Journal Logo

BASIC SCIENCES: Epidemiology

Influence of Marital Status on Physical Activity Levels among Older Adults

PETTEE, KELLEY K.1; BRACH, JENNIFER S.1; KRISKA, ANDREA M.1; BOUDREAU, ROBERT1; RICHARDSON, CAROLINE R.2; COLBERT, LISA H.3; SATTERFIELD, SUZANNE4; VISSER, MARJOLEIN5; HARRIS, TAMARA B.6; AYONAYON, HILSA N.7; NEWMAN, ANNE B.1

Author Information
Medicine & Science in Sports & Exercise: March 2006 - Volume 38 - Issue 3 - p 541-546
doi: 10.1249/01.mss.0000191346.95244.f7
  • Free

Abstract

Regular participation in physical activity can prolong health by offsetting the development of chronic diseases and conditions (2); however, many older adults are inactive (5). Previous studies have investigated factors (6) and identified reasons given by older individuals why they limit or avoid physical activity (10). Nonmedical factors (e.g., living alone) have been identified as important contributors to sedentary behavior (6,10). Past work suggests that married individuals participate in exercise more often than their single counterparts (11,12); however, little is known if a similar relationship exists with regard to total physical activity participation, including both exercise and nonexercise activity.

The Health, Aging, and Body Composition (Health ABC) study offers a distinct opportunity to explore the relationship between marital status and physical activity, not only using the entire cohort, but specifically through a subgroup of married couples that were enrolled together in the study. In Health ABC, both exercise and nonexercise activity were measured, providing an estimate of total physical activity participation.

Health ABC allows these associations to be studied in a group that is more socioeconomically and racially diverse (10,11) than previous study populations. Furthermore, for some participants, both members of the spousal pair underwent the same enrollment procedures and all measured variables were assessed using the same methods. This provided the unique opportunity to examine the relationship of physical activity participation in individual members of a married couple and contrasts previous work where spousal pair data were collected by proxy from a single member of the married couple (12). Moreover, a notable relationship between members of a married couple may suggest a novel approach to utilize when attempting to increase physical activity levels in older individuals.

We hypothesized that marital status would play an important role in physical activity participation and that the activity level of one member of a married couple will be related to the activity level of the other. The purpose of these analyses was to describe the levels and types of activity in relationship to current marital status among older adults and determine if the physical activity levels of the husband was related to the physical activity levels of his wife.

METHODS

Study Population

The Health ABC cohort consisted of 3075 well- functioning white and black men and women aged 70-79 yr recruited between 1997 and 1998. Study participants were recruited from a random sample of white and all African American Medicare beneficiaries residing within each ZIP code from the metropolitan areas surrounding Pittsburgh, PA and Memphis, TN from 1997 to 1998. Black participants were over sampled in attempts to make the whites and blacks equally representative within the study sample. A total of 1548 participants were randomized at the Memphis study site, of which 39.6% were black. At the Pittsburgh study site, 1527 total participants were randomized, of which 43.8% were black. Eligibility criteria for enrollment into the study included no self-reported difficulty walking a quarter mile, walking up 10 steps, or performing basic activities of daily living. Individuals with cancer under treatment or who planned to move in 3 yr were excluded. All participants provided written informed consent and all protocols were approved by the institutional review boards at each respective study site. Demographic factors, physical activity, and anthropometric and prevalent disease measures were assessed at baseline.

Marital Status

Marital status was obtained at the baseline clinic visit through self-report and participants were classified as being either married or nonmarried. Nonmarried individuals included those who reported being never married, widowed, divorced, or separated. The remaining individuals were classified as married.

Physical Activity

Physical activity levels were determined at baseline using a standardized interviewer-administered questionnaire, modeled from commonly used physical activity assessments, including the Minnesota leisure-time physical activity questionnaire (15). Participants reported whether they had participated in a specific activity (e.g., gardening or yardwork) at least 10 times in the past 12 months. If they had participated in a given activity, the frequency and duration spent in the activity during the previous 7 d were determined. The appropriate weighted metabolic equivalent (MET) value for each activity (1) was multiplied by the amount of time spent during the week doing that specific activity. Then, the value in kilocalories per kilogram per week was multiplied by body weight (kg) to create the summary physical activity variable in kilocalories per week.

Exercise was defined as the sum of activity reported during structured, moderate- to high-intensity physical activities and was measured in kilocalories per week. Nonexercise activity was defined as the sum of activity (kcal·wk−1) reported on the questionnaire for unstructured, lower-intensity activities such as household chores, nonexercise walking, stair climbing, paid work, volunteer work, and care-giving. Total physical activity participation was defined as the sum of activity (kcal·wk−1) reported on the questionnaire for both exercise and nonexercise activity.

Physical Activity: Spousal Pairs

Each member of a spousal pair was classified as being either low active or high active based on the Surgeon General's recommendations on physical activity and health (9). Low active included those who reported participating in less than 1000 kcal·wk−1 of exercise activity, whereas high active included those reporting 1000 kcal or more of exercise activity per week.

Confounders

Demographic factors (e.g., age, education, income, gender, race or ethnicity, and geographic area) were considered confounding variables. To obtain a measure of education, participants were queried about the highest grade or year of school that they completed. Based on their response, individuals were categorized into one of three education categories: less than 12 yr, 12 yr, and more than 12 yr. Family income was defined as wages, salaries, social security, or retirement benefits; help from relatives; and rent from property. Individuals were categorized into one of four income categories: less than $10,000, between $10,000 and $25,000, between $25,000 and $50,000, and greater than or equal to $50,000. Body mass index (BMI) and prevalent disease conditions can also affect participation in physical activity and were considered confounding variables. All demographic factors were obtained through self-report at baseline visit. BMI was calculated from height and weight measured with a stadiometer and calibrated balance beam scale, respectively. Prevalent disease was determined by a combination of self-report, specific medication use, and physiologic measures (electrocardiogram, blood pressure, and so on) performed at the baseline clinic visit. Self-reported chronic disease conditions included cancer, stroke, coronary artery disease, diabetes, osteoarthritis, osteoporosis, and peripheral vascular disease. A disease index score was created for each participant, based on the number of present prevalent disease conditions. Individuals were classified with a disease index score of 0 (no prevalent diseases), 1 (one prevalent disease), or ≥ 2 (two or more prevalent disease).

Analytic Procedures

Entire cohort.

Descriptive statistics were used for demographic factors, BMI, physical activity, and disease prevalence in married versus nonmarried participants stratified by gender using the entire Health ABC cohort. Participant age and BMI were normally distributed and mean ± standard deviation for these variables were reported. Total physical activity, exercise, and nonexercise activity were not normally distributed; therefore, medians and 25th and 75th percentiles were reported. Depending on the characteristics of the variable, t-tests, Wilcoxon rank sum, or Chi square (χ2) tests were used to compare married and nonmarried participants, stratified by gender.

Spousal pairs.

Spousal pairs enrolled in Health ABC were identified and all relevant information from each was linked together for each pair (345 pairs). Descriptive statistics were used to compare demographic factors, BMI, physical activity, and disease prevalence between low and high active by gender, for each member of the Health ABC couple. Depending on the characteristics of the variable, analysis of variance (ANOVA), Kruskal-Wallis, or χ2 tests for linear trend were used to describe and compare differences between activity groups.

McNemar's test statistic was calculated to compare each member of a spousal pair with respect to exercise participation. Also, in Health ABC spouse pairs, multivariate logistic regression models (low active vs high active) were used to determine if the husband's exercise status was an important determinant of his wife's exercise status after controlling for confounding factors.

RESULTS

Entire cohort.

The mean age of the entire Health ABC cohort was 73.6 yr. Marital status was missing for 195 individuals and these individuals were excluded from the analyses. For the remaining 2880 participants with complete data, 49.1% were men (72.7% married vs 27.3% not married) and 50.9% were women (37.1% married vs 62.9% not married). When comparing the married and nonmarried men, married men were significantly more educated (P = 0.0004), had a higher income (P < 0.0001), were less likely to be black (P < 0.0001), and reported higher levels of exercise participation (P = 0.0079) than those not married. When comparing married and nonmarried women, married women were significantly younger (P < 0.0001), had lower BMI (P = 0.0015), were more educated (P < 0.0001), had a higher income (P < 0.0001), were less likely to be black (P < 0.0001), and had higher levels of total physical activity participation (P < 0.0001) and nonexercise activity (P < 0.0001) than did nonmarried women. Exercise participation tended to be higher among married women (P = 0.0531). The prevalence of chronic disease was not significantly different between married and nonmarried individuals of either gender (Table 1).

TABLE 1
TABLE 1:
Participant characteristics by marital status, stratifying by gender for individuals enrolled in Health, Aging, and Body Composition study.

Spousal pairs.

To address the issue of whether level of activity was similar among spousal pairs, we classified each member of the couple as low or high active in exercise activity based on the Surgeon General's recommendation of 1000 kcal·wk−1. Among spousal pairs, (Table 2), a greater proportion of men were classified as high active than women (40.3 and 20.3%, respectively, P < 0.0001). In both men and women, those classified as high active reported higher levels of exercise (P < 0.0001) and total physical activity (P < 0.0001) participation than the low active group. No difference was noted in age between high active and low active groups for men and women (P = 0.0969 and P = 0.3014, respectively). As a group, low active men and women were less likely to have a high school education (P = 0.0492 and P = 0.0324, respectively) and had a lower annual income (significant in men only P = 0.0079 and P = 0.0922, respectively) when compared with the high active group. Also, more than two thirds of men and women who were classified as high active were from Pittsburgh (65.5 and 70.0%, respectively) even though only 54% of the spousal pairs were from that city. In both the men and women, no difference was found in the prevalent disease index score between exercise groups (Table 2).

TABLE 2
TABLE 2:
Participant characteristics by exercise group, stratifying by gender for 345 spousal pairs (N = 690) enrolled in health ABC.

The relationship of exercise participation among the Health ABC married couples was statistically significant (P < 0.0001). A total of 179 spousal pairs (51.9%) had both members of the couple classified as low active, 43 (12.5%) pairs as both being high active, and 123 pairs (35.6%) were discordant with regard to activity participation. Based on McNemar's test, the reported exercise participation between husband and wife was statistically significant (38.7073; P < 0.0001) (Fig. 1).

FIGURE 1
FIGURE 1:
Spousal pairs (N = 690 of 345 pairs) by activity group. McNemar's test statistic (38.7073; P < 0.0001).

In the Health ABC spousal pairs, compared with inactive men (low active), active (high active) men were almost three times as likely to have a similarly active spouse (Table 3). The model only modestly attenuated when fully adjusted for covariates. In the unadjusted model (model 1), the active husband was 2.97 (95% CI = 1.73, 5.10) times more likely to have an active wife. The odds ratio decreased to 2.49 (95% CI = 1.41, 4.42), but remained statistically significant, after adjustment for demographic factors, BMI, and prevalent disease status. Results were similar when determining if the wife's exercise status was an important determinant of her husband's exercise status [OR = 2.48 (1.40-4.38)]. When examining the relationship stratified by race, trends were similar. The association, however, was not statistically significant in the black married couples, most likely because of a lack of statistical power (N = 79).

TABLE 3
TABLE 3:
Logistic regression for the female physical activity level within a spousal pair (N = 690; 345 pairs).

DISCUSSION

In this large, socioeconomically and racially diverse cohort of older adults, married individuals were more likely to be active than their single counterparts. Furthermore, the present work demonstrated a strong relationship between individual members of a married couple. Compared with the low active husband, a highly active husband was more likely to also have a similarly active wife. This relationship was observed in both the unadjusted and adjusted models, confirming earlier studies that also found exercise status of one partner was related to whether the other member of the couple also exercises (11,12).

Unlike previous studies, the Health ABC has self-report measures of both nonexercise and exercise activity. Although similar, the terms physical activity and exercise have different meanings. Physical activity is defined as any bodily movement produced by skeletal muscles that result in increased energy expenditure (4). Therefore, things such as housework, gardening, and occupational activity may all be considered types of physical activity. On the other hand, exercise is defined as planned, structured, and repetitive bodily movements done to improve or maintain one or more components of physical fitness (4). When differentiating between the two, exercise is a type of physical activity, but not all physical activity is considered exercise.

In the entire Health ABC cohort, married men participated in higher levels of total (not statistically significant), exercise (P = 0.0079), and nonexercise activity (not statistically significant) than their nonmarried counterparts. More strikingly, married women reported higher levels of total (P < 0.0001), exercise (P = 0.0531), and nonexercise (P < 0.0001) activity. This supports previous literature that suggests for a more accurate assessment of a woman's participation in physical activity, total physical activity, including lower-intensity, nonexercise activities should be included (13). In addition, this findings adds to previous literature (11,12), demonstrating that not only does marital status have a positive effect on exercise participation among older individuals, but also with regard to overall participation in total physical activity, especially among women.

When interpreting the findings, a number of limitations need to be considered. Physical activity data were obtained through self-report methods and can be influenced by a number of factors, such as the inability to accurately recall personal participation in physical activities. Lower-intensity, nonexercise activities (e.g., household chores and care-giving) are harder to recall than higher-intensity activities (7,14). To account for the potential error in recall of lower-intensity activities, we classified spousal pairs based only on reported exercise participation to ensure the most accurate results. Future work, including an objective measure of physical activity, is needed to determine whether a relationship also exists between members of a married couple with regard to total and nonexercise activity. It should be noted, however, that a separate analysis yielded similar results (Kendall coefficient of concordance (0.21; P < 0.0001) when using reported exercise and total physical activity to create three physical activity groups (inactive, lifestyle active, exercisers) (3). Also, the questionnaire used to assess physical activity levels has limited psychometric properties; however, a questionnaire from which it was modeled has been previously validated (15). Furthermore, it should be noted that the questionnaire used in the Health ABC study has been shown to have predictive validity with respect to mobility limitation (16).

It is important to note that physical activity levels in the Health ABC study were measured using a past week questionnaire. Past week assessment tools are subject to issues of seasonality, especially when assessing individuals living in climates that experience fluctuations in temperature, precipitation, and number of daylight hours that occur with the changing seasons (7,8). However, data were collected over the course of the year and did not occur during one or two seasons.

Data from only one point in time, the baseline visit, was used in the analyses. Although informative, this cross-sectional study analysis does not provide information pertaining to the direction of association or allow us to determine whether this relationship between couples persists over time as the cohort ages. We also do not know whether the spousal pairs were both lifelong exercisers or had become physically active together. Furthermore, individuals enrolled in this study were well-functioning men and women 70-79 yr of age. Again, these results may not be generalizable to all. A population including older and sicker individuals may include couples where one member of the pair is extremely inactive because of a particular disease while the other gets a lot of activity because of care-giving responsibilities. This phenomenon could result in a higher discordance rate between spousal pairs with regard to participation in physical activity. Future work is needed to determine if the notable relationship that was found between members of Health ABC spousal pairs continues to exist as the couple ages and develops a greater number of chronic conditions.

The nonmarried group represents a broad category including never married, widowed, divorced, and separated individuals and physical activity levels may vary between groups. Physical activity levels were further examined in these groups, stratified by gender. In men, no statistically significant differences were found between groups for any of the physical activity variables. In women, however, a statistically significant difference was seen in reported exercise levels, with widowed women reporting the lowest levels (P < 0.01).

Findings from the present study found that the husband's physical activity participation status was related to the physical activity participation status of his wife. An active husband was nearly three times more likely to also have an active wife. These results suggest that spousal pair-based interventions, including both members of a married couple, may be an innovative approach to improve both physical activity participation and compliance among older adults.

This study was funded by National Institute on Aging contracts NO1-AG-20-2101, NO1-AG-20-2103, and NO1-AG-20-2106.

REFERENCES

1. Ainsworth, B. E., W. L. Haskell, M. C. Whitt, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med. Sci. Sports Exerc. 32:S498-S516, 2000.
2. Blair, S. N., and S. Brodney. Effects of physical activity and obesity on morbidity and mortality: current evidence and research issues. Med. Sci. Sports Exerc. 31:S646-S662, 1999.
3. Brach, J. S., E. M. Simonsick, S. Kritchevsky, K. Yaffe, and A. B. Newman. Lifestyle activity and exercise: the association with physical function in the Health Aging, and Body Composition (Health ABC) study. J. Am. Geriatr. Soc. 52:502-509, 2004.
4. Caspersen, C. J., K. E. Powell, and G. M. Christenson. Physical activity, exercise, and fitness: definitions and distinctions for health related research. Public Health Rep. 100:126-131, 1985.
5. DiPietro, L. Physical activity in aging: changes in patterns and their relationship to health and function. J. Gerentol. A: Biol. Sci. Med. Sci. 56A:12-22, 2001.
6. Kaplan, M. S., J. T. Newsom, B. H. McFarland, and L. Lu. Demographic and psychosocial correlates of physical activity in later life. Am. J. Prev. Med. 21:306-312, 2001.
7. Kriska, A. M., and C. J. Caspersen. Introduction to the collection of physical activity questionnaires: a collection of physical activity questionnaires for health-related research. Med. Sci. Sports Exerc., 29, Kriska and Caspersen (Eds.). Centers for Disease Control and Prevention, 1997, pp. S5-S9.
8. Matthews, C. E., P. S. Freedson, J. R. Hebert, et al. Seasonal variation in household, occupational, and leisure time physical activity: longitudinal analyses from the seasonal variation of blood cholesterol study. Am. J. Epidemiol. 153:172-183, 2001.
9. Physical Activity and Health: A Report of the Surgeon General, Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, President's Council on Physical Fitness and Sports, 1996-8.
10. Satariano, W. A., T. J. Haight, and I. B. Tager. Reasons given by older people for limitation or avoidance of leisure-time physical activity. J. Am. Geriatr. Soc. 48:505-512, 2000.
11. Satariano, W. A., T. J. Haight, and I. B. Tager. Living arrangements and participation in leisure-time physical activities in an older population. Journal of Aging and Health 14:427-451, 2002.
12. Schone, B. S., and R. M. Weinick. Health-related behaviors and the benefits of marriage for elderly persons. Gerontologist 38:618-2027, 1998.
13. Sternfeld, B., B. E. Ainsworth, and C. P. Quesenberry. Physical activity patterns in a diverse population of women. Prev. Med. 28:313-323, 1999.
14. Taylor, C., T. Coffey, K. Berra, R. Iaffaldo, K. Casey, and W. Haskell. Seven-day activity and self-report compared to a direct measure of physical activity. Am. J. Epidemiol. 120:818-824, 1984.
15. Taylor, H. L., D. R. Jacobs, B. Shucker, J. Knudsen, A. S. Leon, and G. DeBacker. A questionnaire for the assessment of leisure-time physical activities. J. Chron. Dis. 31:741-775, 1978.
16. Visser, M., E. M. Simonsick, L. H. Colbert, et al. Type and intensity of activity and risk of mobility limitation: the mediating role of muscle parameters. J. Am. Geriatr. Soc. 53:762-770, 2005.
Keywords:

EXERCISE; MARRIAGE; SPOUSAL PAIRS; AGING

©2006The American College of Sports Medicine