The feasibility of rapid baseline objective physical activity measurement in a natural experimental study of a commuting population

Background: Studies of the effects of environmental interventions on physical activity should include valid measures of physical activity before and after the intervention. Baseline data collection can be difficult when the timetable for introduction of an intervention is outside researchers' control. This paper reports and reflects on the practical issues, challenges and results of rapid baseline objective physical activity measurement using accelerometers distributed by post in a natural experimental study. Methods: A sample of working adults enrolling for the Commuting and Health in Cambridge study and expressing willingness to wear an activity monitor was selected to undertake baseline accelerometer assessment. Each selected participant received a study pack by post containing the core study questionnaire and an accelerometer to wear for seven consecutive days, and was asked to return their accelerometer and completed questionnaire in person or by post using the prepaid special delivery envelope provided. If a pack was not returned within two weeks of issue, a reminder was sent to the participant. Each participant received up to five reminders by various methods including letter, email, telephone and letter sent by recorded delivery. Results: 95% of participants registering for the study were willing in principle to undertake accelerometer assessment. Using a pool of 221 accelerometers, we achieved a total of 714 issues of accelerometers to participants during a six month period. 116 (16%) participants declined to use the accelerometer after receiving it. Three accelerometers failed, 45 (6% of 714) were lost and many were returned with insufficient data recorded, resulted in 109 (15%) participants re-wearing their accelerometer for a second week of measurement. 550 (77%) participants completed data collection, 478 (87% of 550) to the required standard. A total of 694 reminders were issued to retrieve unreturned accelerometers. More than 90% of accelerometers were retrieved after a maximum of two reminders. Conclusions: It is feasible to use accelerometers to collect baseline objective physical activity data by post from a large number of participants in a limited time period. However, a substantial pool of devices is required and researchers need to be prepared to make significant efforts to recover some of the devices.