High-Risk Nonsteroidal Anti-inflammatory Drug Prescribing: Variation by Documentation of Chronic Kidney Disease Status

Background Chronic kidney disease (CKD) affects over 10% of people globally. Despite its prevalence, CKD is persistently under-recognized and under-documented in healthcare settings. Previous studies have shown that patients who meet laboratory-based criteria for CKD but have no formal CKD diagnosis in their medical record are less likely to receive guideline-directed medical care for CKD. However, it is unclear whether CKD documentation status influences receipt of medications that may be high-risk for patients with CKD, including nonsteroidal anti-inflammatory drugs (NSAIDs). Objective We evaluated variation in NSAID prescribing based on CKD documentation status over a 10-year period, hypothesizing that patients who met diagnostic criteria for CKD but had no corresponding ICD diagnosis on record were more likely to receive long-term prescription NSAIDs. Methods This retrospective cohort study used electronic health record (EHR) data from the TriNetX Research Network representing 67 U.S. healthcare organizations (2013–2023). Adult patients with CKD were classified by CKD documentation status (diagnosis on record vs. criteria for CKD but no diagnosis on record). We applied logistic regression to assess the relationship between CKD documentation status and receipt of long-term NSAIDs, adjusting for demographics, comorbidities, and year. Results Among 1,777,336 patients, 52.9% met diagnostic criteria for CKD but had no CKD diagnosis on record. Overall, 9.0% of patients were prescribed long-term NSAIDs (6.3% of patients with a CKD diagnosis on record and 11.5% of those without). Patients with no CKD diagnosis on record were at 1.76 [95% CI: 1.74–1.79] times greater odds of receiving long-term NSAIDs compared to patients with a diagnosis on record. Female sex and Black/African American race were independently associated with higher odds of receiving long-term NSAIDs (1.47 [95% CI:1.46–1.49] and 1.37 [95% CI:1.30–1.44], respectively). Conclusions Across 67 US healthcare organizations, more than half of patients with laboratory-based evidence of CKD did not have a CKD diagnosis on record in the EHR. Absence of CKD documentation was associated with increased likelihood of receiving long-term NSAIDs. These findings suggest improved EHR documentation as a target for reducing potentially harmful prescribing in patients with CKD.