Text message reminders to improve questionnaire response rates in RCTs: findings from three randomised sub-studies

Background

Valid treatment effect estimates in the analysis of RCTs using patient reported outcomes depend on adequate response rates. Losses to follow-up are often high, and inexpensive ways to improve retention are much sought.

Aim

To assess the effectiveness of reminders sent by SMS text messages before or after questionnaire distribution on patient response rate and time to response in a mental health trial population.

Methods

Three randomised sub-studies were embedded in the UK ACUDep trial at three follow-up points. 523 patients of 755 in the main trial consented to being contacted by text message and were randomised to a pre-questionnaire reminder or no reminder at 3 months, a pre-reminder or post-reminder at 6 months and a post-reminder or no reminder at 9 months. Chi square tests and time-to event analyses were used to assess attrition between groups.

Results

Return rates for pre-reminder SMS were not significantly different at 3 months compared to no reminder (82.9% vs 84.7%, p=.580), but showed significantly lower response rates at 6 months compared to post-reminders (75.2% vs 83.2%, p=.025). Return rates following post-reminders did not significantly differ at 9 months from no reminders (77.1% vs 78.5%, p=.691). Median times to response ranged from 18 to 25 days, with only returns at 6 months being significantly superior for patients receiving post-reminders (log-rank test p=.044).

Conclusions

Overall, SMS text reminders did not appear to substantially improve patient response rates, although the pattern of effects for reminders sent before or after questionnaire distribution was inconclusive.