Simple steps to develop trial follow-up procedures

Ona McCarthy, Rebecca S. French, Ian Roberts, Caroline Free

Loss to follow-up in randomised controlled trials reduces statistical power and increases the potential for bias. Almost half of all trials fail to achieve their follow-up target. Statistical methods have been described for handling losses to follow-up and systematic reviews have identified interventions that increase follow-up. However, there is little guidance on how to develop practical follow-up procedures. This paper describes the development of follow-up procedures in a pilot randomised controlled trial of a sexual health intervention that required participants to provide and return questionnaires and chlamydia test samples in the post. We identified effective methods to increase follow-up from systematic reviews. We developed and tested prototype procedures to identify barriers to follow-up completion. We asked trial participants about their views on our follow-up procedures and revised the methods accordingly. We identified 17 strategies to increase follow-up and employed all but five. We found that some postal test kits do not fit through letterboxes and that that the test instructions were complicated. After identifying the appropriate sized test kit and simplifying the instructions, we obtained user opinions. Users wanted kits to be sent in coloured envelopes (so that they could identify them easily), with simple instructions and questionnaires and wanted to be notified before we sent the kits. We achieved 92 % (183/200) overall follow-up for the postal questionnaire at 1 month and 82 % (163/200) at 12 months. We achieved 86 % (171/200) overall follow-up for the postal chlamydia test at 3 months and 80 % (160/200) at 12 months. By using established methods to increase follow-up, testing prototype procedures and seeking user opinions, we achieved higher follow-up than previous sexual health trials. However, it is not possible to determine if the increase in response was due to our follow-up procedures. Current Controlled Trials ISRCTN02304709 Date of registration: 27 March 2013.