Putting the patient back in clinical significance

Abstract

The present study introduced a modernized approach to Jacobson and Truax's (1991) methods of estimating treatment effects on individual-level (a) movement from the clinical to the normative range and (b) reliable change on posttraumatic stress disorder (PTSD) severity. Participants were 450 trauma-exposed women (M age = 39.2 years, SD = 8.9, range: 18-65 years) who presented to seven geographically diverse community mental health and substance use treatment centers. Data from 53 of these women, none of whom met the criteria for full or subthreshold PTSD, were used to establish the normative range. Using moderated nonlinear factor analysis (MNLFA) scale scoring, which weights symptoms by their clinical relevance, a significantly larger proportion of participants moved into the normative range for PTSD severity scores and/or exhibited reliable changes after treatment compared to the same individuals' movement when using symptom counts. Further, approximately 24% of the participants showed discrepant judgments on reliable change indices (RCI) between MNLFA scores and symptom counts, likely due to the false assumption that the standard error of measurement is equal for all levels of underlying PTSD severity when estimating RCIs with symptom counts. An MNLFA approach to estimating underlying PTSD severity can provide clinically meaningful information about individual-level change without the de facto assumption that PTSD symptoms have equivalent weight. Study implications are discussed with regard to a joint emphasis on (a) measurement models that highlight differential symptom weighting and (b) treatment-arm differences in individual-level outcomes rather than the current overemphasis of treatment-arm differences on group-averaged trajectories.