OBJECTIVES: Case-identifying algorithms are essential for real-world evidence studies, and single-payer health care systems in Canada provide comprehensive, population-level data to support algorithm development and validation. Although case-identifying algorithms are abundant in the literature, there are few centralized resources to support their implementation. Therefore, a targeted literature review (TLR) was conducted to identify and summarize published, validated case-identifying algorithms developed using Canadian administrative health data.

METHODS: A targeted search was conducted in Ovid MEDLINE, PubMed, and publications lists of Canadian administrative health databases. English-language studies were included if they used Canadian data and reported ≥1 of the following performance measures: sensitivity, specificity, positive predictive value, or negative predictive value. No date restrictions were applied. Three independent reviewers screened all identified abstracts and extracted study and algorithm details from eligible studies. Studies were categorized by individual disease and disease areas based on the International Classification of Diseases (ICD)-10 classification system.

RESULTS: A total of 270 studies were identified, of which 89 met the eligibility criteria and were summarized. The majority of studies were based in the province of Ontario (61.8%), followed by Manitoba (12.4%). The average study period spanned 8.8 years, and 61.8% were published after 2015 (following ICD-10 implementation). Studies covered 15 (68.2%) of the 22 ICD-10 disease areas, most frequently mental and neurodevelopmental disorders (10.1%) and circulatory disorders (10.1%). Studies focused on chronic kidney disease, diabetes, hypertension, and rheumatic diseases each represented 3.4% of included studies.

CONCLUSIONS: This study summarized identified algorithms that were developed using comprehensive, population-level data, and used both ICD-9 and ICD-10 coding systems, which allow for applicability to other regions using similar diagnosis coding systems. The results of this TLR will be summarized in a publicly available resource, which, given the abundance of existing algorithms, will support the identification and implementation of validated algorithms for future real-world evidence studies.