A calling for standardised completion criteria in weight management

The criteria for participant completion of a weight management programme (WMP) is arbitrary. Programme commissioners (WMP purchasers) will frequently establish the percentage of attendance that classifies programme completion (e.g. 70% attendance). Differential criteria for WMP completion make it impossible for researchers, practitioners and policy makers to conclude what constitutes an effective programme and what factors predict WMP completion. This study exemplifies the impact of variable completion status on 1) BMI reduction, 2) volume of completers and 3) predictors of completion. Secondary data was obtained from MoreLife – a UK-based, community WMP for children (aged 4-17 years). 2948 children attended between 2009-2014 (Age 10.4±2.8 years, BMI 26.0±5.7kg/m2, Standardised BMI (BMI SDS) 2.48±0.87 units, White 70.3%). Separate analyses were conducted for research aims 1-2, and aim 3. Programme completion was adjusted incrementally by 10% (i.e. 10%, 20% attendance etc…) for research aims 1-2. The volume of programme completers and change in BMI SDS was calculated at each increment of the completion criteria (0-100%). For aim 3, programme completion was defined using five classifications from previous WMP studies (e.g. 50% sessions attended). Multivariable logistic regression determined participant and programme variables predictive of programme completion. Percentage difference between the odds ratio of the original model (completion = 70% attendance) and the four subsequent models was calculated. The volume of participants completing the programme decreased in a linear manner (r = -0.99, p = 0.00) when completion classification became more stringent (i.e. 70-100% attendance). Conversely, the change in BMI SDS became incrementally greater (r = 0.98, p = 0.00). Predictors of completion varied by up to 24.2% in certain variables (e.g. Programme Intake Period) when using five different completion classifications. Statistical significance of the predictor variables were reliant on completion classification (e.g. WMP Group Size was significant in two of five models). The volume of completers and change in BMI SDS were strongly associated with programme completion classification. Poor programme outcomes (e.g. minimal change in BMI SDS) can be masked by (un)demanding completion criteria. Moreover, completion criteria mediates participant and programme characteristics predictive of programme completion. Standardised completion criteria are called for.