OBJECTIVE: We report a review of outcomes in schizophrenia in the twenty-first century, replicating and extending work undertaken by the late Richard Warner in his seminal book, "Recovery from Schizophrenia: Psychiatry and Political Economy" (1985;2004).

METHOD: Warner's methods were followed as closely as possible. Only observational/naturalistic studies were included. Six scientific databases were searched from 2000 to 2020. 6,640 records were retrieved. 47 met inclusion criteria.

RESULTS: Overall, complete recovery is higher in this study than in Warner's (37.75% cf 20.4%), especially for first episode psychosis (FEP) (57.1% cf 20.7%). Clinical recovery, annualized remission rate (ARR), and employment outcomes were significantly superior for first episode psychosis compared with multiple episode psychosis (MEP). ARR shows a trend toward reduction over time, from 2.2 before the financial crash of 2008 to 1.6 after (t = 1.85 df 40 p = .07). The decline is statistically significant for the MEP group (t = 2.32 df18 p = .03). There were no differences in outcome by region, sample characteristics, outcome measures used, or quality of studies. Heterogeneity of clinical outcome measures across the literature makes evidence synthesis difficult. Weak and inconsistent reporting of functional and employment outcomes mean that findings lack meaning with respect to lived experience.

CONCLUSION: Future research strategies should aim to reduce heterogeneity in clinical outcome measures and to increase the emphasis on capture and reporting of more sophisticated measures of social and functional outcome. Outcome domains should be disaggregated rather than conflated into unitary recovery constructs.