Skip to main content

Advertisement

Social determinants of health and disparities in prenatal care utilization during the Great Recession period 2005-2010

Article metrics

Abstract

Background

Early, regular prenatal care utilization is an important strategy for improving maternal and infant health outcomes. The purpose of this study is to better understand contributing factors to disparate prenatal care utilization outcomes among women of different racial/ethnic and social status groups before, during, and after the Great Recession (December 2007–June 2009).

Methods

Data from 678,235 Washington (WA) and Florida (FL) birth certificates were linked to community and state characteristic data to carry out cross-sectional pooled time series analyses with institutional review board approval for human subjects’ research. Predictors of on-time as compared to late or non-entry to prenatal care utilization (late/no prenatal care utilization) were identified and compared among pregnant women. Also explored was a simulated triadic relationship among time (within recession-related periods), social characteristics, and prenatal care utilization by clustering individual predictors into three scenarios representing low, average, and high degrees of social disadvantage.

Results

Individual and community indicators of need (e.g., maternal Medicaid enrollment, unemployment rate) increased during the Recession. Associations between late/no prenatal care utilization and individual-level characteristics (including disparate associations among race/ethnicity groups) did not shift greatly with young maternal age and having less than a high school education remaining the largest contributors to late/no prenatal care utilization. In contrast, individual maternal enrollment in a supplemental nutrition program for women, infants, and children (WIC) exhibited a protective association against late/no prenatal care utilization. The magnitude of association between community-level partisan voting patterns and expenditures on some maternal child health programs increased in non-beneficial directions. Simulated scenarios show a high combined impact on prenatal care utilization among women who have multiple disadvantages.

Conclusions

Our findings provide a compelling picture of the important roles that individual characteristics—particularly low education and young age—play in late/no prenatal care utilization among pregnant women. Targeted outreach to individuals with high disadvantage characteristics, particularly those with multiple disadvantages, may help to increase first trimester entry to utilization of prenatal care. Finally, WIC may have played a valuable role in reducing late/no prenatal care utilization, and its effectiveness during the Great Recession as a policy-based approach to reducing late/no prenatal care utilization should be further explored.

Background

During the Great Recession in the United States (U.S.), indicators of need—such as the percent of children in poverty, unemployment rates, and consumer distress—increased [1,2,3]. Historically, Black and Hispanic populations have had higher rates of unemployment compared to Whites, and during and after the Great Recession these disparate rates were maintained [4]. All ethnic groups experienced increases in unemployment during the Great Recession, but Blacks continued to have the highest unemployment rates, Whites had the lowest, and Hispanics fell between the two [4]. At the same time, community-level safety net resources, including many maternal and child health programs provided by local community health departments (LHDs), experienced cuts which may have contributed to increased difficulties among pregnant women in accessing prenatal care —particularly during the earlier phases of the Great Recession and before federal stimulus funds became available [5,6,7].

Early (within the first trimester) and regular PNC is known to be an important strategy for improving health outcomes for mothers and infants [8, 9]. Improved birth weight and decreased risk of preterm delivery are two of the most significant benefits of early and ongoing utilization of prenatal care [9, 10]. Infants born to women who do not receive prenatal care are three times more likely to have a low birth weight and five times more likely to die than infants born to mothers who receive prenatal care [11, 12]. Improved infant health outcomes associated with early utilization of prenatal care have both quality of life and cost implications. Average medical costs for a premature or low birth weight infant during the first year of life are about $55,393, whereas annual costs for a newborn without complications averages $5085 [13].

Racial/ethnic disparities in timing of entry to prenatal care utilization are well documented and persistent in the U.S.—despite improvements in recent years and national attention to disparity elimination as a primary goal of the U.S. Department of Health and Human Services (DHHS) Healthy People program [8, 14]. Disparities are widely recognized to be complex and multi-faceted at many levels. Their existence ranges from differences that are apparent at the individual level to health outcomes that represent macro-social differences in political ideologies and wealth distribution [15, 16]. Previous research has found that persistent disparities associated with prenatal care utilization are predominantly related to social determinants of health including social circumstances, access to medical care, and behavioral patterns [17].

Changes in individual and community resources during the Great Recession raise questions as to whether existing disparity relationships—defined as differences in rates of early (first trimester) as opposed to late/no entry to prenatal care utilization among different socioeconomic groups—might also be influenced. During a recession in the early 1980’s, Fisher, LoGerfo, and Daling [18] found increases in late entry to prenatal care utilization in Washington (WA) State. In that study, the authors found specific increases among those who resided in low income census tracts (compared to high); however they did not explore differential increases among race/ethnicity groups [18]. Using established methods for analyzing disparities [19,20,21], we also recently found rates of late/no prenatal care utilization increased among some groups during the Great Recession (December 2007–June 2009) in WA and Florida (FL) [22, 23]). For example, our study showed that prior to the Great Recession, 15.3% of White and 20.9% of Black mothers in WA received late or no prenatal care. During the Great Recession, rates of late/no prenatal care utilization increased for both groups—to 17.4 and 28.4%, respectively. The steeper increase among Black mothers yielded a 26.8% increase in disparity in outcomes in relation to White mothers [22]. We have also confirmed the presence of prenatal care utilization outcome disparities in WA and FL (the same study population used in this study) among groups defined by race/ethnicity and other social status characteristics (e.g., education, insurance status, age, marital status) [22, 23].

As a result of these preliminary findings and questions, the purpose of this study was to better understand contributing factors to disparate prenatal care utilization outcomes among women of different racial/ethnic & social status groups before, during, & after the Great Recession (December 2007–June 2009). Our hypothesis was that both individual and social characteristics would play important roles in whether and when pregnant women accessed prenatal care (within the first trimester as opposed to after the first trimester or not at all (late/no prenatal care utilization)) and that relative contributions of community/social characteristics would change during the course of the recession as these inputs varied based on the economy and investments in maternal and child health programs.

Methods

Study design

In this study we assembled and linked a variety of individual and community-level indicators to better understand factors contributing to disparities in timing of entry to prenatal care utilization among women of different racial/ethnic backgrounds and social status groups before, during, and after the Great Recession (2005–2010). Predictors of entry later than first trimester, including non-entry to prenatal care utilization were identified and compared using a cross-sectional pooled time series design. Particular attention was focused on indicators that may have changed during the Recession, such as unemployment rate, partisan voting patterns, or per capita local health department (LHD) expenditures on a supplemental nutrition program for women, infants, and children (WIC) and other maternal child health programs (Table 1).

Table 1 Covariates for regression models

Three recession-related time periods were defined as (1) Baseline Period #0 before the Recession (January 2005–March 2007), (2) Recession Period #1 (December 2007–June 2009—as officially defined by the National Bureau of Economic Research) [6], and (3) Recession Period #2 (July 2009–December 2010) [22]. Per this definition, Recession Period #2 encompasses the months and years after the official Recession Period (#1) during which community-level economic indicators such as unemployment continued to be elevated above baseline (Period #0) levels [4, 33, 34].

In a second analytic phase, we examined a simulated triadic relationship among time, degree of social disadvantage, and late/no entry to prenatal care utilization during three recession-related time periods among pregnant women of different race/ethnicity groups to compare predicted probabilities of late/no prenatal care utilization for three representative scenarios of social disadvantage (“high,” “average,” and “low”). The measures and rationale for each of the scenarios were informed by theory and existing research are further described below in “measures” and Table 2.

Table 2 Social Disadvantage Status Characteristic Constellations

Data and study population

De-identified data from all birth certificates from WA and FL for the years 2005–2010 were retrieved through data-sharing agreements with the Departments of Health (DOH) in FL and WA with institutional review board approval from the University of WA and the FL State Department of Health. These states were selected for inclusion as both experienced a tremendous downturn in economic markers during the Great Recession and both had comparable LHD expenditure data available for the study time period [3, 5, 6, 32,33,34]. The LHD and community data derive from publicly available datasets and have been incorporated into recent maternal and child health-focused studies [26, 32]. Individual birth certificates were linked to county/Local Health Jurisdiction (LHJ)/LHD data using maternal county of residence. All data were cross-sectional and secondary.

The study population consisted of 678,235 individual pregnant women having their first singleton live birth (492,691 in FL; 185,544 in WA) who resided in the 102 LHJs in WA and FL. Non-first time births were excluded to reduce the issues of repeated measures if women had more than one birth during the study period as linking of maternal data between years was not possible. Multiple births were also excluded (only singletons were kept) as multiple births are associated with increased risk of preterm birth, low birth weight, and infant mortality. LHJs follow county lines in FL and in WA, and in WA, three LHJs were comprised from multiple counties. The study was limited to women whose infants had complete birth certificate information on race/ethnicity, maternal county of residence, and timing of entry to prenatal care utilization. For all individual level variables included in this analysis missing-ness was less than 1.0% with the exception of payment source for delivery which was 0.51% in FL and 2.70% in WA (1.11% overall) and maternal WIC utilization which was missing 1.23% of the time in FL and 9.34% of the time in WA (overall missing = 3.48%).

Measures

Predictors for the main outcome of entering prenatal care during as compared to after the first trimester of pregnancy (or not at all) were examined. To measure this outcome, a binary variable, based on continuous birth certificate data, was created with “0” indicating those who entered prenatal care during the first trimester and “1” indicating those who entered prenatal care after the first trimester of pregnancy, or who did not utilize prenatal care at all. The authors chose to combine late and non-entry to prenatal care utilization to be parsimonious and to focus the analysis on characteristics of women who entered prenatal care during the first trimester care (the widely accepted standard of care) as compared to those who entered late or not at all. Covariates were selected based on conceptual and previous research linking them to maternal and child health outcomes—individual, community and LHD expenditure measures and state dummy variables were included. Table 1 provides a complete list of these covariates and related literature supporting their incorporation.

To facilitate estimation of combined effects of social disadvantage during the second part of the analysis, individual characteristics found to be related to late/no prenatal care utilization were grouped into scenarios representative of low, average, or high social disadvantage (Table 2) [35]. The authors chose to do this as people have multiple identities and risks [16, 17]. While complex, this step helps to capture the additive (cumulative) impacts of relative advantage or disadvantage. Characteristics representative of an “average” scenario were defined based on majority (modal) population characteristics in the study population. Not all possible characteristics included in scenarios (e.g., maternal age 20–24) as they were defined to represent extreme ends of the social advantage/disadvantage spectrum in the United States.

The low social disadvantage scenario was specified with characteristics associated with “best” outcomes in a previous study using similar data [22]. In our regression models these groups were the referents. The average disadvantage scenario was defined based on majority/modal population characteristics. Fewer characteristics were defined for the average scenario as there was not a clear majority with regard to marital status and insurance type at the time of delivery. The high disadvantage scenario was defined as those individual-level characteristics most associated with late/no prenatal care utilization. In this scenario, while maternal age < 14 is the age most highly associated with late/no prenatal care utilization, we substituted maternal age 15–19 as it occurs much more frequently and is also associated with increased risk and poor outcomes.

Analysis

We carried out analyses in two phases (1) regression model specification to identify predictors of late/no entry to prenatal care utilization for each recession-related period; and (2) estimation of predicted probabilities for race/ethnicity groups for the three social disadvantage scenarios (low, average, and high) at Recession Periods #0, #1, and #2.

Phase 1: Regression model specification

Using a pooled cross-sectional time series design, multivariate linear probability regression models (LPMs) were estimated to identify which covariates were predictive of late/no prenatal care utilization for the total study population (WA + FL) during Recession Periods #0, #1, and #2. LPMs were chosen to allow for more readily interpretable results of both analytic phases; results from logistic regression models are similar and are provided in Appendix 2 in Table 8. Models were adjusted first for individual, then community, and finally LHD expenditure covariates described above and in Table 1. We conducted all analyses using STATA version 12 [36]. Clustering of individuals within LHJs was addressed using robust standard errors (SEs), correcting for effects of geographically clustered [37] and for the inherent heteroscedasticity in LPMs. Entry to prenatal care utilization by definition occurs at some point during the nine-month course of pregnancy—because of this proximate relationship, no time lags were introduced into the economic data. A value of P < .05 was used to establish statistical significance. Model specification included running models with each of the available LHD expenditure variables. Final preferred model selection was informed by comparing results of Akaike Information Criteria (AIC) and Bayesian Information Criteria (BIC) tests for specified models with the lowest AIC/BIC selected [38].

Phase 2: Calculation of predicted probabilities for three social disadvantage scenarios

Following regression modeling, in the second analytic phase, we estimated the predicted probability an individual has of late/no prenatal care utilization given a set of fixed characteristics using the post-estimation margins command in Stata [37, 39]. Values for individual covariate characteristics were set for each of the three social disadvantage scenarios—low, average, and high—and predicted probabilities of late/no prenatal care utilization were calculated for White, Black, and Hispanic subpopulations. This approach facilitated practical interpretation of the combined effects of social status characteristics that tend to cluster together along the range of social advantage/disadvantage. In these calculations, non-specified variables were assessed at their actual observed values [37, 39]. Predicted probability of late/no prenatal care utilization was estimated for the total study population as well as for each state by specifying state dummy variables within scenarios.

Results

Profile of women who entered PNC late and summary of economic indicators

The characteristics of the study population are presented in Table 3. Women who entered prenatal care late or not at all (compared to those who entered in the first trimester) were younger (twice as likely to be teenagers), less likely to be married, slightly more likely to be foreign-born, and (of those who could have finished high school) almost twice as likely to have less than a high school (HS) education (9.49% vs. 5.11%). They were also nearly twice as likely to be on Medicaid and had a higher rate of WIC utilization.

Table 3 Comparison of demographic and perinatal characteristics: study population for FL, WA, total study population, and United States (n/% unless otherwise indicated) by timing of entry to prenatal care utilization (for entire study period)a,l

During the study period, unemployment increased dramatically in both states (Table 4). FL unemployment rates more than doubled by Period #1 and then tripled by Period #2 from baseline. In WA, unemployment increased, but not as dramatically—from 5.14% (SD 0.94%) at baseline to 6.61% (SD 2.20%) during Period #1 and to 9.71% (SD 1.54%) during Period #2. WIC enrollments and Medicaid as a proportion of payers also increased in both states, but more in FL than in WA for both indicators. Per capita LHD expenditures varied widely in both states, but mean expenditures had an overall trend toward decreased per capita spending for family planning (FP) and for a composite of maternal/infant/child/adolescent (MICA) service lines [26, 32]. We also combined FP and MICA to create the 2MCH expenditure variable (combined expenditures for two maternal and child health (MCH) services —FP and MICA) in our regression models (Table 1) in both states over the course of the study period. Among LHDs in FL, per capita 2MCH expenditures decreased from $8.79 (SD $5.67) during the baseline period to $8.18 (SD $5.54) during Period #1 and to $7.84 (SD $5.11) during Period #2. In contrast to LHD decreases in 2MCH expenditures, WIC expenditures among LHDs generally increased during the study period in both states—from $4.10 (SD = $1.98) during the baseline period to $4.55 (SD = $2.30) during Period #1 and $5.02 (SD = $2.60) during Period #2.

Table 4 Descriptive summary of economic indicators before and during recession periods in WA and FL (Total Study Population)

Phase 1: Regression models results within and between periods

Table 5 summarizes the results of all final models (for Recession Periods #0, #1, and #2).. Only minor variations in coefficient magnitudes were found among individual-level categorical characteristics within model steps or across study periods. For example, the difference in probability of late/no prenatal care utilization for Black mothers (compared to the White reference group) was positive during all steps and periods and increased only slightly over time (from 0.032 to 0.037). All individual-level coefficients were positive with the exception of maternal WIC enrollment—which exhibited a relatively stable negative coefficient (− 0.010 to − 0.012). The largest magnitude individual-level predictors were young age (age < 14 and to a lesser degree age 15–19) and having less than a HS education. Those aged 14 years and younger had a 0.259 to 0.262 greater probability of late/no prenatal care utilization compared to the referent group (age 30–34), while those age 15–19 had a 0.087 to 0.097 greater probability of late/no prenatal care utilization than the referent group. Women who had less than a HS education had a 0.061 to 0.084 greater probability of late/no prenatal care utilization compared to women with at least some college. Having Medicaid or being uninsured (self-pay) were also significant positive predictors during both Recession Periods #1 and #2, but not during the Baseline Period.

Table 5 Final Late/No Prenatal Care Utilization Linear Regression Models for Baseline, Period 1 and Period 2 (controlled for 102 LHD Clusters)

Three continuous community level variables were significantly associated with late/no prenatal care utilization: (1) per capita MDs (negative coefficient, only significant during the Baseline Period); (2) maternal residence in a high poverty LHJ (negative coefficient, significant during Periods #1 and #2 but not during the Baseline); and (3) percent of LHJ residents voting Republican in a national election (positive coefficient significant during all model steps and time periods—increasing from 0.001 at Baseline to 0.002 during Periods #1 and #2 in the final models). In terms of LHD expenditures, per capita WIC expenditures were negative for late/no prenatal care utilization but not significant at any time period. However, the 2MCH coefficient representing LHD FP and MICA expenditures was positive during each time period (Baseline Period #0 = 0.0012, Period #1 = 0.0019, Period #2 = 0.0025) and significant during Periods #1 and #2. The state dummy variable was not significant.

Phase 2: Predicted probability results and comparisons

Results of predicted probability calculations for each of the three social disadvantage scenarios (low, average, high) and race/ethnicity are summarized in Table 6. The predicted values represent the expected probability or expected percentage of individuals (i.e. 0.033 = 3.3%) experiencing late/no care in each group defined by the scenarios in Table 2 and provide a sense of the levels of late/no care experienced by each category of disadvantage and ethnicity/race. Those with combined social characteristics associated with low social disadvantage would be less likely to enter prenatal care late or not at all (range = 0.033 to 0.076) than those with average social status (range = 0.116 to 0.163) for all race/ethnicity groups at all time periods. Those with characteristics representing a high degree of social disadvantage would be the most likely to enter prenatal care late or not at all for all race/ethnicity groups at all time periods (range = 0.379 to 0.482). Differences between race/ethnicity groups within social disadvantage scenarios were much smaller within as compared to between scenarios (the difference between Hispanics and Whites is non-significant and the difference between Blacks and Whites is significant at about 0.03).

Table 6 Predicted Probability of Late/No Prenatal Care Utilization in Total Study Population for Low, Average, and High Social Status Characteristics

Discussion

During the Great Recession, we found individual and social characteristics to play important roles in whether and when pregnant women accessed prenatal care. Indicators of need (e.g., maternal Medicaid enrollment, unemployment rate) increased during the Recession in both study states. Young maternal age and having less than a HS education were found to be the largest individual-level contributors to late/no prenatal care utilization among pregnant women in WA and FL during all three recession-related periods. Relative contributions of individual-level predictors were found to exhibit minimal variation across time periods. Simulated scenarios show a high combined impact on prenatal care utilization among women who have multiple disadvantages. Associations between community (particularly percent of the community voting Republican) and LHD expenditure variables and late/no prenatal care utilization revealed variation over time (compared to Baseline Period #0) and increases in the non-beneficial directions. In contrast, individual maternal enrollment in a supplemental nutrition program for women, infants, and children (WIC) exhibited a protective association against late/no prenatal care utilization.

Previous research on the effect of recessions and/or unemployment on maternal and child health outcomes has been used to study a variety of populations as well as outcomes. Among studies that specifically address recessions and MCH outcomes, most found recessions (usually measured by time and/or unemployment rate) to be negatively associated with timing of entry to prenatal care and birth weight and positively associated with infant mortality [18, 43,44,45,46,47,48,49,50,51]. Race/ethnicity and other individual level characteristics (i.e. maternal education) were rarely taken into consideration in published studies related to past recessions [46, 52]. This paper adds to this body of research by carrying out analyses during the most recent global recession (The Great Recession). Our finding that rates of late/no prenatal care utilization increased during the Great Recession are consistent with previous research. Our explorations of individual and community level contributors to late/no prenatal care utilization extend this research and help set the stage for future research as to whether targeted outreach to individuals with high disadvantage characteristics, particularly those with multiple disadvantages, may help to increase first trimester entry to prenatal care utilization.

In this study, evidence also emerged that WIC may have contributed to reductions in late/no prenatal care utilization over the course of the included recession periods—even in the face of increasing local need. In addition, WIC may have been more effective at reducing late/no prenatal care utilization than the other maternal and child health safety net programs for which we had LHD expenditure data. This finding suggests that the increased WIC enrollment and related increases in local WIC expenditures observed over the course of the Recession may have been particularly beneficial and protective against late/no prenatal care utilization among disadvantaged populations. WIC was the only safety net program for which both individual and community level data were available. It is possible that more nuanced effects among high-need populations targeted by family planning and/or MICA programs with decreasing expenditures were missed; alternatively, results may reflect the general decline in LHD expenditures. WIC may represent a useful policy-based approach to reducing late/no prenatal care utilization and should be further explored.

Regarding LHD expenditures, our findings are consistent with Bekemeier, Yang, Dunbar, Pantazis, and Grembowski [26] who found (using the same LHD expenditure data) that WIC did and 2MCH did not follow changes in local need during the Recession. In our case, LHD expenditures on WIC services were negatively predictive of late/no prenatal care utilization, but not significant at any point. Our findings related to 2MCH were also consistent with Bekemeier et al. [26]. The coefficient size for 2MCH increased over time and was positive rather than negative as might be expected of a maternal and child health program. When considered from the perspective of a $10 increase in per capita maternal and child health expenditures (which would be unlikely as 2MCH budgets generally decreased during the Recession but is a useful example), the probability of late/no prenatal care utilization increased over the course of the study period from 0.01 (0.001 × 10) to 0.03 (0.003 × 10). This is about the same difference in probability observed between Black and White women. During this same time need increased and 2MCH budgets decreased, indicating that the observed increased association may be related to increases in level of need and LHDs stretched to essentially do ‘more with less’ during this study period [53]. Further exploration would be beneficial to understanding this association.

We also identified partisan voting patterns as playing a predictive role in late/no prenatal care utilization. We had included these variables because of prior work by Oakman, et al. [27]. This may an interesting line of inquiry with ongoing shifts in voting patterns and partisan preferences in the United States and beyond.

In the second analytic phase, innovative use of predicted probability methods clearly demonstrate an increased likelihood of late/no prenatal care utilization among women with higher degrees of social disadvantage (Tables 2 and 6). There was little change in these relationships despite changes in need and resources over the course of the Great Recession. While only small changes in coefficient size of race/ethnicity variables were observed in regression modeling and some covariates consistently contributed to a larger degree than others (e.g., education and age were larger contributors than foreign-born status or marital status), the effects of combined social disadvantage become more readily visible when viewed in terms of predicted probability of late/no prenatal care utilization. In these scenarios disparate relationships in prenatal care utilization among Black versus White race/ethnicity groups were maintained, within each level of social disadvantage--with Whites being least likely and Blacks being most likely to enter prenatal care late or not at all. Hispanics consistently fell between Whites and Blacks, though once individual characteristics were controlled for, the difference between Whites and Hispanics was non-significant in these scenarios. These findings demonstrate the cumulative effects of advantage and disadvantage as described by Braveman et al. [35] and Pearlin et al. [25]. Results also suggest that efforts to reduce late/no prenatal care utilization may need to be tailored to best meet the needs of diverse populations based on individual, intermediate, and community characteristics.

Limitations

There were limitations to this study and some are associated with the review of secondary data (missing or inaccurate). First, while we limited analysis to first-time mothers with singleton births (to reduce issues of repeated measures and increased infant health risks associated with multiple births), generalizability of our results may be limited. In particular, due to the nature of the dataset we were unable to fully address intermediate factors (e.g. distance from a health facility) which undoubtedly play a role in access to prenatal care utilization. Second, we focused the analysis on a binary variable (first trimester entry to prenatal care utilization versus late/no prenatal care utilization) instead of breaking prenatal care utilization into multiple categories. We chose this approach in the interest of parsimony since our analyses focused on differentiating characteristics of women who entered prenatal care utilization during the first trimester care from those who entered late or not at all. This binary approach may have underestimated the impacts of later stages or non-entry to prenatal care utilization and more refined measures of care should be explored in future studies. Third, WA and FL both had heavy economic downturns during the Great Recession and lumping them in the modeling may not have captured key differences or differential impacts within states. To allow for consideration of individual states’ results, we included state-only models for reference in Appendix 3 in Table 9 and Appendix 4 in Table 10. While no significant state-level differences were identified in the models of the total population, demographic differences with WA and FL may have influenced state-level model results. Third, we used 2008 presidential voting data for both Recession Periods #1 and #2, and there may be better measures that would more effectively describe the differences in policy-making than what the partisan voting covariate identifies. Finally, not all WIC expenditures in each state were represented in our models—only those that were expended by LHDs. Some LHJs may have alternative providers of WIC and other maternal and child health services. The non-significant associations that we identified with LHD WIC expenditures may be due in part to this as well as to the fact that WIC is a targeted, need-based program while our study population represented all pregnant women and not only those with need and/or who were eligible.

Conclusions

In this study we found that—while individual and community indicators of need increased during the recession—relative contributions of individual predictors as social determinants of health and disparities remained largely consistent over the course of the Great Recession. Young maternal age and low maternal education were the largest magnitude individual predictors of late/no prenatal care utilization during all three recession-related periods. Community and LHD expenditure variables exhibited greater variation—over time, percent voting Republican and 2MCH were both increasingly associated with late/no prenatal care utilization in a non-beneficial direction, while WIC enrollment at the individual level appears to have been protective against late/no prenatal care utilization. These associations should all be further explored. Clustering of individual predictors into low, average, and high social disadvantage scenarios clearly demonstrated the disparate combined probability of late/no entry to prenatal care, as well as persistent racial/ethnic disparity within each level of social advantage/disadvantage. Our findings provide a compelling rationale for targeted outreach to pregnant women with high disadvantage characteristics—particularly those with low education and young age. WIC may represent an effective policy-based approach to reducing disparities in late/no prenatal care utilization and its effects during the Great Recession should be further explored.

Availability of data and materials

The data that support the findings of this study are available from the Washington State Department of Health and the Florida State Department of Health but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of the State Departments of Health in Washington and Florida.

Abbreviations

2MCH:

combined FP and MICA expenditures

ACA:

Affordable Care Act

AIC:

Akaike information criteria

ARRA:

American Recovery and Reinvestment Act

BIC:

Bayesian information criteria

BW/LBW:

Birth weight/Low birth weight

CBSA:

Core-based statistical area

CHAMPUS:

Civilian Health and Medical Program of the Uniform Services

CONSORT:

Consolidated Standards of Reporting Trials

DFB:

Difference from best

DHHS:

Department of Health and Human Services

DOH:

Department of Health

FL:

State of Florida

FM:

Family medicine

FP:

Family planning

FY:

Fiscal year

GA:

Gestational age

GED:

General education diploma

GP:

General practitioner

GR:

Great Recession (December 2007–June 2009)

HP:

Healthy People

HS:

High school

IID:

Increase in disparity

IM:

Infant mortality

IRB:

Institutional Review Board

LBW:

Low birth weight (< 2500 g)

LHD:

Local health department

LHJ:

Local health jurisdiction

LPM:

Linear probability (regression) model

MCH:

Maternal/Child health

MD:

Medical doctor

MICA:

Maternal, infant, child, and adolescent (service line composite LHD expenditures)

MICH:

Maternal, infant, and child health

NBER:

National Bureau of Economic Research

PHAST:

Public Health Activities and Services Tracking Study

PNC:

Prenatal care

PTB32:

Preterm birth < 32 weeks

PTB37:

Preterm birth < 37 weeks

SD:

Standard deviation

SES:

Socioeconomic status

SNAP:

Supplemental Nutrition Assistance Program

TANF:

Temporary Aid for Needy Families

U.S.:

United States

USDA:

United States Department of Agriculture

VLBW:

Very low birth weight (< 1500 g)

WA:

State of Washington

WIC:

Special Supplemental Nutrition Program for Women, Infants, and Children

References

  1. 1.

    Bitler M, Hoynes H, Kuku E. Child poverty, the great recession, and the social safety net in the United States. J Policy Anal Manage. 2011;36(2):358–89.

  2. 2.

    Pilkauskas NV, Currie J, Garfinkel I. The great recession, public transfers, and material hardship. Soc Serv Rev. 2012;86(3):401–27.

  3. 3.

    CredAbility Consumer Distress Index. CredAbility Consumer Distress Index. (n.d.). https://fred.stlouisfed.org/series/CCDIHBQ156N. Accessed 6 Nov 2018.

  4. 4.

    Bureau of Labor Statistics. The Recession of 2007-2009: BLS spotlight on statistics. 2012. https://www.bls.gov/spotlight/2012/recession/. Accessed 26 August 2014

  5. 5.

    National Association of City and County Health Officials. Local health department job losses and program cuts: findings from the January 2011 survey and 2010 National Profile Study (research brief). Washington, D.C: National Association of City and County Health Officials; 2011. https://www.marlerblog.com/uploads/image/ESS-Research-Brief-June-2011-revised-1.pdf. Accessed 27 May 2014

  6. 6.

    Business Cycle Dating Committee, National Bureau of Economic Research. Business Cycle Dating Committee, National Bureau of Economic Research. 2010. http://www.nber.org/cycles/sept2010.html. Accessed 14 December 2018.

  7. 7.

    Willard R, Shah GH, Leep C, Ku L. Impact of the 2008-2010 economic recession on local health departments. J Public Health Manag Pract. 2012;18(2):106–14. https://doi.org/10.1097/PHH.0b013e3182461cf2.

  8. 8.

    Howell EA. Reducing disparities in severe maternal morbidity and mortality. Clin Obstet Gynecol. 2018;61(2):387–99. https://doi.org/10.1097/GRF.0000000000000349.

  9. 9.

    Till SR, Everetts D, Haas DM. Incentives for increasing prenatal care use by women in order to improve maternal and neonatal outcomes. Cochrane Database Syst Rev. 2015;(12):CD009916. https://doi.org/10.1002/14651858.CD009916.pub2.

  10. 10.

    United States Department of Health and Human Services (USDHHS). Preventing infant mortality fact sheet. 2006. https://minorityhealth.hhs.gov/omh/browse.aspx?lvl=4&lvlid=23. Accessed 14 December 2018.

  11. 11.

    Child Trends Databank. Late or no prenatal care. Child trends databank. 2014. https://www.childtrends.org/?indicators=late-or-no-prenatal-care. Accessed 25 August 2014.

  12. 12.

    Office on Women’s Health, U. S. D. of H. and H. S. Prenatal care fact sheet. 2012. https://www.womenshealth.gov/a-z-topics/prenatal-care. Accessed 14 December 2018.

  13. 13.

    March of Dimes. Premature babies cost employers $12.7 billion annually | March of Dimes. 2014. https://www.marchofdimes.org/news/premature-babies-cost-employers-127-billion-annually.aspx. Accessed 25 August 2014.

  14. 14.

    Healthy People 2010. Healthy People 2010: Objectives for improving health. (n.d.). http://healthypeople.gov/2010/. Accessed 27 May 2014.

  15. 15.

    Healthy People 2020. Healthy People 2020: About disparities. (n.d.). https://www.healthypeople.gov/2020/about/foundation-health-measures/Disparities. Accessed 14 December 2018.

  16. 16.

    Schulz A, Northridge ME. Social determinants of health: implications for environmental health promotion. Health Educ Behav. 2004;31(4):455–71. https://doi.org/10.1177/1090198104265598.

  17. 17.

    Bryant AS, Worjoloh A, Caughey AB, Washington AE. Racial/ethnic disparities in obstetric outcomes and care: prevalence and determinants. Am J Obstet Gynecol. 2010;202(4):335–43. https://doi.org/10.1016/j.ajog.2009.10.864.

  18. 18.

    Fisher ES, LoGerfo JP, Daling JR. Prenatal care and pregnancy outcomes during the recession: the Washington state experience. Am J Public Health. 1985;75(8):866–9.

  19. 19.

    Keppel K, Pamuk E, Lynch J, Carter-Pokras O, Kim I, Mays V, et al. Methodological issues in measuring health disparities. Vital Health Stat. 2005;2(141):1–16.

  20. 20.

    Keppel KG, Pearcy JN, Heron MP. Is there progress toward eliminating racial/ethnic disparities in the leading causes of death? Public Health Rep. 2010;125(5):689–97. https://doi.org/10.1177/003335491012500511.

  21. 21.

    Keppel KG, Pearcy JN, Klein RJ. Measuring progress in Healthy People 2010. Healthy People 2010 Stat Notes. 2004;(25):1–16 From the Centers for Disease Control and Prevention/National Center for Health Statistics. https://www.ncbi.nlm.nih.gov/pubmed/15446274. Accessed 6 Nov 2018.

  22. 22.

    Blakeney E. The Great Recession and health disparities: A study of maternal and child health outcomes in Washington and Florida. University of Washington; 2014. Available from https://search.proquest.com/pqdtglobal/docview/1652875181/3BB1917E2B734BFBPQ/30?accountid=14784

  23. 23.

    Abu-Rish Blakeney, E. (2015). Relationship between the Great Recession and widening maternal and child health disparities: Findings from Washington and Florida. Presented at the 2015 APHA Annual Meeting & Expo (Oct. 31 - Nov. 4, 2015), APHA. Retrieved from https://apha.confex.com/apha/143am/webprogram/Paper336158.h

  24. 24.

    Hing E, Hsiao C-J. State variability in supply of office-based primary care providers: United States, 2012. NCHS Data Brief. 2014;(151):1–8 PMID: 24813076.

  25. 25.

    Pearlin LI, Schieman S, Fazio EM, Meersman SC. Stress, health, and the life course: some conceptual perspectives. J Health Soc Behav. 2005;46(2):205–19.

  26. 26.

    Bekemeier B, Yang Y, Dunbar M, Pantazis A, Grembowski D. Targeted health department expenditures benefit birth outcomes at the county level. Am J Prev Med. 2014;46(6):569–77. https://doi.org/10.1016/j.amepre.2014.01.023.

  27. 27.

    Oakman TS, Blendon RJ, Campbell AL, Zaslavsky AM, Benson JM. A partisan divide on the uninsured. Health Aff. 2010;29(4):706–11. https://doi.org/10.1377/hlthaff.2009.1019.

  28. 28.

    US Census Bureau, D.I.D. Small Area Income & Poverty Estimates (SAIPE) main page. (n.d.). https://www.census.gov/programs-surveys/saipe.html. Accessed 2 September 2014.

  29. 29.

    Kelleher C, Timoney A, Friel S, McKeown D. Indicators of deprivation, voting patterns, and health status at area level in the Republic of Ireland. J Epidemiol Community Health. 2002;56(1):36–44. https://doi.org/10.1136/jech.56.1.36.

  30. 30.

    Kondrichin SV, Lester D. Voting conservative and mortality. Percept Mot Skills. 1998;87(2):466. https://doi.org/10.2466/pms.1998.87.2.466.

  31. 31.

    Burkey, ML. Gini coefficients for the 2000 Census. 2006. http://main.burkeyacademy.com/home/gini-coefficients. Accessed 2 September 2014.

  32. 32.

    Bekemeier B, Dunbar M, Bryan M, Morris ME. Local health departments and specific maternal and child health expenditures: relationships between spending and need. J Public Health Manag Prac. 2012;18(6):615–22. https://doi.org/10.1097/PHH.0b013e31825d9764.

  33. 33.

    Federal Reserve Economic Data. FRED Economic Data. (n.d.). http://research.stlouisfed.org/fred2/. Accessed 27 May 2014.

  34. 34.

    Washington State Board of Health. Health impact reviews—Proposed cuts to health care and human services programs. 2009. http://sboh.wa.gov/OurWork/HealthImpactReviews/Archive. Accessed 6 November 2018.

  35. 35.

    Braveman PA, Heck K, Egerter S, Marchi KS, Dominguez TP, Cubbin, et al. The role of socioeconomic factors in black-white disparities in preterm birth. Am J Public Health. 2014:e1–9. https://doi.org/10.2105/AJPH.2014.302008.

  36. 36.

    StataCorp. Stata statistical software: Release 12. 2011. https://www.stata.com/support/faqs/resources/citing-software-documentation-faqs/. Accessed 14 December 2018.

  37. 37.

    Stata.com. Stata 13 help for margins. (n.d.). https://www.stata.com/help.cgi?margins. Accessed 14 December 2018.

  38. 38.

    Dziak J, Coffman D, Lanza S, & Li R. Sensitivity and specificity of information criteria. The Methodology Center. 2012. https://methodology.psu.edu/media/techreports/12-119.pdf. Accessed 26 August 2014.

  39. 39.

    Williams R. Using the margins command to estimate and interpret adjusted predictions and marginal effects. Stata J. 2012;12(2):308–31.

  40. 40.

    Division of Vital Statistics. Vital statistics data available online. 2014. https://www.cdc.gov/nchs/data_access/Vitalstatsonline.htm. Accessed 14 December 2018.

  41. 41.

    Livingston G, & Cohn D. The new demography of American motherhood. Pew Research Center: Social & Demographic Trends. 2010. http://www.pewsocialtrends.org/2010/05/06/the-new-demography-of-american-motherhood/. Accessed 27 May 2014.

  42. 42.

    Markus AR, Andres E, West KD, Garro N, Pellegrini C. Medicaid covered births, 2008 through 2010, in the context of the implementation of health reform. Womens Health Issues. 2013;23(5):e273–80. https://doi.org/10.1016/j.whi.2013.06.006.

  43. 43.

    Catalano R, Goldman-Mellor S, Saxton K, Margerison-Zilko C, Subbaraman M, LeWinn K, Anderson E. The health effects of economic decline. Annu Rev Public Health. 2011;32:431–50. https://doi.org/10.1146/annurev-publhealth-031210-101146.

  44. 44.

    Dooley D, Prause J. Birth weight and mothers’ adverse employment change. J Health Soc Behav. 2005;46(2) (June):141–55. https://doi.org/10.1177/002214650504600202.

  45. 45.

    Zilko CEM. Economic contraction and birth outcomes: an integrative review. Hum Reprod Update. 2010;16(4). https://doi.org/10.1093/humupd/dmp059.

  46. 46.

    Oberg CN. The great Recession’s impact on children. Matern Child Health J. 2011;15(5) (July):553–4. https://doi.org/10.1007/s10995-011-0807-8.

  47. 47.

    Catalano R, Hansen HT, Hartig T. The ecological effect of unemployment on the incidence of very low birthweight in Norway and Sweden. J Health Soc Behav. 1999;40(4) (December):422–8. https://doi.org/10.2307/2676334.

  48. 48.

    Catalano R, Serxner S. The effect of ambient threats to employment on low birthweight. J Health Soc Behav. 1992;33(4) (December):363–77. https://doi.org/10.2307/2137314.

  49. 49.

    Dehejia R, Lleras-Muney A. Booms, busts, and babies’ health. Q J Econ. 2004;119(3):1091–130. https://doi.org/10.1162/0033553041502216.

  50. 50.

    Gerdtham UG, Ruhm CJ. Deaths rise in good economic times: evidence from the OECD. Econ Hum Biol. 2006;4(3):298–316. https://doi.org/10.1016/j.ehb.2006.04.001.

  51. 51.

    Neumayer E. Recessions lower (some) mortality rates: evidence from Germany. Soc Sci Med. 2004;58(6):1037–47. https://doi.org/10.1016/j.socscimed.2004.02.018.

  52. 52.

    Margerison-Zilko CE, Catalano R, Hubbard A, Ahern J. Maternal exposure to unexpected economic contraction and birth weight for gestational age. Epidemiology. 2011;22(6):855–8. https://doi.org/10.1097/EDE.0b013e318230a66e.

  53. 53.

    Bekemeier B, Chen ALT, Kawakyu N, Yang Y. Local public health resource allocation: Limited choices and strategic decisions. Am J Prev Med. 2013;45(6):769–75. https://doi.org/10.1016/j.amepre.2013.08.009.

Download references

Acknowledgements

The authors would like to acknowledge valuable feedback provided by Andrew Dannenberg and Chris Adolph on this manuscript as it was being developed. We also appreciate the assistance of Erich von Abele for his professional writing services who assisted with final copyediting and formatting.

Funding

Funding for this study was provided by University of Washington’s ITHS TL1 Predoctoral Training Grant (TL1 TR000422), a University of Washington School of Nursing Hahn Endowed Fellowship, and NIH National Heart, Lung, and Blood Institute K12 (#5K12HL137940) as part of the UW Implementation Science Training Program. The first author received all funding for this study. TL1 funding was used to support the authors time, to obtain research supplies (e.g., statistical programs, data retrieval fees), and research related travel. Fellowship funds were used to support costs of research supplies after the completion of TL1 funding. K12 funds were used to support the authors time during manuscript finalization and revision.

Author information

EB formulated the study questions, retrieved, cleaned, and analyzed the data and interpreted results. JH provided expert statistical assistance. All authors (EB, JH, BB, BZ) made substantial contributions to the conception and interpretation of the data as well as read and approved the final manuscript.

Authors’ information

EB is a Research Assistant Professor at the University of Washington School of Nursing. She has a bachelor’s degree in Politics and Environmental Studies (Whitman College), a bachelor’s degree in Nursing (Johns Hopkins), a master’s degree in International Education (New York University), and a PhD in Nursing Science with a focus on social science statistics (University of Washington). Dr. Blakeney led this project as part of her doctoral studies (during and on the Great Recession) and has extensive training in health services research methods. JH is a Sociology Professor with deep expertise in research methods and statistics at the University of Washington in the Department of Sociology. He has spent the past twenty-five years studying sociology of health over the life course—often focusing on the role of family, peer group, and neighborhood context on various outcomes. BB is a Professor in the School of Nursing at the University of Washington. Her research is focused on public health systems with a particular interest in the structures and practices of state and local health departments in relation to health outcomes and reducing disparities. BZ is Professor in the School of Nursing at the University of Washington. Her research is focused on collaborative approaches to improving quality and safety of patient care within healthcare systems.

Correspondence to Erin L. Blakeney.

Ethics declarations

Ethics approval and consent to participate

This project was approved by the University of Washington Human Subjects Division (HSD #42509), the Florida State Department of Health, and the Washington State Department of Health.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendices

Appendix 1

Table 7 Predicted Probability of Late/No PNC by State and Total Study Population for Low, Average, and High Social Status Characteristics

Appendix 2

Table 8 Logit Model Results of Late/No PNC in Total Study Population (WA + FL)

Appendix 3

Table 9 WA-only Regression Model Results

Appendix 4

Table 10 FL-only Regression Model Results

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Blakeney, E.L., Herting, J.R., Bekemeier, B. et al. Social determinants of health and disparities in prenatal care utilization during the Great Recession period 2005-2010. BMC Pregnancy Childbirth 19, 390 (2019) doi:10.1186/s12884-019-2486-1

Download citation

Keywords

  • WIC
  • Prenatal care utilization
  • Great recession
  • Disparities
  • Partisan voting patterns
  • Social behavior