Review of Primary Care-Based Physical Activity Intervention Studies

Key Physical Activity Outcome and Length of Follow-Up

Studies used a wide range of physical activity outcome variables, with varying definitions of sedentary lifestyle, regular and moderate activity, and vigorous exercise. All studies relied on patient self-reports of physical activity, with one study using activity monitors and one including a measure of physical fitness (VO2 max) to corroborate self-report. In both studies, the significant intervention effect on the self-reported physical activity outcomes was corroborated by this additional activity monitor or fitness data. Length of follow-up ranged from 4 weeks to 4 years, with 7 studies reporting follow-ups of 12 months or longer. Few studies reported on outcomes other than physical activity or the multiple risk factors associated with the intervention. Three studies assessed body mass index, one assessed cholesterol and blood pressure, 2 assessed quality of life, and one reported on cost-effectiveness.

Although the majority of studies did not specifically report on subject attrition, it was possible to calculate attrition rates for all studies. Attrition was moderately high and ranged from 1% to 44% (median = 18%) at follow-ups of fewer than 12 months, and 20% to 56% (median = 41%) at longer-term follow-ups. Only 4 of the 11 studies with attrition rates of 15% or higher used intention-to-treat analyses or imputation methods to address attrition. Those patients who dropped out tended to be less educated, more likely to smoke, and in poorer health.

RE-AIM Criteria

Eleven of 15 studies reported on the percentage of the eligible population who agreed to participate (Reach), ranging from 35% to 100% (median = 74%), but only 3 of those described whether enrolled subjects were representative of the larger population from which the sample was drawn. In all 3 studies, participants differed from nonparticipants on demographic variables such as sex and smoking status, raising concerns about the representativeness of the samples studied. In contrast, only 3 studies reported on the percentage of eligible primary care settings and providers who agreed to participate in the study (Adoption).

The authors of 12 studies reported at least some data on intervention implementation. The most common implementation measures were whether patients attended a visit or whether physicians delivered advice. Implementation for these relatively crude indices was variable (range = 30%-100%) but generally high, especially for delivery of advice (80%-100%).

In terms of efficacy, of the 10 studies reporting 0- to 11-month postintervention outcomes, 7 reported statistically significant physical activity outcomes. Regarding maintenance, of the 7 studies reporting 12-month or longer postintervention outcomes, 3 reported statistically significant outcomes. As in other health behavior change areas, it appears challenging to maintain initial treatment effects. Two of the 3 studies reporting both short-term and long-term results^32,33 found significant efficacy results at initial follow-ups, but in all 3 studies results were no longer significant at later follow-up.

Discussion

Brief primary care-based physical activity interventions are effective in producing moderate short-term improvements in self-reported physical activity levels. Of the 10 studies reporting short-term outcomes, 7 reported statistically significant results. A more detailed evaluation of intervention characteristics revealed a number of factors associated with successful outcomes, including brief interventions (of 3 to 10 minutes) that focused on physical activity only, were tailored to patient characteristics and preferences, and included supplemental written materials. In contrast, longer interventions focusing on multiple risk factors that did not include written patient materials did not achieve significant short-term results for physical activity. Short-term improvements on physical activity were observed across all types of interventionists, including physicians, nurses, and combinations of physicians and health educators.

There are some intervention components that seem logical that we cannot recommend. Unlike the literature in other health behavior change areas,^34,35 the use of follow-up supports after the initial intervention did not increase the likelihood of positive short-term results. Another surprise was that theory-based interventions did not appear more successful than those not explicitly theory based; specifically, interventions using the transtheoretical model were not effective. None of the studies that focused solely on older adults (those aged 50 years and older) achieved positive short-term results, indicating that this is an area in need of further attention. The issue of long-term effectiveness is difficult to evaluate. Only 7 studies reported long-term outcomes, and only 3 of those achieved significant results.

RE-AIM Conclusions

Compared with the general health promotion literature, more studies reported on the Reach and representativeness of participants than is typically observed,^36-38 and these results were encouraging. Primary care-based physical activity interventions are an effective means of reaching a large segment of sedentary adults, although the representativeness of study samples remains unclear. Men, smokers, and older adults appear less likely to participate in physical activity interventions.