I’m happy to share that our paper describing medicalcoder has been published in JAMIA Open:
DeWitt PE, Russell S, Feinstein JA, Rebull MN, Bennett TD. “medicalcoder: a unified and longitudinally aware framework for International Classification of Diseases code-based comorbidity assessment in R.” JAMIA Open. 2026; 9(5):ooag182. doi:10.1093/jamiaopen/ooag182.
Comorbidity algorithms based on International Classification of Diseases (ICD) codes are widely used to describe patient populations and support risk adjustment. Applying them to real clinical data can be awkward: datasets may contain both ICD-9 and ICD-10 codes, and some algorithms depend on details such as whether a diagnosis was present on admission or marked as the primary diagnosis. Many existing tools also work one encounter at a time, while health records often need to be considered across time.
The paper describes how medicalcoder brings these tasks into one R interface. It includes an internal ICD code database and implementations of variants of the Charlson, Elixhauser, and Pediatric Complex Chronic Conditions (PCCC) algorithms. The package accepts full or compact ICD codes, supports mixed ICD-9/ICD-10 data, and can use present-on-admission and primary-diagnosis indicators. Comorbidity flags can be calculated for individual encounters or carried forward cumulatively across a patient’s record.
We also report that medicalcoder runs with R 3.5.0 or later and does not require additional packages, an internet connection, or external data files. Its outputs are consistent with published reference implementations. The longitudinal option can identify additional flags when a carry-forward assumption is appropriate, while respecting algorithm logic for present-on- admission and technology-dependent conditions.
The article is available through JAMIA Open. The package is available on CRAN and GitHub.