BACKGROUND: Healthcare databases are increasingly used for medication and device postmarket safety studies. Electronic health records (EHRs) provide a rich source of structured and unstructured data, but changes in the data over time and the potential impact on study results must be considered. On October 1, 2015, the US transitioned from the use of ICD-9-CM to ICD-10-CM coding. We conducted a study that developed algorithms including ICD-9-CM codes to identify intrauterine device (IUD)-related uterine perforation and expulsion.
OBJECTIVES: The purpose of this analysis was to assess the reliability of algorithms developed that were validated with ICD-9-CM codes, and then transitioned to ICD-10-CM codes, under the assumption of no temporal trends that would affect the incidence of the outcomes of interest.
METHODS: Four sites in the US participated: three Kaiser Permanente sites—Northern California, Southern California, Washington—and Regenstrief Institute in Indiana. The proportion of patients with an IUD who had a uterine perforation or IUD expulsion in the 12 months before and 12 months after implementation of ICD-10-CM coding was assessed. For this evaluation, the previously validated, site-specific algorithms for uterine perforation and IUD expulsion—developed using structured data (e.g., ICD-9-CM, Current Procedural Terminology codes, National Drug Codes) and unstructured data (Natural Language Processing terms)—were utilized, replacing ICD-9-CM with the mapped ICD-10-CM codes.
RESULTS: The number of IUD insertions per year at each site ranged from 3,074 to 49,866. The number of uterine perforation events per year ranged 0.06%-0.33%, and IUD expulsions ranged 0.50%-0.91%. In the data pooled across sites, the proportion of women with an IUD and a uterine perforation was 0.11% in the 12 months before ICD-10-CM code implementation and 0.12% in the 12 months after. The proportion of women with an IUD expulsion in the 12 months before and 12 months after ICD-10-CM code implementation was 0.72% and 0.77%, respectively. At each research site, the proportions before and after ICD-10-CM code implementation were relatively consistent, except at the smallest site, where estimates were more variable but based on a small number of events. At this site, all potential cases were manually reviewed to verify case status.
CONCLUSIONS: Algorithms developed and validated for uterine perforation and IUD expulsion showed no evidence of change due to transition from ICD-9-CM to ICD-10-CM codes, resulting in a consistent proportion of women classified as experiencing these outcomes.
