Diagnostic assessment of first-episode psychosis patients and adherence to German S3 guidelines: a retrospective cross-sectional study.

Abstract

BACKGROUND: Guidelines for diagnostic work-up in first-episode psychosis (FEP) vary worldwide. The German DGPPN S3 guidelines recommend a comprehensive work-up, including neuroimaging. However, real-world adherence to these recommendations remains unclear. This study examined guideline adherence in a Swiss tertiary psychiatric hospital.

Methods: We conducted a retrospective cross-sectional study analysing electronic healthcare records of first-episode psychosis patients hospitalised with an ICD-10 chapter F2 diagnosis for the first time between October 2022 and September 2023. We assessed adherence to recommended mandatory DGPPN S3 assessments - neurological examinations, blood analyses, drug screening, MRI - and evaluated completion of optional assessments such as EEG and lumbar puncture.

Results: A total of 68 first-episode psychosis patients were included from 364 patients screened: 44 (64.7%) were men; their median age was 29 (IQR: 23-33) years; 35 (51.5%) were involuntary admissions. Nearly all patients (n = 66 or 97.1%) received thorough neurological examinations and blood analyses, while 56 (82.4%) underwent drug screening (with 35 [44.6%] testing positive for cannabis). MRI was conducted in 38 (55.9%) cases. Non-completion of MRI was mainly due to patient refusal (n = 10 or 14.7%) or early discharge (n = 16 or 23.5%). Optional EEG and lumbar puncture were less frequently performed: in 26 (38.2%) and 4 (5.9%) patients, respectively.

Conclusions: Overall, guideline adherence was high, particularly for essential diagnostic procedures. However, only around half of the sample underwent MRI imaging, largely because of patient refusal or patient-requested discharge prior to completion of the suggested assessment. These findings highlight the need for optimised diagnostic workflows and enhanced patient education strategies to improve guideline adherence in FEP assessment.