Automated CCI® Score Calculator
Year: 2026
Type: preprint
Abstract: Postoperative complications affect approximately one in four surgical patients, contributing to patient morbidity, as well as prolonged hospital stays and increased healthcare costs. It is therefore crucial to measure and assess these complications accurately for evaluating surgical outcomes and benchmarking quality of perioperative care. While the Clavien-Dindo classification (CDC) is the current standard, it is only used in practice to capture the most severe complication according to the invasiveness of its management (ranging from supportive measures (grade I) to invasive interventions (grade III), and ultimately death (grade V)). The Comprehensive Complication Index® (CCI®) provides a more comprehensive assessment, taking into account each postoperative complication and their severity in weights – derived from their respective CDC grades. The total CCI® score is a modified sum of all weights, ranging from 0 (no complication) to 100 (death). Nonetheless, its use in routine care is limited by the time and expertise required to classify each postoperative complication according to all therapeutic procedures, whereas the CDC is easier to define. A solution to improve efficiency in CCI® determination is automating it from electronic health record (EHR) data. While unsupervised machine learning techniques including artificial intelligence are increasingly developed within institutional EHR, their performance across languages is limited and the data security issues are an issue. On the other hand, structured administrative EHR are routinely generated for billing and documentation, including summaries with all diagnoses and interventions, classified with standardized labels and codes. Our project responds directly to this opportunity: we aim to (1) develop and validate a rule-based, transparent, auditable, and locally deployable program for automated CCI® generation from structured EHR documents of one hospital, exchangeable to the reference standard CCI® (i.e., generated by two clinical experts in surgery), (2) validate externally (other hospital, other language) that the performance of the program is exchangeable to the reference standard. If successful, our project will facilitate systematic outcome measurement, benchmarking, quality improvement initiatives, and surgical research.
Authors: Joana Rodrigues Ribeiro, Milo Puhan
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7204652731
