Abstract
Abstract
Psychological treatment for health anxiety is delivered across medical and psychological services, but apparent benefit depends on the comparator, outcome and assessment time. This systematic review examined these distinctions using source-linked reanalysis. PubMed and Europe PMC searches through 22 September 2026 identified 8,799 unique records, supplemented by registry and bounded citation searches. Retrieval and adjudication were closed on 1 October 2026. Forty-three primary efficacy reports were organized into 42 conservative synthesis units; uncertain overlap precluded treating these as 42 independent trials. Forty-five numerical results were reproduced in Stata, with repeated assessments and alternative estimands kept distinct. Selected observed-endpoint contrasts for internet cognitive behavioural therapy ranged from a substantial benefit against a discussion forum (Hedges g −1.60, 95% CI −2.10 to −1.10) to an imprecise difference against behavioural stress management (−0.26, −0.58 to 0.06). In medical services, the CHAMP trial reported a 12-month model-based Health Anxiety Inventory difference of −2.97 points (−4.37 to −1.57) against usual care. These examples do not constitute a pooled treatment ranking. Eligible studies without usable outcomes remained in the narrative synthesis. Limited database coverage, unavailable reports, single-author AI-assisted screening and result-specific bias constrain completeness and certainty. Psychological treatments should be interpreted against their actual alternatives and clinical settings; symptom improvement alone does not establish improved functioning or reduced healthcare use.