Identifying appropriate fit models for special populations, such as the aged or the diseased, is critical for evaluating wearable health monitoring systems. Having suitable fit is essential in vital signal monitoring towards reducing motion artifacts and to facilitate reliable signal quality. In this study, 3D-scanned data of an obese population was used to case-select a fit model representative of sleep apnea patients. Key bivariate measurements were used as the basis in selecting a suitable model, and further optimized using additional multivariate measurements and body shape. The chosen model was reverse engineered and the torso region was 3D printed for fit testing garment prototypes and for determining optimal placement of sensors and other embedded hardware components. The approach can be generalized for other garment types.