Familial hypercholesterolaemia (FH) is an inherited lipid disorder characterised by raised LDL-C and increased risk of premature atherosclerotic cardiovascular disease. Despite effective treatments, FH remains substantially underdiagnosed. Electronic health records (EHRs) enable systematic case-finding, but evidence on their effectiveness remains limited. This review aimed to evaluate EHR-based strategies for FH identification.

Seven databases and grey literature were systematically searched for relevant studies. Eligible studies reported on systematic EHR-based case-finding in adults (≥18 years). Meta-analysis of FH prevalence was conducted using random-effects modelling. Risk of bias was assessed using ROBINS-I; evidence certainty with GRADE.

Of 831 citations screened, 12 eligible studies were included, including three from a prior review. Case-finding approaches included traditional diagnostic criteria (Simon–Broome, DLCN, MEDPED), hybrid models, and machine-learning algorithms (FAMCAT, FIND FH, TARB-Ex). FH prevalence estimates varied: 1.2% (95% CI 0.0%–3.0%; p=0.06) in general population studies, 41% (95% CI 2%–90%; p=0.02) in high-risk CVD populations, and 15% (95% CI 2%–34%; p=0.00) in genetically confirmed cohorts. Novel algorithmic approaches such as FAMCAT 2 and incorporating EHR-genomic data models demonstrated superior performance to traditional criteria. Secondary outcomes were inconsistently reported, though cholesterol levels at diagnosis were consistently higher in probable/confirmed FH, and markedly elevated in genetically confirmed cohorts. Certainty of evidence was moderate due to heterogeneity, non-randomised design, and potential publication bias.

Algorithmic/genomics augmented EHR-based methods can enhance FH identification, but evidence remains limited. Standardised, scalable approaches validated in diverse populations are required to inform equitable FH screening and policy development.