Parent-Child Agreement Concerning Psychopathology: The Differential Impact of Parent Stress, Parent-Child Discrepancies, and Response Bias on Inpatient Treatment Outcome

It is well-established that informant agreement on self-report measures of psychopathology is quite low (Achenbach, McConaughy, & Howell, 1987). However, it is only recently that researchers have endeavored to derive meaning out of discrepant data. The current study is part of this burgeoning area of the literature by examining the differential impact of parent stress, parent-child discrepancies, and response bias on inpatient treatment outcome. A conceptual model of treatment outcome and parent-child response bias is also proposed.

This was an archival study of a database from an inpatient psychiatric setting. The sample consisted of 77 preadolescents (ages 9 to 12 years old) who had completed the Personality Inventory for Youth (PIY; Lachar & Gruber, 1995), whose parents had completed the Personality Inventory for Children (PIC-2; Lachar & Gruber, 2001) and Parenting Stress Index (PSI; Abidin, 1995), and whose therapist had completed the Children's Psychiatric Symptom Rating Scale (CPSRS; Pogge, 2009) at admission and discharge. Positive treatment outcome was defined as positive change on the CPSRS from admission to discharge. Correlational and multiple regression analyses investigated whether parent stress components of the PSI, parent-child discrepancies between the PIY/PIC-2 clinical scales, and response bias features of the PIY and PIC-2, significantly predicted inpatient treatment outcome.

Descriptive analyses demonstrated that the preadolescents in this study improved overall during hospitalization across symptom domains. However, there was little support for the hypothesis that the predictors significantly impacted treatment outcome; very few of the PSI scales, PIY/-PIC-2 clinical discrepancies, and response bias features, significantly predicted change on the CPSRS. Multiple regression analyses indicated that there was a positive correlation between the predictors and global change on the CPSRS (r = .38, p = .03), as they together accounted for 14.1% of the variance of the full model (r² = .14). However, no predictor was significant on its own so the differential contribution of the predictors is unknown. The predictors did not significantly predict Length of Stay (LOS) either together or individually.

The largely insignificant results may have been due to limitations related to small sample size, uncontrolled confounds, and single-item assessment of treatment outcome. An important implication of this study is that a clinician should use clinical judgment with a specific case that takes into account elevations of parent stress, parent-child discrepancies, and response bias in order to design a targeted plan. The overall goal for the future is to develop a framework for diagnosing and treating clinical problems in order to provide the most effective treatment for every person.