(557) Accuracy of ICD-10 Codes for Placenta Accreta Spectrum: Frequency and Reasons for Miscoding

Objective: The ICD-10 codes for placenta accreta spectrum (PAS) were introduced in 2015, facilitating improved research and quality tracking of this important source of maternal morbidity. However, the validity of these codes has not been thoroughly evaluated. We aimed to identify the frequency and reasons for PAS miscoding.

Study Design: We conducted a case series analysis of all deliveries associated with an ICD-10 code for PAS in the two large healthcare systems (17 hospitals, including 2 PAS referral centers) in our state between September 2015 and December 2023. Medical records were reviewed by three reviewers to determine whether PAS was present using the FIGO clinical and/or histopathologic criteria. If PAS was excluded, the reason for misapplication of the code was assigned.

Results: There were 394 deliveries associated with ICD-10 codes for PAS during the study period, with most (n=156, 68%) occurring at a PAS referral center; 40% (95% CI 35-45%) were false positive PAS. Many demographic and delivery factors were associated with a false PAS diagnosis including younger maternal age at delivery, a lower number of prior cesarean deliveries, absence of placenta previa, and delivery without hysterectomy, and later gestational age at delivery (Table). The most common cause (42.9%) for a false positive was hemorrhage due to another etiology not meeting clinical or histopathologic criteria for PAS (Figure). In 22.4% of false positives, there was antenatal suspicion of PAS, but PAS was not seen at delivery. In 22.4% of false positives, the code was erroneously applied.

Conclusion: In this study, 40% of cases with an ICD-10 code for PAS did not meet clinical and/or histopathologic criteria for PAS. This underscores the need for improved clinician understanding of the clinical PAS criteria and improved coding practices for PAS. Additionally, studies using ICD-10 to identify cases of PAS should consider the possibility of bias introduced by a significant false positive rate.