA did not differ by White vs. non-White PLWH [t = 0.12, p = 0.91] [35]. In a systematic review, authors found that the dropout rate from PA programs was not moderated by race or ethnicity among PLWH [69]. Authors of a narrative review found Black women living with HIV experienced barriers to engaging in PA due to cultural perceptions, addressing the intersection between gender, race and culture [72]. An observational intervention study reported the proportion of Black African PLWH among a group non-adherent with PA (73% vs. 9% White British) and the White British (46% vs. 27% Black African) in a group adherent to PA were significantly higher. Authors suggested further investigation on the relationship between race and PA among PLWH [57]. In a systematic review, authors could not ascertain a relationship between race and PA among PLWH among the included studies as one study showed higher PA among the non-White PLWH, while another showed the opposite and four studies showed no association between race and engagement in PA or exercise [70].

Discussion
To our knowledge, this is the first scoping review that examined the nature and extent of evidence on the relationship between SDOH and engagement in PA or exercise among adults living with HIV. Our findings indicate that gender was the most common SDOH among included articles, followed by social support and coping skills, and income and social status. Most included articles (76%) explored the relationship between SDOH and PA or exercise in PLWH as a primary aim, whereas the rest of the articles (24%) reported SDOH as a secondary aim. Among the 23 quantitative studies, the majority (78%) used a cross-sectional design that prevented establishing a causal relationship between the SDOH of interest and engagement in PA or exercise. Eight qualitative studies highlighted socio-environmental factors as barriers and facilitators of PA or exercise, with financial cost, social support, cultural context, and physical environment (e.g., gym environment, safe place, and weather) as the most common SDOH influencing engagement in PA or exercise among PLWH.
Despite the explicit focus of articles examining relationships between different SDOH and PA or exercise among PLWH, none of the articles used a SDOH framework to inform their analytical approach. In addition, a large number of studies examined only a single SDOH [34,37,39,40,[43][44][45]49,50,53,56,[64][65][66], while only one systematic review evaluated eight SDOHs identified in the PHAC framework [70]. Identifying more than one SDOH is important to evaluate the intersection between different determinants. Although, some studies evaluated more than one SDOH in relation to PA or exercise (Table 2), separate results were reported for each determinant. In two studies, women from low SES [37] and Black African women from disadvantaged communities were found to engage in low PA or exercise [51]. None of the other studies appeared to report the impact of intersectionality on PA or exercise. This scoping review indicates an evidence gap in the literature as childhood experiences, biology and genetic endowment, and access to health services were not reported in the included articles, highlighting the need for future research to address these determinants and and the impact of combinations of SDOH among PLWH engaging in PA or exercise.
Results from individual studies in this scoping review suggest that women living with HIV engage in less PA compared to men living with HIV. Only one study reported higher PA among women [42], whereas other studies reported non-significant findings. However, women were underrepresented in most of the included studies. Gender remains an important SDOH to consider in the context of PA and exercise among PLWH. While gender was the most common SDOH addressed in the review, genders other than men and women were underrepresented. Only five studies included other genders [46,49,56,65,66], among which one study reported combined female and transgender data [56] (Table 1). Moreover, some studies utilized sex and gender interchangeably. Transgender groups are at higher risk of HIV, specifically transgender women have 49 times the odds of having HIV compared to the general population [73]. Thus, future studies should consider other genders in addition to men and women on engagement in PA or exercise among PLWH.
Our review suggests no association between unhealthy behaviours such as smoking and alcohol use with engagement in PA. Smoking and alcohol consumption are not considered as healthy behaviors because of their adverse impacts on physical and mental health [74,75]. Interestingly, a cross-sectional study from Brazil reported a higher level of PA among young, middle aged and older adults who were weekly alcohol consumers [76]. We recommend evaluating the relationship between healthy or unhealthy behaviours and PA or exercise in larger population-based studies involving PLWH.
Our findings suggest that stigma associated with culture and race/racism were underrepresented in the literature, with only 12% and 15% of included studies addressing these SDOH, respectively. While one systematic review was inconclusive regarding the relationship between race and PA [70], authors of one narrative review suggested engagement in PA was lower among Black women due to perceived barriers [72]. Some literature was suggestive of an association between White race with higher PA in general population. For example, one study conducted a secondary data analysis from the National Health and Nutrition Examination Survey of 9472 adolescent and young adult respondents from 2007 through 2016 reported White race and higher income were associated with greater PA in adjusted models [77]. The Centre for Disease Control and Prevention (CDC) conducted a Behavioral Risk Factor Surveillance System (BRFSS) among 52 United States jurisdictions from 2017 to 2020, and showed lower prevalence of physical inactivity outside of work among non-Hispanic White (23%) compared to non-Hispanic Black (30%) participants [78]. Based on the 2003-2009 American Time Use Surveys (ATUS) data, Saffer et al. (2019) showed non-work PA was significantly lower among Blacks relative to non-Hispanic Whites [79]. Conversely, total work related PA was significantly higher among Black peoples compared to White and Asian peoples given the context that Black individuals were more likely to work in blue-collar and physically demanding occupations [80]. Although these findings involve the general population, the association between race and PA engagement among PLWH needs further evaluation. We recommend examining the intersection between race, gender, employment, education, and culture among PLWH to better understand engagement in PA or exercise.
This review identified gaps in the literature related to SDOH and engagement in PA or exercise among HIV. Definitions of PA and exercise varied significantly across the studies. None of the included studies adopted a SDOH framework as the theoretical basis for their work. Additionally, authors used a variety of self-reported questionnaires and objective measures of PA and exercise. Considering the diversity in PA definition and assessments, and benefits and limitations of objective versus self-report of PA, researchers should consider the inclusion of an operational definition of PA or exercise in future work and provide a rationale for the selection of the PA or exercise assessment tools. Similarly, authors examining SDOH should better define the determinant(s) of interest and the research context. Clearly defined PA, exercise and SDOH variables using a conceptual framework to guide the analytical work and standardized and validated outcome measures should improve the interpretation of evidence in this field and foster standardized approaches in future research. Furthermore, the majority of studies were cross-sectional in nature making it difficult to identify the causal relationship between SDOH and PA or exericse. Future longitudinal or cohort studies among PLWH, may help to further understand the strength and direction of the relationships and whether PA or exercise is a moderator or mediator of health outcomes among PLWH. Another major gap in the evidence is the absence of research on the impacts of the intersection between multiple SDOH on PA and exercise. Finally, many studies included small sample sizes with unclear transferability to the broader HIV population, highlighting the need for a richer understanding of the collective influence that a combination of SDOHs have on engaging in PA or exercise among PLWH.
Other research has explored the relationship between SDOH and physical activity or exercise among other chronic illness populations. This body of work has also documented that the ability for individuals to engage in exercise is affected by cost, health status and disability [80]. Social determinants research in other chronic conditions found similar barriers and facilitators of engagement in PA. For example, people living with diabetes with a low socioeconomic status and lower education level, found it difficult to engage in PA regularly [81]. In a study conducted with adults living with multiple chronic conditions, an unsafe environment, which attributes to low socioeconomic status and housing were correlated with lower engagement in PA [82]. One qualitative study among people with type 1 diabetes who progressed to end-stage renal disease identified lower socioeconomic status, lower education level, and female gender as barriers to PA [83]. Social support and inclusion were also determinants and found to facilitate and encourage engagement in adults living with chronic conditions [84].
Strengths of this scoping review includes our rigorous adherence to well-established scoping review methodology [24,25,27], our broad selection criteria involving a range of study designs and methodologies, and our team based approach involving independent review and multiple piloting of the inclusion and data extraction process. We did not assess quality of the included studies, as this is not a requirement of scoping reviews but rather to provide an overview of the evidence [25]. Potential limitations of our review is that we may have missed relevant articles despite our extensive search strategy of multiple databases. We contacted the corresponding authors of titles and abstracts without full-text articles and waited up to 4 weeks before excluding them. We only included English language articles because of the cost and time associated in translating foreign materials. Despite these limitations, this review provides important insight into the nature of extent of evidence on the role of SDOH on engagement in PA or exercise among PLWH.

Conclusions
This scoping review characterized the evidence examining the role of SDOH on PA or exercise among PLWH. Gender was the most commonly reported SDOH in the literature with the majority of studies addressing only one SDOH. Results from individual studies suggest that social support is a facilitator, and financial constraints and costs are a barrier to engagement in PA or exercise among PLWH. Results of this review may help to inform clinicians, researchers, and policy makers to better understand the role of SDOH as potential barriers or facilitators while promoting or evaluating PA or exercise with PLWH. Future research should consider the intersection between different determinants to better understand the combined impact of different determinants on engagement in PA or exercise.