87 Randomized controlled trial of a social network intervention 3 Data management After inclusion, a unique project number will be allocated to each participant. The key to these project numbers will be stored in a password protected file. All paper and digital data will be stored encoded in an online validated data management environment – Castor EDC – by a research assistant. All stored data will be double checked, corrected (if necessary) and signed. Only principal investigators and trained research associates will have access to Castor EDC and research files. Paper questionnaires, other data and informed consents will be stored in a locked cabinet at the participating site. Data will be stored for 15 years according to legislation. Storage of data will be supervised by the principal investigator in compliance with the Dutch Personal Data Protection Act. The study will be conducted based on the quality handbook of the Amsterdam Public Health (APH) research institute. Data monitoring is provided by independent researchers at the ArkinMental Health Institute. An audit can be initiated by an independent researcher of the Amsterdam University Medical Center throughout the course of the study. During an audit, the independent researcher will investigate whether the execution of the study is based on the quality standards of the APH research institute. Statistical analysis An intention-to-treat analysis will be included in the data analyses of the primary outcome measure (mental wellbeing) and secondary outcome measures (psychiatric functioning, criminal recidivism, substance abuse, quality of life, social network, social support, loneliness, and self-sufficiency). Besides, subgroup analyses will be conducted on the protocol completers sample. Generalized linear mixed models (GLMM) in Statistical Package for the Social Sciences (SPSS) will be used to model all continuous outcome measures, considering the distributional characteristics of the data. If the normal distribution assumption is violated, resampling methods will be used to transform data into a normal distribution or appropriate non-parametric tests will be used. Generalized estimating equations (GEE) will be used to analyze the categorical outcome measures. The main between-group effect (TAU versus FNC) and the interaction effect (group * time) will be analyzed whilst correcting for baseline assessment. Outcome data of all assessments will be included in the analyses to examine the differences in the development in time between both groups. The criminal recidivism data will be analyzed using Survival Analysis and corrected for time at risk. Missing data will be addressed using multiple imputation or statistical techniques that control for missingness-at-random.
RkJQdWJsaXNoZXIy MjY0ODMw