Data analysesStatistical methodology. Structural equation modeling (SEM) was used to test the mathematical viability of an omnibus multivariate model of predictors of treatment adherenceamong individuals greater than one year post renal transplant. Neurocognitive Abilities,Everyday Problem-Solving, Depression, Self-Efficacy, and Medication Adherence served aslatent variables represented by scores on the specific (observed) measures described above. AsSEM is a confirmatory form of analysis, use of this technique allowed us to test the theorizedcausal model seen in Fig 1. Various goodness of fit indices are used to assess model fit to datain structural equation modeling. As recommended by O’Rourke and Hatcher [49], we reportincremental, absolute, and parsimonious goodness of fit indices.The Comparative Fit Index (CFI) provides a measure of incremental or relative fit. The CFIis thought to be the best index in covariance structure analyses, with little sampling variabilityfor a relative fit index [67]. Values greater than .94 indicate good fit between the proposedmodel and actual data [68]