Exercise Programs Help Reduce Anxiety in Patients with Chronic Illnesses

Abstract & Commentary

By Dónal P. O'Mathùna, PhD . Dr. O'Mathùna is Senior Lecturer in Ethics, Decision- Making & Evidence, School of Nursing, Dublin City University, Ireland; he reports no financial relationship to this field of study.

Synopsis: A meta-analysis of 40 randomized controlled trials found evidence of reduced anxiety among patients with chronic illnesses who exercised regularly. The impact of several study variables was investigated as potential contributors to the variability found in different studies.

Source: Herring MP, et al. The effect of exercise training on anxiety symptoms among patients: A systematic review. Arch Intern Med 2010;170:321-331.

A systematic review was conducted by searching the Physical Activity Guidelines for Americans Scientific Database from January 1995 to August 2007. This search was supplemented by further searches of MEDLINE, Web of Science, PsycINFO, and Google Scholar up to the end of 2008. One goal of the systematic review was to provide an overall effect size for patients with chronic illnesses. The second goal was to examine several variables including the type or dose of exercise training and whether these accounted for some of the variation found in the results.

The inclusion criteria for studies were: 1) written in English, 2) participants were sedentary adults with a chronic illness, 3) random assignment to either an exercise intervention lasting at least 3 weeks or an intervention lacking exercise, and 4) an anxiety outcome that was measured pre- and post-exercise training. The search identified 135 potential studies, of which 40 met all inclusion criteria.

The most common conditions among the patients in the included studies were cardiovascular disease, fibromyalgia, multiple sclerosis, psychological conditions, or cancer. The exercise training programs averaged 3 days per week, 42 minutes per session, and lasted 16 weeks. For all outcomes, an effect size was calculated based on standard statistical methods and gave positive values for reductions in anxiety.

A total of 75 different effects were derived from the 40 studies and these were included in the meta-analysis. A random effects model was used for the meta-analysis to take account of heterogeneity among the studies. The overall mean effect size was 0.29 with a 95% confidence interval (CI) of 0.23-0.36. Since the CI did not include 0, the result was statistically significant.

To address the second goal of the systematic review, regression analysis was conducted on six primary moderators: exercise session duration, program length, time frame of anxiety report, fitness level change, intervention type, and comparison type. The first three items were found to be independently related to effect size. Exercise sessions lasting longer than 30 minutes had larger effects than those lasting 10-30 minutes. Exercise programs of 3-12 weeks resulted in significantly larger reductions in anxiety than those lasting more than 12 weeks. The effect size was smaller when anxiety levels were reported for "right now" or during the past week, compared to those measuring anxiety over periods exceeding the previous week.

Several secondary moderators were examined to provide descriptive data about plausible influences on outcomes. Only two were discussed in detail: illness type and adherence. Exercise significantly impacted anxiety levels in those with all types of illness except multiple sclerosis. Adherence was not found to moderate the anxiety outcomes, but this may have been because adherence data were not collected in many studies.


Anxiety often accompanies chronic illness. However, health care professionals may not recognize symptoms of anxiety or provide suggestions for the relief of anxiety. Failure to address anxiety can negatively impact treatment outcomes and increase the suffering of patients with chronic illness. Pharmacological and cognitive behavioral therapies for anxiety are available, but interest in complementary interventions is common. Such interventions include a variety of relaxation programs. Exercise training may also help reduce anxiety and provide a number of other health benefits.

This systematic review provides clear evidence of the beneficial effects of exercise for those with chronic illnesses (overall mean effect size = 0.29). What is more difficult to interpret is how much benefit these programs are likely to provide. The "effect size" is one way of representing the difference between the mean of the treatment group and that of the control group.1 It is commonly used in meta-analyses when a variety of methods are used to measure the same outcome. An effect size is sometimes viewed as small if its value is 0.20 or less, high if 0.80 or greater, and medium if in between.

A meta-analysis provides a summary statistic for outcomes tested in several trials. One of the challenges with having several trials addressing the same outcome is how to account for varying estimates of effects. This meta-analysis found fairly consistent results in that 66 of the 75 effects calculated for the trials were greater than zero. Meta-analysis provides further opportunities to investigate such variability and generate hypotheses.2 These calculations do not demonstrate causation, but can provide suggestions about factors that can guide future research and make suggestions about factors that are important for clinical developments.

Based on these results, exercise programs had better outcomes when they involved individual sessions lasting longer than 30 minutes. Programs of 3-12 weeks duration had better outcomes, but the reasons for this are unclear. The reviewers speculated that it might be connected to adherence. However, about one-third of the studies did not report information on adherence, which would be important for future studies to record. The time frame over which participants were asked to evaluate their anxiety was also correlated to effect size. Eighty percent of the studies asked about anxiety over the previous week, while anxiety measured over a longer time frame showed greater reductions. The reviewers suggested that researchers should ask about longer-term anxiety levels as this could confirm if exercise has longer-lasting effects. Other nonsignificant findings were also interesting. For example, little difference in effect size was found for age, gender, or whether exercise guidelines met contemporary intensity recommendations.

This meta-analysis was carefully designed and conducted according to the QUOROM statement.3 These guidelines are intended to improve the quality of reports of meta-analyses, just as the CONSORT statement is designed to improve the quality of RCT reports. The meta-analysis provides clear evidence for the recommendation of exercise programs for those with many chronic illnesses. However, further research is needed before many of the details of such programs can be reliably translated into clinical guidelines.


1. Valentine JC, Cooper H. Effect Size Substantive Interpretation Guidelines: Issues in the Interpretation of Effect Sizes. Washington, DC: What Works Clearinghouse; 2003.

2. Rosenthal R, DiMatteo MR. Meta-analysis: Recent developments in quantitative methods for literature reviews. Annu Rev Psychol 2001;52:59-82.

3. Moher D, et al. Improving the quality of reports of meta-analyses of randomised controlled trials: The QUOROM statement: Quality of reporting of meta-analyses. Lancet 1999;354:1896-1900.