The mean age of the 73 participants was 50.3 years (sd=6.3, range 40 to 72), with a mean age of onset of schizophrenia at 25.6 years (sd=11.7). A total of 47% were female. The racial/ethic distribution included 45% white, non-Hispanic, 26% African-American, non-Hispanic, 15% Hispanic, and 8% Asian or Native American. The mean level of educational attainment was 12.3 years (sd=2.1). A total of 62% of participants had never married, and only 3% were currently working for pay. In regard to living situation, 70% of participants were in some sort of supervised living (e.g., board and care home) with the remainder in private residences. In terms of psychopathology, the mean PANSS Positive Syndrome Score was 14.6 (6.2), PANSS Negative Syndrome Score was 15.5 (5.2), and the Calgary Depression Rating Scale was 4.4 (4.5). Symptom severity on the PANSS subscales was in the mild-moderate range. In terms of the severity of depression in the sample, 23.3% of patients had scores on the CDSS of 7 or greater, which is indicative of the presence of a major depressive episode.
Behavioral Activation and Avoidance
The mean score on the Behavioral Activation subscale was 23.2 (sd=7.8) and the Avoidance subscale was 21.0 (sd=10.4). Avoidance and Activation subscales were significantly correlated with each other (r=0.345,p=0.003), such that greater avoidance was associated with reduced activation. With regard to demographic and basic illness characteristics, there were no significant relationships between Activation or Avoidance scales and age, education, gender, ethnicity, or marital status. Similarly, there were no significant relationships with either age of onset of illness or medication dose (Chlorpromazine equivalents).
With respect to psychopathologic symptoms, Activation and Avoidance revealed somewhat different patterns of association. High severity on PANSS Positive and PANSS Negative subscales were each significantly and positively correlated with the Avoidance scale, but not Activation. In contrast, greater severity of depressive symptoms (CDSS Scores) was correlated with the Activation subscale, but not Avoidance. Using r-to-Z transformations, the differences in correlations for PANSS Negative Score (z=2.49, p=0.012) and CDSS Score (z=4.33,p<0.001) were significant, but not PANSS Positive (z=1.09,p=0.273).
As with symptoms, Activation and Avoidance were differentially related to aspects of functioning. Higher scores on global cognitive ability (RBANS score), functional capacity (UPSA score) and observer-rated functioning (SLOF Score) associated with lower Avoidance, but no significant relationships were evident with Activation. In contrast, better subjective functioning, as measured by the RAS was associated with greater Activation but not Avoidance. The difference between correlations, via r-to-Z transformation, was significant for RAS Scores (z=5.17,p<0.001), but not for RBANS Scores (z=1.13, p=0.259), SLOF-Functioning (z=1.51,p=0.129) or UPSA Scores (z=1.90,p=0.057).
We examined whether Avoidance and Activation were significant predictors of the three functional indicators (UPSA, SLOF-Functioning, and RAS Scores) in three hierarchical regressions; in each, the first step included the PANSS Negative Scale Score, PANSS Positive Scale Scores, and CDSS Scores, and the second step included the Activation and Avoidance subscales. With UPSA as the dependent variable, additional variance accounted for by Activation and Avoidance was not significant [r2 Change = 0.018, F-change(2,67)=0.889, p=0.412]. With SLOF as the dependent variable, the additional variance was non-significant [r2 Change = 0.065, F-change(2,67)=2.52, p=0.089]. However, with RAS Scores, the second block was significant [r2 Change = 0.079, F-change(2,67)=6.92, p=0.002], and within this block, Behavioral Activation was a significant predictor (Beta=0.197, t=2.2, p=0.033) of subjective functioning (RAS Score) beyond that associated with symptoms but Avoidance was not (Beta=−0.2, t=−0.3, p=0.793).
Finally, we examined whether Activation and Avoidance predicted significant variation in observer-rated functioning, using two separate hierarchical regressions adjusting for functional capacity (UPSA Scores) and cognitive ability (RBANS Total Score). In the model with UPSA entered in the first step, the second step with both Activation and Avoidance was not significant ([r2 Change = 0.043, F-change(2,67)=1.5, p=0.228]). In a second regression analysis, we adjusted for global cognitive ability in the first step (RBANS score) and found that that the addition of Behavioral Activation and Avoidance in the second step was associated with a significant change in variance accounted for ([r2 Change = 0.100, F-change(2,67)=3.3, p=0.045]). Inspection of the subscales revealed that Avoidance was a significant predictor of SLOF scores after adjusting for cognitive ability (Beta=−0.351, t=2.5, p=0.015) and Behavioral Activation was not significant (b=0.137, s.e.=0.239, t=0.572, p=0.569).