Home > RNJ > 2006 > January/February > Testing a Model of Post-Stroke Exercise Behavior

Testing a Model of Post-Stroke Exercise Behavior
Marianne Shaughnessy, PhD CRNP Barbara M. Resnick, PhD CRNP Richard F. Macko, MD

Stroke is the leading cause of disability in older Americans, and survivors tend to be sedentary, thereby risking loss of functional gains achieved during rehabilitation and increasing cardiovascular risk. Studies of motivation to exercise in older adults suggest that self-efficacy and outcome expectations are key determinants of initiation and adherence to exercise programs. This study tested a theoretical model of physical activity in stroke survivors. A survey of exercise beliefs and patterns was sent to National Stroke Association stroke support groups. Responses from 312 stroke survivors (mean age 63 years, 57% female, 70% White) indicated that only 31% exercised four times weekly. Self-efficacy and outcome expectations for exercise, before exercise history, and physician recommendation all directly and indirectly influenced self-reported exercise behavior and accounted for 33% of the total variance in exercise behavior. Model testing supported the theory and the model fit the data. Interventions to strengthen self-efficacy and outcome expectations for exercise, along with reminders for clinicians to encourage regular exercise programs, may increase the likelihood of initiating and maintaining an exercise program, potentially improving physical function and cardiovascular fitness in this population.

Stroke is the leading cause of disability in older Americans, affecting approximately 750,000 people each year (American Heart Association [AHA], 2004). The consequences of chronic neurological deficits can markedly impair quality of life and functional independence after stroke for the approximately 4.5 million stroke survivors in the United States (AHA). Following completion of conventional rehabilitation, patients adopt or return to a sedentary lifestyle (Lofgren, Nyberg, Mattson, & Gustafson,1999; Paolucci et al., 2001). Reduced activity levels can lead to profound cardiovascular deconditioning and disuse atrophy that may compound disability. The resulting loss of the functional gains achieved during rehabilitation begins a cycle of disuse that leads to worsening activity intolerance (Macko et al., 2001). Finally, because many stroke patients are older, their neurological deficits and disuse atrophy are compounded by age-related declines in overall fitness and muscle mass (Ryan, Dobrovolny, Silver, Smith, & Macko, 2000), making it even harder to prevent or compensate for functional declines after stroke. Recent evidence suggests that exercise may ameliorate some of these declines in functionally impaired stroke survivors (Macko et al.; Silver, Macko, Forrester, Goldberg, & Smith, 2000; Smith, Silver, Goldberg, & Macko, 1999). However, a lack of consensus statements on the safety and efficacy of exercise after stroke may limit physician recommendations and patient awareness of the potential health benefits of exercise programs. Little is known about actual exercise behavior and its determinants in this population.

Numerous factors negatively influence physical activity in older adults, including lack of motivation (Dishman, 1994), social issues and cultural expectations (Resnick, 1999), environmental factors such as insufficient room to walk (Dishman), coexisting disease states (Morey, Pieper, & Cornoni-Huntley, 1998), fear of falling (Hill, Schwarz, Kalogeropoulos, & Gibson, 1996), unpleasant sensations associated with exercise (Morey, 1998), and lack of knowledge about the benefits of exercise on the part of the older adult (Dishman; Resnick & Spellbring, 2000). The theory of self-efficacy states that specific efficacy expectations affect behavior, motivational level, thought patterns, and emotional reactions in response to any situation (Bandura, 1977). Perceptions and beliefs affecting exercise behavior include self-efficacy expectations (i.e., individuals’ judgment of their capabilities to organize and execute courses of action) and outcome expectations (i.e., beliefs that if a certain behavior is performed, there will be a specific outcome or benefit). These beliefs are essential to the adoption and maintenance of exercise behavior in older adults (Bandura; Jette et al., 1998) and have been demonstrated to be strongly associated with exercise behavior in other older and disabled groups (McAuley, 1993). Less research has focused on the relationship between self-efficacy and outcome expectations related to exercise and physical activity in the stroke-impaired population.

This survey-based study examines the influences of exercise perceptions and beliefs within the context of other clinical, demographic, and psychosocial variables on the self-reported physical activity and exercise patterns of stroke survivors. The purpose of this study was to test a hypothesized model that considered the relationship between self-efficacy and outcome expectations for exercise, demographic variables, exercise history before stroke, and physician influences on physical activity in patients after stroke.



The survey was distributed as part of a mailing from the National Stroke Association (NSA) to 1,200 stroke support group facilitators throughout the United States and Canada. A cover letter from the investigator requested distribution to any stroke survivor who would be able to complete the survey. Return of a completed survey implied consent to participate in the study. An interactive version of the survey was also posted on the NSA Web site.


A single-page survey instrument assessing extent of the effect of stroke on beliefs and perceptions regarding exercise and other demographic, social, and behavioral variables was developed by the authors in consultation with a multidisciplinary team of experts in the clinical care of stroke patients. Demographic and descriptive information was obtained including age, gender, race, time since stroke, self-reported premorbid and current exercise activity, the influence of fatigue on daily activity, and whether the participant had been told to exercise by the primary healthcare provider. Exercise activity was defined as participation in a physical activity of at least 20 minutes duration that caused sweating or increases in respiratory or heart rate (American College of Sports Medicine, 1998). Participants were asked to describe how often they engaged in this type of activity on a Likert-type scale, with 1 = never, 2 = less than once per week, 3 = 1–3 times per week, and 4 = 4 or more days per week. This survey was meant to assess the frequency of self-reported physical activity adherent to current public health recommendations.

Two distinct scales were used to assess beliefs regarding exercise. The Short Self-Efficacy for Exercise scale (SSEE) consists of 4 items that focus on challenges associated with exercise for post-stroke patients. Participants were asked to respond to how strongly they believed they were capable of exercising in light of challenges such as pain, fatigue, depression, and exercising alone with response items ranging from 1 = not confident to 5 = very confident. There was evidence of internal consistency with an (alpha) = 0.86 and with confirmatory factor analysis (Resnick & Nigg, 2003). The Short Outcome Expectations for Exercise (SOEE) comprises 5 items that assess respondents’ agreement with statements on the potential benefits of exercise (e.g., enjoyment of activity, improved well-being, mood, alertness, and endurance) with responses ranging from 1 = strongly disagree to 5 = strongly agree. There was support for the internal consistency of the SOEE with an (alpha) = 0.90 and confirmatory factor analysis (Resnick & Nigg). There was evidence of construct validity in stroke survivors based on significant lambda values (or item loadings) for all items onto their respective constructs. Likewise, there was evidence of construct validity of the SSEE and the SOEE; self-efficacy expectations significantly influenced exercise and accounted for 13% of the variance in exercise; and outcome expectations explained an additional 2% of the variance in exercise (Shaughnessy, Resnick, & Macko, 2004).

Data Analysis

Structural equation modeling was selected to test the relationships among all the variables because all paths are estimated and evaluated simultaneously, and specific variables (e.g., self-efficacy and outcome expectations) can be analyzed as both outcome and predictor variables (Bollen, 1989). Descriptive statistics and bivariate correlations were used to describe the sample and evaluate relationships among variables. Model testing was done using Amos 4.0 (Arbuckle, 1997). The sample covariance matrix was used as input and a maximum likelihood solution sought. The c2 statistic, the normed fit index (NFI), and Steigers Root Mean Square Error of Approximation (RMSEA) were used to estimate model fit. The larger the probability associated with the c2, the better the fit of the model to the data (Bollen; Loehlin, 1998; Pedhazur, 1991). The NFI tests the hypothesized model against a reasonable baseline model and ideally should be 1.0. The baseline model is created by the statistical program and represents a model that does not fit the data (Pedhazur). The RMSEA is a population-based index and consequently is insensitive to sample size. An RMSEA of <0.10 is considered good, and <0.05 is very good (Bollen). Path significance was based on the Critical Ratio (CR), which is the parameter estimate divided by an estimate of the standard error. A CR > 2 in absolute value was considered significant (Pedhazur). A p < .05 level of significance was used for all analyses.


A total of 321 stroke patients completed the survey. Analysis was performed on the 312 surveys for which there were complete data. The clinical and demographic characteristics of the study sample are shown in Table 1. Table 2 contains premorbid and current reported exercise activity. Table 3 summarizes the scores on the SSEE and SOEE, and the participant responses to queries on the effect of fatigue and physician recommendations to exercise. Following stroke, there was a decrease in reports of regular exercise from 45% to 31%, and those who reported rare or no exercise activity climbed from 15% to 27% of respondents. The majority of the participants (69%) stated that they did not perform as much exercise as they would like to, and 84% reported that they would be interested in an exercise program if one were available.

Bivariate correlations between self-reported exercise behavior and its potential determinants are shown in Table 4, and results of model testing with parameter estimates and R2 are shown in Figure 1. Ten of the 16 hypothesized paths were statistically significant. Self-efficacy expectations, outcome expectations, exercise behavior before stroke, and recommendation by a physician to exercise all directly influenced exercise and accounted for 33% of the variance in self-reported exercise behavior. Those individuals with stronger self-efficacy and outcome expectations, exercise activity before stroke, and a recommendation to exercise by a physician were more likely to exercise. Fatigue, recommendation by a physician, and age indirectly influenced exercise through self-efficacy expectations. Respondents who were older and reported higher levels of fatigue were more likely to have lower self-efficacy for exercise. Fatigue, race, and self-efficacy significantly influenced outcome expectations accounting for 34% of the variance in outcome expectations.

The model fit the data with a c2 of 38 (degrees of freedom = 20, [c2]/df = 1.9, p = .01). The NFI was 0.99, and RMSEA was 0.05. In an attempt to test a more parsimonious model, nonsignificant paths were removed from the model and the trimmed model was retested. There was no improvement in model fit when the nonsignificant paths were included (c2 difference = 6, df difference = 2, p > .05).


As reported by other researchers (Jette et al., 1998; McAuley, 1993; Resnick & Spellbring, 2000), self-efficacy and outcome expectations exert an influence on physical activity and exercise behavior in a group disabled by stroke. In this model, self-efficacy and outcome expectations for exercise, physician recommendation to exercise, and exercise history before stroke explained 33% of the variance in exercise behavior in this sample of stroke survivors. These findings also support the theoretical model that self-efficacy expectations are considered to have a greater influence on behavior than outcome expectations (Resnick & Spellbring). As previously reported in nonneurological populations (Hill et al., 1996; Jette et al.; Resnick & Spellbring, 2000), outcome expectations provided an important, independent influence on exercise behavior. Moreover, findings from qualitative studies (Resnick & Nigg, 2003; Sharon, Hennessey, Brandon, & Boyette,1997) indicate that older adults identify the health and psychological benefits of exercise as strongly influencing exercise behavior. Developing specific, directed behavioral interventions to strengthen self-efficacy and outcome expectations may significantly improve exercise behavior in older adults after stroke. Further studies are needed to determine the extent to which these beliefs can be modified to change attitudes and behavior related to exercise across the continuum of after stroke care.

Fatigue, physician recommendation, and age were the only variables to directly influence self-efficacy expectations; and self-efficacy expectations, fatigue, and race influenced outcome expectations. Age, race, and physician recommendation had indirect effects on exercise behavior through self-efficacy and outcome expectations. In this study, in which ages ranged from 21 to 91 years, age influenced self-efficacy expectations as anticipated, with older patients reporting less self-efficacy for exercise. Race was also noted in this sample to influence outcome expectations, with White participants more likely to believe in the outcomes of exercise. Older age and African American race, both of which are linked to an increased risk of stroke (AHA, 2004), may also predispose patients to lower self-efficacy and poorer outcome expectations for exercise and therefore poorer outcomes. Hence, both of these demographic variables may identify populations that may most benefit from interventions that strengthen their beliefs and improve exercise behavior.

Fatigue influenced beliefs about ability to exercise and the outcomes of exercise. Fatigue is a well-documented consequence of stroke that can occur at any time during the recovery process, is widely prevalent, and persists over time. Fatigue is multifactorial, related to cardiovascular deconditioning, increased energy cost for gait deficits, or psychological factors including depression (Michael, 2002). Regardless of the cause, fatigue is an important barrier to physical activity that may be modifiable with exercise. Clearly, education is needed to modify knowledge and beliefs about the benefits of exercise to alleviate fatigue and improve overall health status.

Being told to exercise by a primary healthcare provider had an important influence on outcome expectations and directly influenced exercise behavior. This adds to the current base of evidence that counseling to exercise by a healthcare provider can influence patient behavior. Prior research related to health behaviors in older adults has demonstrated that when healthcare providers talk with their older patients about healthcare practices, they can influence this behavior. The Centers for Disease Control and Prevention (CDC) reported data from the National Health Interview Survey of 1998, and only 52% of respondents 50 years and older reported they had been asked about physical activity (CDC, 2002). In a study of older adults, the 48% who reported receiving a physician recommendation to exercise were sedentary or overweight, suggesting that physician advice is targeted only at the highest risk groups (Damush, Stewart, Mills, King, & Ritter, 1999). Similarly, only 45% of the respondents of this survey received a recommendation to exercise. Specific motivational interventions given by healthcare providers to encourage exercise have been noted to change behavior and help older adults initiate and adhere to regular exercise programs (Haber & Looney, 2000; Marcus et al., 1997).

This study further reveals that many stroke survivors perform minimal, if any, exercise on a regular basis. A majority of respondents reported they wanted to exercise more and would if opportunities were available. The identified determinants of self-reported exercise were self-efficacy for exercise, exercise history before stroke, outcome expectations of exercise, and recommendation from a physician. Three out of these four determinants are potentially modifiable. This raises the possibility that interventions designed to educate stroke survivors regarding outcome expectations and to strengthen self-efficacy may improve exercise behaviors. Further, a concerted effort to educate healthcare providers may result in more consistent recommendations for stroke survivors to exercise. In patients who have suffered a stroke, exercise has been demonstrated as an effective means of improving cardiovascular fitness reserve, muscle strength, and endurance (Macko et al., 2001; Ryan, 2000; Silver et al., 2000; Smith et al., 1999). Future research must assess and enhance the modifiable influences on exercise.


The findings of this study are limited by sampling bias. These respondents attended stroke support groups. This indicates they are both mobile and motivated to seek support and information. Such respondents are probably more likely to be engaged in health promoting activities. Moreover, only a small number of the available participants responded, with 236 paper responses from the mailing and 85 from the NSA Web site. This sample was also relatively homogeneous, with the majority of the sample being White, female, and unmarried, which may affect the generalizability of the findings. The study was cross-sectional, not a prospective, longitudinal examination of exercise beliefs and patterns of behavior. Prospective studies of diverse and well-characterized stroke cohorts are needed to elucidate influences on physical activity patterns in stroke survivors.


The variables in this model accounted for 33% of the variance in self-reported exercise behavior. Many other potential variables not included in the model may help to explain the lack of exercise behavior in this sample, including cognitive impairment, depression, social phobias associated with disabilities, fear of falling, lack of specific recommendations regarding optimal training modalities and their safety (Resnick & Spellbring, 2000), lack of access to appropriate facilities (Nies, Vollman, & Cook, 1998), or the effect of unpleasant sensations such as pain associated with exercise (Nies et al.; Resnick & Spellbring). Continued research incorporating these additional factors into the model may help to further explain exercise behavior in patients after stroke.

The U.S. Preventive Services Task Force suggests that a lack of physical activity plays a major role in poor health status and excess disability (2002). Healthy People 2010 (U.S. Department of Health and Human Services, 2000) outlines a global recommendation for all Americans to increase their levels of physical activity, suggesting a minimum of 30 minutes on most, if not all, days of the week. The aging of the American population and the resulting expected increase in stroke survivors accentuates the need for further research to define strategies to eliminate barriers to exercise to increase cardiovascular health, fitness, and function in this population.


The authors acknowledge the contributions to this article made by Dr. Steven Kittner and Dr. Denise Orwig at the University of Maryland in the development of the initial survey instrument; Dr. Daniel Hanley of Johns Hopkins University Medical Center Department of Neurology, and the leadership and staff of the National Stroke Association for their assistance in distribution of the original surveys and subsequent interactive online versions; Mona Choi, doctoral candidate at the University of Maryland Baltimore School of Nursing for assistance with data entry; and Kathleen Michael, PhD RN, for editorial assistance.

This study was sponsored by a Veterans Administration Post Doctoral Nursing Research Fellowship, University of Maryland, Baltimore School of Nursing designated research initiative funds, Baltimore VA Geriatrics Research, Education and Clinical Center, and the Claude D. Pepper Older Adult Independence Center.

About the Authors

Marianne Shaughnessy, PhD CRNP, is associate director of education and evaluation at the Baltimore Veterans Administration Medical Center, Geriatrics Research, Education, and Clinical Center, and assistant professor at the University of Maryland School of Nursing, Baltimore, MD. Address correspondence to her at Baltimore VA Medical Center (BT/18/GR), 10 N. Greene Street, Baltimore, MD 21201, or to mshaughn@grecc.umaryland.edu.

Barbara M. Resnick, PhD CRNP, is a professor at the University of Maryland School of Nursing.

Richard F. Macko, MD, is associate director of research at the Baltimore Veterans Administration Medical Center, Geriatrics Research, Education, and Clinical Center, and a professor in the department of neurology at the University of Maryland School of Medicine.


American College of Sports Medicine. (1998). The recommended quantity and quality of exercise for developing and maintaining cardiorespiratory and muscular fitness and flexibility in healthy adults. Medicine and Science in Sports & Exercise, 30, 975–991.

American Heart Association. (2004). Heart disease and stroke statistics—2005 update. Dallas: Author.

Arbuckle, J. (1997). Amos users’ guide version 3.6. Chicago: Small Waters Corporation.

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191–215.

Bollen, K. (1989). Structural equations with latent variables. New York: John Wiley & Sons.

Centers for Disease Control and Prevention. (2002). Prevalence of healthcare providers asking older adults about their physical activity levels—United States, 1998. Morbidity & Mortality Weekly Report, 51(19), 412–414.

Damush, T. M., Stewart, A. L., Mills, K. M., King, A. C., & Ritter, P. L. (1999). Prevalence and correlates of physician recommendations to exercise among older adults. Journals of Gerontology: Series A, Biological Sciences and Medical Sciences, 54(8), M423–M427.

Dishman, R. (1994). Motivating older adults to exercise. Southern Medical Journal, 87, S79–S82.

Haber, D., & Looney, C. (2000). Health contract calendars: A tool for health professionals with older adults. Gerontologist, 40(2), 235–239.

Hill, K. D., Schwarz, J. A., Kalogeropoulos, A. J., & Gibson, S. J. (1996). Fear of falling revisited. Archives of Physical Medicine and Rehabilitation, 77(10), 1025–1029.

Jette, A. M., Rooks, D., Lachman, M., Lin, T. H., Levenson, C., Heislein, D., et al. (1998). Home-based resistance training: Predictors of participation and adherence. Gerontologist, 38(4), 412–421.

Loehlin, J. (1998). Latent variable models. Mahwah, NJ: Lawrence Erlbaum Associates.

Lofgren, B., Nyberg, L., Mattson, M. & Gustafson, Y. (1999). Three years after in-patient stroke rehabilitation: A follow up study. Cerebrovascular Disease, 9, 163–170.

Macko, R. F., Smith, G. V., Dobrovolny, C. L., Sorkin, J. D., Goldberg, A. P., & Silver, K. H. (2001). Treadmill training improves fitness reserve in chronic stroke patients. Archives of Physical Medicine and Rehabilitation, 82(7), 879–884.

Marcus, B. H., Goldstein, M. G., Jette, A., Simkin-Silverman, L., Pinto, B. M., Milan, F., et al. (1997). Training physicians to conduct physical activity counseling. Preventive Medicine, 26(3), 382–388.

McAuley, E. (1993). Self-efficacy and the maintenance of exercise participation in older adults. Journal of Behavioral Medicine, 16(1), 103–113.

Michael, K. (2002). Fatigue and stroke. Rehabilitation Nursing, 27(3), 89–94, 103.

Morey, M. C., Pieper, C. F., & Cornoni-Huntley, J. (1998). Is there a threshold between peak oxygen uptake and self-reported physical functioning in older adults? Medicine and Science in Sports and Exercise, 30(8), 1223–1229.

Nies, M. A., Vollman, M., & Cook, T. (1998). Facilitators, barriers, and strategies for exercise in European-American women in the community. Public Health Nursing, 15(4), 263–272.

Paolucci, S., Grasso, M. G., Antonucci, G., Bragoni, M., Morelli, D., Cioro, P., et al. (2001). Mobility status after inpatient stroke rehabilitation: 1 year follow up and prognostic factors. Archives of Physical Medicine and Rehabilitation, 82, 2–8.

Pedhazur, E. S. L. (1991). Measurement, design & analysis: An integrated approach. Mahwah, NJ: Lawrence Erlbaum Associates.

Resnick, B. (1999). Motivation to perform activities of daily living in the institutionalized older adult: Can a leopard change its spots? Journal of Advanced Nursing, 29, 792–799.

Resnick, B., & Nigg, C. (2003). Testing a theoretical model of exercise behavior for older adults. Nursing Research, 52(2), 80–88.

Resnick, B., & Spellbring, A. M. (2000). Understanding what motivates older adults to exercise. Journal of Gerontological Nursing, 26(3), 34–42.

Resnick, B., Zimmerman, S. I., Orwig, D., Furstenberg, A. L., & Magaziner, J. (2000). Outcome expectations for exercise scale: Utility and psychometrics. Journals of Gerontolgoy. Series B, Psychological Sciences and Social Sciences, 55(6), S352–356.

Ryan, A., Dobrovolny, L., Silver, K. H., Smith, G. V., & Macko, R. F. (2000). Cardiovascular fitness after stroke: Role of muscle mass, and gait deficit severity. Journal of Stroke and Cerebrovascular Disease, 9, 1–8.

Sharon, B., Hennessey, C., Brandon, J. & Boyette, L. (1997). Older adults’ experiences of a strength training program. The Journal of Nutrition, Health and Aging, 1: 103–108.

Shaughnessy, M., Resnick, B., & Macko, R. F. (2004). Reliability and validity testing of the short self-efficacy and outcomes expectations for exercise scales in stroke survivors. Journal of Stroke and Cerebrovascular Disease, 13(5), 214–219.

Silver, K. H., Macko, R. F., Forrester, L. W., Goldberg, A. P., & Smith, G. V. (2000). Effects of aerobic treadmill training on gait velocity, cadence, and gait symmetry in chronic hemiparetic stroke: A preliminary report. Neurorehabilitation and Neural Repair, 14(1), 65–71.

Smith, G. V., Silver, K. H., Goldberg, A. P., & Macko, R. F. (1999). “Task-oriented” exercise improves hamstring strength and spastic reflexes in chronic stroke patients. Stroke, 30(10), 2112–2118.

U.S. Department of Health and Human Services. (2000). Healthy people 2010. Rockville, MD: Author.

U.S. Preventive Services Task Force. (2002). Behavioral counseling in primary care to promote physical activity: Recommendations and rationale. Annals of Internal Medicine, 137, 205–207.