Summary
This paper presents PheKB (Phenotype KnowledgeBase), a centralised online platform enabling researchers to develop, share, and validate electronic phenotype algorithms for mining electronic health record data. The work demonstrates that phenotype algorithms—particularly those using ICD codes, medications, and natural language processing—achieve high positive predictive values (median 96–97.5%) and are transportable across different healthcare systems, supporting reproducible, research-grade phenotyping at scale.
UK applicability
Whilst this tool was developed and evaluated within United States healthcare systems, the underlying methodology for algorithm transportability and validation could be relevant to UK biobank and NHS data research initiatives, though implementation would require adaptation to UK data standards and NHS information governance frameworks.
Key measures
Positive predictive values (PPVs) for cases and controls at authoring vs. implementation sites; algorithm component frequencies (ICD codes, medications, natural language processing); number of algorithm views and reuses
Outcomes reported
The study reported the status and impact of PheKB, an online repository containing 30 finalised and 62 in-development phenotype algorithms, and analysed the performance of these algorithms when applied across different healthcare institutions.
Topic tags
Dig deeper with Pulse AI.
Pulse AI has read the whole catalogue. Ask about this record, its theme, or how the findings apply to UK farming and policy — every answer cites the underlying studies.