Citation

Gattellari M, Ward JE. J. Public Health (Oxford) 2004; 26(4): 384-387.

Affiliation

School of Public Health and Community Medicine, University of New South Wales & Centre for Research, Evidence, Management and Surveillance, South Western Sydney Area Health Service, Australia. Melina.Gattellari@swsahs.nsw.gov.au

Copyright

(Copyright © 2004, Oxford University Press)

DOI

10.1093/pubmed/fdh177

PMID

15598859

Abstract

BACKGROUND: Self-administered questionnaires are commonly used in experimental studies to elicit quality of life or other outcomes. Hence, achieving an acceptable level of follow-up from patients is critical to minimizing bias. Many methods for maximizing follow-up remain untested. It is also unclear what level of follow-up is required to prevent bias being introduced. METHODS: We recruited 246 men from general practice surgeries in Sydney, Australia. These 246 men were randomized to receive a covering letter with their follow-up questionnaire either advising of a deadline to reply (Deadline, n = 126) or a standard letter without a deadline (No Deadline, n = 120). Four standardized reminder prompts subsequently were administered. We calculated interim response rates and the final proportion of follow-up questionnaires received according to group. We also compared scores on two main outcomes, namely, knowledge and decisional conflict at each time when reminder prompts were administered. RESULTS: One hundred and twelve (88.9%) men in the Deadline group returned their follow-up questionnaires compared with 102 (85.0%) men in the No Deadline group. This difference was not statistically significant [odds ratio = 1.41, 95% confidence interval (CI) = 0.67-2.99; p = 0.36]. Time to response also was not significantly affected by cover letter received (hazard ratio = 0.96; 95% CI = 0.73-1.25; p = 0.76). Results of the original RCT were similar in terms of direction and effect size at all times irrespective of when reminder prompts were administered. CONCLUSION: The addition of a deadline adds no further impact in improving response rates from male patients compared with an unspecified letter. Despite the accepted wisdom that higher response protects against bias, differences in outcomes were consistent throughout the post-test data collection period.

Language: en