Authors: Sarah B. Lieber (1Division of Rheumatology, Hospital for Special Surgery, New York, NY, USA; 2Department of Medicine, Weill Cornell Medicine, New York, NY, USA), Jerad Moxley (2Department of Medicine, Weill Cornell Medicine, New York, NY, USA; 3Division of Geriatrics and Palliative Medicine, Weill Cornell Medicine, New York, NY, USA), Lisa A. Mandl (1Division of Rheumatology, Hospital for Special Surgery, New York, NY, USA; 2Department of Medicine, Weill Cornell Medicine, New York, NY, USA), M. Carrington Reid (2Department of Medicine, Weill Cornell Medicine, New York, NY, USA; 3Division of Geriatrics and Palliative Medicine, Weill Cornell Medicine, New York, NY, USA), Sara J. Czaja (2Department of Medicine, Weill Cornell Medicine, New York, NY, USA; 3Division of Geriatrics and Palliative Medicine, Weill Cornell Medicine, New York, NY, USA)
Categories: Article, chronic conditions, mediation analysis, physical activity, self-rated health
Source: Journal of applied gerontology : the official journal of the Southern Gerontological Society
Authors: Sarah B. Lieber, Jerad Moxley, Lisa A. Mandl, M. Carrington Reid, Sara J. Czaja
How negative self-perceptions of aging relate to physical activity (PA) in older adults with arthritis is unclear. We examined whether general health mediated the relationship between Awareness of Age-Related change (AARC) losses and PA. We analyzed baseline data from a randomized controlled trial of a PA intervention for adults ≥60 years who self-reported PA, AARC, general health, pain, and social support. We evaluated point-biserial correlations between PA and other factors in participants with self-reported arthritis and developed a mediation model incorporating AARC losses, general health, and PA. We observed significant correlations between PA and general health, social support, and AARC losses in this physically active sample. General health mediated the effect of AARC losses on PA. While negative self-perceptions of aging were associated with less PA in older adults with arthritis, their impact was attenuated by general health. PA interventions for this population may benefit by targeting perceived general health.
Physical activity (PA) is routinely recommended for adults with arthritis, including osteoarthritis and inflammatory arthritis such as rheumatoid arthritis (England et al., 2023; Gwinnutt et al., 2023; Kolasinski et al., 2020; Piercy et al., 2018). In addition to the benefits of PA in the general population, PA is associated with beneficial arthritis-specific effects, including reduced pain and improved physical functioning (Gwinnutt et al., 2022). Nevertheless, physical inactivity is common among mid-aged and older adults with arthritis (Austin et al., 2012). This may be increasingly true for individuals with arthritis as they age (Chmelo et al., 2013).
Commonly reported barriers to engaging in PA among adults with osteoarthritis and rheumatoid arthritis include pain, functional limitations, potential harms or perceived ineffectiveness of PA, other chronic conditions, and limited motivation or social support (Kanavaki et al., 2017; Veldhuijzen van Zanten et al., 2015). Older (vs. mid-aged) adults with osteoarthritis may be more likely to believe that painful arthritis is inherent to aging and do not expect to engage regularly in activities, such as walking, with advancing age (Appelt et al., 2007). Older adults with rheumatoid arthritis also have cited perceived poor general health as an impediment to PA engagement (Lange et al., 2019). Given the increasing incidence and prevalence of arthritis in older adults, the complex relationships between perceptions of aging and PA, including the impact of general health, represent an area that has received limited attention and deserves further exploration among older individuals with arthritis to inform future efforts at improving PA in this population.
In this exploratory study, we examined associations between PA and positive and negative self-perceptions of aging, general health, and individual characteristics (e.g., sociodemographic features and multimorbidity) among older adults with arthritis. Based on our findings, we also investigated whether general health mediated the relationship between negative self-perceptions of aging and PA, hypothesizing that perceived health status of older adults with arthritis could mediate the effect of an individual’s Awareness of Age-Related Change (AARC) losses and their ability and/or motivation to engage in PA (Kaspar et al., 2019).
This cross-sectional study analyzed baseline data from a randomized controlled trial (RCT) of a digital intervention designed to enhance PA versus an educational control at two sites. The protocol for the RCT has been described previously (Lieber et al., 2024). Written informed consent was obtained from all participants prior to enrollment.
This study was approved by the Weill Cornell MedicineInstitutional Review Board (#1806019352) and registered on ClinicalTrials.gov (ClinicalTrials.gov identifier NCT03538158; date of registration May 25, 2018).
Participants were ≥60 years of age, English speaking, demonstrated sixth-grade reading comprehension, and passed the Telephone Interview for Cognitive Status (Brandt et al., 1988). Those with cognitive (score <26 on the Mini Mental Status Exam (Folstein et al., 1975)) or visual (corrected or uncorrected visual acuity <20/40) impairment, ongoing participation in a structured physical exercise regimen, or other health conditions impacting participation in a PA intervention were excluded. Participants were recruited through classified advertisements, virtual flyers, community newsletters disseminated at senior centers and geriatric clinics, virtual presentations, and social media advertisements. Only participants who self-reported arthritis of any type were included in the current analysis.
Assessments included in the current analysis are detailed below.
Self-perceptions of aging were drawn from the 10-item short form of the AARC questionnaire, including gains and losses subscales addressing the impact of aging on one’s own health and physical functioning, cognitive functioning, interpersonal relations, social-cognitive/social-emotional functioning, and lifestyle engagement domains (Kaspar et al., 2019). Responses range from 1 (“not at all”) to 5 (“very much”). A mean for each subscale is reported, with a higher score indicating more perceived positive or negative age-related changes.
General health was drawn from the first question of the 36-Item Short Form Survey Instrument (SF-36): “In general would you say your health is…?” Responses range from 1 (“excellent”) to 5 (“poor”) (Ware et al., 1995).
PA was drawn from the Global Physical Activity Questionnaire (World Health Organization, 2005), a 16-item instrument encompassing activity performed at work, during travel to and from places, and recreational activities, along with sedentary behavior. PA levels are tabulated and transformed into metabolic equivalent (MET)-minutes/week; 150 minutes of moderate-intensity physical activity, 75 minutes of vigorous-intensity physical activity, or 600 met-minutes of moderate and vigorous physical activity per week meets United States (US) Centers for Disease Control and the World Health Organization PA guidelines (Piercy et al., 2018; World Health Organization, 2020).
Pain level and pain interference were drawn from questions 21 and 22 of the 36-Item Short Form Survey Instrument (SF-36) (Ware et al., 1995): “How much bodily pain have you had during the past 4 weeks?” and “During the past 4 weeks, how much did pain interfere with your normal work (including both work outside the home and housework)?” Responses range from 1 (“none”) to 6 (“very severe”) and 1 (“not at all”) to 5 (“extremely”) for questions 21 and 22, respectively. An average of both questions is reported.
Social support was drawn from the 12-Item Interpersonal Support Evaluation List, which includes 3 subscales (i.e., appraisal, belonging, and tangible) (Cohen et al., 1985). The score for each item ranges from 1 (“definitely false”) to 4 (“definitely true”), with reverse coding for some items and a total possible score of 48, with a higher score indicating more social support.
Sociodemographic and related features, including age, gender, race, ethnicity, and chronic conditions, were self-reported. Multimorbidity was defined by ≥2 chronic conditions, including arthritis.
Sample characteristics were tabulated using descriptive statistics. Point-biserial correlations were used to evaluate the relationships between self-reported PA and self-reported sociodemographic characteristics, multimorbidity, self-perceptions of aging, general health, pain level and interference, and social support. Negative self-perceptions of aging (independent variable), general health (mediator), and PA (dependent variable) were included in a mediation model, with race, ethnicity, and social support as covariates. The mediation model is displayed in Figure 1. For model estimation, 5000 bootstrap samples were employed to attain 95% confidence intervals, which were deemed significant if they did not overlap with 0. Analyses were performed with SPSS Statistics 29.
The sample consisted of 99 participants with self-reported arthritis at baseline. The sample’s mean age was 70.5 years (standard deviation [SD] 7.0), 81.8% were women, and 17.2% and 13.1% self-identified as Black or African American and Hispanic or Latino, respectively (Table 1). Multimorbidity was observed in 72.7%. Mean AARC gains and losses subdomain scores were 4.3 (SD 0.7) and 2.3 (SD 0.7), respectively. Mean general health and social support scores were 2.6 (SD 0.7) and 25.4 (SD 6.7), respectively. Mean total physical activity level was 2368.7 MET-minutes/week (SD 5822.7 MET-minutes/week), exceeding the generally accepted thresholds for adequate PA in the older adult population (Piercy et al., 2018; World Health Organization, 2020).
Moderately strong, statistically significant point-biserial correlations were observed between PA and general health (r = −0.29; p < .01), social support (r = 0.27; p < .01), and AARC losses (r = −0.27; p < .01) (Table 2).
The mediation model relating AARC losses and self-reported PA, as mediated by self-reported general health, with race and ethnicity and social support as covariates, is presented in Table 3 and Figure 1. AARC losses had a significant direct effect on self-reported general health (unstandardized ß = 0.25, 95% confidence interval (CI) [0.05, 0.45]). Self-reported general health had a significant direct effect on self-reported PA (unstandardized ß = −2315.12, 95% CI [−4184.8, −445.44]); in contrast, AARC losses was not significantly associated with self-reported PA (unstandardized ß = −471.06, 95% CI [−2320.80, 1378.69]). Overall, general health significantly mediated the negative effect of greater AARC losses on PA, that is, for each 1-unit increase in AARC losses, PA decreased by 582.6 met-minutes/week through the effect of general health on PA (unstandardized ß = −582.61; 95% confidence interval [−1502.88, −24.82]).
In this exploratory study, we investigated the relationship between self-perceptions of aging and PA, and the mediating effect of self-reported general health, among older adults with arthritis. Negative self-perceptions of aging were associated with less PA in older adults with arthritis in simple correlation models, but the impact of these negative self-perceptions was mediated by general health. These findings suggest that general health is an important mediator of the relationship between negative self-perceptions of aging and PA in older adults with arthritis.
PA confers multiple universal benefits, including protective effects against cardiovascular disease, cerebrovascular disease, diabetes mellitus, and some forms of cancer; PA also promotes enhanced cognitive and mental health and quality of life (Committee on the Public Health Dimensions of Cognitive Aging, Board on Health Sciences Policy, Institute of Medicine, 2015; Langhammer et al., 2018). In addition, PA confers numerous specific benefits to adults with arthritis. Multiple forms of PA, including aerobic and strength training, have been associated with less pain and enhanced physical functioning and health-related quality of life in individuals with osteoarthritis (Gwinnutt et al., 2022). Likewise, aerobic and strength training have been associated with less pain, better physical function, and lower disease activity in individuals with rheumatoid arthritis (Gwinnutt et al., 2022). Thus, the US Centers for Disease Control (Piercy et al., 2018), the American College of Rheumatology (England et al., 2023; Kolasinski et al., 2020), and the European Alliance of Associations for Rheumatology (Gwinnutt et al., 2023) have consistently recommended PA for adults with various arthritis phenotypes. Despite these recommendations, PA levels among older adults with arthritis remain suboptimal. For instance, in a recent cross-sectional study of 3,343 US adults ≥65 years of age with rheumatic diseases, low levels of PA were self-reported by 33% and 31% of participants with osteoarthritis and rheumatoid arthritis, respectively (Kumthekar et al., 2023). Thus, these and other data provide strong support for efforts designed to improve PA among those with arthritis (Sokka et al., 2008). Consistent with previous reports in the general population of older adults (Lohne-Seiler et al., 2014; Opdal et al., 2020), our findings support a correlation between PA and general health that may extend to other favorable outcomes in this population.
Our findings augment prior observations on the association of self-perceptions of aging with PA (Nakamura et al., 2022; Wurm et al., 2010) in the general population. For instance, in a longitudinal analysis of Health and Retirement Study participants with a mean age of 65 years (N = 13,752), those in the highest (relative to lowest) quartile of aging satisfaction according to 8 items compiled from the Philadelphia Geriatric Center Morale Scale and the Berlin Aging Study were found to have higher likelihood of frequent self-reported PA, as defined by > 1 time/week of vigorous or moderate PA (risk ratio 1.23, 95% CI 1.12–1.34) (Baltes & Mayer, 2001; Lawton, 1975; Nakamura et al., 2022). While adults with osteoarthritis have cited negative self-perceptions pertaining to their health, aging, and PA as a barrier to PA in semi-structured interviews (McKevitt et al., 2022), there has been limited evaluation of self-perceptions of aging in relation to PA using validated instruments in adults with arthritis. Interestingly, we found that negative, but not positive self-perceptions of aging were correlated with PA in older adults with arthritis. Similarly, a recent meta-analysis showed that the AARC losses (but not gains) subdomain was significantly associated with lesser physical wellbeing in the general population (Sabatini et al., 2020). In another study, positive attitudes toward aging were associated with better self-report physical health in the general population (Bryant et al., 2012), possibly due to differences in related, but distinct constructs of physical health. Further studies are needed in older adults with arthritis to clarify the relationship between self-perceptions of aging and PA, which could inform patient-centered PA interventions in this growing population.
Our findings further suggest that self-reported general health is a significant mediator of the relationship between negative self-perceptions of aging and PA in older adults with arthritis. We hypothesize that perceived health status of older adults with arthritis may mediate the effect of an individual’s AARC losses and their ability and/or motivation to engage in PA. For instance, negative self-perceptions of aging have been associated with higher likelihood of delays in seeking healthcare (Sun & Smith, 2017) such that older adults with arthritis with more negative self-perceptions of aging may have fewer encounters with healthcare providers and receive less counseling on PA. Nevertheless, alternative pathways linking self-perceptions of aging, PA, and perceived health status have been In a previous investigation of participants in the Ageing in Spain Longitudinal Study Pilot Survey with a mean age of 64.84 years (N = 1124), cognitive functioning and PA were found to mediate the relationship between self-perceptions of aging and perceived health status (Fernández-Jiménez et al., 2024). Similarly, among German adults 65–85 years of age with multimorbidity (N = 309), the relationship between positive self-perceptions of aging and self-rated health was mediated by PA (Beyer et al., 2015). Our framework would suggest that interventions to improve PA in older adults with arthritis may need to prioritize targeting perceived general health over self-perceptions of aging.
Our study has limitations. Due to the cross-sectional study design, causal inferences cannot be made. However, the directionality of our mediation analysis is supported by prior observations suggesting that negative self-perception of aging is associated with worsening physical functioning in older adults (Sargent-Cox et al., 2012). Although the AARC questionnaire does not include a general health domain, some of the items may be related to or impacted by an individual’s self-perception of general health, potentially explaining some of the mediating effect. Diagnosis of arthritis was self-reported, and we were unable to distinguish among arthritis types or sites of arthritis involvement. Given enrollment at two US-based academic centers, our findings may not be generalizable to individuals from other countries or those who are community-based. Participants in the study were physically active by self-report and had a high level of educational attainment and may not be representative of all older adults. Self-reported PA may be discordant from objectively measured PA (Prince et al., 2008), and self-reported health may differ from physician-reported health status. Collection of complimentary data on objective PA may be an informative future direction.
Our results underscore that self-perception of general health may be a significant mediator of the relationship between negative self-perceptions of aging and PA, suggesting that attention to perceived general health such as through self-care interventions and practices may enhance PA promotion efforts in older adults with arthritis (Narasimhan et al., 2023). Further research is needed in diverse longitudinal cohorts to clarify the potential mechanistic pathways linking self-perceptions of aging, general health, and PA to inform the development of patient-centered PA interventions.