Caregiver screen improves identification of youth at high risk for psychosis



Using a formal, validated caregiver screening tool in addition to an adolescent’s screening results appears to provide more accurate identification of adolescents who are at clinical high risk for psychosis, a new study in Schizophrenia Research suggested.

Because the Structured Interview for Psychosis Risk Syndromes is resource-intensive to screen youth for attenuated psychosis syndrome, several shorter screening tools often are used with adolescents, including the 12-item Prime Screen-Revised (PS-R) that tests for attenuated or psychotic-like symptoms. Yet clinicians also might seek parent or guardian input, in light of the median age of psychosis of 22 years.

Noting that "parent input appears to be a valued but unstandardized component of (clinical high risk for psychosis) assessment," Emily Kline and her associates at the University of Maryland, Baltimore, sought to test a modified version of the PS-R for caregivers (CGPS-R).

Although they found that the PS-R and the CGPS-R were not correlated with each other, "both measures demonstrated moderate to large correlations with clinician ratings on the [Scale of Prodromal Symptoms]" and improved the overall accuracy of positive cases based on the Structured Interview for Psychosis Risk Syndromes, reported Ms. Kline, a doctoral candidate at the university, and her associates (Schizophr. Res. 2013;147:147-52).

They recruited 52 youths (54% female) who were a mean age of 15 years (SD, 2.09), all within the ages of 12-22 and currently receiving mental health services, and their caregivers to test the caregiver-modified screening tool. About 50% of the youth participants were African American, 37% were white, and 13% were categorized as more than one race or "other."

Among the caregivers, 63% of those filling out the screens were mothers, 13% were fathers, 10% were mothers and fathers, 10% were grandparents, and 4% were another caregiver. The modification of the PS-R for caregivers involved only replacing "I," "me," and "my" with "he/she," "him/her," and "his/her."

After all participants and their caregivers had completed the PS-R and CGPS-R screens, respectively, the youth and caregivers were evaluated by staff with the Kiddie Schedule for Affective Disorders and Schizophrenia, present and lifetime version (K-SADS-PL). Then the youths only were evaluated using the Structured Interview for Psychosis Risk Syndromes. Finally, Ms. Kline’s team compared the results of the Structured Interview for Psychosis Risk Syndromes with the PS-R only, the PS-R with the CGPS-R, and the CGPS-R only.

The researchers specifically used the sum of positive symptoms ratings within the Scale of Prodromal Symptoms part of the Structured Interview for Psychosis Risk Syndromes since the five positive symptoms "are most central to the risk syndrome diagnoses," they wrote.

Because the PS-R authors recommend a threshold of at least two responses of five or six on a Likert-type agreement scale of zero to six, the investigators used a threshold of at least four responses of five or six when the PS-R and CGPS-R results were "summed" in a combined screen to identify "positive" cases. Based on Structured Interview for Psychosis Risk Syndromes classification, 27 of the 52 participants (52%) were identified as positive cases for psychosis risk syndrome or current psychosis (six participants).

Comparing the youth screening (PS-R) results with those of the caregivers (CGPS-R), agreement was poor (r = 0.09), though the caregiver screen did correlate with the clinician Scale of Prodromal Symptoms ratings (r = 0.41, P < .01). When the researchers compared the accuracy of the PS-R to that of the PS-R and CGPS-R together in identifying youth as positive cases according to Structured Interview for Psychosis Risk Syndromes, the accuracy and specificity improved at cost to sensitivity.

The PS-R had an overall accuracy of 71%, with a specificity of 0.60 and a positive predictive value of 0.69. These numbers improved when the CGPS-R was added to the PS-R, for 75% accuracy, 0.76 specificity, and 0.77 positive predictive value, albeit dropping sensitivity from 0.81 with the PS-R to 0.74 with both tools and "resulting in a potentially greater number of [clinically high risk] youth ‘missed’ by the screener," they wrote.

Screening items with the most agreement among adolescents, caregivers, and clinicians were those related to auditory hallucinations, while items related to being controlled by an outside force showed poor agreement between caregivers and youth "but were highly predictive of Scale of Prodromal Symptoms ratings and/or [Structured Interview for Psychosis Risk Syndromes] diagnosis."

The authors discussed several possible reasons for the disparities between the youth and caregivers, including differing interpretations of the items, concern about stigma for revealing some information, caregivers’ misinterpretation of youths’ behavior, youths’ mental illness–induced distorted perceptions, caregivers’ own psychopathology, and the inability of caregivers to observe certain items in youths.


Next Article: