Physical activity behavior in older women with rheumatoid arthritis




To identify physical activity behaviors of older (≥60 years) women with rheumatoid arthritis, and to determine if physical activity behavior differed by demographic characteristics or disease duration.


Cognitively intact, ambulatory women with rheumatoid arthritis seen at an urban university arthritis center (n = 185) responded to a survey that included demographics, number of years since RA diagnosis, and physical activity assessed with the Yale Physical Activity Survey.


Participants had a mean age of 70 years and a mean disease duration of 17.6 years. Mean total of physical activity, including low-, moderate-, and high-intensity activity, was 23 hours/week, with 47 different physical activities identified. Housework comprised 67% of total physical activity time, whereas leisure activities and planned exercise comprised only 15% and 10%, respectively. Sixty percent of the women reported participation in some type of vigorous activity, and 88% reported they had done leisure walking in the past month. Age was negatively associated with scores on the activity summary index (r = –0.195, P < 0.01), and there was a significant difference on the activity summary index by employment status, with women in the work force (n = 47) and those involved in housekeeping (n = 105) scoring significantly higher (P = 0.003) than the women (n = 33) who described themselves as retired (F = 7.81, 2 degrees of freedom, P = 0.001).


Older women with rheumatoid arthritis may participate in a broader array of physical activities than previously assumed, and could benefit from increasing the proportion of moderate or vigorous activity incorporated into their daily routines.


Older women who show radiographic worsening of their rheumatoid arthritis (RA) and develop functional limitations secondary to immobility early in the disease course are at risk of becoming disabled (1). Performance of regular exercise is a widely recommended self-management strategy that has been shown to be effective in the amelioration of RA disease consequences and the preservation of musculoskeletal function in those diagnosed with RA. Recommended exercise strategies for those with RA have progressed from active and passive range of motion and rest (2) to more vigorous forms of physical training programs that could improve the physical performance capacity of those with RA without increasing their disease activity or causing deterioration of their joint status (3–5).

The major problem associated with exercise regimens in both the general population and among persons with RA has been that people don't do them. Up to 70% of older adults report having no regular exercise routine, and 60% report that they have not walked a mile or more in the past year (6). Those persons 75 years of age and older are especially inactive. Even more disturbing are the estimates that leisure-time physical activity rates among those with arthritis are 7% lower than for adults without arthritis (7).

In 1995, exercise specialists from the American College of Sports Medicine and Centers for Disease Control (ACSM/CDC) came together to issue a revised public health directive as evidence mounted that the health benefits (largely cardiovascular) associated with these formal exercise regimens also occurred with accumulated lifestyle physical activity. Their revised recommendation was for the accumulation of 30 minutes of moderate-intensity lifestyle physical activity in short segments of activity on most, if not all, days of the week (8). Although this goal may seem relatively negotiable, it has been estimated that only 22% of adults engage in at least 30 minutes of light to moderate physical activity per day (9), with women over the age of 50 representing the most sedentary segment of the population (10).

Recent data, however, indicate that as many as 75% of women over age 40 may meet the 30 minutes/day guideline when all domains of activity (household, leisure, occupational) are taken into account (11). The incorporation of physical activity into daily household and work activities, as well as leisure-time activities, might be particularly appealing to older women with RA who must overcome fatigue, joint pain, and limited energy resources in addition to the usual activity performance barriers faced by those without a chronic illness. Indeed, the American College of Rheumatology (ACR) embraces the current ACSM/CDC accumulated physical activity recommendation for those with RA (12).

Little is known, however, about the types and amounts of physical activities that may already be incorporated in the lifestyles of those with RA, but it is expected to be low enough such that interventions to increase lifestyle physical activity behavior might greatly impact health status and quality of life, while reducing subsequent health care costs in this population. At least 1 study found that older adults (≥65 years), not selected for arthritis, significantly improved their ability to perform the sit-to-stand test after a 16-week intervention to increase lifestyle physical activity (13). Previous studies of RA populations have predominantly examined the safety and efficacy of structured exercise programs organized in formal facilities (4, 14, 15).

Thus the purpose of this study was to identify the lifestyle physical activity behaviors (household, leisure, and occupational) of community-dwelling older (≥60 years) women with RA and to determine if physical activity behavior differed by demographic characteristics (race, education, marital status, income, employment, and age) or disease duration. Understanding the types of physical activities already practiced by women with RA is the crucial first step in the development of interventions to increase physical activity levels in this population.


This was a descriptive, cross-sectional study performed during the years 2001–2002.


Study subjects were English-speaking cognitively intact women, 60 years of age or older, able to walk household distances, and diagnosed by a board-certified rheumatologist as having rheumatoid arthritis according to the ACR (formerly American Rheumatism Association) 1987 revised criteria (16). Subjects were obtained from the caseloads of 8 rheumatologists practicing at 2 sites within a large urban university medical center, serving patients with widely varied socioeconomic and racial backgrounds.

Medical records of all 500 rheumatology patients at clinic site 1 were reviewed, producing a list of 64 patients who met the study's sex, age, and diagnosis criteria. At clinic site 2, the larger of the 2 sites, a computerized database search yielded 343 subjects who had diagnosis code 714.0 (rheumatoid arthritis) and met the sex and age criteria. The listings from these 2 clinic sites provided an overall sampling frame of 407 subjects, 18% (n = 72) of whom were either non-English speaking, cognitively impaired, or too functionally limited to be included in the study. Of the remaining 335 women, 76 (23%) declined, and 39 (12%) were not able to be reached. Two hundred and twenty subjects agreed to participate, with 185 (84%) completing the survey.

The 185 participants were compared on age and racial characteristics with the 150 women who were either unable to be reached, declined, or withdrew from the study. Participants were significantly younger (70 years versus 73 years) than nonparticipants (t = 3.452, degrees of freedom [df] = 333, P = 0.001), but the 2 groups did not differ in racial composition (χ2 = 7.80, df = 5, P = 0.168). Participant groups from the 2 separate clinic sites were also compared on age, race, educational level, and marital status. No significant differences were found on any of these demographic characteristics, so the 2 groups of participants were combined for analyses.

The 185 participants were predominantly white (72%), not employed (75%), and with a mean age of 70 years (range 60–88 years). For those who offered income information, 46% reported incomes less than $30,000. Although only 45% of the women were married, 61% reported currently living with at least 1 other adult. Just less than half (45%) of the women were college educated. The average RA disease duration was 17.6 years (range 0.6–66 years, SD = 12.6).

Physical activity

A comprehensive description of the 2-part interviewer-administered Yale Physical Activity Survey (YPAS), developed specifically for the measurement of physical activity in epidemiologic studies of older adults, is detailed elsewhere (17). Briefly, part 1 addressed activities performed during a typical week from the past month. The 5 categories of activities contained in part one were organized to sequentially follow how an older person might spend his or her day: housework activities, yard work, care-giving of elders or children, purposeful exercise, and leisure (recreational) activities. Part 2 of the YPAS assessed vigorous activity, leisure walking, moving, standing, and sitting behaviors, with estimations of activity frequency and duration over the past month. These 5 activity dimensions were then scored (adjusted for intensity of activity) and summed to create an activity dimensions summary index (ADSI), expressed as total units for each subject (ranging from 0 to 133), to summarize activity levels and intensity for the past month. Two-week repeatability correlation coefficients on 76 men and women (without arthritis) aged 60–86 ranged from 0.42 to 0.65 (17). In addition, the ADSI was significantly associated with estimated VO2max (per treadmill test; r = 0.58, P = 0.004) and percentage of body fat (r = –0.43, P = 0.03) (17).

In addition to the traditional subscales produced by the YPAS, a set of intensity-specific indices was created for the purposes of this study. All activities with an intensity code <3.0 were defined as low-intensity activities. All activities with an intensity code between 3.0 and 5.5 were designated as moderate intensity. All activities with an intensity code of 6.0 or greater were defined as high intensity.

The score for each intensity-specific index was obtained by summing the number of minutes per week spent engaged in activities within each intensity level. Scores on the high- and moderate-intensity indices were significantly correlated with the ADSI (r = 0.41, P < 0.0001 and r = 0.27, P < 0.0001, respectively), but not the low-intensity index (r = 0.1, P = 0.18), as would be expected.


The study was reviewed and approved by the university Institutional Review Board. An introductory letter was sent to all potentially eligible subjects, explaining the purpose of the study, assuring them of complete confidentiality, and informing them that they would receive a followup telephone call. During the telephone contact, the researcher further explained the study, addressed questions, and screened for age and functional status eligibility using 4 questions regarding self care and ambulation limitations. Followup appointments were made with all eligible and interested subjects.

Participants were given the opportunity to choose the location as well as the data collection method for the survey that would be most convenient for them. Three types of data collection methods were offered: face-to-face survey (n = 54), mail (demographics) and face-to-face (n = 49), or mail and telephone survey (n = 82). The majority (94%) of participants who had a face-to-face survey were met at their regularly scheduled physician appointments on campus. Participants who completed the survey face-to-face were compared with women who completed it over the phone on age, race, and educational attainment relative to the data collection method. No significant differences were found: F = 0.69, df = 2, P = 0.5; χ2 = 9.29, df = 8, P = 0.32; and χ2 = 19.0, df = 16, P = 0.26, respectively. There were no significant differences in scores on the ADSI by survey method.

Data analysis

All data were entered into a personal computer using SPSS version 10 software (SPSS, Chicago, IL). Several data quality checks were conducted, including visual confirmation and examination of descriptive statistics and frequency distributions; outliers were detected and any inaccuracies or omissions in data entry were corrected. Descriptive statistics of central tendency and dispersion were prepared for all variables in the data set. For categorical variables (race, income, educational level, marital status), frequency distributions were examined for range of responses. Scores for the YPAS categorical and intensity indices, as well its summary and dimensional indices, were computed and summarized in frequency distributions with tabulations of descriptive statistics for each. Analysis of variance was used to compare mean scores on the ADSI by demographic characteristics (race, education, marital status, income, and employment). Pearson correlation coefficients were calculated to evaluate the relationship between age, disease duration, and scores on the ADSI.


Weekly physical activity behavior

The women with RA reported participating in 47 different physical activities divided among the 5 categories of the YPAS part 1 (Figure 1). The descriptive data contained in Table 1 is limited to the most frequently reported of the 47 activities. Within the housework category, the 2 most frequently reported activities, light housework and shopping, were of moderate intensity. These were followed by 2 low-intensity activities, food preparation and dishwashing. Close to half the women reported at least some stair climbing with a load, the only high-intensity housework activity. Examination of the number of weekly hours spent performing these activities revealed that women spent most of their time engaged in food preparation and light housework.

Figure 1.

Proportion of total weekly activity time (23 hours) spent engaged in various activities.

Table 1. Hours per week spent engaged in selected physical activities
Physical activityMET value*Participation no. (%)Mean hours/weekSDMinimum hoursMaximum hours
  • *

    One metabolic rate (MET) represents the approximate rate of oxygen consumption of a seated individual at rest.

  • Denotes minutes as opposed to hours.

High intensity≥6.0     
 Stair climbing  with load8.589 (48)0.30.705
 Brisk walking6.072 (40)0.81.308.2
 Cycling6.030 (16)0.20.603.5
 Tennis7.03 (2)2.8m25.6m05.3
Moderate intensity3.0–5.5     
 Light housework3.0168 (91)3.03.5020
 Shopping3.5167 (90)2.92.5012
 Gardening4.545 (24)0.72.5028
 Clearing walks  and driveway5.035 (19)0.20.605.5
 Child care4.013 (7)0.21.1012
 Elder care5.57 (4)0.32.3028
 Leisure walking3.5127 (67)1.83.5035
 Pool exercises3.093 (50)1.01.43010.5
 Weight training3.020 (11)0.20.807.0
Low intensity<3.0     
 Food preparation2.5157 (85)3.53.7022.1
 Dishwashing2.5156 (84)2.22.3010.5

In the yard work category, gardening was the most frequently reported activity, followed by clearing walks and driveways, both moderate-intensity activities. For the care-giving category, childcare was the most frequently reported activity, also of moderate intensity, but more hours were spent caring for elder relatives. Within the planned exercise category, pool exercises (moderate intensity) and brisk walking (high intensity) were the most frequently reported activities, with participation rates of 50% and 40%, respectively. These were followed in frequency by cycling and weight training activities (16% and 11%, respectively). In the leisure activity category, leisure walking was the most frequently reported activity performed, and the largest proportion of time was spent engaged in this moderate-intensity activity. Just 3 women reported playing tennis, the only designated high-intensity physical activity listed in this category.

Overall, women spent most of their weekly activity time doing housework (mean ± SD 15.5 ± 10.5 hours/week), followed by leisure physical activities (3.5 ± 5.2 hours/week), planned exercise (2.4 ± 2.6 hours/week), and yard work (1.2 ± 3.0 hours/week). Minimal time was spent in care-giving activities for this group of women (mean ± SD 0.5 ± 2.5 hours/week). In general, the women spent a mean total of 23 hours per week engaged in lifestyle physical activity. Figure 1 shows that of the total time spent engaged in these activities, 67% was spent doing housework. Further examination (Table 2) revealed that the majority of time spent in these activities was at what was defined as the moderate level of intensity. Table 2 demonstrates that the women in this study averaged 117 minutes per day engaged in accumulated moderate-intensity activities, which exceeds the ACSM/CDC recommendation of 30 minutes per day of moderate-intensity activities. High-intensity activity was most often performed within the context of planned exercise.

Table 2. Minutes per day spent engaged in physical activities by intensity index*
Intensity-Specific Activity IndexMean minutes/daySDMinimum minutesMaximum minutes
  • *

    MET = metabolic rate.

Low-intensity index (<3.0 METs)68.0540342
Moderate-intensity index (3.0–5.5 METs)117.0836.6394
High-intensity index (≥6.0 METs)12.916093

Monthly physical activity behavior

The frequency and duration of the 5 activity dimensions (vigorous activity, moderate-intensity leisure walking, moving, standing, and sitting) performed in the prior month were examined in part 2 of the YPAS. Sixty percent of the women (n = 110) participated in some type of vigorous activity during the past month, and more than half of these (61%, n = 67) participated in vigorous activities 3 or more times per week. Overall, 88% (n = 163) of the women participated in moderate-intensity leisure walking, but of these, only 30% (n = 50) walked at the recommended frequency of 5 or more times per week.

Moving for approximately 3 or more hours per day was reported by 72% of the women. Conversely, most of the women (75%) reported standing (without moving) for <3 hours per day (these were mutually exclusive categories). Close to half (48%) of the women reported sitting for 6 or more hours per day on a typical day during the past month.

Mean scores on the 5 activity dimension indices and the summary index are displayed in Table 3. There were no significant differences on ADSI scores by racial group (white versus nonwhite), education level, marital status, or income (Figure 2). However, there was a significant difference on the ADSI by employment status. A Tukey honestly significantly different test showed that women in the work force (n = 47) and those involved in housekeeping (n = 105) scored significantly higher (P = 0.003) on the ADSI than those 33 women who described themselves as retired (F = 7.81, df = 2, P = 0.001). Age was negatively associated with scores on the ADSI (r = –0.195, P < 0.01). As age increased, there was a decrease in the ADSI. Scores on the ADSI were not significantly correlated with the number of years participants had been diagnosed with RA (r = –0.06, P = 0.46).

Table 3. Activity dimension indices and activity dimensions summary index
Activity dimensions indexPossible rangeActual rangeMeanSD
Activity dimensions summary index0–1337–10840.020.0
Vigorous activity index0–600–6012.514.5
Walking activity index0–480–4813.110.9
Moving activity index0–150–159.63.4
Standing activity index0–100–104.21.9
Sitting activity index0–41–42.60.8
Figure 2.

Mean activity dimensions summary index scores by demographic characteristics. *Means in the same row that do not share superscripts differ at P < 0.01.


This is the first descriptive study of lifestyle physical activity in women older than 60 years with RA. Household, leisure, and planned exercise activities were examined (along with integrated occupational activities) to provide insight into accumulated daily lifestyle physical activity levels. Overall, the women reported spending a mean of 23 hours per week on household, leisure, planned exercise, yard work, and care-giving activities (occupational activities were distributed among these categories according to the nature of the activities performed during work hours). This is 10 hours per week less than reported (33.9 hours per week) by community-dwelling elderly women without musculoskeletal impairments who participated in an earlier study (17).

The women reported participating in 47 different types of physical activities divided among the 5 YPAS categories, but reported spending most of their time engaged in housework activities, much of which was classified as moderate intensity. Leisure activities occupied the second largest amount of activity time after housework, with leisure walking being the overwhelming favorite activity. Overall, 67% (n = 127) of the women reported participating in weekly leisure walking. This preference for leisure walking is consistent with the study by Wilbur et al (18) of working women (77% of whom reported walking for pleasure) and with the findings of others reporting on older populations (11, 19). Anecdotally, the majority of women reported that much of the leisure walking took place within the context of daily life (running errands, walking to public transportation, shopping, etc.), although 40% of the women reported participating in walking as a true leisure activity. However, the actual amount of time spent walking might have reflected the degree of functional limitation experienced by some of the women with lower extremity arthritis.

Pool exercises, brisk walking, and cycling were the most popular types of planned exercises reported by the women with RA. These were also the most commonly reported activities in the survery by Gecht et al (20) of planned exercise behaviors among persons with all types of arthritis. Neuberger et al (21) also studied planned exercise behavior in persons with RA, who reported performing range of motion and muscle strengthening exercises, but these activities were not commonly reported as separate forms of physical activity in these older women with RA.

Further examination revealed that the majority of time spent engaged in lifestyle activities was at the moderate level of intensity. Surprisingly, >75% of the women reported an average of 55 minutes per day in self-reported moderate-level activity, suggesting that they met the ACSM/CDC guideline of 30 minutes or more per day of accumulated moderate-intensity activity (8). This stands in contrast to the 77% of older diabetic adults in a prior study who reported engaging in less than 60 minutes of weekly physical activity (22). That study, however, examined only planned physical activity and did not include measures of household or lifestyle activities (22). The 2 surveys of persons with arthritis reported on the performance or nonperformance of planned habitual physical activity (lifestyle physical activities were not examined), but did not include information on duration, frequency, and intensity of effort in their reported findings (20, 21).

The high percentage of women meeting the new national guideline in this study is consistent with that reported by Brownson et al (11) for community-dwelling older women (72%). However, results must be interpreted with caution, because women with RA might have performed activities in adaptive ways due to functional limitations that both increased the amount of time required to complete the activity and decreased the level of presumed intensity. The assigned intensity codes for all activities in the physical activity survey were validated with older adults who were free of mobility limitations (17). These same activities might have been performed at considerably lower intensity levels by the women with RA.

With respect to the activity dimension indices, the mean score for the vigorous activity index (12.5) was similar to that (12.7) found in an earlier study of healthy elderly men and women (17), but was lower than the mean (17.4) reported by Campbell et al in a study of healthy older women (23). Overall, 60% (n = 110) of the women reported participating in some type of vigorous activity during the past month. Of these, 61% (n = 67) reported participating in vigorous activities 3 or more times per week, enough to reap cardiovascular health benefits.

Prior research examining physical activity in persons with arthritis (RA and osteoarthritis) did not identify intensity level of activity (20, 21). Again, cautious interpretation is necessary, because the reported vigorous experience might only have been an indication of physical deconditioning in these women. For example, Neuberger et al (21) previously found that 63% of her sample of persons with arthritis (including RA and osteoarthritis) had very poor or poor aerobic fitness levels, as measured by bicycle ergometer performance. Thus, if the women in this study had similarly poor fitness levels, activities not generally considered vigorous (e.g., walking 1–2 blocks) might have indeed been experienced by them as such. There is a need for studies that compare aerobic fitness levels to reported physical activity behavior in this population.

The mean score on the leisure walking index (13.1) was lower than the mean scores of healthy older participants found by DiPietro et al (17) and by Campbell et al (23) (17.3 and 16.0, respectively). Because persons with mobility limitations were excluded from the earlier studies, the lower scores found for women with RA on the leisure walking index were most likely due to symptoms of lower extremity arthritis. Overall, 88% (n = 163) of the women participated in walking, but of these, only 30% reported walking at the recommended frequency of 5 or more times per week. Sixty-four percent of women reported walking for 10–30 minutes per walk, as recommended in the ACSM/CDC guideline.

Moving, standing, and sitting behavior indices scores were calculated on the basis of average hours per day spent engaged in the activity. The mean score for the women on the moving index (9.6) was similar to those reported by DiPietro et al (17) and Campbell et al (23) (8.6 and 9.5, respectively). DiPietro et al and Campbell et al reported mean standing index scores of 4.4 and 2.5, respectively, which were similar to the mean (4.2) for women with RA. The mean sitting index score, however, was somewhat higher for women in this study compared with the 2 earlier studies (2.6 versus 2.0 and 2.5, respectively). The higher sitting index scores seen in this study might have reflected exposure to professional occupational therapy or published Arthritis Foundation advice to sit for any activity that might be adapted for sitting (ironing, cooking, etc.) to conserve limited energy. This strategy is widely advocated in both professional and lay literature regarding energy conservation for those with inflammatory conditions.

The participants scored a mean of 40.0 on the ADSI, compared with means of 45.1 and 48.7 for subjects in the in the 2 earlier studies of elderly subjects (17, 23). The lower ADSI scores for women in this study might have been a result of less leisure walking behavior. As noted above, persons with mobility limitations were excluded from the earlier studies. Being older was significantly correlated with lower physical activity levels, which is consistent with findings from previous studies of community-dwelling elders (22, 24, 25). Understandably, retired women scored significantly lower on the ADSI than women working outside the home and those keeping house. Anecdotal reports suggested that many of the retired women who did not do housework were living in retirement communities (often due to functional limitations) or alone at home with hired periodic housekeeping help. Because such a large percentage of activity in these women was attributed to housework, those not keeping house would have had to be more active in one of the other categories to close the gap in activity performance scores.

In summary, women with RA reported participating in 47 different physical activities divided among 5 categories. Overall, the women spent a mean of 23 hours per week on household, leisure, planned exercise, yard work, and care-giving activities, which was less than that seen in other populations of both well and ill older persons. Further examination revealed that the majority of time spent engaged in these activities was at the presumed moderate level of intensity. However, functional limitations might have altered the intensity with which women performed all activities, making it difficult to conclude that this group met national ACSM/CDC guidelines. Even though women reported participation in leisure walking more than any other activity, they reportedly walked less frequently than other populations of older persons.

However, walking duration was not lower than that reported by healthy populations in previous studies. Thus it appears that the women did not walk frequently, but when they did, they tended to walk as long or longer than the durations reported by other community samples. This greater time taken for walking activities might have reflected the greater effort or the slower speed of women with functional limitations.

The major limitation of this study was the exclusive reliance on the self-report measure of physical activity behavior to assess that outcome variable. As with all self-report measures, responses were subject to memory error, which might have been significant in this elder population. Overestimation of time spent performing household tasks is common. Furthermore, all physical activity behavior data were collected during an interview survey, and while this format allowed for clarification of item responses, it might also have promoted biased response patterns, especially in that proportion of participants known to the clinician interviewer. The effect of this arrangement was not estimated with a measure of social desirability of responses.

No formal test of cognition was employed, but every attempt was made during the screening interview to assess appropriateness of responses, thereby excluding obviously impaired women. Additionally, all psychometric testing of the self-report measure of physical activity utilized in this study was performed exclusively on groups of community-dwelling older adults free of mobility limitations. It is not known how this might have affected the reliability and validity of the data. Although the response rate was very reasonable, and the sampling procedure included attempting to reach all known older women with RA on campus, findings from this study may not be generalizable to other populations because the entire sample was recruited from a single university medical practice. Even though every effort was made to recruit all known potential subjects, the final sample consisted primarily of well-educated white women, so findings may not be applicable to women with RA from other (especially non–English-speaking) racial groups.

This study provided some important insights into the nature of physical activity behavior of older women with RA. The lifestyle physical activity types and amounts observed in this study were both similar to and different from those documented for other populations of older adults. While participants were surprisingly active, it was not possible to conclude with certainty that this group of women with RA met the physical activity guidelines set forth by the ACSM/CDC, which calls for the accumulation of 30 minutes of moderate-intensity physical activity in short segments of activity on most, if not all, days of the week. Because persons with RA may not be aware of the activities enjoyed by others with arthritis, health care providers should capitalize on patient care encounters as opportunities to both assess and encourage physical activity within the context of daily life routines to meet the ACSM/CDC recommended activity levels.