Developing an app-based behavioral intervention to help depressed individuals return to work

The Northwestern site will oversee the two phases of the study (design and develop [Phase 1] and empirically test the efficacy [Phase 2] of DRIVEN). Drs. Griffith and Shankman will be responsible for the data analysis, working with Ms. Jackson, the statistician on the project. Data analyses will be presented at group meetings, including consultants, to ensure a comprehensive view of the results. All analyses will be document in code (e.g. R, Mplus) and organized and annotated. Additionally, the Northwestern site will provide significant expertise in implementation science (Co-I Wolf), biostatistics (Co-I Jackson), and coordinate employment-focused interventions for individuals with psychiatric conditions, and organizational behavior. Dr. Shankman will oversee the collection of the psychiatric symptom and diagnostic data. The postdoctoral fellow and professional RAs will be responsible for processing and preparing symptom data for analysis. For the psychiatric interview data, Dr. Shankman will be responsible for evaluating and maintaining rater consistency with gathering both diagnostic and symptom data. Postdoctoral fellows and RAs will administer the psychiatric interviews. MPI Shankman will train each of these individuals by (a) watching videos of model psychiatric cases (recorded by the developers of the psychiatric interview used in the study [SCID]), and (b) conducting practice interviews along with MPI Shankman where the interviewer achieves perfect reliability with the MPI. The MPIs will also oversee maintenance of diagnostic and symptom assessment reliability, by conducting regular meetings (attended by raters), holding regular training sessions with model cases, and regularly monitoring the online database for rater drift. Reliability will be assessed by randomly selecting 10% of interviews and coding interviews based on video recordings every 6 months by study interviewers. Kappa and ICC scores of .80 or higher will be judged reliable. If scores fall below .80, discrepancies will be examined and discussed among the PIs and all study interviewers to address potential drift.