BACKGROUND: Approximately 16% of women who use contraception in the United States use long-acting reversible contraceptives such as an intrauterine device (IUD). Outcomes such as uterine perforation and IUD expulsion might be associated with difficulty in IUD insertion.
OBJECTIVES: Assess difficult IUD insertion as a potential risk factor for uterine perforation and IUD expulsion.
METHODS: A cohort of women receiving IUDs were identified from each of four sites (three Kaiser Permanente sites—Northern California, Southern California, and Washington—and Regenstrief Institute, Indiana) for assessment of exposures covered elsewhere. Structured data (National Drug Codes, International Classification of Diseases 9/10, Current Procedural Terminology) and unstructured data (clinical notes with natural language processing) were used to identify uterine perforation, IUD expulsion, and five possible indicators of difficult IUD insertion: cervical dilation, ultrasound assessment, paracervical block, misoprostol use, and provider note of difficult insertion. The proportion of women with indicators of potentially difficult insertion pooled across sites was assessed overall and by uterine perforation and IUD expulsion along with the Hazard Ratio (HR) for difficult insertion associated with each outcome.
RESULTS: Among 326,658 women with IUD insertions, 29,777 (9.1%) had at least one indicator for a potentially difficult insertion, ranging 3.5%-16.3% across sites. The frequency of each indicator varied across sites. Women who experienced difficult insertion had a higher risk for uterine perforation and IUD expulsion (Hazard Ratio [HR] for uterine perforation: 1.38 [95% CI 1.14, 1.68]; HR for IUD expulsion: 1.16 [95% CI 1.08, 1.24]). This slightly higher proportion of uterine perforation and IUD expulsion among women with difficult insertions was seen at all sites, though there was variability in the specific indicators both across outcomes and by site.
CONCLUSIONS: The prevalence of difficult insertions was similar to published data from prospective studies (8%-9%) and was higher among women with IUD perforations and expulsions. We were able to identify indicators of difficult insertion using structured and unstructured data. Differences in prevalence exist across sites which might suggest differences in modes of data collection, patient characteristics, or clinician practice. We cannot determine the proportion of indicators that represent true difficult insertions.
