Is there a relationship between psychiatrist's clinical assessment of medication non-adherence and treatment outcomes? Implications for practice.

Abstract

Background: There is considerable research on the ramifications of medication non-adherence for adults with psychotic illnesses. Much of which has tightly controlled designs and strict inclusion/exclusion procedures (i.e., it is less "ecologically valid," or consistent with real-world challenges in care). The authors sought to determine predictive relationships between psychiatrists' clinical assessments of non-adherence and treatment outcomes, via a design that would be more applicable to practice.

Method: Multiple regression analyses were conducted on non-adherence, symptom severity upon admission, number of recent hospitalizations, and length of hospital stay. The sample consisted of 182 inpatients with psychotic spectrum disorders and significant risk and vulnerability factors. Non-adherence was measured via the psychiatrists' diagnosis of V15.81. Symptom severity was measured via the 24-item Brief Psychiatric Rating Scale (BPRS-E).

Results: There were null findings on non-adherence and BPRS-E pretest score ( $\chi$ ²= 2, p = 0.16), recent hospitalizations ( $\chi$ ²= 1.2, p = 0.27), and length of stay (β = 0.003, p = 0.97). Higher symptom severity predicted a modestly longer length of stay ( $\chi$ ²= .20, p = 0.007), though Bonferroni correction nullified this finding. White/Caucasian participants were far more likely to be non-adherent than black/African-American participants (t = -8.66; p > .00001).

Conclusions: Null findings suggest the psychiatrist's initial, quick-form assessment of non-adherence may not necessarily presume a poor prognosis. Perhaps, because individuals with severe and chronic psychotic disorders may have greater coping, adaptive, and survival skills than is often assumed. In severely under-resourced hospitals, such second thoughts and more reliable information about adherence and contributing factors may improve treatment outcomes.