Do discharge codes underestimate hospitalisation due to heart failure? Validation study of hospital discharge coding for heart failure

Background Discharge codes are frequently used to describe hospital activity related to heart failure (HF). Objectives To determine whether discharge codes for HF underestimated or overestimated hospital activity related to HF. Design Patients with atrial fibrillation (AF), who commonly have HF, were identified and their case notes reviewed to identify cases of HF missed by discharge codes. Participants and methods Patients admitted between November 1997 and January 1998 with either HF or AF. Identification of HF and AF by ICD10 hospital discharge codes. Identification of additional cases of AF from a central hospital‐wide ECG database. Results We identified 330 cases with an ICD 10 code for HF, of which 43 (13%) were deemed to be miscoded, 32 patients (10%) were classified as possible, 39 (12%) as probable and 216 (65%) as definite HF. Results were similar whether or not HF was the primary discharge diagnosis. We identified 452 patients with AF, of whom 45 (10%) were classified as probable and 193 (43%) as definite HF. 129 (54%) of these cases had no diagnostic discharge code for HF. ICD 10 discharge codes for HF were correct in 77% of cases but identified only 66% of patients with probable or definite HF in this analysis. Screening of other diagnoses would have identified further cases of HF. Conclusions Hospital discharge codes substantially underestimate hospital events related to HF in the UK.