Abstract

Objective: Follow-up data were collected using postal questionnaires and if participants did not respond, then data was collected using telephone interviews. The objectives of this study were to examine, for the two methods, how respondents differed in characteristics and whether the observed treatment difference varied. Study Design and Setting: A large clinical trial of lower back pain. Results: About 60 (98/163) of the nonresponders to postal questionnaire provided data by telephone, which increased the overall response rate by 14 (from 71 to 85). A consistent treatment difference was found across the methods for the outcome measures at 12 months, implying that the observed treatment effect had not been modified. There were some differences between the participants: responders of postal questionnaire were older, likely to be female, white (ethnic origin), not working, with less disability of back pain, compared with those who responded by a telephone interview. At 12 months, there was greater improvement in back pain, disability, and general health for those who responded by postal questionnaires. Conclusion: Researchers should consider the use of more than one method of collecting data as this increases response rate, participant representativeness, and enhances precision of effect estimates. © 2012 Elsevier Inc. All rights reserved.