AUTHORS
Joshua Rivenbark, MD, PhD, Jerome Federspiel, MD, PhD2, Alison Gemmill, PhD, MPH3, Jane A Little, MD4, Elizabeth Simmons, PhD5, Justin Trogdon, PhD6, Lydia Pecker, MD7
AFFILIATIONS
1 Thomas Jefferson University, Philadelphia, Pennsylvania, United States
2 Duke University, Durham, North Carolina, United States
3 University of California Los Angeles, Los Angelos, California, United States
4 University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States
5 University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States
6 University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States
7 Johns Hopkins University, Baltimore, Maryland, United States
ABSTRACT
Introduction
Pregnant people with sickle cell disease (SCD) are at substantially increased risk of severe maternal morbidity (SMM) and mortality relative to the general obstetric population. In addition to the human toll, SMM events have an economic cost, as pregnancies complicated by these events have 1.5-fold higher total health care costs. Despite the associations between SCD and SMM, and SMM and health care costs, the costs of pregnancy with SCD have not been characterized. Here, we describe the health care costs accrued during pregnancy for the Medicaid-enrolled obstetric SCD population relative to a Medicaid-enrolled obstetric control population without SCD.
Methods
This study used national Medicaid insurance claims data, specifically the Transformed Medicaid Statistical Information System Analytic Files from 2018 to 2022, which included Inpatient, Other Services (outpatient), Pharmacy, Managed Care, and Demographic and Eligibility files. Individuals with SCD were identified as those who met the standard case definition for SCD in administrative data (≥ 3 non-pharmacy encounters with an associated ICD-10 CM code for SCD within a 5-year span of time). Individuals who were dual-eligible for Medicare, disenrolled from Medicaid for more than 60 days, or enrolled in managed care were excluded.
Pregnancy episodes were defined by identifying pregnancy outcomes in TAF Inpatient and TAF Other Services files using diagnostic and procedural codes. The first outcome observed for each individual was used to construct a pregnancy episode, which was defined as the full gestation period plus 90 days post-partum. Gestation was determined using ICD-10 CM codes indicating gestational age at time of pregnancy conclusion. Only pregnancy episodes which were entirely within the study window were included. A limit of one pregnancy per person was included.
A control cohort, from a random 5% sample of all individuals enrolled in Medicaid with a pregnancy-related diagnostic or procedural code during the study window, was derived using the same selection and pregnancy episode identification criteria as above.
The primary outcome of interest was total fee-for-service costs accumulated over the course of the pregnancy window, converted to 2022 dollars. Demographic characteristics available included age, race/ethnicity, state of residence, and poverty status.
A coarsened exact matching strategy was used to compare mean and median costs in the SCD cohort to the control cohort with exact matching on race/ethnicity, state, and poverty, and coarsened exact matching on age. Alternative matching strategies were tested as sensitivity analyses. Costs were divided into delivery admission, non-delivery inpatient admissions, outpatient visits, and pharmacy.
Results
There were 413 individuals with SCD and 30,047 individuals in the control cohort who met all selection criteria and had pregnancy episodes defined. The SCD population had a median (25th-75th percentile) age of 24 years (20 – 29); the proportions of reported race/ethnicity were 72.4% non-Hispanic Black, 10.7% Hispanic, 2.7% non-Hispanic White, and 14.3% other or unreported. Reported income was under 100% of the Federal Poverty Level (FPL) for 43.6%, over 100% FPL for 7.3%, and unreported for 49.2%. After coarsened exact matching, there were 401 individuals with SCD and 9,809 individuals in the control cohort that were matched.
Mean (SD) health care costs accrued during pregnancy episodes were $26,956 ($34,058) in the SCD population versus $9,549 ($8,783) in the control cohort. Median (25th-75th percentile) costs were $15,534 ($8,220 – $34,210) in the SCD population versus $7,725 ($4,558 – $12,400) in the control cohort. In a weighted, log-transformed regression model comparing costs between SCD and control, SCD was associated with a 151% increase in costs (p < 0.001). In the SCD population, the greatest proportion of costs were attributable to non-delivery inpatient admissions, followed by delivery admission.
Discussion
People with SCD accrued double the median health costs and nearly triple the mean health care costs during pregnancy when compared to a matched control population. This finding supports ongoing research aimed at improving the health of people with SCD during pregnancy, which may in turn reduce health care costs. Given the high costs associated with pregnancy and SCD, even relatively resource-intensive treatments such as transfusion therapy may be cost-saving or cost-effective.
DISCLOSURES
Joshua Rivenbark, MD, PhD: Agios Pharmaceuticals - Research Funding