Key result
Novel HF screening in hospitalized diabetics yields 96% sensitivity, outperforming standard discharge codes.
Why the study?
The reliability of hospital discharge codes for heart failure, acute myocardial infarction, and stroke compared with adjudicated diagnoses, as well as a scalable adjudication approach, needed to be evaluated in people with diabetes.
How reliable are ICD-10 hospital discharge codes for identifying cardiovascular hospitalizations in people with diabetes, and can a novel screening approach improve accuracy for heart failure?
Population
685 people with diabetes (1274 admissions) at three Australian hospitals during 2018-2020
Comparison
Hospital discharge codes vs adjudicated diagnosis
Design
Population-based pilot validation study
Authors
Loading...
ICD-10 codes may underestimate HF hospitalizations in diabetes; cross-sectional data leave open whether novel screening improves case identification in practice.
Observational (n=685)
Yes
How reliable are ICD-10 hospital discharge codes for identifying cardiovascular hospitalizations in people with diabetes, and can a novel screening approach improve accuracy for heart failure?
Effect estimate: Kappa 0.81 for HF, ≥0.91 for AMI/stroke
ICD-10 codes substantially underestimate heart failure hospitalizations in patients with diabetes, but a novel two-stage screening approach can accurately identify cases with high sensitivity and specificity.
Huynh et al. (2024) conducted an observational in Diabetes (n=685). ICD-10 codes and novel screening approach vs. Adjudicated diagnosis was evaluated on Reliability of hospital discharge codes for HF, AMI and stroke compared with adjudicated diagnosis (Kappa 0.81 for HF, ≥0.91 for AMI/stroke). A novel screening approach for heart failure in hospitalized patients with diabetes yielded 96% sensitivity and 98% specificity, outperforming standard ICD-10 codes (72% sensitivity).