Variations in Health and Health Behaviors by Nativity Among Pregnant Black Women in Philadelphia

Abstract

Objectives. We compared health behaviors and health outcomes among US-born, African-born, and Caribbean-born pregnant Black women and examined whether sociodemographic and psychosocial characteristics explained differences among these population subgroups.

Methods. We analyzed data from a prospective cohort study conducted in Philadelphia, Pennsylvania, with a series of nested logistic regression models predicting tobacco, alcohol, and marijuana use and measures of physical and mental health.

Results. Foreign-born Black women were significantly less likely to engage in substance use and had better self-rated physical and mental health than did native-born Black women. These findings were largely unchanged by adjustment for sociodemographic and psychosocial characteristics. The foreign-born advantage varied by place of birth: it was somewhat stronger for African-born women than for Caribbean-born women.

Conclusions. Further studies are needed to gain a better understanding of the role of immigrant selectivity and other characteristics that contribute to more favorable health behaviors and health outcomes among foreign-born Blacks than among native-born Blacks in the United States.

Studies examining health differences among immigrant subgroups and their native-born counterparts have largely focused on Hispanics. This literature consistently shows that Hispanics born outside the United States have lower mortality rates and better health and reproductive outcomes than do US-born Hispanics.^1–3 These studies also reveal that the protective effect of foreign-born status varies by country of origin.^4,5 For example, Cho et al. found that Mexican immigrants had better self-reported health status, fewer activity limitations, and fewer sick days confined to bed than did persons from Cuba and Central America,⁴ and Hummer et al. reported a similar variation for infant mortality rates.³

Far less is known about the health of foreign-born Black immigrants, who make up an expanding proportion of US immigrants. For example, in 1960 fewer than 1% of Black US residents were foreign born; by 2005 this figure increased to 8%.^6,7 Studying health differentials among native-born and foreign-born Blacks may shed light on factors that contribute to racial health disparities in the United States.

Previous studies showed that health behaviors, health status, and reproductive outcomes were more favorable among foreign-born Blacks than among native-born Blacks.^3,8–10 However, only a handful of studies have examined health status¹¹ or birth outcomes^12–15 among foreign-born Blacks by region of birth. These studies found that foreign-born Black women, whether from the Caribbean¹⁴ or Africa,¹² were less likely than US-born Black women to have low-birth-weight infants. A recent study of 2000 vital records for New York, New York, observed that although the risk of low birth weight was lower among infants of foreign-born Black mothers than among infants of native-born Black mothers, that risk varied by the foreign-born mothers' place of birth: infants born to women from the Dominican Republic had the lowest risk, and infants born to women from Haiti had the highest risk.¹⁵ Proposed explanations for more favorable birth outcomes and better health status among the foreign-born include selective migration, greater social support, and fewer adverse health behaviors.^{4,5,9,13,14,16–18}

We examined the role of nativity in health behaviors and health status among pregnant Black women in Philadelphia, Pennsylvania. We compared health behaviors and status among Black women born in the United States, the Caribbean, and Africa. Our data, collected through extensive face-to-face interviews, allowed us to examine whether individual-level sociodemographic and psychosocial characteristics explained differences in behavior and health among Black women by region of birth.

METHODS

Our data were from the baseline survey of a prospective cohort study, funded by the Centers for Disease Control and Prevention, investigating the role of stress and neighborhood context on racial/ethnic differences in pregnancy outcomes in Philadelphia. We recruited the study population between February 1999 and September 2004 from 12 community-based health centers and hospital-based clinics serving low-income residents. These facilities provided obstetric care to women regardless of ability to pay.

During the study period, 9343 discrete women entered prenatal care, and 8960 (95.9%) were approached by study interviewers and screened for eligibility. Women were considered eligible if they spoke either English or Spanish, lived within the City of Philadelphia, and were at less than 21 weeks of gestation at the time of their first prenatal care visit. Of the 8960 women screened, 5461 (60.9%) were considered eligible. The most common reasons for ineligibility were gestation of 21 weeks or more (31.4%) and language barrier (40.9%). Among those women excluded because of language, approximately 20% were Black.

Of the 5461 eligible women, 4908 (89.9%) provided written informed consent and were then interviewed by experienced community-based female interviewers, who used standardized questionnaires to obtain information about the women's sociodemographic characteristics, health behaviors, health, psychosocial characteristics, social support, and housing.^19,20 Fewer than 1% of these women did not complete the survey.

We limited the sample to women who self-identified as Black or African American and who were born in the United States, the Caribbean, or Africa (n = 3177). We further excluded 76 women for whom information was missing on health status and health behaviors (n = 40) and the explanatory variables of interest (n = 36). The final sample comprised 3101 women.

Health Status and Health Behaviors

Our dependent variables consisted of 2 measures of physical health, 2 measures of mental health, and 3 health behaviors. The 2 physical health measures were self-rated physical health (poor and fair versus good, very good, excellent) and obesity (body mass index [BMI; defined as weight in kilograms divided by height in meters squared] ≥ 30). We calculated BMI from self-reported prepregnancy height and weight. Some evidence shows that self-reported height and weight are reasonable proxies for measured height and weight,²¹ although little is known about the validity of self-reported height and weight among foreign-born populations.

Our mental health indicators were self-reported mental health status (poor and fair versus good, very good, and excellent) and depressive symptoms, measured by the Center for Epidemiologic Studies Depression Scale (CES-D), an instrument widely used to assess depressive symptoms in community samples.²² The CES-D scores range from 0 to 60; higher scores represent greater emotional distress, with standard cutpoints of 16 or higher (possibly depressed) and 23 or higher (probably depressed). We used the higher cutpoint, which is more predictive of depressive disorder and has been recommended for use among pregnant women.^22–24

Health behaviors were evaluated by self-reports of smoking after the woman found out that she was pregnant, a variable that is highly correlated with smoking in the year prior to the pregnancy, and of alcohol and marijuana use in the 12 months prior to the pregnancy.

Individual-Level Characteristics

We assessed several individual-level characteristics known to predict health status and health behaviors. We coded our key explanatory variable, place of birth, as women born in the United States, women born in Africa, and women born in the Caribbean. We added maternal age as a continuous variable and marital status as single, married–cohabiting, and other (divorced, widowed, and separated). Two variables evaluated access to instrumental social support (material support) and exposure to stressful life circumstances. Access to material support was constructed as a dichotomous measure from the following questions: Do you know someone who (1) would take you to the doctor? (2) would loan you $100? (3) would help you with daily chores if you were sick? (4) you could talk to about problems? and (5) would watch your children? A yes response to all 5 questions indicated a high level of material support.

Our composite stress exposure measure incorporated stressful exposures in housing (type, stability, disrepair, crowding, utility cutoffs), interpersonal conflict (arguments, harassments, threats to personal safety), material hardship (food, clothing, medical care), and neighborhood danger (street violence, drug traffic, and so forth). We summed negative indicators in each domain and created a total stress score ranging from 0 to 14.¹⁹

Our 2 indicators of socioeconomic status (SES) were educational attainment (< high school, high school–GED, and > high school) and material hardship. Respondents were asked 5 questions covering how often they were unable to make monthly payments on bills and could not afford food, leisure activities, medical care, and clothing. Responses to each question ranged from 1 (never) to 5 (very often). We summed these responses and coded the hardship scale as 1 (< 7), 2 (7–10), 3 (11–15), and 4 (≥ 16). This material hardship score better captured variation in economic resources than did household income, which was missing for a larger fraction of the sample.

Statistical Analysis

We estimated a series of multilevel logistic regression models to predict our outcomes of interest. Model 1 included only place of birth and age. In model 2, we added other individual-level characteristics to investigate whether differences in these outcomes by place of birth could be explained by individual-level attributes. We used Stata, version 10,²⁵ for all analyses.

Although we present the distribution of individual-level attributes by region of birth (Table 1), our statistical test was derived from a comparison of all native-born women with all foreign-born women, not with foreign-born subgroups.

TABLE 1.

Baseline Characteristics of Native-Born and Foreign-Born Pregnant Black Women: Philadelphia, PA, 1999–2004

	Total (n = 3101), % or Mean (SD)	US Born (n = 2816), % or Mean (SD)	Foreign Born (n = 285), % or Mean (SD)	Pa	Foreign Born
					African Born (n = 106), % or Mean (SD)	Caribbean Born (n = 179), % or Mean (SD)
Self-reported physical health fair or poor	10.7	11.3	4.6	.001	2.9	5.6
Obeseb	25.1	25.8	18.6	.027	17.9	19.0
Self-reported mental health fair or poor	18.4	19.5	7.0	.001	6.6	7.3
Depressive symptomsc	22.0	21.7	25.6	.125	22.6	27.4
Smokingd	20.0	21.7	3.9	.001	1.9	5.0
Alcohol usee	34.3	35.6	22.1	.001	14.2	26.8
Marijuana usee	21.0	22.7	4.2	.001	0.9	6.2
Age, y	24.2 (5.8)	23.9 (5.7)	27.5 (6.4)	.001	28.8 (5.9)	26.7 (6.6)
Marital status				.001
Single	81.5	84.2	55.1		46.2	60.3
Married/cohabiting	16.5	14.0	41.0		50.0	35.8
Other	2.0	1.8	3.9		3.8	3.9
Educational attainment				.001
< High school	33.4	34.7	21.1		12.3	26.3
High school	47.2	47.6	42.8		39.6	44.7
> High school	19.4	17.7	36.1		48.1	29.0
Material hardship indexf				.001
Low	34.9	35.8	25.6		32.1	21.8
Moderate	26.7	27.1	22.9		22.6	22.9
High	21.2	20.9	24.4		21.7	25.1
Very high	17.2	16.2	28.0		23.6	30.2
Material support lowg	16.7	15.9	24.2	.001	24.5	24.0
Objective stress scoreh	3.8 (2.4)	3.9 (2.4)	3.0 (2.3)	.001	2.7 (2.0)	3.1 (2.5)

^aDifference between US-born and foreign-born women.

^bBody mass index of 30 kg/m² or higher, calculated from self-reported prepregnancy height and weight.

^cScore of 23 or higher on Center for Epidemiologic Studies Depression Scale.

^dDuring pregnancy.

^eIn 12 months before pregnancy.

^fDerived from 5 questions (range = 1–5) regarding inability to pay monthly bills and to afford food, leisure activities, medical care, and clothing (low = < 7; moderate = 7–10; high = 11–15; very high ≥ 16).

^gDichotomous measure based on a yes (high) or no (low) response to all of the following: Do you know someone who (1) would take you to the doctor? (2) would loan you $100? (3) would help you with daily chores if you were sick? (4) you could talk to about problems? and (5) would watch your children?

^hComposite measure assessing housing, interpersonal conflict, material hardship, and neighborhood danger. Range = 0–13.

RESULTS

Descriptive characteristics for the full sample and by place of birth are shown in Table 1. About 9% of the women (285 of 3101) were foreign born. Among the foreign born, about one third were born in Africa (n = 106) and two thirds in the Caribbean (n = 179; Table 1).

We detected clear differences in health behaviors, health status, and individual-level explanatory variables between the foreign-born and native-born women. The foreign-born women reported less substance use (alcohol, cigarettes, and marijuana) and better overall health status (less obesity and better self-rated physical and mental health; Table 1). The foreign-born women were also more likely to be married or cohabiting, were better educated, and reported fewer stressful exposures. By contrast, the foreign-born women reported more material hardship and less access to material social support than did the native-born women, resulting in a mixed profile of risk. We observed no significant difference in the probability of scoring high on the CES-D.

Comparisons by region of birth among foreign-born participants revealed some interesting variations. Women born in Africa had somewhat more favorable risk profiles than did women born in the Caribbean. However, the differences between the 2 groups on most variables did not reach statistical significance at the 5% level, except for recent use of alcohol and marijuana, age, and educational attainment (results not shown).

Although women in our sample were somewhat younger and more disadvantaged than Black women who gave birth in Philadelphia in 2000 (as reported in vital statistics birth records), the relative sociodemographic differences between native-born and foreign-born Black women were similar in the 2 data sets. For example, in the 2000 city birth records for Philadelphia, the mean age of native-born Black women was 24.9 years and of foreign-born Black women, 28.6 years. In our sample, the respective mean ages were 23.9 years and 27.5 years. In 2000, about 19% of the native-born Black women and 53% of the foreign-born Black women were married, according to vital statistics data. In our sample, 14% of the native-born Black women and 41% of the foreign-born Black women were either married or cohabiting.

Educational differences exhibited a similar pattern. About 27% of the native-born Black women and 15% of the foreign-born Black women who gave birth in Philadelphia in 2000 had less than a high school education; the respective figures in our sample were 35% and 21%. We observed a similar pattern in relative differences in smoking rates, with 13% of the native-born Black women and 2% of the foreign-born Black women reporting that they smoked during pregnancy in the 2000 birth record data; 22% and 4%, respectively, of our sample reported smoking (tabulations by the authors). Thus, our sample reflected nativity differences that were observed among all births by Black women in Philadelphia during the study period.

Health Behaviors

Foreign-born Black women reported significantly lower rates of tobacco, alcohol, and marijuana use than did native-born Black women. As shown in Table 2, the lower rate of adverse health behaviors among foreign-born Black women persisted after adjustment for individual-level attributes in our nested logistic regression models. For African-born women, the odds ratios (ORs) adjusted for age only were 0.05 for smoking, 0.23 for alcohol use, and 0.04 for marijuana use. The ORs for Caribbean-born women were 0.14, 0.57, and 0.24, respectively. These results indicate that Black women born in Africa were less likely than Caribbean-born women to engage in negative health behaviors, although limitations in sample size prohibit drawing firm conclusions.

TABLE 2.

Logistic Regression Models Predicting Health Behaviors Among Native-Born and Foreign-Born Pregnant Black Women: Philadelphia, PA, 1999–2004

	Smokinga		Alcohol Useb		Marijuana Useb
	Model 1,c OR (95% CI)	Model 2,d OR (95% CI)	Model 1,c OR (95% CI)	Model 2,d OR (95% CI)	Model 1,c OR (95% CI)	Model 2,d OR (95% CI)
Place of birth
United States (Ref)	1.00	1.00	1.00	1.00	1.00	1.00
Africa	0.05 (0.01, 0.19)	0.06 (0.02, 0.27)	0.23 (0.13, 0.40)	0.25 (0.14, 0.44)	0.04 (0.01, 0.26)	0.05 (0.01, 0.37)
Caribbean	0.14 (0.07, 0.29)	0.14 (0.07, 0.29)	0.57 (0.39, 0.79)	0.60 (0.42, 0.86)	0.24 (0.13, 0.44)	0.26 (0.14, 0.49)
Marital status
Single (Ref)	1.00	1.00	1.00	1.00	1.00	1.00
Married/cohabiting	0.50 (0.38, 0.66)	0.75 (0.57, 1.00)	0.85 (0.69, 1.04)	0.97 (0.78, 1.21)	0.55 (0.41, 0.73)	0.72 (0.54, 0.96)
Other	0.84 (0.47, 1.52)	1.07 (0.56, 2.02)	0.93 (0.55, 1.58)	1.01 (0.59, 1.73)	0.68 (0.33, 1.40)	0.77 (0.36, 1.64)
Educational attainment
< High school (Ref)	1.00	1.00	1.00	1.00	1.00	1.00
High school	0.18 (0.13, 0.25)	0.26 (0.19, 0.36)	1.03 (0.83, 1.28)	1.32 (1.04, 1.66)	0.39 (0.29, 0.52)	0.58 (0.43, 0.78)
> High school	0.40 (0.33, 0.49)	0.48 (0.39, 0.59)	0.99 (0.83, 1.18)	1.10 (0.92, 1.32)	0.66 (0.55, 0.80)	0.82 (0.67, 1.01)
Material hardship indexe
Low (Ref)	1.00	1.00	1.00	1.00	1.00	1.00
Moderate	1.61 (1.27, 2.06)	1.39 (1.07, 1.81)	1.88 (1.55, 2.29)	1.67 (1.36, 2.05)	1.91 (1.51, 2.42)	1.60 (1.25, 2.05)
High	1.73 (1.34, 2.22)	1.18 (0.89, 1.57)	1.87 (1.52, 2.31)	1.54 (1.22, 1.93)	2.19 (1.70, 2.80)	1.46 (1.11, 1.92)
Very high	2.00 (1.54, 2.60)	1.18 (0.86, 1.62)	1.73 (1.38, 2.17)	1.33 (1.03, 1.72)	2.35 (1.81, 3.05)	1.35 (1.00, 1.83)
Material supportf
High (Ref)	1.00	1.00	1.00	1.00	1.00	1.00
Low	2.12 (1.71, 2.64)	1.54 (1.21, 1.97)	0.99 (0.81, 1.21)	0.81 (0.65, 1.01)	1.77 (1.43, 2.19)	1.24 (0.98, 1.57)
Objective stress scoreg	1.26 (1.21, 1.31)	1.16 (1.11, 1.22)	1.13 (1.09, 1.16)	1.11 (1.07, 1.16)	1.27 (1.22, 1.31)	1.20 (1.15, 1.25)

Note. CI = confidence interval; OR = odds ratio.

^aDuring pregnancy.

^bIn 12 months before pregnancy.

^cBivariate associations, adjusted for age only.

^dAdjusted for all individual-level explanatory variables shown.

^eDerived from 5 questions (range = 1–5) regarding inability to pay monthly bills and to afford food, leisure activities, medical care, and clothing; (low = < 7; moderate = 7–10; high = 11–15; very high ≥ 16).

^fDichotomous measure based on a yes (high) or no (low) response to all of the following: Do you know someone who (1) would take you to the doctor? (2) would loan you $100? (3) would help you with daily chores if you were sick? (4) you could talk to about problems? and (5) would watch your children?

^gComposite measure assessing housing, interpersonal conflict, material hardship, and neighborhood danger.

In model 2, we introduced socioeconomic and psychosocial characteristics. The ORs for African and Caribbean women relative to native-born women were only slightly attenuated. Thus, observed socioeconomic and psychosocial characteristics did little to explain the birth-origin differences in health behaviors.

Many of the other individual attributes were also significant predictors of these behaviors. For example, women with higher levels of education and high levels of perceived material support were significantly less likely to report use of tobacco and marijuana. Higher levels of exposure to objective stress and material hardship increased the risk of these 2 health behaviors. These findings did not persist for alcohol use after adjustment, except for our objective stress score and material hardship.

Health Status

Table 3 presents the results from the nested logistic regression models for physical and mental health outcomes. Foreign-born status was clearly protective for our 2 physical health measures. With adjustment for age only, the ORs were 0.23 for poor–fair self-rated physical health and 0.45 for obesity for African-born women, and 0.46 and 0.55, respectively, for Caribbean-born women. As with health behaviors, the foreign-born advantage was somewhat more pronounced for Black women born in Africa than for those born in the Caribbean. These differences persisted nearly unchanged with the addition of controls for individual-level attributes (model 2).

TABLE 3.

Logistic Regression Models Predicting Obesity, Self-Rated Physical and Mental Health, and Depressive Symptoms Among Native-Born and Foreign-Born Pregnant Black Women: Philadelphia, PA, 1999–2004

	Obesea		Self-Rated Physical Health Fair–Poor		Self-Rated Mental Health Fair–Poor		Depressive Symptomsb
	Model 1,c OR (95% CI)	Model 2,d OR (95% CI)	Model 1,c OR (95% CI)	Model 2,d OR (95% CI)	Model 1,c OR (95% CI)	Model 2,d OR (95% CI)	Model 1,c OR (95% CI)	Model 2,d OR (95% CI)
Place of birth
United States (Ref)	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00
Africa	0.45 (0.27, 0.76)	0.41 (0.24, 0.70)	0.23 (0.07, 0.72)	0.27 (0.08, 0.88)	0.27 (0.13, 0.59)	0.38 (0.17, 0.84)	1.04 (0.65, 1.66)	1.52 (0.92, 2.52)
Caribbean	0.55 (0.37, 0.82)	0.50 (0.33, 0.74)	0.46 (0.24, 0.89)	0.49 (0.25, 0.95)	0.31 (0.18, 0.55)	0.36 (0.20, 0.65)	1.35 (0.96, 1.90)	1.61 (1.11, 2.33)
Marital status
Single (Ref)	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00
Married/cohabiting	1.04 (0.83, 1.29)	1.12 (0.89, 1.40)	0.89 (0.64, 1.23)	1.12 (0.80, 1.57)	0.73 (0.55, 0.95)	1.00 (0.75, 1.32)	0.67 (0.52, 0.87)	0.73 (0.56, 0.97)
Other	0.71 (0.39, 1.28)	0.72 (0.39, 1.30)	1.26 (0.58, 2.71)	1.42 (0.65, 3.10)	0.95 (0.50, 1.82)	1.13 (0.57, 2.23)	1.19 (0.67, 2.12)	1.29 (0.71, 2.36)
Educational attainment
< High school (Ref)	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00
High school	0.91 (0.71, 1.16)	0.84 (0.65, 1.09)	0.59 (0.41, 0.83)	0.80 (0.55, 1.16)	0.55 (0.42, 0.77)	0.96 (0.71, 1.29)	0.67 (0.52, 0.86)	0.96 (0.73, 1.26)
> High school	1.10 (0.91, 1.33)	1.02 (0.83, 1.24)	0.71 (0.55, 0.92)	0.87 (0.66, 1.13)	0.63 (0.51, 0.77)	0.88 (0.71, 1.10)	0.69 (0.57, 0.84)	0.90 (0.74, 1.11)
Material hardship indexe
Low (Ref)	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00
Moderate	1.03 (0.83, 1.28)	1.05 (0.85, 1.32)	1.60 (1.17, 2.19)	1.36 (0.98, 1.88)	1.46 (1.13, 1.88)	1.10 (0.84, 1.44)	1.61 (1.27, 2.04)	1.25 (0.97, 1.60)
High	1.08 (0.86, 1.35)	1.11 (0.87, 1.42)	1.91 (1.38, 2.64)	1.38 (0.98, 1.96)	1.94 (1.49, 2.51)	1.18 (0.88, 1.57)	2.03 (1.59, 2.60)	1.27 (0.97, 1.66)
Very high	0.96 (0.75, 1.22)	1.03 (0.78, 1.36)	1.92 (1.36, 2.70)	1.25 (0.85, 1.85)	2.62 (2.01, 3.42)	1.36 (1.00, 1.85)	3.19 (2.49, 4.09)	1.68 (1.27, 2.24)
Material supportf
High (Ref)	1.00	1.00	1.00	1.00	1.00	1.00	1.00	1.00
Low	0.85 (0.67, 1.07)	0.92 (0.72, 1.17)	2.01 (1.54, 2.62)	1.60 (1.20, 2.13)	2.56 (2.07, 3.16)	1.76 (1.39, 2.23)	2.39 (1.95, 2.93)	1.60 (1.28, 2.00)
Objective stress scoreg	1.00 (0.97, 1.04)	1.01 (0.97, 1.06)	1.17 (1.12, 1.22)	1.09 (1.03, 1.15)	1.27 (1.22, 1.32)	1.16 (1.11, 1.22)	1.24 (1.20, 1.29)	1.15 (1.10, 1.20)
Health behaviors
Smokingh	0.68 (0.55, 0.84)	0.62 (0.49, 0.78)	1.70 (1.34, 2.25)	1.17 (0.87, 1.56)	2.58 (2.10, 3.17)	1.56 (1.23, 1.97)	1.68 (1.37, 2.05)	1.11 (0.88, 1.40)
Alcohol usei	1.04 (0.88, 1.24)	1.10 (0.91, 1.33)	1.42 (1.13, 1.80)	1.15 (0.89, 1.49)	1.51 (1.25, 1.82)	1.04 (0.83, 1.29)	1.63 (1.37, 1.94)	1.36 (1.11, 1.65)
Marijuana usei	0.83 (0.67, 1.03)	0.83 (0.66, 1.05)	1.68 (1.30, 2.16)	1.14 (0.85, 1.53)	2.41 (1.97, 2.95)	1.50 (1.18, 1.90)	1.89 (1.56, 2.30)	1.25 (0.99, 1.57)

Note. CI = confidence interval; OR = odds ratio.

^aBody mass of index 30 kg/m² or higher, calculated from self-reported prepregnancy height and weight.

^bScore of 23 or higher on Center for Epidemiologic Studies Depression Scale.

^cBivariate associations, adjusted for age only.

^dAdjusted for all individual-level explanatory variables shown.

^eDerived from 5 questions (range = 1–5) regarding inability to pay monthly bills and to afford food, leisure activities, medical care, and clothing (low = < 7; moderate = 7–10; high = 11–15; very high ≥ 16).

^fDichotomous measure based on a yes (high) or no (low) response to all of the following: Do you know someone who (1) would take you to the doctor? (2) would loan you $100? (3) would help you with daily chores if you were sick? (4) you could talk to about problems? and (5) would watch your children?

^gComposite measure assessing housing, interpersonal conflict, material hardship, and neighborhood danger.

^hDuring pregnancy.

ⁱIn 12 months before pregnancy.

Of the other individual-level characteristics, higher levels of material hardship and exposure to objective stress and a lower level of perceived material support increased the risk of poor and fair physical health. Each of the adverse health behaviors was also significantly associated with poor and fair self-rated health in the unadjusted model 1, but none remained significant after adjustment for maternal economic and psychosocial characteristics. By contrast, only smoking, in addition to nativity, was a significant predictor of obesity—: women who reported smoking during pregnancy were significantly less likely to be obese.

The results for our 2 mental health measures were contradictory: foreign-born Black women were not significantly less likely than native-born Black women to report high levels of depressive symptoms in the CES-D. However, foreign-born women were significantly less likely to report their mental health status as poor. We speculate that this inconsistency occurred because each measure taps into different dimensions of mental health or because women from other cultures may interpret the questions on the CES-D or the self-rated mental health measure differently from women born in the United States.

As with the 2 physical health measures, foreign-born Black women rated their mental health significantly better than did native-born Black women. The crude OR for self-rated poor and fair mental health was 0.27 for African-born women and 0.31 for Caribbean-born women. These associations remained essentially unchanged after adjustment for individual-level characteristics. In addition to nativity, a low level of material support and higher levels of stress exposure significantly increased the risk of poor and fair self-reported mental health, as did the use of tobacco and marijuana.

By contrast, higher levels of education and being married or cohabiting significantly lowered the risk of depressive symptoms, after adjustment for other socioeconomic attributes, psychosocial factors, and health behaviors. High levels of material hardship, low levels of perceived material support, and higher levels of stress exposure and substance use increased the risk of depressive symptoms. However, the correlations we observed with health behaviors did not establish the direction of causality. It may be that poor mental health or high levels of depressive symptoms predicted smoking and marijuana use rather than the reverse.

DISCUSSION

Our results are consistent with previous findings that foreign-born Black women are significantly less likely than native-born Black women to engage in negative health behaviors and more likely than were native-born Black women to report good physical and mental health. The nativity advantage we observed remained largely unchanged after adjustment for sociodemographic and psychosocial characteristics. Furthermore, we found that the foreign-born advantage varied by place of birth: the advantage was somewhat stronger for African-born women than for Caribbean-born women. Other studies have documented similar gradients in hypertension,²⁶ functional and self-care limitations,²⁷ and self-rated health.¹¹

Theories of immigrant health advantage have focused on the roles of social support, socioeconomic status, and selective migration. It is commonly hypothesized that the cultural context of immigrants differs from that of their US-born counterparts in promoting norms and values that strengthen familial ties and social support networks and protect migrants from engaging in negative health behaviors.^4,16 This explanation is consistent with evidence showing that the foreign born are less likely than the native born to smoke²⁸ and that recent immigrants have more favorable BMIs than do their native-born counterparts.²⁹ These patterns have also been documented among foreign-born Black US residents, who are less likely than native-born Blacks to smoke and to be obese.^28–30 Our results are similar to the results of those previous studies.

However, less attention has been paid to the possibility that immigration itself can disrupt existing social ties in countries of origin and can lead to diminished access to emotional and instrumental support at least temporarily, which in turn may increase levels of stress and social isolation until new social networks are activated.¹⁶ Our results are consistent with the possibility that immigration may disrupt social networks. Foreign-born Black women were significantly more likely than native-born Black women to report that they had reduced access to material support. Access to material support in turn was consistently associated with improved health behaviors and self-reported physical and mental health status. However, adjustment for this difference did not explain the better health behaviors and physical and mental health status of immigrants.

Numerous studies have documented significant associations between various measures of SES, health behaviors, and reproductive and other health outcomes.^2,3,31 In our sample, the foreign-born Black women were better educated than the native-born Black women, but foreign-born women reported greater material hardship. Although both of these SES indicators were significant predictors of substance use, neither explained the differences between native-born and foreign-born women in these behaviors. For physical and mental health outcomes, both SES measures were only weakly associated with physical and mental health status in this relatively young sample.

One of the most common theoretical explanations for better health among foreign-born women than among native-born US women is selective migration. Immigrants do not represent a random sample of the population in their country of origin; instead they consist of individuals who are positively selected on the basis of health and other observed and unobserved characteristics. Although it was not possible to assess the extent of health selection in our sample, such selection was likely to be present, and may have varied by place of birth.

The consistent advantage we observed of African-born women over Caribbean-born women could have been related to differential patterns of immigrant selectivity. Simply stated, health selectivity may be higher among immigrants from Africa than from the Caribbean. African immigrants are more likely to enter the United States on diversity visas and employment-based preferences, although they are also more likely than were Caribbean immigrants to come as refugees.^6,32 In our sample, approximately half of the African-born women came from Liberia, and among this group, many may have come as refugees. By contrast, Caribbean immigrants are more likely to obtain visas on the basis of family ties, for example, as immediate relatives of US residents.^6,11 In our sample, almost 40% of the Caribbean-born women were born in Jamaica, and about 30% were born in Haiti. Unfortunately, we did not have information on visa type for the immigrant women in our sample. Further studies are needed to gain a better understanding of the role of immigrant selectivity and other characteristics that contribute to more favorable health behaviors and health outcomes among foreign-born Black women than among native-born Black women.

Although health selectivity may account for the regional differences in health and health behaviors among Black women, these differences may also arise from different perceptions of discrimination and marginalization. Prior to immigration, Caribbean women may have closer contact with US society via family ties and frequent travel back and forth between the United States and their country of origin. It follows, therefore, that Caribbean-born women may be more perceptive about acts of discrimination than are African-born women, who have had less exposure to US society because of distance. In essence, the Caribbean-born women may be more likely than African-born women to internalize experiences of discrimination, thus leading to poorer health and health behaviors. In keeping with this theory, the US-born Black women may be most likely to experience the cumulative consequences of discrimination and marginalization over their lifetime, leading to high levels of stress and its health-eroding consequences.^33,34

Limitations

Measures of health behaviors and health status, including BMI, were derived from self-reports. Normative perceptions of physical and mental health and the reporting of height and weight may vary by country of origin. Such reporting differences may also influence reporting of substance use, especially the use of illegal substances such as marijuana.

Several authors have proposed that over time, acculturation to the US environment leads to negative changes in health behaviors and diet such that immigrant health status and behaviors begin to resemble those of the native born.^29,35,36 We were unable to test this hypothesis in our sample. Finally, our sample was limited to foreign-born and native-born Black women living in an inner city, and thus our results are most relevant for similar contexts.

Conclusions

We found significant nativity differentials in health behaviors and health status among Black women who sought prenatal care in community-based health centers and hospital-based clinics in Philadelphia from 1999–2004. African-born and Caribbean-born Black women were significantly less likely than were native-born Black women to smoke, drink alcohol, and use marijuana, and foreign-born women reported better health status than did native-born women. Among the foreign born, women born in Africa were somewhat healthier than women born in the Caribbean and were less likely to engage in negative health behaviors. However, sample sizes prevented firm conclusions.

The foreign-born advantage was not explained by observed sociodemographic or psychosocial characteristics. Further studies are needed to gain a better understanding of the role various hypothesized mechanisms, including immigrant selectivity, play in producing the more favorable health behaviors and health outcomes among foreign-born Blacks than among native-born Blacks in the United States.