The Effect of Survey Nonresponse on Quitline Abstinence Rates: Implications for Practice.

INTRODUCTION

Quitline outcome studies are used to maintain and improve the effectiveness of these evidence-based cessation services. Nonresponse has the potential to bias survey results and many US and Canadian quitlines are reporting survey response rates below 50%. This study examines the effect of nonresponse bias on quit rates in three state quitline populations. Results provide implications of nonresponse bias for quitline practice.

METHODS

Quit status, defined as abstinent for 30 days or more 7 months after registering for services, was collected from Minnesota, Hawaii, and Florida quitline participants that responded to a survey. We assigned each responder to a wave based on the number of contacts required to obtain a survey response.

RESULTS

The latest two responder groups had the lowest quit rates within each state, although results were not statistically significant. Quit rates in the latest responder wave (Wave 6) were between 4% and 13% points lower than the earliest responders (Wave 1). The cumulative quit rates show what the quit rate would have been had the study ended after the corresponding wave. In all four studies, the cumulative quit rate was lowest in Wave 6.

CONCLUSION

To increase accuracy of quit rates, quitlines should focus on increasing survey response rates. Suggestions for improving survey response rates are provided.