How Reliable Are Patient-reported Outcomes in Schizophrenia?

Share this content:
The primary outcome in this study was reliability of patient-ratings on subjective well-being.
The primary outcome in this study was reliability of patient-ratings on subjective well-being.

It is well documented that cognitive deficits in patients diagnosed with schizophrenia extend across several domains of executive function. In addition to a combination of positive (hallucinations, delusions) and negative (apathy, low motivation, “flat affect”) symptoms, many affected individuals also experience cognitive symptoms such as impaired learning and memory, attention, and cognitive flexibility.

Patient-reported outcomes play a significant role in clinical research in patients diagnosed with schizophrenia, but diminished cognitive function and poor self-recognition performance may indeed limit the capacity of the affected persons to perform on these self-ratings measures.

Researchers at the University of Toronto and the Keio University School of Medicine in Tokyo examined test-retest reliability of patient-reported outcomes measures in 50 outpatient study participants who were diagnosed with schizophrenia or schizoaffective disorder. Subjective well-being was assessed by using the Subjective Well-being under Neuroleptics scale-Short form (SWNS). The findings indicate that poor reliability of patient-ratings on subjective well-being is associated with cognitive impairment and disorganization, even when the ratings were performed back to back. This research was published in Psychiatry Research.

Although the participants showed adequate subjective well-being (ie, the mean SWNS total scores were above 80 at baseline and at the second measure), the identical total scores were observed in only 8% of individuals. A difference of >10 points on a total score was observed in 16% of patients. “A 10 point score change is regarded as clinically meaningful in subjective well-being,” according to the authors. It is important to note that differences of this magnitude were observed when the two tests were administered “at essentially the same time.” Only 22% of patients scored within a 1 point score difference between the two ratings on all 20 items of SWNS.

Age, educational attainment, duration of illness, antipsychotic medication dose, the PANSS (Positive and Negative Symptoms Scale) five factor and total scores, and BNA (Brief Neurocognitive Assessment) symbol-coding, letter-number span and global Z scores did not correlate with the observed difference in the SWNS total scores. In other words, the relationship between the overall score difference and these clinical characteristics was found not be significant.

However, with regard to the differences at the individual item level, “The number of [the SWNS] items with >1 point score difference was significantly correlated with the PANSS disorganized factor and total scores, and significantly and inversely correlated with the BNA letter-number span and global Z scores,” investigators wrote. In other words, the relationship between cognitive impairment, disorganization, and global severity of illness and the individual item scores was found to be significant.

Even though investigators used only a single self-rating scale in this study, and the sample size was relatively small, the authors suggest that “Patient-reported outcomes should be interpreted with caution in patients with schizophrenia manifesting prominent disorganization and/or cognitive impairment."


Takeuchi H, Fervaha G, Remington G. Reliability of a patient-reported outcome measure in schizophrenia: results from back-to-back self-ratings. Psychiatry Res. 2016;244:415-419.

You must be a registered member of Psychiatry Advisor to post a comment.

Sign Up for Free e-newsletters