The Relationship Between Patient Perceived Pain and the Numerical Rating System: A Prospective Validation Using Electronic Health Record Data

Lisiane Pruinelli,Minh Nguyen,Nicholas Stalter,Steven G Johnson,Gyorgy Simon,Stuart Speedie,Sisi Ma
DOI: https://doi.org/10.3233/SHTI240153
2024-07-24
Abstract:We developed a method of using the Clinically Aligned Pain Assessment (CAPA) measures to reconstruct the Numeric Rating System (NRS). We used an observational retrospective cohort study design with prospective validation using de-identified adult patient data derived from a major health system. Data between 2011-2017 were used for development and 2018-2020 for validation. All included patients had at least one NRS and CAPA measurement at the same time. An ordinal regression model was built with CAPA components to predict NRS scores. We identified 6,414 and 3,543 simultaneous NRS-CAPA pairs in the development and validation dataset, respectively. All CAPA components were significantly related to NRS, with RMSE of 1.938 and Somers' D of 0.803 on the development dataset, and RMSE of 2.1 and Somers' D of 0.74 when prospectively validated. Our model was capable of accurately reconstructing NRS based on CAPA and was exact when the NRS was [0,7].
What problem does this paper attempt to address?