Underreporting of osteoporosis: an analysis of clinical coding and the effects of a quality improvement trial

Osteoporotic fractures are underrepresented in administrative data. In the present study, only 1 of 94 admissions for osteoporotic fracture was coded correctly, mostly due to inadequate clinical documentation. Following education of junior doctors, 36% of osteoporotic fractures were coded correctly, with coding of femur fractures particularly amenable to improvement.

Osteoporotic fractures are underrepresented in hospital administrative data. This study analysed the accuracy of clinical coding of osteoporotic fractures and prospectively evaluated the effectiveness of an educational intervention for junior medical officers (JMO) to improve documentation and coding accuracy.

Coding accuracy was measured in 279 consecutive inpatients deemed to have suffered an osteoporotic fracture based on predetermined criteria. Subsequently, an education programme for JMO was delivered, focusing on osteoporotic fracture criteria and clinical coding requirements. A 10-week post-intervention analysis was conducted to determine improvement in coding accuracy using baseline data as a historical control.

At baseline, 1 of 94 (1.1%) admissions meeting the criteria for osteoporotic fracture received the correct ICD-10 M80 primary diagnosis code. Post-intervention, coding accuracy improved markedly, with 14 out of 39 (36%) osteoporotic fracture admissions coded correctly. However, this improvement varied by site, with 77% of femur fractures, but only 10% of radius and 13% of humerus fractures receiving the correct primary clinical code. Coding accuracy decreased with time, from 42% in the first 5 weeks post-education to 27% in the subsequent 5 weeks.

Osteoporotic fractures are severely underrepresented in hospital administrative data due to inadequate documentation leading to inaccurate coding. Whilst education programmes can improve coding accuracy, long-term efficacy is limited by multiple factors. Since clinical coding forms an important part of public health resource allocation, automated system-based solutions are required to ensure accurate statistical capture of the true burden of osteoporosis and associated fractures.