Appraisal of the references supporting the STOPP/START.version 3 criteria

The strength of evidence underlying the STOPP/START version 3 (SS.v3) criteria has not been reported and remains controversial, potentially limiting their clinical implementation. This study aimed at appraising the references provided for the SS.v3 criteria.

All references cited for the 190 SS.v3 criteria were evaluated and discussed by a clinical pharmacist-geriatrician pair and a geriatrician expert in literature appraisal who reached agreement. The main endpoint was the evidence level classified by data study design (level I: randomised controlled trials; II: cohort studies; III: case-control; IV: case series; V: expert opinion). References were classified as inappropriate if they showed mismatch, lacked supporting content, or were clinically discordant with the criterion.

Of the 454 SS.v3 references, 322 are reviews (STOPP 70%, START 71%), either systematic (STOPP 29%, START 44%) or narrative (STOPP 41%, START 26%). Evidence is level I (STOPP 28%, START 64%), level II (STOPP 13%, START 8%), levels III-IV-V (STOPP 45%, START 19%), or downgraded because of inappropriateness (STOPP 15%, START 8%). At the level of the 190 criteria, the best evidence is level I (STOPP 43%, START 88%), level II (STOPP 13%, START 7%), levels III-IV-V (STOPP 35%, START 4%), while 10 criteria (STOPP 8%, START 0%) lack appropriate supporting reference.

Fewer than half of the STOPP criteria are supported by high-level evidence, unlike most START criteria. A reassessment of many STOPP criteria and their references is warranted. Clinicians should be aware of these limitations when involving in shared decision-making with older patients.