Innovations In Clinical Neuroscience

NOV-DEC 2017

A peer-reviewed, evidence-based journal for clinicians in the field of neuroscience

Issue link:

Contents of this Issue


Page 55 of 83

56 ICNS INNOVATIONS IN CLINICAL NEUROSCIENCE November-December 2017 • Volume 14 • Number 11–12 R E V I E W weighted UPSM coefficients had generalizable utility, yielding transformed PANSS factors with high specificity while retaining good levels of correlation with standard PANSS factors. APPLICATION OF UPSM TRANSFORM TO PANSS SEVERIT Y SCORES AT BASELINE In an analysis applying the UPSM PANSS factor transform to pooled baseline five study data, UPSM transformed PANSS factor scores at baseline were found to correlate well with their respective standard (Marder) PANSS factor scores at baseline, with Pearson's correlations of transformed versus standard PANSS factors in the moderate-to-high range for PANSS positive (0.53), disorganized (0.84), negative symptoms, apathy/avolition (0.74), negative symptoms, deficit of expression (0.76), hostility (0.95), anxiety (0.75), and depression (0.84). Between-factor correlations for both standard (Marder) PANSS factors and transformed PANSS factors were low at baseline. For example, the correlation between Marder PANSS positive and negative factors was -0.03 at baseline. Inverse correlations were noted for transformed PANSS positive and negative factors (-0.18 for negative- apathy/avolition; -0.26 for negative-deficit of expression). The concurrent validity of selected standard and transformed PANSS factor severity scores at baseline was assessed based on Pearson's correlations. The correlation at baseline between MADRS total score and standard (Marder) PANSS depression/anxiety factor and transformed PANSS depression factor was 0.56 and 0.53, respectively. The correlation at baseline between the NSA scale score and standard PANSS negative symptom score was 0.68, and the correlation with the NSA score was 0.53 for the transformed PANSS negative-apathy/avolition subscore and 0.55 for transformed PANSS negative-deficit of expression subscore. STRUCTURE OF (BASELINE) SYMPTOM SEVERIT Y VERSUS STRUCTURE OF SYMPTOM CHANGE As summarized above, baseline severity scores exhibit low levels of between-factor correlation using both standard (Marder) and transformed PANSS factors. However, when measuring change in symptom severity during treatment, there is a notable difference between standard PANSS factors, where change across factors is found to be highly correlated (factors exhibit pseudospecificity), and transformed PANSS factors, which continue to have the same low levels of between-factor correlation observed at baseline (i.e., factors may be characterized as exhibiting ergodicity). Table 1 summarizes, in an analysis of pooled five-study data, Pearson's correlations between PANSS severity scores at Week 6-endpoint, which provide additional confirmation of this difference. In the top panel of Table 1, correlations among standard (Marder) PANSS factors at Week 6 range from 0.34 to 0.68. In contrast, correlations among transformed PANSS factors is notably lower (-0.22–0.20, excluding correlations between negative subscores and depression and anxiety subscores). In an analysis of Week 6 data from Study 233 (1 of the 5-pooled studies), 22 transformed PANSS factor severity scores showed high levels of correlation with the respective standard (Marder) PANSS factors, with correlations ranging from 0.73 to 0.94 (Table 2). Transformed PANSS factor severity scores at Week 6 also showed moderate levels of correlation between the NSA total score and the negative-apathy/ avolition subscore (0.53) and the negative deficit of expression subscore (0.49). Similarly, the transformed PANSS factor depression severity score showed moderate levels of correlation with the MADRS total score (0.63, Table 2). PERSPECTIVE We have briefly summarized here a factor analytic procedure that generated weighted coefficients which, when applied to each of the 30 PANSS items utilized in schizophrenia clinical trials, yields transformed PANSS factors that meet four important criteria: 1) they have high face validity (i.e., they exhibit medium-to-large correlations with the standard [Marder] PANSS factors, and thus appear to be measuring the same clinical symptom TABLE 1. Pearson's correlations among PANSS factor severity scores at Week 6 Endpoint (pooled 5-study data) PANSS: Positive and Negative Syndrome Scale; MPOS: Marder positive symptoms; MDIS: Marder disorganized thinking; MNEG: Marder negative symptoms; MHOS: Marder hostility/excitement; MDEP: Marder depression/anxiety; UPSM: uncorrelated PANSS score matrix; UPSM-POS: positive symptoms; UPSM-DIS: disorganized thinking; UPSM-NAA: negative-apathy/ avolition; UPSM-NDE: negative-deficit of expression; UPSM-DEP: depression symptoms; UPSM-ANX: anxiety symptoms

Articles in this issue

Archives of this issue

view archives of Innovations In Clinical Neuroscience - NOV-DEC 2017