Attrition and augmentation biases in time series analysis: Evaluation of clinical programs