Skip to main content
NCBI home page
JAMA Network logoLink to JAMA Network
. 2026 Sep 22;9(9):e2635321. doi: 10.1001/jamanetworkopen.2026.35321

Decomposition of Black and White Infant Mortality Rates

Monica J Alexander 1,2,✉, Rohan Alexander 1,3, Leslie Root 4
PMCID: PMC13598697  PMID: 42771344

This cross-sectional study uses 22 years of mortality data to quantify gestational age distribution and mortality rates in infants born to Black mothers compared with infants born to White mothers.

Key Points

Question

How has the relative contribution of the 2 components of the Black and White infant mortality gap—gestational age composition and within–gestational age mortality risk—varied over time and for different causes of death?

Findings

In this cross-sectional study of 57 870 726 live births and 366 384 infant deaths from 2003 to 2024, most of the mortality gap between infants born to Black and White mothers was accounted for by differences in gestational age at birth; however, the share of the mortality risk component increased from 21% to 32%. The gap increased after 2020, with sudden unexpected infant death showing the largest share increase of any cause.

Meaning

These findings suggest that closing the infant mortality gap requires addressing both preterm birth disparities and growing differences in within–gestational age mortality risk, particularly for sudden unexpected infant death.

Abstract

Importance

The infant mortality rate (IMR) of the non-Hispanic Black population in the US remains more than twice that of the non-Hispanic White population. Understanding this disparity in terms of differences in gestational age distribution, mortality risk within gestational age categories, and specific causes of death is essential for targeting interventions.

Objective

To decompose the Black and White infant mortality gap into gestational age composition and mortality risk components and to identify the causes of death that may be factors in recent changes.

Design, Setting, and Participants

This cross-sectional study analyzed Period Linked Birth/Infant Death Records from the US National Center for Health Statistics from January 1, 2003, through December 31, 2024. Participants included all live births to non-Hispanic Black mothers and non-Hispanic White mothers with known gestational age and plausible birth weight for gestational age.

Exposures

Maternal race and ethnicity of non-Hispanic Black or non-Hispanic White.

Main Outcomes and Measures

The primary measures were the composition effect (differences in gestational age distribution) and rate effect (differences in within–gestational age mortality) from a Kitagawa decomposition of the mortality difference between infants born to Black and to White mothers, overall and by specific cause of death.

Results

There were 57 870 726 live births and 336 384 infant deaths in the non-Hispanic Black (141 651 [38.7%] of all deaths) and non-Hispanic White (224 733 [61.2%] of all deaths) populations between 2003 and 2024. The Black IMR declined from 13.2 (95% CI, 12.9-13.5) to 10.7 (95% CI, 10.5-11.0) deaths per 1000 live births between 2003 and 2024, while the White IMR declined from 5.6 (95% CI, 5.5-5.7) to 4.3 (95% CI, 4.2-4.4) deaths per 1000 live births. The absolute gap decreased from 7.7 to 6.0 deaths per 1000 live births between 2003 and 2019, but then increased to 6.4 (95% CI, 6.1-6.8) deaths per 1000 live births. The gestational age composition effect accounted for a majority of the difference across study years, but the rate effect increased as a share of the gap from 21% (95% CI, 18%-24%) in 2003 to 32% (95% CI, 28%-35%) in 2024. Between 2020 and 2024, the share of sudden unexpected infant death (SUID) to the gap showed the largest increase of any single cause (0.18 deaths per 1000 live births; SUID gap in 2024, 1.56 [95% CI, 1.42-1.71] deaths per 1000 live births).

Conclusions and Relevance

In this cross-sectional study of IMRs, gestational age composition differences accounted for the largest share of the Black and White infant mortality gap. However, the increased share of the mortality risk component over time underscores the need for targeted interventions beyond preterm birth prevention.

Introduction

The infant mortality rate (IMR) for infants born to non-Hispanic Black mothers in the US is more than twice that of infants born to non-Hispanic White mothers, a disparity that has persisted for decades.1 After years of decrease, the gap in Black and White IMR increased after 2020, as Black infant mortality increased while White infant mortality remained stable.2 Recent trend analyses have identified sudden unexpected infant death (SUID) as a notable exception to the long-term declines, with SUID rates increasing for the non-Hispanic Black population and the Black and White SUID disparity ratio increasing from 2.2 in 2017 to 2.8 in 2020.2,3

The IMR for infants born to non-Hispanic Black mothers is higher for 2 reasons. First, these infants are more likely to be born preterm, which increases the risk of dying in the first year—an age composition effect. Second, these infants face higher mortality risk within the same gestational age categories, reflecting a rate effect.4,5 To quantify the share of the age composition and rate effects, we applied the Kitagawa method,6 a demographic decomposition technique, to 22 years of infant mortality data. We also investigated the main causes of death to characterize the recent increase in the gap in Black and White IMR.

Methods

This cross-sectional study was determined by the University of Colorado Institutional Review Board to not involve human participants. The board waived the requirement for informed consent because the study used existing data that did not contain identifiable private information. This study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies.

Data were obtained from publicly available Period Linked Birth/Infant Death Records from the National Center for Health Statistics from January 1, 2003, through December 31, 2024.7 These files link each infant death certificate to its corresponding birth certificate, covering approximately 99% of infant deaths. The analysis was restricted to births to non-Hispanic Black mothers and non-Hispanic White mothers with known gestational age. These racial and ethnic categories are produced by the National Center for Health Statistics and were drawn from birth certificate information. Records with implausible combinations of birth weight and gestational age were excluded using the inclusion criteria described by Alexander et al8 (eAppendix 1 in Supplement 1). Gestational age was classified into 7 categories: extremely preterm (<28 weeks), very preterm (28-31 weeks), moderately preterm (32-33 weeks), late preterm (34-36 weeks), early term (37-38 weeks), full or late term (39-41 weeks), and postterm (≥42 weeks).

Statistical Analysis

The Black and White IMR difference was decomposed using the Kitagawa method,6 to partition the gap into an age composition effect (differences in the gestational age distribution) and a rate effect (differences in mortality within gestational age groups). We also performed this decomposition on each cause of death separately. To do so, we used International Classification of Disease, Tenth Revision (ICD-10) codes to group deaths into the top 13 top categories by underlying cause, with remaining deaths classified as other. By construction, summing the cause-specific composition effects recovers the overall composition effect, and similarly for rate effects. Full details and formulas for the decomposition are provided in eAppendix 2 in Supplement 1. We obtained 95% CIs for all rates and decomposition results using a Monte Carlo procedure (1000 iterations) that resampled cell-level death counts from a Poisson distribution, with the mean equal to the observed deaths, recomputed the decomposition, and provided the 2.5% and 97.5% quantiles. Sensitivity analyses were conducted using different gestational age classifications as well as a cause of death decomposition that split the combined SUID category into its 3 ICD-10 components (eAppendix 3 and eTable 4 in Supplement 1).

Analyses were conducted in R, version 4.5 (R Project for Statistical Computing).9 Reproducible code and data are available in eAppendix 4 in Supplement 1. Statistical significance was defined as 95% CIs excluding 1.00.

Results

There were 57 870 726 live births and 366 384 infant deaths in the non-Hispanic Black (141 651 [38.7%] of all deaths) and non-Hispanic White (224 733 [61.2% of all deaths) populations from January 1, 2003, to December 31, 2024 (eTable 1 in Supplement 1). Infants born to non-Hispanic Black mothers were more often born preterm (<37 weeks) compared with non-Hispanic White mothers (16.9% vs 10.5% over the study period). The Black IMR declined from 13.2 (95% CI, 12.9-13.5) to 10.7 (95% CI, 10.5-11.0) deaths per 1000 live births between 2003 and 2024 (19% reduction); the White IMR declined from 5.6 (95% CI, 5.5-5.7) to 4.3 (95% CI, 4.2-4.4) deaths per 1000 live births (23% reduction). The absolute Black and White infant mortality rate gap decreased from 7.7 to 6.0 deaths per 1000 live births between 2003 and 2019 but increased to 6.4 (95% CI, 6.1-6.8) by 2024, as Black infant mortality increased while White infant mortality continued to decline (Figure 1A).

Figure 1. Line Graph of Infant Mortality by Race and Bar Graphs of Decomposition of Composition and Rate Effects, 2003-2024.

Two-panel figure: infant mortality line chart and decomposition bar chart, 2003 to 2025. Panel A titled Infant mortality rates by race. Horizontal axis labeled Year with ticks from 2003 through 2025. Vertical axis labeled Deaths per 1000 live births, ranging from 0 to 15 with labeled ticks at 0, 5, 10, and 15. Two time series lines: a dark gray line labeled Black at the far right, and a bright blue line labeled White at the far right. Each line has a narrow, same-color shaded band surrounding it. The Black line starts near 13 point 4 in 2003, remains near 13 point 4 to 13 point 5 through about 2005, declines to about 12 point 9 by 2008, to about 12 point 4 in 2009, then to about 11 point 2 in 2010 to 2011, and gradually to about 10 point 2 by 2020; it stays near 10 point 2 in 2021, rises to about 10 point 6 in 2022, and ends near 10 point 7 in 2024. The White line starts near 5 point 6 in 2003, stays near 5 point 5 to 5 point 6 through about 2007, declines to about 5 point 2 in 2009, to about 5 point 0 in 2011, then gradually to about 4 point 4 by 2020; it remains near 4 point 3 to 4 point 5 from 2021 through 2024. A vertical dashed line at year 2020 crosses the full plot height. Panel B titled Decomposition of the gap into composition and rate effects. Horizontal axis labeled Year with ticks from 2003 through 2025. Vertical axis labeled Share of the gap, deaths per 1000 live births, ranging from 0 to 8 with labeled ticks at 0, 2, 4, 6, and 8. For each year, two adjacent bars with error bars: orange bars for Composition and blue-gray bars for Rate, with a legend near the upper center. Orange bars are consistently higher, near 6 point 0 to 6 point 3 in 2003 to 2005, decreasing to about 5 point 4 to 5 point 6 by 2008 to 2009, near 5 point 0 to 5 point 2 around 2010 to 2012, near 4 point 9 to 5 point 1 from about 2013 to 2019, about 4 point 7 in 2020, about 4 point 5 in 2021, about 4 point 2 in 2022, and about 4 point 3 to 4 point 4 in 2023 to 2025. Blue-gray bars range roughly from about 1 point 1 to 1 point 8 before 2020, then increase after 2021 to about 1 point 8 to 2 point 1 by 2022 to 2025. A vertical dashed line at 2020 appears in this panel as well.

Open in a new tab

A, Shaded regions of the line graph indicate 95% CIs; dashed line, 2020. B, Composition effect represents differences in gestational age distribution; rate effect, differences in within–gestational age mortality; error bars, 95% CI; dashed line, 2020.

In every year of the study period, the gestational age composition effect accounted for the majority of the gap in Black and White IMR. However, the mortality rate effect increased as a share of the gap, from 21% (95% CI, 18%-24%) in 2003 to 32% (95% CI, 28%-35%) in 2024. This shift increased beginning in 2020: between 2020 and 2024, the share of the rate effect increased from 1.42 (95% CI, 1.15-1.69) to 2.04 (95% CI, 1.75-2.33) deaths per 1000 live births, while the share of the age composition effect declined from 4.49 (95% CI, 4.39-4.60) to 4.40 (95% CI, 4.29-4.52) deaths per 1000 live births (Figure 1B). Sensitivity analyses showed similar results using more granular gestational age categories (eTable 4 in Supplement 1). Across study years, the earliest gestational age category (<28 weeks) accounted for over half the overall gap in IMR (in 2003, 5.29 [95% CI, 5.06-5.51] deaths per 1000 live births of a total gap of 7.7; in 2024, 3.51 [95% CI 3.31-3.72] deaths per 1000 live births of a total gap of 6.4) (eTables 2 and 3 in Supplement 1).

The Table shows cause-specific decomposition for 2024. SUID was the largest single cause contributor to the IMR gap (1.56 [95% CI, 1.42-1.71] deaths per 1000 live births). SUID deaths include those from sudden infant death syndrome, accidental suffocation and strangulation in bed, and other causes that remain unexplained even after investigation. Its contribution was due almost entirely to the rate effect (1.44 [95% CI, 1.30-1.59 of 1.56] deaths per 1000 live births). Short gestation or low birth weight was the second-largest contributor (1.44 [95% CI, 1.30-1.57] deaths per 1000 live births), but this was primarily an age composition effect.

Table. Cause-Specific Decomposition of the Gaps in Black and White IMR, With Kitagawa Decomposition for 2024.

Cause of death IMR gap, deaths per 1000 live births (95% CI) Composition effect, deaths per 1000 live births (95% CI)a Rate effect, deaths per 1000 live births (95% CI)b Rate effect, %c
All causes 6.44 (6.14 to 6.75) 4.40 (4.29 to 4.52) 2.04 (1.75 to 2.33) 32
Sudden unexpected infant death 1.56 (1.42 to 1.71) 0.12 (0.10 to 0.14) 1.44 (1.30 to 1.59) 92
Short gestation or LBW 1.44 (1.30 to 1.57) 1.25 (1.19 to 1.30) 0.19 (0.07 to 0.29) 13
Otherd 1.09 (0.95 to 1.23) 1.01 (0.96 to 1.06) 0.07 (−0.06 to 0.20) 7
Infection 0.52 (0.44 to 0.60) 0.35 (0.32 to 0.38) 0.17 (0.10 to 0.23) 32
Maternal complication 0.45 (0.37 to 0.52) 0.44 (0.41 to 0.47) 0.01 (−0.05 to 0.07) 1
Congenital malformation 0.42 (0.30 to 0.53) 0.44 (0.41 to 0.47) −0.01 (−0.13 to 0.10) −3
Respiratory distress 0.18 (0.14 to 0.24) 0.20 (0.18 to 0.22) −0.02 (−0.06 to 0.03) −9
Placenta, cord, or membranes 0.17 (0.11 to 0.22) 0.19 (0.17 to 0.21) −0.02 (−0.07 to 0.02) −15
Necrotizing enterocolitis 0.14 (0.10 to 0.18) 0.12 (0.10 to 0.13) 0.02 (−0.01 to 0.06) 17
Circulatory disease 0.13 (0.09 to 0.18) 0.06 (0.04 to 0.07) 0.08 (0.04 to 0.12) 59
Hypoxia and birth asphyxia 0.11 (0.06 to 0.15) 0.09 (0.08 to 0.11) 0.01 (−0.02 to 0.05) 14
Assault (homicide) 0.09 (0.06 to 0.13) 0.01 (0.00 to 0.01) 0.09 (0.05 to 0.12) 91
Unintentional injury 0.08 (0.05 to 0.11) 0.01 (0.00 to 0.02) 0.07 (0.04 to 0.11) 89
Hemorrhage 0.06 (0.03 to 0.10) 0.12 (0.10 to 0.13) −0.05 (−0.09 to −0.02) −89
Open in a new tab

Abbreviations: IMR, infant mortality rate; LBW, low birth weight.

a

Composition effect represents differences in gestational age distribution.

b

Rate effect indicates differences in within–gestational age mortality.

c

Rate percentage indicates the ratio of the rate to the overall gap.

d

Other represents any remaining cause of death.

Over the study period, SUID death had the largest rate effect of any cause, and this effect increased over time (Figure 2). From 2020 to 2024, the SUID rate effect increased from 1.26 (95% CI, 1.14-1.38) to 1.44 (95% CI, 1.30-1.59) deaths per 1000 live births, reflecting increasing disparities in within–gestational age SUID mortality rates rather than changes in the gestational age distribution. In the post–COVID-19 pandemic period (2020 and later), the share from SUID accounted for the largest increase of any single cause to the gap in IMR (0.18 deaths per 1000 live births; SUID gap in 2024, 1.56 [95% CI, 1.42-1.71] deaths per 1000 live births) (Figure 2). This increase accounted for 35% of the total net change (eFigure 1 in Supplement 1). Splitting the combined SUID category into its 3 ICD-10 components showed that ill-defined or unspecified (R99) had the largest increase, although sudden infant death syndrome and accidental suffocation and strangulation in bed also showed increases since 2015 (eFigure 2 in Supplement 1).

Figure 2. Line Graphs of Rate Effect (Within–Gestational Age Mortality Risk) in the Mortality Gap Between Infants Born to Black and White Mothers by Cause of Death, 2003-2024.

Line chart of rate effect by year and cause, with a marked rise after 2020. Single-panel line chart with shaded confidence bands for six cause categories. Horizontal axis labeled Year, with tick labels from 2003 through 2025. Vertical axis labeled Rate effect, deaths per 1000 live births, with labeled ticks at minus zero point five, zero, zero point five, one point zero, one point five, and two point zero. A vertical dashed reference line at 2020 intersects the plot area. At upper right, a boxed legend maps line colors to labels: dark teal, Sudden unexpected infant death; orange, Short gestation or L B W; bright blue, Infection; magenta, Maternal complication; light green, Congenital malformation; gray, Respiratory distress. The dark teal series runs highest across all years, near zero point nine to one point zero in 2003 to 2004, dipping to about zero point seven around 2011, rising to about one point zero by 2019, then increasing sharply after the 2020 dashed line to roughly one point three in 2021 and about one point four to one point five by 2022 to 2024; its shaded band widens slightly in later years. The orange series stays positive around zero point three to zero point four in the mid 2000s, declines toward about zero point two by 2011 to 2014, fluctuates near zero point two to zero point three through 2019, then varies around roughly zero point one to zero point three after 2020. The bright blue series remains low and positive, roughly zero point one to zero point two, with small peaks near 2006 and around 2009, and values near about zero point one to zero point two after 2020. The magenta series oscillates close to zero, ranging approximately from minus zero point one to plus zero point one, with small positive bumps around 2004 and 2017 and near-zero values after 2020. The light green series lies mostly below zero, around minus zero point one to minus zero point two, dipping more negative near 2020 to about minus zero point two, then trending upward toward about minus zero point one by 2024. The gray series stays closest to zero, slightly negative to slightly positive, generally between about minus zero point zero five and plus zero point zero five across the full period. Each series has a same-color semi-transparent band surrounding the line indicating uncertainty.

Open in a new tab

LBW indicates low birth weight. Shaded regions indicate 95% CIs; dashed line, 2020.

Discussion

In this analysis of 22 years of national linked birth-infant death data, we found that while gestational age composition differences remain the largest factor in the gap in Black and White IMR, the mortality risk component has increased and now accounts for nearly one-third of the gap. The increase in the IMR gap since 2020 was due in part to increasing SUID disparities.

The increasing rate effect indicates that within–gestational age mortality differences are increasing relative to composition differences. This finding suggests that even if preterm birth disparities were eliminated, a large mortality gap would remain. The concentration of a 2020 through 2024 increase in disparities in SUID warrants further investigation. This increase coincided with the COVID-19 pandemic, a respiratory pandemic, indicating a possible etiology for the large increase in SUID’s rate effect. The increase also paralleled racially disparate increases in maternal mortality during the pandemic.10 The dual burden of increasing racial disparities in both maternal and infant deaths underscores the interconnected nature of perinatal health inequities and the need for comprehensive strategies spanning the pregnancy to infancy continuum.

Limitations

Here we examined only mothers who were non-Hispanic Black and non-Hispanic White, but infant mortality and particularly SUID death rates are also high among children born to non-Hispanic American Indian and Alaska Native and non-Hispanic women of more than 1 race.11 The Kitagawa method assumes additive effects and cannot capture interactions between composition and rate differences. Cause of death classification relies on death certificate coding, which may vary in accuracy across populations. Other demographic and socioeconomic characteristics, such as maternal age, parity, and educational attainment, could change differently over time in the non-Hispanic Black and non-Hispanic White populations. Future work could extend this analysis to investigate how these factors may contribute to the changing gap in IMR.

Conclusions

In this cross-sectional study of infant mortality, gestational age composition differences continued to account for the largest share of the gap in Black and White IMR. However, the within–gestational age mortality risk component increased as a share of the overall gap from 2003 to 2024. The increased mortality risk component highlights the need for continued efforts to reduce preterm birth disparities, alongside targeted interventions addressing the increasing within–gestational age mortality risk differences, particularly for SUID.

Supplement 1.

eAppendix 1. Data exclusions

eTable 1. Record exclusions, 2003-2024

eFigure 1. Change in each cause’s contribution to the Black-White IMR gap between 2020 and 2024 (deaths per 1,000 live births)

eFigure 2. Rate effect (within-gestational-age mortality risk) contributions to the Black-White gap by the three ICD-10 components of Sudden Unexpected Infant Death, 2003-2024

eTable 2. Per-gestational-age decomposition of the Black-White infant mortality for 2003.

eTable 3. Per-gestational-age decomposition of the Black-White infant mortality for 2024

eAppendix 2. Supplemental methods details

eAppendix 3. Sensitivity analyses

eTable 4. Comparison of 7-category (main) and 16-category gestational-age decomposition, 2024

eAppendix 4. Reproducible code and data

jamanetwopen-e2635321-s001.pdf (479KB, pdf)
Supplement 2.

Data Sharing Statement

jamanetwopen-e2635321-s002.pdf (16KB, pdf)

References

  • 1.Ely DM, Driscoll AK. Infant mortality in the United States, 2022: data from the Period Linked Birth/Infant Death File. Natl Vital Stat Rep. 2024;73(5):1-20. doi: 10.15620/cdc/157006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Wolf ER, Rivara FP, Sen A, Woolf SH. Sudden unexpected infant death and disparities in infant mortality in the US, 1999-2022. JAMA Pediatr. 2025;179(3):344-346. doi: 10.1001/jamapediatrics.2024.6200 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Shapiro-Mendoza CK, Woodworth KR, Cottengim CR, et al. Sudden unexpected infant deaths: 2015-2020. Pediatrics. 2023;151(4):e2022058820. doi: 10.1542/peds.2022-058820 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Alexander M, Root L. Competing effects on the average age of infant death. Demography. 2022;59(2):587-605. doi: 10.1215/00703370-9779784 [DOI] [PubMed] [Google Scholar]
  • 5.Sosnaud B. Decomposition analysis of disparities in infant mortality rates across 27 US states. Demogr Res. 2024;50(40):1185-1222. doi: 10.4054/DemRes.2024.50.40 [DOI] [Google Scholar]
  • 6.Kitagawa EM. Components of a difference between two rates. J Am Stat Assoc. 1955;50(272):1168-1194. doi: 10.1080/01621459.1955.10501299 [DOI] [Google Scholar]
  • 7.National Center for Health Statistics . User Guide to the 2022 Period/2021 Cohort Linked Birth/Infant Death Public Use File. National Center for Health Statistics; 2023. https://ftp.cdc.gov/pub/Health_Statistics/NCHS/Dataset_Documentation/DVS/period-cohort-linked/22PE21CO_linkedUG.pdf [Google Scholar]
  • 8.Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol. 1996;87(2):163-168. doi: 10.1016/0029-7844(95)00386-X [DOI] [PubMed] [Google Scholar]
  • 9.R Core Team . R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing; 2025. https://www.R-project.org/ [Google Scholar]
  • 10.Park RY, Bilinski A, Parks RM, Flaxman S. Trends in maternal, fetal, and infant mortality in the US, 2000-2023. JAMA Pediatr. 2025;179(7):765-772. doi: 10.1001/jamapediatrics.2025.0440 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Shapiro-Mendoza CK, Cottengim CR, Erck Lambert AB, Geary S, Barfield WD, Parks SE. New data on sudden unexpected infant deaths by cause and race and ethnicity: 2015–2022. Pediatrics. 2025;156(1):e2024069558. doi: 10.1542/peds.2024-069558 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1.

eAppendix 1. Data exclusions

eTable 1. Record exclusions, 2003-2024

eFigure 1. Change in each cause’s contribution to the Black-White IMR gap between 2020 and 2024 (deaths per 1,000 live births)

eFigure 2. Rate effect (within-gestational-age mortality risk) contributions to the Black-White gap by the three ICD-10 components of Sudden Unexpected Infant Death, 2003-2024

eTable 2. Per-gestational-age decomposition of the Black-White infant mortality for 2003.

eTable 3. Per-gestational-age decomposition of the Black-White infant mortality for 2024

eAppendix 2. Supplemental methods details

eAppendix 3. Sensitivity analyses

eTable 4. Comparison of 7-category (main) and 16-category gestational-age decomposition, 2024

eAppendix 4. Reproducible code and data

jamanetwopen-e2635321-s001.pdf (479KB, pdf)
Supplement 2.

Data Sharing Statement

jamanetwopen-e2635321-s002.pdf (16KB, pdf)

Articles from JAMA Network Open are provided here courtesy of American Medical Association

RESOURCES