PERSonalised Incentives for Supporting Tobacco cessation (PERSIST) among healthcare employees: a randomised controlled trial protocol 347 10 of <10 ppm is considered indicative of abstinence.23 CO quantification is a reliable and non-invasive method to ascertain recent smoking and the preferred method in populations using nicotine replacement therapy.23 In case of disagreement between self-reported abstinence and CO reading, abstinence is defined by the latter. Secondary outcomes are smoking abstinence directly after the group-based smoking cessation training (t=0), and after three months (t=3), after six months (t=6) and after fifteen months (t=15). The same methods and definitions to determine continuous abstinence are used for all outcomes. Tertiary outcomes are the incremental cost-effectiveness ratio and the incremental cost utility ratio, where the first is calculated using the total costs per quitter and the latter by the costs and self-reported quality of life (EQ-5D). Sample size The CATCH trial, which employed similar paying schemes as the standard arm of the current trial, reported 41.1% abstinence at twelve months follow up in the intervention arm, versus 26.4% among participants who only attended the training sessions.13 The PERSIST trial will be embedded in health care institutions, many of which are currently implementing smoke-free policies, which is expected to be an extra external motivator for cessation. As such the validated abstinence percentages are expected to be slightly higher than in the CATCH trial, i.e. 45% and 30%, respectively. We expect that personalisation of the incentives will further increase effectiveness to approximately 50%. In order to detect this difference in validated sustained abstinence (i.e. 50% vs. 30%) with a power of 0.8 and an alpha of 0.05 and using 1:1 randomisation, a total of at least 186 participants is needed. Taking into account a 15% loss to follow-up due to unexpected employee turnover,11, 24 220 participants are required. Further taking into account the potential maximum deviation from a 1:1 allocation ratio which may arise due to the stratified block randomisation procedure used, the minimum sample size required will increase in proportion to the number of different hospitals included in the trial. As the trial aims to estimate the effect of personalised incentives compared to no incentives and not to estimate the effect of each individual scheme, skewed allocation among the schemes is not an issue and is therefore not accounted for in the required sample size.
RkJQdWJsaXNoZXIy MTk4NDMw