Our systematic review of articles on single risk factor interventions to promote physical activity among patients affected by at least 1 chronic disease in primary care settings had seemingly conflicting results. Two studies concluded that the interventions evaluated had no effect on levels of physical activity, while the other reported a positive short-term effect using a somewhat intensive intervention.
The 3 studies in our review differed with respect to the instruments used to quantify physical activity and the nature and intensity of interventions. The studies used different validated questionnaires (Godin’s questionnaire,50,51
the International Physical Activity Questionnaire,52
and the Short Questionnaire to Assess Health-enhancing Physical Activity53
). To assess the level of activity more accurately, objective measuring devices, such as pedometers or accelerometers, could be used along with the questionnaires. As to the nature and intensity of interventions, we observed that Little et al,46
who measured a short-term (1 month) effect, evaluated an intervention that went beyond simple counseling by a general practitioner to integrate other health professionals and psychological techniques aimed at changing patients’ behaviour. The study by van Sluijs et al,48
however, also used these interventions without demonstrating a short- or long-term effect. In that study,48
contact with a counselor was only by telephone. Perhaps face-to-face counseling would have been more effective. These 2 studies were also based on different behavioural theories; the first on the theory of planned behaviour and the other on stage-of-change theory. Godin et al have argued that level of physical activity would be measured better by a model taking into account, not in sequence but simultaneously, both intention to engage in activity in the near future and recent past participation in activity.50,51
The different nature and intensity of interventions in the 3 studies might partly explain their different results. Studies evaluating more intensive interventions are needed.
Critical reviews of studies evaluating interventions to promote physical activity among more general primary care populations also had varying results. Four reviews reported there was not enough evidence to conclude that the interventions were effective34–37
; 7 had more positive results.38–43,54
Variation in the nature and intensity of interventions in these studies might explain the differences in results here also. Most of the authors agreed on the need for further high-quality research. Interestingly, 2 reviews that reported positive effects concluded that studies in which interventions were tailored to participants’ characteristics or that considered behaviour change strategies were effective.41,42
The article by van Sluijs et al48
reported that the study population as a whole had significantly increased their level of physical activity at 1-year follow up. As discussed in the article, increases in activity levels among control groups have often been observed in randomized controlled trials promoting physical activity.48,55–57
The authors proposed some explanations for this finding. First, they assessed whether there was a measurement effect. By comparing a group measured in 4 ways with a group measured in 2 ways at 6-month follow-up, they were able to demonstrate a positive measurement effect on level of activity. This might explain the observed increase in activity levels among control subjects. Second, the Hawthorne effect, which is the tendency of people who are singled out for special attention to perform better merely because of the expectations created by the situation,58
might have contributed to this increase as well. Third, general practitioners in the control arm might have offered more advice on activity to their patients than is usually the case. This study stresses the challenges of doing randomized controlled trials on promoting physical activity and the importance of taking biases, such as the Hawthorne and measurement effects as well as contamination, into account in designing studies and interpreting results.
Further research is needed to identify which kinds of interventions could be effective over the long term among chronically ill patients. Collaborative nurse-physician practices might be good settings in which to develop such interventions. In addition to patients with chronic conditions related to cardiovascular disease, future studies could include patients with other problems such as osteoarthritis or asthma.
The main limitation of a systematic review is its inability to include all the relevant literature and all the unpublished material on the subject. We realize that some articles could have been missed during our searches, even though our search strategy using all the terms as key words favoured a more exhaustive literature review. Our hand search was another way to help us identify additional relevant articles. Restricting the search to articles published in French or English was also a limitation.
Other interventions could have been used to promote physical activity among these primary care patients within a framework of programs targeting other behaviour (eg, smoking, reducing weight). As a preliminary step, we preferred to focus only on single risk factor interventions to document their isolated effect because trying to change behaviour in several areas at the same time might have made a difference to each behaviour change.
There are insufficient data at this time to assess the effectiveness of single risk factor interventions to promote physical activity among patients affected by at least 1 chronic disease in primary care. We reviewed 3 studies: 2 concluded that the interventions had no effect on level of activity, while the other reported a short-term positive effect. The intervention deemed effective was based on the theory of planned behaviour and integrated physican-nurse counseling. Further research is needed to identify which kinds of interventions would be effective over the long term among chronically ill patients.