Racial Disparities in Infant Mortality Remain but are Decreasing in Arkansas and the US, 1995-2022

The Infant mortality rate or IMR is routinely expressed as deaths per 1,000 live births up through the first 12 months of life.  Worldwide, the IMR has been improving for the last 75 years, from about 125/1,000 to 25/1,000, [1] but it differs widely from one country to another.  According to the World Health Organization, it ranges from under 2/1,000 in countries like San Marino, Estonia, Japan, Norway, Singapore, Slovenia, Finland, and Belarus, to over 60/1,000 in Mali, the Democratic Republic of the Congo, Niger, Guinea, South Sudan, Somalia, Nigeria, Central African Republic, and Sierra Leone [2].  Improvements in the IMR are largely attributed to improved access to care, implementation of infant vaccination programs, and better nutrition.  Generally speaking, the IMR is lower in wealthier countries, but there are numerous exceptions [3]. 

In the United States, per the Centers for Disease Control and Prevention (CDC), the IMR in 2022 was 5.6 deaths per 1,000 live births, and not surprisingly, the rates differed as much as 3-fold by state [4]. The highest rate was 9.1/1,000 in Mississippi and the lowest was 3.3/1,000 in Massachusetts.  Arkansas has the 3^rd highest infant mortality in the US at 7.7/1,000 according to the CDC [4]. 

It has been well described that there are differences of IMR in different socioeconomic and racial subgroups in the US and that addressing social determinants of health are key in reducing disparities in IMR [5]. We performed this analysis using data which are publicly available from the CDC to demonstrate the changes of IMR over time in the US and Arkansas and to identify subgroups where opportunities for improvement are most evident.

Data for analysis were assembled from the CDC WONDER (Wide-ranging Online Data for Epidemiological Research) data repository (available online at cdc.wonder.gov) [6]. At that location there are 5 separate queryable infant mortality databases which in sum cover the years from 1995-2022.  Numbers of infant deaths (as defined by a death in the 12 months of life) and total births were assembled for every 5-year maternal age range and racial strata for each year in the databases and concatenated into a single analytic database covering 1995-2022.  Stratum-specific death rates per 1,000 births were calculated in Excel.  Excel was also used to generate tables and figures and to calculate linear regression statistics and P-values. 

Due to confidentiality concerns relating to small numbers, age-specific data for races other than African American or White were almost entirely suppressed in single year data.  As a result, limited analysis is presented for races other than African American and White.  Further, because data for African American women <= 19 or over 35 years of age were mostly suppressed for the same reasons, age groups were lumped into <=24, 25-29, and 30+ years of age to allow direct comparison. The WONDER data set also suppresses other likely important variables relating to socioeconomic status such as income, maternal education, marital status, gestational month of prenatal care onset, and infant birthweight. 

During 1995-2022, racial designations varied in the 5 CDC Wonder datasets.  To make them directly comparable, persons listed as Chinese, Filipino, and Other Asian race as coded in 1995-2002 and 2019-2022 were recoded as Asian or Pacific Islander for this analysis.

Figure 1 depicts the IMRs for persons of White and African American race for the US and Arkansas from 1995-2022.  All groups depicted have linear trends that are decreasing during the time period (slopes less than 0 and P values < 0.05).

 Figure 1: Infant Mortality rates by Race in Arkansas and the US, 1995-2022 

The slope of the decrease in IMR among African Americans in Arkansas (-0.071 deaths per 1,000 births per year, 95% CI (-.0078--0.135)) is smaller than the slope of the decrease in the IMR among African Americans in the US (-0.176 deaths per 1,000 births per year, 95% CI (-0.158--0.196)).  Further, it seems that in the most recent data from 2014-2022 AR rates seem to diverge from US rates and may even be increasing but this result is not significant (P=0.27). 

Among Whites, the slope of the decrease in IMR is not statistically different between AR and the US when the entire timeframe is examined, but Whites in Arkansas has IMRs on average 1.6 per 1,000 births higher than Whites in the US.  Further it appears that the decreasing trend of IMR among Whites in AR may be slowing in the most recent data from 2012-2022. 

Figure 2 depicts race specific IMRs from Arkansas versus the US by race.  Arkansas has higher IMRs than the U.S. in every racial category.  All of these differences in race-specific IMRs between Arkansas and US are highly statistically significant (P < 0.05).  The largest absolute difference is in Whites (6.59 vs 4.65 deaths per 1,000 births).  African Americans and American Indian/Alaskan Natives in Arkansas have higher IMRs than the other races depicted in the figure (P < 0.05).

Figure 2: Infant Mortality rates by Race, Arkansas versus the US 2017-2022. 

Figure 3 depicts IMRs by race and age of mother in Arkansas.  Over the entire study period from 1995-2022, African American women had an IMR of 12.5 deaths/1,000 infants and White women had an IMR of 6.6 deaths/1,000 infants (P for difference in means < 0.000001).

 Figure 3: Infant Mortality Rates by Race in Arkansas 1995-2022. 

For African American women, IMRs do not appear to differ by age, (12.3 for women <=24 years old, 12.7 for women 25-29 years old, and 12.5 for women 30 years of age or older) (P>0.47 in all cases).  As for time trends in African American women aged 30 years or more, their IMR is decreasing 0.12 per year from 1995-2022 (P=0.02).  In African American women aged 25-29 years we observed a decrease of 0.09 per year, but it didn’t quite meet significance criteria (P=0.055).  In African American women of younger ages, there is no significant trend in IMR (P=0.70).  

In contrast, for White women in Arkansas, it is clear that women <= 24 years old have IMRs that are higher than women 25-29 years old (8.2 versus 5.4 deaths per 1,000 infants) and that women 25-29 years old have lower IMRs than older white women (5.4 versus 6.0 deaths per 1,000 infants) (P < 0.004 in all cases).  Time trends in White women demonstrate that IMRs are decreasing at 0.05 deaths per 1,000 infants per year in women < 24 (P=0.006), increasing in women 25-29 by 0.04 deaths per 1,000 infants per year (P=0.006), and not changing in older white women (P=0.99).

This is a secondary data analysis of a public dataset.  As a result, it is subject to the typical limitations associated with public datasets including missing and most notably suppressed data.  For example, this analysis is unable to perform subgroup analyses in non-White or non-African American races and omits important variables relating to socioeconomic status such as income, maternal education, marital status, gestational month of prenatal care onset, and infant birthweight as these variables are suppressed from subgroup analyses for confidentiality concerns. 

This analysis contributes to the literature by focusing on Arkansas, a state challenged by IMRs higher than the national average, and comparing it to the rest of the US to identify demographic factors that may be associated with the excess.  It incorporates the most recently available data on the CDC website. 

On a positive note, this analysis clearly demonstrates decreasing IMRs over time among White and African American infants in AR and the US.  However, using US rates as a baseline, there clearly is room for improvement in the IMRs in all racial and age groups in Arkansas.  This analysis may further help policy makers help refine and target programs to reduce disparities where the opportunities to improve IMRs in certain age and racial subgroups are most evident.  Subgroups with higher IMRs in Arkansas include African American women in general (and in particular those under 30 years of age), American Indian/Native Alaskan women, Native Hawaiian/Pacific Islander women, and White women aged 24-29 years of age. 

Even though we identify racial subgroups at higher risk, it is important to note that the majority of risk for infant mortality is not truly genetic.  Rather, increased risk is likely conferred primarily through a variety of social determinants of health which comingle with race such as poverty, education, access to care, gestational month of onset of prenatal care, nutrition, smoking and pre-gestational diabetes [7,8].  Our study, due to data suppression rules is not able to address these issues directly. 

Interested investigators could improve on our analysis with a collaborative effort of multiple birth defect registries.  These registries would have the aforementioned socioeconomic status related data that this analysis lacks and could provide greater granularity in characterizing the problem.  Addressing the problem, however, is far more complicated, and will take time.  It no doubt requires a system level approach focusing on numerous social determinants of health as the paper by Reno and Hyder identify [5].

This report echoes previous reports demonstrating dramatic differences in IMR by race with African American infants having rates roughly twice as high as White infants.  Compared to White children, American Indian/Alaskan Native and Native Hawaiian/Pacific Islander children also have higher IMRs.  Asian children have lower IMRs than Whites. Arkansas has higher IMRs in all racial subgroups than the US.

1. https://www.macrotrends.net/global-metrics/countries/wld/world/infant-mortality-rate.
2. https://www.who.int/data/gho/data/indicators/indicator-details/GHO/infant-mortality-rate-(probability-of-dying-between-birth-and-age-1-per-1000-live-births).
3. Tresserras R, Canela J, Alvarez J, Sentis J, Salleras L (1992) Infant mortality, per capita income, and adult illiteracy: an ecological approach. Am J Public Health 82: 435-438.
4. https://www.cdc.gov/maternal-infant-health/infant-mortality/index.html..
5. Reno R, Hyder A (2018) The Evidence Base for Social Determinants of Health as Risk Factors for Infant Mortality: A Systematic Scoping Review. Journal of Health Care for the Poor and Underserved29: 1188-1208. 
6. https://wonder.cdc.gov/lbd.html.
7. Orischak M, Fru DN, Kelly E, DeFranco EA (2022) Social determinants of infant mortality amongst births to non-Hispanic Black women. American Journal of Obstetrics & Gynecology 226: S706.
8. Kim D, Saada A (2013) The social determinants of infant mortality and birth outcomes in Western developed nations: a cross-country systematic review. Int J Environ Res Public Health 10: 2296-2335.