International incidence of psychotic disorders, 2002-17: a systematic review and meta-analysis.

BACKGROUND

The last comprehensive systematic review of the incidence of psychotic disorders was published in 2004. New epidemiological data from different settings now permit a broader understanding of global variation. We examined the variation in psychosis by demographic characteristics and study method.

METHODS

For this systematic review and meta-analysis, we searched PubMed, Embase, Web of Science, PsycINFO, and bibliographies, and directly contacted first authors. We sought to obtain citations of original research published between Jan 1, 2002, and Dec 31, 2017, on incidence of non-organic adult-onset psychotic disorder. We included papers that were published or in grey literature and had no language restrictions. Data were extracted from published reports, where possible, by sex, age, and ethnic group. Quality of yield was assessed. Data were assessed using univariable random-effects meta-analysis and meta-regression. We registered our systematic review on PROSPERO, number CRD42018086800.

FINDINGS

From 56 721 records identified, 177 met inclusion criteria. The pooled incidence of all psychotic disorders was 26·6 per 100 000 person-years (95% CI 22·0-31·7). Heterogeneity was high (I≥98·5%). Men were at higher risk of all psychotic disorders (incidence rate ratio 1·44 [1·27-1·62]) and non-affective disorders (1·60 [1·44-1·77]) than women, but not affective psychotic disorders (0·87 [0·75-1·00]). Ethnic minorities were also at excess risk of all psychotic disorders (1·75 [1·53-2·00]), including non-affective disorders (1·71 [1·40-2·09]). Meta-regression revealed that population registers reported higher rates of non-affective disorders (9·64 [2·72-31·82]), schizophrenia (2·51 [1·24-5·21]), and bipolar disorder (4·53 [2·41-8·51]) than first contact study designs.

INTERPRETATION

We found marked variation in incidence of psychotic disorders by personal characteristics and place. Some geographical variation could be partially explained by differences in case ascertainment methods.

FUNDING

None.