PERSonalised Incentives for Supporting Tobacco cessation (PERSIST) among healthcare employees: evaluation and lessons learned 377 11 of the design by the researchers.29 Incorporating potential participants in the design of a study might help overcome such mismatches in beliefs between researchers and participants. Strengths and limitations In addition to the limitations pointed out in the interviews, the PERSIST trial has a relatively high lost-to-follow-up rate. Most likely this was due to the fact that those who started smoking again did not complete the questionnaires anymore. This was addressed by considering those lost-to-follow-up as non-abstinent, in accordance with the Russell standard.18 The PERSIST trial and the interview study were further limited by underrepresentation of lower-educated persons, although many lower-educated persons work in hospitals and tobacco use is more common among lower-educated persons in the Netherlands.30 Finally, CO measurements only detect smoking up to 24 hours prior to the measurement, and may not be ideal to capture sustained abstinence. However, informing participants that there will be CO measurement will increase the validity of self-reported abstinence.31 The interview study also likely suffered from selection bias. Less opinionated individuals might have been less interested in participating. The recruitment took place within and around a smoke-free zone, therefore individuals might have felt ‘caught’ and thus be less willing to participate. In addition, there could be recall bias, as the interview study took place more than a year after the smoking cessation program had been implemented. All interviews were conducted at one of the participating hospitals for practical reasons. Therefore, there could be differences between employers that were not captured in our interviews. Finally, it is important to realize that those who did not sign up for interviews might overlap with those who did not sign up for smoking cessation programs at the workplace. Hence, there is a risk that we may lack information about how to reach this group, as they can be inherently different from those who agreed to participate in the interviews. Recommendations for practice and directions for future research Despite these limitations, several recommendations for future research and implementation can be derived from this study. Recruitment issues such as those experienced in PERSIST are not unique and are common in smoking cessation research and public health. The first practical advice to employers would be to
RkJQdWJsaXNoZXIy MTk4NDMw