Sociodemographic Factors Associated with Antenatal Steroid Exposure among Late Preterm Births

As the risk of preterm birth is known to vary widely, we aimed to determine if antenatal steroid exposure among LPBs varied based on sociodemographic characteristics. We hypothesized that sociodemographic factors may influence a provider’s clinical judgment of a patient’s likelihood of preterm birth and, therefore, also be associated with antenatal steroid use. This cross-sectional analysis used the 2021 US natality data and included non-anomalous, liveborn, hospital-based singleton births at ≥34 weeks of gestation to mothers without diabetes, a cohort similar to those in the ALPS Trial. The following sociodemographic factors were compared among those who received vs. did not receive steroids using chi-square tests: age, race (as designated and categorized on the birth certificate), ethnicity, primary pay for the delivery, marital status, and education. In 2021, 237,025 late preterm births met eligibility criteria, of which 17.3% were exposed to antenatal steroids. Among the following sociodemographic factors, the odds of antenatal steroid receipt were lower compared to the reference majority population: 1) Black (adjusted odds ratio (aOR) 0.88 (95% CI 0.81, 0.96)) and Native Hawaiian or Other Pacific Islander (NHOPI) (aOR 0.58 (95% CI 0.43, 0.79) compared to White race; 2) less than high school education (aOR 0.76 (95% CI 0.72, 0.81)) or high school education (aOR 0.87 (95% CI 0.83, 0.91)) compared to post-secondary education; and 3) public (aOR 0.83 (95% CI 0.78, 0.87)) or no insurance (aOR 0.68 (95% CI 0.61, 0.77)) compared to private insurance. Age and marital status were not associated with steroid receipt. Despite no known differential treatment effects of antenatal steroids by sociodemographic factors, steroid exposure among LBPs varied significantly among races, ethnicities, payers, and education levels in the US. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 24, 2024. ; https://doi.org/10.1101/2024.01.22.24301562doi: medRxiv preprint

In 2016, the Antenatal Late Preterm Steroid Trial demonstrated the benefit of antenatal steroids in reducing respiratory morbidity among late preterm births (“LPB,” 34-36 weeks of gestation).1 Prior studies have shown that this trial and its dissemination resulted in increased steroid use in the late preterm period in the US, though adoption was not uniform.2–4 As the risk of preterm birth is known to vary widely, we aimed to determine if antenatal steroid exposure among LPBs varied based on sociodemographic characteristics.5 We hypothesized that sociodemographic factors may influence a provider’s clinical judgment of a patient’s likelihood of preterm birth and, therefore, also be associated with antenatal steroid use. Study Design: This cross-sectional analysis used the 2021 US natality data and included non-anomalous, liveborn, hospital-based singleton births at ≥34 weeks of gestation to mothers without diabetes, a cohort similar to those in the ALPS Trial.1 The primary outcome was the receipt of steroids, as reported on the birth certificate. The following sociodemographic factors were compared among those who received vs. did not receive steroids using chi-square tests: age, race (as designated and categorized on the birth certificate), ethnicity, primary pay for the delivery, marital status, and education. A multivariable logistic regression model was constructed with all the sociodemographic and two clinical factors hypothesized to influence clinical decision-making (parity and week of gestation at delivery). Models used cluster robust standard errors based on county of birth. Missing data were included in the model as “unknown.” P-values <0.05 were considered statistically significant. The Mass General Brigham institutional review board categorized this as non-human subjects research.

In 2021, 237,025 late preterm births met eligibility criteria, of which 17.3% were exposed to antenatal steroids. Among the following sociodemographic factors, the odds of antenatal steroid receipt were lower compared to the reference majority population (Table 1): 1) Black race . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 24, 2024. ; https://doi.org/10.1101/2024.01.22.24301562doi: medRxiv preprint (adjusted odds ratio (aOR) 0.88 (95% CI 0.81, 0.96)) and Native Hawaiian or Other Pacific Islander (NHOPI) (aOR 0.58 (95% CI 0.43, 0.79) compared to White race; 2) less than high school education (aOR 0.76 (95% CI 0.72, 0.81)) or high school education (aOR 0.87 (95% CI 0.83, 0.91)) compared to post-secondary education; and 3) Medicaid (aOR 0.83 (95% CI 0.78, 0.87)) or no insurance (aOR 0.68 (95% CI 0.61, 0.77)) compared to private insurance. Age and marital status were not associated with steroid receipt. Of the two clinical factors included, nulliparity and earlier gestational ages were associated with increased odds of steroid receipt.

Despite no known differential treatment effects of antenatal steroids by sociodemographic factors, steroid exposure among LBPs varied significantly among races, ethnicities, payers, and education levels in the US. Many of the same subgroups with lower odds of steroid receipt have the highest preterm birth rates.5 There are several possible explanations, including 1) provider assessment of baseline risk of preterm delivery does not influence the likelihood of steroid receipt, 2) underlying biases affect clinical judgment that results in disparate use; 3) patients have a varying willingness to receive intervention; or 4) differential clinical and structural (e.g., distance to hospital) factors influence the opportunity to intervene (e.g., patients presenting “too late” to receive the intervention). The study used US birth certificate data; thus, it was limited in its ability to observe the exact timing of steroid administration or other clinical factors that may have affected clinical decision-making. Additional studies examining the drivers of these sociodemographic differences, such as patient perception of risk, clinician judgment, and hospital or regional factors affecting steroid use, are needed to inform policy solutions to address disparities in antenatal steroid use, especially among populations at the highest risk for preterm birth. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 24, 2024. ; https://doi.org/10.1101/2024.01.22.24301562doi: medRxiv preprint

1. Gyamfi-Bannerman C, Thom EA, Blackwell SC, et al. Antenatal Betamethasone for Women at Risk for Late Preterm Delivery. New England Journal of Medicine. 2016;374(14):1311- 1320. doi:10.1056/NEJMoa1516783 2. Clapp MA, Melamed A, Freret TS, James KE, Gyamfi-Bannerman C, Kaimal AJ. US Incidence of Late-Preterm Steroid Use and Associated Neonatal Respiratory Morbidity After Publication of the Antenatal Late Preterm Steroids Trial, 2015-2017. JAMA Netw Open. 2022;5(5):e2212702. doi:10.1001/jamanetworkopen.2022.12702 3. Freret TS, Cohen JL, Gyamfi-Bannerman C, et al. Regional Variation in Antenatal Late Preterm Steroid Use following the ALPS Trial. medRxiv. 2023;2023.05.25.23290522. doi:https://doi.org/10.1101/2023.05.25.23290522 4. Society for Maternal-Fetal Medicine (SMFM) Publications Committee. Implementation of the use of antenatal corticosteroids in the late preterm birth period in women at risk for preterm delivery. Am J Obstet Gynecol. 2016;215(2):B13-15. doi:10.1016/j.ajog.2016.03.013 5. Manuck TA. Racial and ethnic differences in preterm birth: A complex, multifactorial problem. Semin Perinatol. 2017;41(8):511-518. doi:10.1053/j.semperi.2017.08.010 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 24, 2024. ; https://doi.org/10.1101/2024.01.22.24301562doi: medRxiv preprint Table 1: Antenatal Steroid Exposure by Sociodemographic Factors Characteristics Late Preterm Births Potentially Eligible for Antenatal Steroids 1 N=237,025 No Steroid Receipt N=195,967 Steroid Receipt 2 N=41,058 Adjusted Odds Ratio (95% CI) for Steroid Receipt 3 Age, years <18 3110 (1.6%) 519 (1.3%) 0.93 (0.83, 1.04) 18-34 150925 (77.0%) 31679 (77.2%) Reference ≥35 41932 (21.4%) 8860 (21.6%) 0.97 (0.93, 1.01) Race 4 AIAN 2209 (1.1%) 439 (1.1%) 0.96 (0.74, 1.26) Asian 10371 (5.3%) 2103 (5.1%) 0.83 (0.73, 0.95)** Black 41212 (21.0%) 8127 (19.8%) 0.88 (0.81, 0.96)** More than 1 race 5741 (2.9%) 1273 (3.1%) 1.00 (0.88, 1.15) NHOPI 968 (0.5%) 118 (0.3%) 0.59 (0.43, 0.79)*** White 135466 (69.1%) 28998 (70.6%) Reference Ethnicity Hispanic 53160 (27.1%) 8547 (20.8%) 0.74 (0.66, 0.84)*** Non-Hispanic 140920 (71.9%) 32080 (78.1%) Reference Unknown 1887 (1.0%) 431 (1.0%) 1.09 (0.81, 1.47) Marital status Married 89456 (45.6%) 20115 (49.0%) Reference Not married 86157 (44.0%) 17100 (41.6%) 1.03 (0.99, 1.07) Unknown 20354 (10.4%) 3843 (9.4%) 0.96 (0.77, 1.20) Education Less than high school 29197 (14.9%) 4533 (11.0%) 0.76 (0.72, 0.81)*** High school 58723 (30.0%) 11108 (27.1%) 0.87 (0.83, 0.91)*** Post-secondary 104724 (53.4%) 24758 (60.3%) Reference Unknown 3323 (1.7%) 659 (1.6%) 0.87 (0.68, 1.11) Payer Medicaid 84326 (43.0%) 20787 (50.6%) 0.83 (0.78, 0.87)*** Other 6144 (3.1%) 1288 (3.1%) 0.90 (0.78, 1.05) Private 97419 (49.7%) 17795 (43.3%) Reference Self-pay 6642 (3.4%) 935 (2.3%) 0.68 (0.61, 0.77)*** Unknown 1436 (0.7%) 253 (0.6%) 0.75 (0.59, 0.95)* Parity Nulliparity 57667 (29.4%) 13668 (33.3%) 1.13 (1.08, 1.17)*** Multiparity 137473 (70.2%) 27263 (66.4%) Reference Unknown 827 (0.4%) 127 (0.3%) 0.79 (0.54, 1.15) Gestational Age at Birth (week) 34 weeks 30905 (15.8%) 12139 (29.6%) 2.85 (2.71, 3.01)*** 35 weeks 53214 (27.2%) 13119 (32.0%) 1.78 (1.72, 1.85)*** 36 weeks 111848 (57.1%) 15800 (38.5%) Reference AIAN, American Indian or Alaskan Native; NHOPI, Native Hawaiian or Other Pacific Islander. 1 “Late Preterm Birth” included births between 34-36 weeks of gestation. “Eligible” births included non-anomalous, liveborn, hospital-based singleton births at ≥34 weeks of gestation to mothers without diabetes. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 24, 2024. ; https://doi.org/10.1101/2024.01.22.24301562doi: medRxiv preprint 2 P-value comparisons between the “no steroid receipt” and “steroid receipt” were <0.001 for all characteristics. 3 The population majority was set as the reference group for each category. P-value designations: * p<0.05, **p<0.01, ***p<0.001. 4 The race variable (‘mrace6’) was pre-defined into the 6 categories in the US natality data. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 24, 2024. ; https://doi.org/10.1101/2024.01.22.24301562doi: medRxiv preprint