BACKGROUND: The purpose of our study was to assess how the transition from ICD-9-CM to ICD-10-CM may impact fall circumstance rates using a sample of inpatient records for Massachusetts adults ages 65 and older admitted for a fall-related injury.
METHODS: We used the Massachusetts FY2012 Inpatient Hospital Discharge Database maintained by the Center for Health Information and Analysis (CHIA) to identify cases for the review. Our sample was selected from Massachusetts residents ages 65 and older with a fall-related external cause code assigned to their medical record. Cases from all acute care hospitals were sampled in approximately the same proportion with which they made up the percent of fall-related injury cases for this age group. The final sample included 659 records. We hired professional medical records coders to review each record and assign both ICD-9-CM and ICD-10-CM external cause of injury codes. Coders used a coding manual developed for the review and were blinded to the extent possible to hospital-assigned ICD-9-CM external cause codes. We calculated agreement ratios (ICD-10-CM count/ICD-9-CM count) based on project-assigned codes for fall circumstance categories.
RESULTS: Based on preliminary analysis, agreement ratios were very high for many fall circumstance categories, including falls from: stairs or curb (99%), a bed or chair (100%), a building, bridge or other structure (100%), and a ladder or scaffolding (100%). Fall circumstances in which there was less agreement included falls from other furniture (131%), other falls between levels (75%), falls with strike against object (63%), other falls (242%), and unspecified falls (54%). Potential limitations: agreement ratios in this study may have been closer to 100% because ICD-9-CM and ICD-10-CM codes were assigned by the same coder. Codes assigned by project coders may differ from those assigned by hospital-based coders. The sample was a subset of Massachusetts residents, those ages 65 and older, and may not be generalizable to all ages or to other injury mechanisms.
CONCLUSIONS: Dual-coded data that includes external cause codes for both ICD-9-CM and ICD-10-CM can be used to identify injury categories whose rates may be impacted by the transition to ICD-10-CM codes. While rates for some fall circumstances may not be significantly impacted by the transition to ICD-10-CM, rates of other fall circumstances may be signficantly lower or higher.