UC Davis

OBJECTIVE: To evaluate the accuracy of Current Procedural Terminology (CPT) codes used to bill inpatient second- and third-trimester labor induction abortion relative to clinically confirmed encounters at a Northern California academic medical center. STUDY DESIGN: We conducted a two-phase retrospective chart review. In phase one, we assessed whether encounters billed using CPT codes recommended for labor induction abortion (59850, 59851, 59855, 59856) from 2014 to 2024 represented true labor induction abortions based on electronic medical record (EMR) review using predefined clinical criteria. In phase two, we assessed how actual labor induction abortions were coded through identification of clinically confirmed encounters through manual review of labor and delivery unit paper logs from January to December 2024 and examined the CPT codes applied to these encounters, including use of vaginal delivery codes permitted by professional guidance for gestations ≥ 20 weeks. RESULTS: Of 57 encounters billed using labor induction abortion-specific CPT codes over the 11-year period in phase one, 26 (46%) represented clinically confirmed labor induction abortions. Manual review in phase two identified 47 such procedures in 2024 alone; of these, only 5 (11%) were coded using labor induction abortion-specific CPT codes. Including vaginal delivery coding permitted by professional guidance, 20 of 47 cases (43%) were coded appropriately. CONCLUSIONS: CPT coding for inpatient second- and third-trimester labor induction abortion fails to reliably capture the true incidence of these procedures and demonstrates substantial inconsistency and misclassification, limiting utility of data for abortion research, surveillance, and financial analyses. IMPLICATIONS: CPT coding and billing for second- and third-trimester labor induction abortion at one institution is highly inaccurate and does not capture the true incidence. Assuming this finding is representative of national trends, billing datasets for these procedures are likely unreliable and have limited utility for research and policy.