Table 2.

Comparison of MCID estimations for improvement on the PANSS by linking PANSS and global impressions scores among published studies on the topic.

Study	Population	Linking Method	Evaluation of Change	Baseline PANSS	MCIDa
				Mean (SD)	Absolute Value	Percent of Baselineb
Current Study	Single study, Multiple sites, Enrolled Schizophrenics, Excluded other psychotic disorders and first episode, inpatient/outpatient, n=1442	EPL	change in CGI-S	75.6 (17.6)	15.3	34
Schennach-Wolff et al12	Multicenter follow-up program, Enrolled all psychotic disorders, inpatient only, n=398	EPL	CGI-I	--	5	17
Levine et al11	Pooled sample of 4 drug trials, first episode, acute, chronic, Schizophrenia and Schizoaffective disorder, n=2698	EPL	CGI-I	--	--	22
Leucht et al14,32	Pooled sample of 7 drug trials, Enrolled all psychotic disorders, included only hospitalized “florid” psychosis, n=4091	EPL	CGI-I, Change in CGI-S	94 (19)	15	24
Cramer et al15	Single sample at multiple VA sites, Schizophrenia resistant to at least 2 medications, n=423	ANOVA	5-level change scalec	91 (15)	--	21

^aMCID of the PANSS as calculated by anchor based methods which link change in PANSS to a measure of global clinical change. For studies using the CGI-I, MCID estimated as the change in PANSS associated with CGI-I=3 (Minimally Improved). Reported estimates are averaged over the time periods analyzed.

^bThe MCID as a percent change from baseline. Cramer et al¹⁵ is the only paper specifically accounting for the minimum PANSS total score of 30 points when calculating percent change.

^cCramer et al¹⁵ used a 5-level clinician rated global clinical change scale.

Abbreviations: PANSS, Positive and Negative Syndrome Scale; MCID, Minimum Clinically Important Difference; SD, Standard Deviation; CGI-S, Clinical Global Impressions Scale – Severity; CGI-I, Clinical Global Impressions Scale – Improvement; EPL, Equipercentile Linking