Talukder Animesh, Kougianou Ioanna, Healy Colm, Lång Ulla, Kieseppä Valentina, Jalbrzikowski Maria, O'Hare Kirstie, Kelleher Ian
Centre for Clinical Brain Sciences, Division of Psychiatry, University of Edinburgh, Edinburgh, UK.
School of Medicine, University College Dublin, Dublin, Ireland.
Psychol Med. 2025 Feb 12;55:e46. doi: 10.1017/S0033291724003520.
Psychosis prediction has been a key focus of psychiatry research for over 20 years. The two dominant approaches to identifying psychosis risk have been the clinical high-risk (CHR) and the familial high-risk (FHR) approaches. To date, the real-world sensitivity of these approaches - that is, the proportion of all future psychotic disorders in the population that they identify - has not been systematically reviewed.
We systematically reviewed and meta-analysed studies in MEDLINE, Embase, PsychINFO, and Web of Science (from inception until September 2024) that reported data on the sensitivity of CHR and FHR approaches - i.e., individuals with a psychosis diagnosis preceded by a CHR diagnosis or a history of parental psychosis (PROSPERO: CRD42024542268).
We identified four CHR studies and four FHR studies reporting relevant data. The pooled estimate of the sensitivity of the CHR approach was 6.7% (95% CI: 1.5-15.0%) and of the FHR approach was 6.5% (95% CI: 4.4-8.9%). There was a high level of heterogeneity between studies. Most FHR studies had a low risk of bias, but most CHR studies had a high risk of bias.
Pooled data suggest that CHR and FHR approaches, each, capture only about 6-7% of future psychotic disorders. These findings demonstrate the need for additional approaches to identify risk for psychosis.
20多年来，精神病预测一直是精神病学研究的重点。识别精神病风险的两种主要方法是临床高危（CHR）方法和家族高危（FHR）方法。迄今为止，这些方法在现实世界中的敏感性——即它们所识别出的人群中所有未来精神病性障碍的比例——尚未得到系统评价。
我们系统检索并荟萃分析了MEDLINE、Embase、PsychINFO和Web of Science（从数据库建立至2024年9月）中报告CHR和FHR方法敏感性数据的研究（国际前瞻性系统评价注册库：CRD42024542268）。
我们确定了四项报告相关数据的CHR研究和四项FHR研究。CHR方法敏感性的合并估计值为6.7%（95%置信区间：1.5-15.0%），FHR方法敏感性的合并估计值为6.5%（95%置信区间：4.4-8.9%）。研究之间存在高度异质性。大多数FHR研究的偏倚风险较低，但大多数CHR研究的偏倚风险较高。
汇总数据表明，CHR和FHR方法各自仅能识别约6-7%的未来精神病性障碍。这些发现表明需要其他方法来识别精神病风险。
