Defining Treatment as Usual for Attenuated Psychosis Syndrome: A Survey of Community Practitioners

We targeted for participation a national sample of clinical psychologists (N=500), psychiatrists (N=500), and general practitioners (GPs) (N=500). The sample of psychologists was drawn from the membership directory of the American Psychological Association Center for Workforce Studies. Samples of psychiatrists and GPs were obtained from the Physician Masterfile of the American Medical Association (AMA) through their database licensee, Direct Medical Data. This inclusive list of all U.S. physicians contains the names of individuals who are not necessarily members of the AMA or the APA. Use of this list rather than a list of AMA and APA members served to make our sample more representative of U.S. psychiatrists and other physicians.

The study was approved by the University of Hawaii at Manoa Institutional Review Board, the American Psychological Association Committee on Workforce Studies, and Direct Medical Data. Data were collected in 2008.

We created vignettes about four fictional individuals (Diego, Paul, Mike, and Claire). Each vignette described an individual with fully psychotic, attenuated-psychotic, or no psychotic symptoms, for a total of 12 vignettes. Five graduate students trained to administer the most widely used psychosis risk diagnostic instrument reviewed the vignettes. They estimated the Global Assessment of Functioning (GAF) score of the individual depicted in each vignette and compared the vignettes to a set of criteria that are used to develop vignettes. Graduate student raters agreed on which conditions were represented by all vignettes, and GAF scores fell in the intended range. The vignettes then were reviewed by 11 experts in the field of psychotic disorders. At least 80% of the expert raters described each vignette as representing the intended conditions well or very well. [Excerpts of the vignettes are available online as a data supplement to this report.]

A checklist of 50 treatment components followed each vignette. The checklist comprised pharmacologic and psychosocial interventions reflecting a broad mix of strategies. The treatment checklist was compiled from portions of other lists, including a measure assessing beliefs about interventions (10), a checklist created by the Hawaii Department of Health (11), the practice guidelines of the International Early Psychosis Association (8), and consultation with mental health professionals.

Data collection took place via a mixed-mode survey using mail and a Web site. Surveys were mailed to potential participants with an option to participate online via a link printed on the survey booklet. Efforts were made to present the information as similarly as possible across formats.

Participants were mailed a prenotice letter and one week later received the survey along with a cover letter and postage-paid return envelope. Each survey packet contained vignettes depicting three individuals and questions following each vignette. Twelve versions of the survey were created. Each survey included one vignette describing psychosis, one describing attenuated symptoms of psychosis, and one describing no psychosis. The order of the level of symptoms depicted in vignettes was counterbalanced to control for order effects. Each vignette was represented equally across the 12 survey versions. Surveys were put in repeat order and assigned sequentially to individuals on the mailing list.

Participants received instructions to read each vignette, review the checklist that followed, and mark interventions that would be helpful for the character. They were also asked to review the checklist again and mark the interventions that were contraindicated for the individual depicted in the vignette.

The sample contained 293 usable responses, with a response rate of 20%. Of these respondents, 128 (44%) were psychologists, 94 (32%) were psychiatrists, and 68 (23%) were GPs. Three respondents did not identify their field of practice. Missing data across variables ranged from 1% to 8% and were excluded listwise per analysis.

The mean±SD age of respondents was 52.6±10.9 years; 153 (52%) were male, and 247 (84%) were white. All participants reported that they were licensed to practice in their field. They reported an average of 21.1±10.8 years as active practitioners. With respect to the counterbalancing of vignettes, no significant differences in patterns of responding were detected for character or order of presentation within the survey.

Each of the 50 treatment components included in the checklist elicited a variable number of “helpful” or “harmful” endorsements. Because of the study’s aims of investigating treatment as usual, the 15 treatment components that elicited over 30% (N>90) of endorsements in either direction (helpful or harmful) were selected for further analysis. [Figures depicting response choices are included in the data supplement to this report.]

Chi square statistics were calculated to compare the frequency with which these components were selected as helpful or harmful or not selected for the attenuated-symptoms vignettes versus the no-symptoms and fully psychotic vignettes. Significant overall differences in endorsement rates among the three symptom conditions were found for ten treatment components and the no-treatment option (Table 1).

Post hoc comparisons examined significant differences between treatment component endorsements for each symptom condition. All treatments selected as helpful by over 30% of respondents were chosen significantly more frequently for the attenuated-symptoms condition than for the no-symptoms condition. Significantly more practitioners rated cognitive coping as helpful for the attenuated-symptoms vignettes than for the psychosis vignettes. Antipsychotic medication, family support groups, and case management were more frequently considered to be helpful for the psychosis vignettes than for the attenuated-symptoms vignettes.

Among the six treatment components identified as harmful by over 30% of respondents, significant overall differences in endorsement rates between the symptom conditions were found for electroconvulsive therapy (ECT), hypnosis, and no treatment. Practitioners rated hypnosis and no treatment as harmful significantly more often for the attenuated-symptoms vignettes than for the no-symptoms vignettes. ECT was considered harmful significantly more often by practitioners responding to the attenuated-symptoms vignettes than to the psychosis vignettes.

Among the medication treatment options for the attenuated-symptoms case, antipsychotics were recommended more often than other medication classes. Antidepressants were chosen as helpful for attenuated-symptoms cases by 67 (23%) respondents and as harmful by 35 (12%) respondents. Antianxiety medications were rated as helpful by 33 participants (11%) and as harmful by 34 participants (12%). Sedatives were recommended by 12 respondents (4%) and advised against by 73 respondents (25%).

No pattern of differences emerged upon examination of attenuated-symptoms treatment recommendations across the three provider types. GPs endorsed fewer treatment components than psychiatrists and psychologists. Psychiatrists more frequently rated cognitive coping and psychoeducation as helpful compared with psychologists and GPs, and psychologists rated psychoeducation as helpful more frequently compared with GPs. Psychologists more frequently rated case management as helpful compared with the other two groups and more frequently rated ECT as harmful compared with psychiatrists. The groups did not differ in endorsements of antipsychotic medication.

The treatment components that were rated helpful most frequently provide insight into treatment as usual for attenuated psychosis syndrome. Antipsychotic medication was the component chosen most commonly by practitioners responding to attenuated-symptoms vignettes. This finding is consistent with findings that 25% of participants meeting research criteria for high-risk status were receiving antipsychotic medications at study intake (12) and that providers tend to consider vignettes of attenuated psychosis syndrome to represent a psychotic disorder (6).

In light of these results, it is relevant to note that few trials have provided evidence for the efficacy of antipsychotic medications for the treatment of attenuated symptoms (7). Experts in the field have discouraged the prescription of antipsychotics for these patients pending further safety and efficacy testing (12). Given the particular vulnerability of young people to motor- and obesity-related side effects (9) and the lack of agreement regarding reliable diagnostic criteria for attenuated or high-risk states, prescribers might be well-advised to take a conservative stance when considering antipsychotic medications.

Cognitive coping was rated second highest among the treatment components considered helpful for the attenuated psychosis syndrome vignettes. To date, two randomized controlled trials have supported the effectiveness of enhancing coping skills through cognitive therapy for preventing or delaying the onset of psychosis among high-risk individuals (7,13). In addition, experts recommend cognitive therapy for attenuated psychosis syndrome, in part because of its low side-effect profile (12). Cognitive therapy may be well-suited to the treatment of attenuated symptoms, given its emphasis on challenging distorted perceptions, ignoring “unhelpful” thoughts, and actively reframing stressors and cognitive distortions early in the progression of illness, when insight is relatively intact.

Family involvement and psychoeducation were the third and fourth most frequently recommended treatments for the attenuated-symptoms vignettes. These components echo the recommendations of the International Early Psychosis Association. Given the typical timing of psychosis onset, families are likely to play a large role in helping to support treatment and ensure their relative’s safety. Psychoeducation, often neglected in traditional clinical settings, may be a particularly effective tool for increasing client and family engagement (14).

Comparing responses for the attenuated-symptoms and no-symptoms cases suggests that practitioners approached treatment for these cases with very different conceptualizations. Practitioners named restraint, catharsis, and hypnosis as harmful significantly more frequently for the attenuated-symptoms case than for the no-symptoms case. These findings may reflect a perception that individuals with attenuated symptoms could be more vulnerable to iatrogenic effects. Finally, the choice of no treatment was considered harmful significantly more often for the attenuated-symptoms condition than for the no-symptoms condition, implying that practitioners viewed the character with attenuated symptoms as having a greater need of services.

Comparisons of responses for attenuated-symptoms versus psychosis vignettes suggested that practitioners approach treatment for these conditions fairly similarly, although several differences emerged. Intensive treatment options (antipsychotic medications, support groups, and case management) were recommended more often for characters presenting with psychosis than with attenuated symptoms. Cognitive coping was selected more often for attenuated symptoms than for psychosis, possibly reflecting practitioners’ opinion that the greater insight associated with attenuated symptoms would enable the effectiveness of cognitive techniques. With regard to treatments judged to be harmful, few significant differences emerged between the attenuated-symptoms and psychosis vignettes. ECT was selected as harmful more often for attenuated symptoms than for psychosis, perhaps reflecting a perception of this treatment as appropriate for only the most severe and treatment-refractory cases, consistent with best-practice guidelines for schizophrenia (14).

The findings of this study were limited by the overall response rate for practitioners of 20%, with differing rates among the types of practitioners surveyed. It is likely that respondents differed from nonrespondents, which may limit the generalizability of findings. The response was likely influenced by variables such as practitioner experience, type, setting, and familiarity with psychosis. A majority of the sample was white and older, raising the possibility that responding was related to those characteristics as well. The age of the participants was fairly representative of the members of the APA (www.apa.org/workforce/publications/02-member/table-1.pdf) and the AMA (www.ama-assn.org/assets/meeting/2011a/a11-clrpd-reports.pdf); however, white practitioners were overrepresented.

The practitioner response rates were very consistent with at least one previous mailed survey design, in which response rates for psychologists, psychiatrists, and GPs were 40%, 35%, and 20%, respectively (15).

Despite the likelihood that the sample was not precisely representative, we believe that the sample was sufficient to achieve the aims of the study. Lower response rates may be due at least in part to nonresponse from clinicians unfamiliar with or uninterested in early psychosis. Given that these practitioners may be less likely to treat such patients, our sample may actually yield a more informed, if less representative, opinion.

The use of vignettes to survey clinical practices also has limitations in that vignettes lack the complexity of real-life cases and may elicit optimized responses. The treatment checklist had limitations as well because it contained such a large number of response options that several treatment components were rarely selected.

Updated, concise best-practice recommendations reflecting the most recent research about treatment of attenuated psychosis syndrome would be helpful to community practitioners who encounter this client population. Diagnostic reliability may represent an obstacle to the dissemination of best practices. Practitioners might benefit from training on the identification and treatment of attenuated symptoms. The potential inclusion of an attenuated psychosis syndrome category as an “area for future research” in the forthcoming DSM-5 may encourage investigators to continue to refine the relation of this syndrome to full-threshold psychotic disorders, thus enabling the development of empirically supported interventions for this high-risk group.

Funding for this study was provided in part by grant R03MH076846 from the National Institute of Mental Health; the Mental Hygiene Administration, Maryland Department of Health and Mental Hygiene, and Baltimore Mental Health Systems; a Research Seed Funding Initiative grant from the University of Maryland, Baltimore County; and the Division of Child and Adolescent Psychiatry, University of Maryland.

The authors report no competing interests.