Neonatal Outcomes for Immigrant vs. Native-Born ... - Springer Link

Matern Child Health J (2011) 15:269–279 DOI 10.1007/s10995-010-0612-9

Neonatal Outcomes for Immigrant vs. Native-Born Mothers in Taiwan: An Epidemiological Paradox Sudha Xirasagar • Jung-Chung Fu • Jihong Liu • Janice C. Probst • Duey-Perng Lin

Published online: 30 April 2010 Ó Springer Science+Business Media, LLC 2010

Abstract In Taiwan, immigrant women by marriage face social discrimination due to widespread impressions that they give birth to low birth weight, high-risk, high cost babies due to their lower socioeconomic status, shorter stature, and lower pre-pregnant weight than native-born Taiwanese women. This study compared crude and adjusted birth outcomes of immigrant mothers (Chinese and Vietnamese) relative to native-born Taiwanese, and tested for the phenomenon of an epidemiological paradox of favorable neonatal outcomes among immigrants. Data from patient charts of all singleton live births during 2002–2007, weighing C500 and \4,000 g, and C20 weeks gestational age at a regional hospital in Kaohsiung, Taiwan, were used. Multiple regression analysis was used to test the hypothesis controlling for maternal characteristics (demographics, national origin, obstetric and prenatal factors) and neonatal

S. Xirasagar (&) J. C. Probst Department of Health Services Policy and Management, Arnold School of Public Health, University of South Carolina, 800 Sumter Street, Columbia, SC 29208, USA e-mail: [email protected] J. C. Probst e-mail: [email protected] J.-C. Fu D.-P. Lin Department of Obstetrics and Gynecology, Kaohsiung Municipal United Hospital, Kaohsiung, Taiwan ROC e-mail: [email protected] D.-P. Lin e-mail: [email protected] J. Liu Department of Epidemiology and Biostatistics, Arnold School of Public Health, University of South Carolina, Columbia, SC, USA e-mail: [email protected]

characteristics (birth weight, gestational age). Of 3,267 births satisfying the inclusion criteria, 19.0% were to Chinese and Vietnamese mothers. Crude birth weight was lowest for Taiwanese mothers, who also had the highest rate of preterm delivery (\37 weeks). The adjusted birth weight for Chinese and Vietnamese mothers was 87.7 and 74.7 g higher, respectively than native-born Taiwanese (both P \ 0.001). Chinese and Vietnamese mothers also had lower odds of preterm birth (ORs 0.46 and 0.47, respectively). Findings support paradoxically better neonatal outcomes among Chinese and Vietnamese immigrant mothers in Taiwan. Findings can be used to initiate public education to reverse the widespread negative perceptions and attitudes towards immigrant spouses in Taiwan. Keywords Birth weight Preterm delivery Immigrant birth outcomes Epidemiological paradox

Introduction With rapid economic development and westernization, Taiwan has experienced a steep fertility decline to 1.07 per woman in 2008 [1]. This demographic change has occurred concomitant with rapid economic growth, rapid increases in women’s educational levels, [2] and a redefinition of women’s self-identity and socio-cultural roles, with a widespread psychosocial shift in women’s attitudes to marriage, a change that has significantly lagged among men. Thus, there is a significant unmet demand for spouses among Taiwanese men, particularly men in farming and low-income occupations. With Taiwan being a relatively developed country in the region, it is the immigration destination of choice for neighboring countries, both economic migration and by marriage. Marriages and births

123

270

involving overseas women have increased dramatically, particularly mainland Chinese and Vietnamese: 13.8% of marriages [3] and 13% of live births in 2006 (compared to 5.12% in 1998) [4]. Documenting immigrant mothers’ birth outcomes is a public health priority to ensure evidencebased policies and health promotion efforts, and to meet the needs of culturally and linguistically diverse immigrant populations [5]. Our study is also motivated by social justice concerns. There is a widespread attitude of discrimination both within households and in the larger society toward immigrant wives, [6] particularly toward Vietnamese. The latter are generally shorter and thinner than Taiwanese, widely attributed to poor nutrition and inferior ethnicity, and assumed to result in low birth weight, high-risk babies requiring significant expenditure of resources. There are no documented studies on public discrimination. However, enough credible reports exist that prompted a US State Department human rights citation regarding the social treatment of immigrant wives in Taiwan, calling upon the government to address the issue [6]. Our study represents a step forward in that direction, seeking to document scientifically valid evidence based on empirical data on a robust sample of immigrant and native-born Taiwanese neonates. Our study controls for the documented biological and socio-economic variables driving neonatal outcomes. Systematic research can provide policy makers the evidence base needed to educate the public to counter affective discriminatory attitudes. Nationwide crude rates show better birth outcomes for immigrant relative to native Taiwanese mothers: low birth weight rates of 5 and 6.9%, and preterm birth rates of 6.7 and 9.3%, respectively [7]. Taiwanese mothers are generally older and at higher risk of hypertension and diabetes. Counterbalancing these risk factors are immigrant mothers’ risks due to lower education and SES [8]. SES negatively impacts birth weight [9, 10]. Therefore, adjusting for confounding variables can unveil the true differences in neonatal outcomes. In addition to the domestic public health and social purposes of the study, our work fills an international knowledge gap regarding the ‘‘epidemiological paradox,’’ a frequent finding that some immigrant groups have better neonatal outcomes despite lower SES in many countries [11]. In the US, better neonatal outcomes are documented among immigrants (rates of infant mortality, low birthweight, and preterm birth) [11, 12], However, outcomes vary by the mother’s birth country, with notably inconsistent associations of SES. Mexican-born women with lower SES have better birth outcomes than US-born Mexicanorigin women [13–16] but foreign-born East–Asian and Asian Indian mothers, despite their higher average SES than the general US population experience higher rates of

123

Matern Child Health J (2011) 15:269–279

low birth weight and preterm birth than their US-born counterparts, [17]. This study investigated the neonatal outcomes of immigrant versus native-born Taiwanese mothers. Most documented studies are based on secondary data sources such as birth certificate or medical claims data which lack data on key risk factors, such as prenatal care, medical conditions, and behavioral factors (smoking and alcohol). This study used primary data on these variables extracted from patient records, and therefore makes a significant scientific contribution to the international literature on the ‘‘epidemiological paradox.’’

Methods We conducted a retrospective, hospital-based clinicalepidemiological study of all births that took place at the Kaohsiung Municipal Hospital (where two of the authors are employed) between January 1, 2002 and December 31, 2007. Kaohsiung county is a relatively less developed region in southern Taiwan with disproportionately greater population engaged in farming and low SES occupations and disproportionately more immigrant mothers. The period was selected because marriages with foreign brides accelerated since 2000, and the 6-year period accommodated our estimated sample needs for adequate statistical power. Sample inclusion criteria were, singleton live births, birth weight C500 and\4,000 g, gestational age C20 weeks, and maternal country of birth being Taiwan, Vietnam or mainland China. Of total 3,401 live births during the study period, 3,286 (96.6%) satisfied the study inclusion criteria, of which 19 were excluded due to chart error (case recorded as primiparous with history of cesarean delivery). The final analytic sample consisted of 3,267 births. Chart review was conducted by a specialist ob/gyn assisted by a bachelorsqualified nurse trained in chart review, and able to translate into English any relevant nursing notes that are occasionally documented in Mandarin. Data were extracted from patient charts on neonatal birth weight and gestational age at delivery, on maternal demographic, medical, behavioral, and obstetric variables and paternal age. Maternal demographics include age (\20, 20–35, [35 years), height (cm), pregravid weight (kg), net weight gain during pregnancy (kg), country of birth (Taiwan, Vietnam, and China), education (below high school, high school graduate, and [high school), marital status (single or currently married), parity (primi or multiparous), smoking during pregnancy, and alcohol consumption. Per documented literature, we defined low birth weight \2,500 g and preterm birth \37 completed gestational weeks [18]. Gestational age was determined by the last menstrual period evaluated with sonographic fetal size, the


latter superseding LMP in discrepant cases [19]. Pregnancy diabetes included preexisting diabetes and gestational diabetes (per National Diabetes Data Group guidelines using 50 and 100 g loading glucose tolerance tests [20, 21]. Hypertensive disorder was recorded if two blood pressure readings at least 6 h apart were greater than 140/90 mmHg, either systolic or diastolic, regardless of pre-pregnant or intra-pregnant onset [22]. Net weight gain is the difference between pregnancy weight gain and neonatal birth weight [23]. Maternal birth history (parity and history of cesarean) was coded into three mutually exclusive categories, primiparous, multiparous with no history of cesarean delivery, and multiparous with history of cesarean. Paternal age was the only paternal variable. Adequacy of prenatal care was categorized into four groups based on a modified Kotelchuck index calculation [24], the ratio of actual to expected number of prenatal visits for gestational age per American College of Obstetrics and Gynecology recommendations, without the Kotelchuk adjustment for month of initializing prenatal care (data not available). The four categories of prenatal care adequacy are: inadequate \50% of expected visits, intermediate 50–79%, adequate 80–109%, and adequate plus C110% [25]. Data on parity, pre-pregnant weight, educational attainment, smoking during pregnancy, and alcohol use were self-reported. Marital status (single or currently married) and birth place were sourced from official personal identification documents. Single includes never married, divorced, widowed, or cohabiting women. The key neonatal outcomes of interest are birth weight, odds of low birth weight, and odds of preterm birth. The key research objective was to explore associations between maternal national origin and the neonatal outcomes of birth weight, odds of low birth weight, and odds of pre-term delivery. Bivariate statistics (chi-square/Student t tests) were examined. Multiple regression analyses were used to compare adjusted birth weights by ethnic origin, and logistic regression to study preterm birth and low birth weight. Analyses were performed with SPSS version 15, with statistical significance set at the 0.05 level. The study was approved by the Institutional Review Boards of the Kaohsiung Municipal United Hospital and the University of South Carolina.

Results Maternal Demographics and Risk Factors Table 1 shows the distribution of study cases by demographic characteristics, medical risk factors and prenatal care. Taiwanese were the largest group, 81.0%, followed

271

by mainland Chinese 10.4%, and Vietnamese 8.6%. Both Chinese and Vietnamese mothers are over-represented in our sample compared to the national distribution, 5.1 and 5.8%, respectively among total 392,246 births in Taiwan between January 2006 and December 2007. (These rates are calculated from administrative claims data procured from the National Health Insurance Research Institute for a different study). The mean pre-pregnant weight (57.1 kg) was highest for Taiwanese. Maternal height of Taiwanese mothers (159.4 cm) was similar to Chinese but greater than Vietnamese. Mean maternal age was 29.5 years (±5), with Taiwanese mothers being significantly older. Immigrant mothers were less likely to be single than Taiwanese, although the majority (92.7%) were married. Taiwanese mothers show higher educational levels: 51.7% have a college degree, compared to Chinese, 12.6%, and Vietnamese, 3.2%. Vietnamese have the highest proportion with less than high school education. Taiwanese mothers have the highest rates of smoking and history of cesarean delivery. The distribution of adequacy of prenatal care was similar among the three ethnic groups. Neonatal Outcomes and Maternal Complications Table 2 shows the distribution of neonatal outcomes and maternal complications by maternal national origin. Mean birth weight was highest for babies of mainland Chinese, 3222.4 g (±372.8), significantly higher than of Taiwanese, 3126.9 g (±411.8), and Vietnamese 3163.0 g (±368.9). Preterm birth rates were also most favorable for Chinese, 3.8%, relative to Taiwanese (7.7%; P \ 0.01), and Vietnamese (4.6%, P = 0.58; P = 0.0584 for Vietnamese relative to Taiwanese). The respective rates of low birth weight were 3.5, 5.7, and 2.5%. Taiwanese women delivered at significantly lower mean gestational age, 38.4 (±1.6) weeks than Chinese and Taiwanese. Taiwanese women also had significantly higher incidence of hypertension and diabetes (17.6 and 6.9%, respectively) than Chinese (10.5 and 5.1%) and Vietnamese mothers (7.0 and 2.1%). Adjusted Differences in Neonatal Outcomes by Maternal National Origin Because the sample had a very low prevalence of smoking (including zero prevalence among Vietnamese), smoking status was excluded from regression analysis. The remaining demographic, obstetric and medical risk factors were examined for associations with birth weight (as a continuous variable) using multiple linear regression analysis. Starting with the full model, insignificant variables were sequentially deleted one at a time based on

123

272


Table 1 Parental demographics and maternal risk characteristics of newborns of native-born Taiwanese and immigrant Chinese and Vietnamese mothers in Kaohsiung Municipal United Hospital, 2002–2007 Taiwanese n = 2646 (81.0%) Mean ± SD/No. (%)

Mainland Chinese n = 341 (10.4%) Mean ± SD/No. (%)

Vietnamese n = 280 (8.6%) Mean ± SD/No. (%)

Total n = 3267 (100%) Mean ± SD/No. (%)

Parental demographics Maternal age, y/o

30.3 ± 4.8

28.0* ± 4.1

24.2**,*** ± 4.0

29.5 ± 5.0

Paternal age, y/o

33.4 ± 5.2

37.9* ± 6.1

38.5*** ± 6.3

34.4 ± 5.8

Height, cm

159.4 ± 5.1

158.9 ± 4.8

155.6**,*** ± 4.4

159.1 ± 5.2

Pregravid weight, Kg

57.1 ± 9.8

52.6* ± 7.3

48.9**,*** ± 6.5

55.9 ± 9.7

BMI, Kg/m*m

22.5 ± 3.6

20.8* ± 2.5

20.2**,*** ± 2.5

22.1 ± 3.5

Net weight gain, Kg

9.3 ± 3.5

9.9* ± 3.4

9.3** ± 3.3

9.4 ± 3.5

\High school

186 (7.0%)

174 (51.0%)

212 (75.7%)

572 (17.5%)

High school

1,091 (41.2%)

124 (36.4%)

59 (21.1%)

1,274 (39.0%)

[High school

1,369 (51.7%)

43 (12.6%)

9 (3.2%)

1,241 (43.5%)

227 (8.6%)

9 (2.6%)

3 (1.1%)

239 (7.3%)

Education, n (%)à

Single mother, n (%)à Medical risk factors, n (%)à Maternal birth history Primiparous

1,208 (45.7%)

190 (55.7%)

168 (60.0%)

1,566 (47.9%)

Multiparous and history of CS

499 (18.9%)

39 (11.4%)

23 (8.2%)

561 (17.2%)

Multiparous w/o history of CS

939 (35.5%)

112 (32.8%)

89 (31.8%)

1,140 (34.9%)

Pregnancy loss, n (%)§

326 (12.3%)

52 (15.2%)

24 (8.6%)

402 (12.3%)

Smoking, n (%)à

84 (3.2%)

4 (1.2%)

0 (0%)

88 (2.7%)

Inadequate

725 (27.4%)

106 (31.1%)

77 (27.5%)

908 (27.8%)

Intermediate

1,136 (42.9%)

145 (42.5%)

119 (42.5%)

1,400 (42.9%)

Adequate

718 (27.1%)

89 (26.1%)

77 (27.5%)

884 (27.1%)

Adequate plus

67 (2.5%)

1 (0.3%)

7 (2.5%)

75 (2.3%)

Hypertensive disorderà

459 (17.3%)

34 (10.0%)

19 (6.8%)

512 (15.75%)

Diabetesà

168 (6.3%)

14 (4.1%)

6 (2.1%)

188 (5.8%)

Prenatal careà

Maternal complications

* Significant difference at 0.05 level between Taiwanese and mainland Chinese ** Significant difference at 0.05 level between mainland Chinese and Vietnamese *** Significant difference at 0.05 level between Taiwanese and Vietnamese Net weight gain: weight gain during pregnancy minus neonatal birth weight à Significant difference in distribution by ethnic group at p, 0.01 level. All categorical variables were evaluated for differences in ethnic distribution using Pearson Chi-Square tests § p value for Pearson Chi-Square test is 0.04 for distribution among ethnic groups

P-values. Table 3 presents the final model retaining statistically significant variables which collectively explain 35% of the variance in neonatal birth weight (R2 = 0.35; adjusted R2 = 0.349). The adjusted mean birth weights for Chinese and Vietnamese mothers were higher than for Taiwanese by 87.7 and 74.7 g, respectively (P \ 0.000 and P \ 0.001). Among the control variables, the significant covariates were maternal age ([35 years), height, diabetes, hypertensive disorder, primiparous birth (lower birth weight), multiparous mother with history of cesarean delivery (higher birth weight), prenatal care adequacy, pre-

123

pregnant weight, fetal gender, and gestational age, all directions consistent with the documented literature. The adjusted odds of preterm birth are lower for Chinese and Vietnamese relative to Taiwanese, OR = 0.46 and 0.47, respectively. Among the control variables, maternal age ([35 years), education (lower odds at higher education levels), hypertensive disorder (increased odds), net weight gain, and adequacy of prenatal care were significant, largely consistent with the documented literature. Regarding low birth weight, Chinese and Vietnamese had lower unadjusted odds relative to Taiwanese (0.59 and 0.42,


273

Table 2 Birth outcomes of Taiwanese and immigrant mothers from mainland China and Vietnam N

Taiwanese 2,646

Mainland Chinese 341

Vietnamese 280

3,126.9 ± 411.8

3,222.4 ± 372.8*

3,163.0 ± 368.9**

38.4 ± 1.6

38.8 ± 1.4*

Neonatal characteristics Birthweight, mean (±SD), g Gestational week, mean (±SD) N (%)

N (%)

Preterm birth

205 (7.7%)

13 (3.8%)*

Low birth weight

151 (5.7%)

12 (3.5%)

1,343 (50.8%)

175 (51.3%)

Male neonate, n (%)

38.8 ± 1.4*** N (%) 13 (4.6%) 7 (2.5%)*** 158 (56.4%)

* Significant difference between Taiwanese and mainland Chinese at P \ 0.05 ** Significant difference between mainland Chinese and Vietnamese at P \ 0.05 *** Significant difference between Taiwanese and Vietnamese at P \ 0.05

respectively). The study was inadequately powered to assess adjusted associations (Table not presented).

Discussion Our study finds relatively better neonatal outcomes for immigrant Chinese and Vietnamese mothers relative to native-born Taiwanese. Crude birth weights, on average are similar among Taiwanese and Vietnamese, both lower than Chinese. Adjusted birth weights for Chinese and Vietnamese mothers are significantly higher than for Taiwanese, and their odds of pre-term delivery are lower. Examining the inter-ethnic differences in biological predictors of birth weight, we find that mainland Chinese women are similar to native Taiwanese women on maternal height and pre-pregnant weight, both greater than Vietnamese. Both Chinese and Taiwanese are also better placed than Vietnamese on all other maternal demographic factors. Taiwanese women are at a disadvantage on age, hypertensive disorder, diabetes, and smoking, with the Chinese being intermediate between the remaining two groups. While unadjusted birth weights are on average similar for Taiwanese and Vietnamese, on adjusting for risk factors, Vietnamese have higher mean birth weight as well as lower risk of pre-term delivery. The pattern of prenatal care adequacy was similar across the three ethnic groups. Our findings confirm that prenatal care adequacy predicts birth weight and preterm delivery, with the best outcomes for women with adequate prenatal care relative to inadequate, intermediate and ‘‘adequate plus’’ categories. (‘‘Adequate plus’’ prenatal care visits are more likely to be abnormal pregnancies with a higher risk of adverse outcomes, lower birth weight and pre-term delivery). Our finding of better neonatal outcomes among immigrant women is consistent with the documented

epidemiological paradox of better neonatal outcomes among Mexican and other Hispanic subgroups in the US relative to native-born women of the same ethnicity [13, 26–28]. The neonatal advantage due to having a foreignborn mother among blacks and Hispanics does not hold among foreign-born whites and Asians [29], a finding that is attributed by past authors to ethnic heterogeneity (and corresponding cultural and behavioral heterogeneity) among the generic census classification of ‘‘Asian’’ For example, a subset of Asian–Americans, Korean–Americans show better birth outcomes among foreign-born compared to US-born Korean–Americans [30]. Our study based on primary data identified national origin accurately and provides robust evidence for the epidemiologic paradox in birth outcomes. In general, low SES and education are associated with adverse birth outcomes [31]. After adjusting for these factors, our study finds that immigrant groups have favorable birth outcomes relative to native-born women. Our finding provides robust, primary data-driven validation of the generic observation that immigrant groups have better birth outcomes [16]. In the US Hispanic women, despite less prenatal care and low SES have better birth outcomes than native-born white women [14, 15, 32–35]. Studies from other developed countries have documented a similar epidemiological paradox of heavier neonates borne by immigrant mothers hailing from less developed countries with low socio-economic conditions relative to native-born mothers, notably studies from Greece, Italy, and the US [11, 36–38]. A notable exception is that of Vietnamese immigrant mothers in Australia, who had 263 g lighter babies than Australian-born women, unadjusted for any confounders [38]. Considering our study findings, it appears plausible that lack of adjustment for biological, socio-economic and prenatal care factors may account for some of the divergent findings in international studies.

123

274


Table 3 Adjusted associations of maternal national origin with neonatal birth weight and odds of pre-term birth—Final models retaining significant variables (n = 3,267 live births) Birth weight (g)

Preterm delivery odds

Unadjusted

Adjusted

Unadjusted

Adjusted

Chinese

95.6*** 49.9, 141.2

87.7*** 50.0, 125.4

0.47** 0.27, 0.84

0.46* 0.25, 0.87

Vietnamese

36.1

74.7***

0.58**

0.47*

-13.7, 85.9

32.2, 117.3

0.33, 1.03

0.24, 0.91

National origin Taiwanese (reference)

Maternal age Aged 20–35 (ref.) Aged \ 20 Aged [ 35

-14.9

0.59

-89.2, 59.5

0.18, 1.92

36.8*

2.1***

2.5, 71.2

1.50, 2.92

–

0.66

Education \High school (ref.) High school

0.42, 1.05 [High school

–

0.53** 0.33, 0.84

Maternal birth history Primipara (ref.) Multipara with H/O cesarean

110.0***

–

77.2, 142.8 Multipara w/o H/O cesarean

35.6** 10.1, 61.0

Diabetes

101.4***

–

52.1, 150.7 Height

6.1***

–

3.7, 8.5 Hypertension Prepregnant weight

-74.0***

1.50*

-106.1, -41.9

1.07, 2.09

8.6*** 7.2, 9.9

–

Gestational weeks

134.8***

–

Male

127.2, 142.3 97.8***

–

75.3, 120.3 Net weight gain

–

Unmarried

-85.0***

0.95* 0.91, 0.99 –

-130.1, -39.8 Adequacy of prenatal care Prenatal care adequate (reference) Prenatal care inadequate

-38.0* -69.0, -7.0

0.40, 0.83

Prenatal care intermediate

-29.2*

0.40***

-57.2, -1.2

0.28, 0.56

123

0.58**


275

Table 3 continued Birth weight (g) Unadjusted Prenatal care adequate plus

Preterm delivery odds Adjusted

Unadjusted

Adjusted

-39.3

5.53***

-118.0, 39.3

3.25, 9.40

Constant

3126.9***

-3564.0***

R-square

0.005

0.35

Note: Numbers in parentheses are 95% confidence intervals * 0.01 \ p \ 0.05, ** 0.001 \ p \ 0.01, *** p \ 0.001

The neonatal outcomes for Vietnamese and Chinese mothers in our study are superior to the prevailing outcomes in their native countries. In China, the incidence of low birth weight was 4.6, 3.9 urban and 4.8% rural [39], compared to 3.4% for Chinese immigrants to Taiwan. Similarly, in Vietnam, regional studies show low birth weight rates of 7.9% urban and 12.5% rural [40]. The World Health Organization global survey reported a low birth weight rate of 5.2% and preterm birth rate of 13.6% in Vietnam [41]. Both rates are much higher than Vietnamese immigrants’ rates of 2.4 and 4.5%, respectively in our study. The factors driving the epidemiological paradox in Taiwan could be a combination of cultural norms in mothers’ country of origin [37], the protective effect of equal access to prenatal care, and selection effects. Conservative cultural norms regarding women’s smoking and alcohol use may mediate part of the favorable outcomes among Vietnamese. For example, Latin-American immigrants have a lower prevalence of smoking, alcohol and illicit drug use during pregnancy relative to US-born Latinas, which may reflect conformity to native cultural norms that may be protective [42–44]. In our study immigrants have almost zero prevalence of smoking. In a separate regression analysis restricted to the Taiwanese sub-sample, smoking was associated with 123.5 g lighter babies after adjusting for demographic and medical risk factors (results not presented), which may partly account for better neonatal outcomes among Vietnamese and Chinese. A second explanation for the epidemiological paradox in Taiwan may be equal access to prenatal care for immigrant women (regardless of citizenship status) under Taiwan’s universal coverage single-payer health system. Prenatal care is associated with neonatal outcomes [45.] In many countries, immigrants lack health insurance resulting in fewer prenatal visits [46]. In our study all three ethnic groups had similar levels of prenatal care. Prenatal care is free in Taiwan, and delivery is covered after a 4-month stay. Universal insurance coverage eliminates prenatal care disparities, which is reflected in our data, and may contribute to immigrants’ favorable outcomes in Taiwan despite their much lower educational status.

A third plausible explanation is the selection effect, due to higher propensity of the relatively prosperous among a country’s people to legally migrate to more developed countries. This is because they have better economic resources to migrate and have better access to emigration opportunities. Therefore, those who migrate may be relatively healthier than their domestically-anchored compatriots. The literature supports the selection effect. Mexican immigrants to the US have better health status than those who returned to Mexico [47, 48]. In conclusion, in our study Vietnamese, and to a less extent, Chinese mothers have the relative advantages of younger age and lack of smoking, advantages that are reinforced by full access to healthcare, and possibly by the psychosocial benefits of migration to a higher living standard. These factors may collectively mediate their better birth outcomes. Taiwan’s immigrant access to medical care and low maternal parity contrasts with the US where Mexican-born immigrants are often uninsured, have fewer prenatal visits than the native born population, higher parity, and lower education [31]. The factors that contribute to better birth outcomes for immigrants in Taiwan may be different from those operating in other countries.

Policy Implications for Taiwan and International Significance of the Study Our study provides a scientifically robust evidence base for the Taiwanese government to start a public conversation about the healthy neonatal outcomes of immigrant mothers, particularly Vietnamese who face social discrimination the most in Taiwan. While adjusted neonatal outcomes are better for Vietnamese, the worst case evidence, crude rates show that even their unadjusted outcomes are no worse than those of Taiwanese women. These findings could help shift the general public’s attitudes from an affective reactionary mode to a rational conversation based on evidence, and help mitigate the prevailing social discrimination against immigrant spouses. Immigrant spouses are socially handicapped to begin with, by language difficulties in case

123

276


of Vietnamese, and by the relatively disadvantaged families they marry into, poor, socially excluded, and rural. Our findings can help the government of Taiwan to address the human rights citations on a sound evidence-based footing. Internationally, our study is relevant for many countries with large East Asian immigrant populations such as the US, Australia and some European countries, providing robust validation of the tentative findings documented for immigrant East Asian mothers, usually based on birth registry data that lack information on many key demographic and medical risk factors.

Study Limitations One study limitation is the use of data from a single urban hospital, which may jeopardize generalization to the national population. The selection of hospital was driven by the researchers’ access to the charts being employees of the hospital (compliant with IRB protections). This disadvantage should be balanced against the increased statistical power enabled by overrepresentation of immigrant births due to higher proportion of immigrant wives in the relatively less prosperous region of southern Taiwan. The data source also enabled a stronger scientific contribution because of the richness of variables available from the primary data source which enabled adjustment for key clinical, behavioral and prenatal care history. In national claims data, these items are either not available (e.g. smoking, prenatal care history, maternal height, pre-pregnant weight), or not reliably coded (e.g. maternal complications). Our sample over-represents immigrant mothers— 19% compared to 11% nationally. However, it is unlikely that the study’s essential conclusions are jeopardized. Our sample mean birth weights for the three groups are comparable to the national population means for these subgroups. We calculated these rates from nationwide birth certificate data for the same period, using a dataset

procured from the National Health Insurance Research Institute for a different study (results not yet published). Our study sample to national population comparisons are as follows: 3,127 g versus 3,154 g, respectively for Taiwanese, 3,222 g versus 3,263 g for Chinese, and 3,163 g versus 3,180 g for Vietnamese mothers. Similar to our sample, the population mean for Chinese is significantly higher than that of Taiwanese and Vietnamese, and the difference between Taiwanese and Vietnamese is not significantly different. Therefore, despite our localized data source, the study conclusions may generalize to the population. Another study limitation is the lack of adequate statistical power to assess the role of behavioral factors such as smoking and alcohol use. Having used the data on all study-eligible births during a 6-year period, we could not overcome this study limitation.

Conclusion Adjusted for confounding variables, immigrant mothers of Chinese and Vietnamese origin have heavier babies and a reduced risk of preterm birth. The evidence from Taiwan supports the existence of an epidemiological paradox in birth outcomes for immigrants. Without adjustments for confounders, Vietnamese mothers’ babies have outcomes similar to Taiwanese mothers, and when adjusted for confounders, they have better outcomes. These findings provide an adequate evidence base for the Taiwanese government to initiate a public education initiative to address the human rights issue of social discrimination against immigrant women by marriage.

Appendix See Table 4.

Table 4 Full model with all independent variables: adjusted associations between maternal national origin and birth weight/Pre-term birth (coefficient estimates and odd ratios (95% CI) Birth weight (g) Unadjusted National origin


n = 3,267

Unadjusted

Adjusted

n = 3,267

Taiwanese (reference) Chinese Vietnamese Maternal age Aged 20–35 (reference)

123

95.6**

78**

0.5*

0.5*

49.9, 141.2

33.6, 122.3

0.3, 0.8

0.2, 0.9

36.1

59.1*

0.6

0.4*

-13.7, 85.9

6.1, 112.0

0.3, 1.0

0.2, 0.9


277

Table 4 continued Birth weight (g) Unadjusted Aged \ 20 Aged [ 35


Unadjusted

Adjusted

-24.1

0.0

-119.5, 71.3

0.0

32.4

1.8**

-6.4, 71.2

1.2, 2.7

-24.9

0.7

Education \High school (reference) High school Above high

-66.0, 16.1

0.4, 1.1

-11.3 -54.4, 31.7

0.6* 0.3, 0.98

119.0**

1.2

84.9, 153.2

0.81, 1.7

Maternal birth history Primipara (reference) Multipara with history of cesarean Multipara w/o history of cesarean Diabetes Height Hypertension

35.1*

0.8

8.3, 61.9

0.5, 1.1

94**

0.8

43.1, 145.0

0.4, 1.4

6.0**

0.98

3.4, 8.5

0.9, 1.0

-74.0**

1.6*

-107.1, -40.9

1.1, 2.3

Prepregnant weight

8.8**

1.0

Gestational weeks

7.4, 10.2 139.8**

0.98, 1.0 x

131.8, 147.9

x

104.5**

1.2

Male

81.3, 127.6

0.9, 1.6

0.9

1.03*

-1.5, 3.4

1.00, 1.06

1.6

0.96

-1.9, 5.1

0.9, 1.0

-2.7

1.1

-38.6, 32.2

0.7, 1.6

-148.6

x

-350.7, 53.6

x

Prenatal care inadequate

-38.5* -70.9, -6.0

0.5** 0.4, 0.8

Prenatal care intermediate

-25.7

0.4**

Father’s age Net weight gain Pregnancy loss Unmarried Adequacy of prenatal care Prenatal care adequate (reference)

Prenatal care adequate plus

-54.2, 2.8

0.3, 0.6

-30.2

5.4**

-109.0, 48.6

3.1, 9.4

Constant

3,126.9

-3,783

R-square

0.005

0.35

* p \ 0.05, ** p \ 0.01

123

278

References 1. Wang, L. (2009). Change of population structure from total fertility rate perspective. Council for Economic Planning and Development. Accessed June 15, 2009. Available at http://www. cepd.gov.tw/m1.aspx?sNo=0011563&ex=%20&ic=. 2. National statistics. (2009). Accessed March 8, 2009. Available at http://ebas1.ebas.gov.tw/pxweb/Dialog/Searchpx2.asp?SQ=Q &lang=9. 3. Ministry of Interior. (2008). Number of marriage and divorce registration by nationality in Taiwan-Fuchien Area. Accessed June 15, 2008. Available at http://www.moi.gov.tw/stat/index.asp. 4. Ministry of Interior. (2008). Foreign and mainland Chinese spouse Life Survey. Accessed January 20, 2008. Available at http://www.ris.gov.tw/ch4/0930617.html. 5. Liao, C.-C., & Chan, W.-Y. (2005). Concerns over the life and health of alien brides from Southeastern Asia and Mainland China. Tzu Chi Nursing Journal, 4, 12–16. 6. Unites States Department of State, Bureau of Democracy, Human Rights and Labor Report on Taiwan dated March 11, 2008. Accessed April 2, 2010 at http://www.state.gov/g/drl/rls/hrrpt/ 2007/100517.htm. 7. Bureau of Health Promotion. (2008). 2003 Statistics of birth reporting system. Accessed March 8, 2008. Available at http://www.bhp.doh.gov.tw/download/themeParkId=542/960301/ 960301.htm. 8. Wang, H.-H., Chung, U.-L., Chou, P.-H., & Chiang, Y.-P. (2006). Physical and mental pressure—a survey on pregnant women in Taiwan who originally came from Southeast Asia. Journal of Health Management, 4, 89–101. 9. Sung, J., Taylor, B., & Blumenthal, D. (1994). Maternal factor, birthweight and racial differences in infant mortality: a Georgian population-based study. Journal of the National Medical Association, 86, 437–443. 10. Shmueli, A., & Cullen, M. R. (1999). Birth weight, maternal age, and education: new observations from Connecticut and Virginia. Yale Journal of Biological Medicine, 72, 245–258. 11. Gould, J. B., Madan, A., Qin, C., & Chavez, G. (2003). Perinatal outcomes in two dissimilar immigrant populations in the United States: A dual epidemiologic paradox. Pediatrics, 111, e676– e682. 12. de la Rosa, I. A. (2002). Perinatal outcomes among Mexican Americans: A review of an epidemiological paradox. Ethnicity and Disease, 12, 480–487. 13. Centers for Disease Control, Prevention. (2002). State-specific trends in U.S. live births to women born outside the 50 states, the District of Columbia—United States, 1990 and 2000. MMWR Morbidity and Mortality Weekly Report, 51, 1091–1095. 14. Crump, C., Lipsky, S., & Mueller, B. (1999). Adverse birth outcomes among Mexican–Americans: are US-born women at greater risk than Mexico-born women? Ethnicity & Health, 4, 29–34. 15. English, P., Kharrazi, M., & Guendelman, S. (1997). Pregnancy outcomes and risk factors in Mexican Americans: the effect of language use and mother’s birthplace. Ethnicity and Disease, 7, 229–240. 16. Hessol, N. A., & Fuentes-Afflick, E. (2000). The perinatal advantage of Mexican-origin Latina women. Annals of Epidemiology, 10, 516–523. 17. Madan, A., Holland, S., Humbert, J. E., & Benitz, W. E. (2002). Racial differences in birth weight of term infants in a northern California population. Journal of Perinatology, 22, 230–235. 18. Martin, J. A., Hamilton, B. E., Sutton, P. D., Ventura, S. J., Menacker, F., Kirmeyer, S., et al. (2009). Births: Final Data for 2006. National Vital Statistics Reports, 57.

123

Matern Child Health J (2011) 15:269–279 19. Kim, H., Chang, K., Yang, J., Yang, S., Lee, H., & Ryu, H. (2002). Clinical outcomes of pregnancy with one elevated glucose tolerance test value. International Journal of Obstetrics & Gynecology, 78, 132–138. 20. National Diabetes Data Group. (1979). Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes, 28, 1039–1057. 21. Cheng, Y. W., Block-Kurbisch, I., & Caughey, A. B. (2009). Carpenter-Coustan criteria compared with the national diabetes data group thresholds for gestational diabetes mellitus. Obstetrics and Gynecology, 114(2 Pt 1), 326–332. 22. Oken, E., Ning, Y., Rifas-Shiman, S. L., Rich-Edwards, J. W., Olsen, S. F., & Gillman, M. W. (2007). Diet during pregnancy and risk of preeclampsia or gestational hypertension. Annals of Epidemiology, 17, 663–668. 23. Moraes, C. L., Amorim, A. R., & Reichenheim, M. E. (2006). Gestational weight gain differentials in the presence of intimate partner violence. International Journal of Obstetrics & Gynecology, 95, 254–260. 24. Kotelchuck, M. (1994). An evaluation of the Kessner adequacy of prenatal care index and a proposed adequacy of prenatal care utilization index. American Journal of Public Health, 84, 1414–1420. 25. Kotelchuck, M. (2009). Overview of adequacy of prenatal care utilization index. Accessed October 6, 2009. Available at http:// www.mchlibrary.info/databases/HSNRCPDFs/Overview_APC UIndex.pdf. 26. Acevedo-Garcia, D., Soobader, M. J., & Berkman, L. F. (2007). Low birthweight among US Hispanic/Latino subgroups: the effect of maternal foreign-born status and education. Social Science and Medicine, 65, 2503–2516. 27. Callister, L. C., & Birkhead, A. (2002). Acculturation and perinatal outcomes in Mexican immigrant childbearing women: An integrative review. Journal of Perinatal & Neonatal Nursing, 16, 22–38. 28. Guendelman, S., Buekens, P., Blondel, B., Kaminski, M., Notzon, F. C., & Masuy-Stroobant, G. (1999). Birth outcomes of immigrant women in the United States, France, and Belgium. Maternal and Child Health Journal, 3, 177–187. 29. Acevedo-Garcia, D., Soobader, M.-J., & Berkman, L. F. (2005). The differential effect of foreign-born status on low birth weight by race/ethnicity and education. Pediatrics, 115, e20–e30. 30. Cho, Y., Song, S.-E., & Frisbie, W. P. (2005). Adverse birth outcomes among Korean Americans: The impactof nativity and social proximity to other Koreans. Population Research and Policy Review, 24, 263–282. 31. Berkowitz, G. S., & Papiernik, E. (1993). Epidemiology of preterm birth. Epidemiologic Reviews, 15, 414–443. 32. Miletic, T., & Stoini, E. (2005). Influence of maternal pregravid weight, height and body mass index on birth weight of male and female newborns. Collegium Antropologicum, 29, 263–266. 33. Buekens, P., Notzon, F., Kotelchuck, M., & Wilcox, A. (2000). Why do Mexican Americans give birth to few low-birth-weight infants? American Journal of Epidemiology, 152, 347–351. 34. Fuentes-Afflick, E., Hessol, N. A., & Perez-Stable, E. J. (1999). Testing the epidemiologic paradox of low birth weight in Latinos. Archives of Pediatrics and Adolescent Medicine, 153, 147–153. 35. Rosenberg, T. J., Raggio, T. P., & Chiasson, M. A. (2005). A further examination of the ‘‘epidemiologic paradox’’: Birth outcomes among Latinas. Journal of National Medical Association, 97, 550–556. 36. Auger, N., Luo, Z. C., Platt, R. W., & Daniel, M. (2008). Do mother’s education and foreign born status interact to influence birth outcomes? Clarifying the epidemiological paradox and the healthy migrant effect. Journal of Epidemiology and Community Health, 62, 402–409.

Matern Child Health J (2011) 15:269–279 37. Triantafyllidis, G., Christopoulou, C., & Damianaki, D. (2009). Gestational age, birthweight and feeding policy differences between Greek and immigrant women in a public nursey. International conference on health promoting hospital and health services, 2009, Crete, Greece, p. p153. 38. Chan, A., Roder, D., & Macharper, T. (1988). Obstetric profiles of immigrant women from non-English speaking countries in South Australia, 1981–1983. Australian and New Zealand Journal of Obstetrics and Gynaecology, 28, 90–95. 39. Yu, D. M., Zhao, L. Y., Liu, A. D., Yu, W. T., Jia, F. M., & Zhang, J. G. (2007). et al. [Incidence of low birth weight of neonates and the influencing factors in China]. Zhonghua Yu Fang Yi Xue Za Zhi, 41(Suppl), 150–154. 40. Dinh, P. H., To, T. H., Vuong, T. H., Hojer, B., & Persson, L. A. (1996). Maternal factors influencing the occurrence of low birthweight in northern Vietnam. Annals of Tropical Paediatrics, 16, 327–333. 41. WHO. (2002). Meeting of Advisory Group on maternal nutrition and low birthweight, December 4–6, 2002, Geneva. 42. Cho, Y., & Hummer, R. A. (2001). Disability status differentials across fifteen Asian and Pacific Islander groups and the effect of nativity and duration of residence in the US. Social Biology, 48, 171–195.

279 43. Collins, J. W., Jr., & Shay, D. K. (1994). Prevalence of low birth weight among Hispanic infants with United States-born and foreign-born mothers: the effect of urban poverty. American Journal of Epidemiology, 139, 184–192. 44. Lee, S. K., Sobal, J., & Frongillo, E. A., Jr. (2000). Acculturation and health in Korean Americans. Social Science and Medicine, 51, 159–173. 45. Fuentes-Afflick, E., & Hessol, N. A. (1997). Impact of Asian ethnicity and national origin on infant birth weight. American Journal of Epidemiology, 145, 148–155. 46. Thamer, M., Richard, C., Casebeer, A. W., & Ray, N. F. (1997). Health insurance coverage among foreign-born US residents: the impact of race, ethnicity, and length of residence. American Journal of Public Health, 87, 96–102. 47. Soldo, B., Wong, R., & Palloni, A. (2002). Migrant health selection: Evidence from Mexico and the US. Population Association of America Annual Meeting, May 9–11, 2002, Atlanta, GA. 48. Scribner, R. (1996). Paradox as paradigm–the health outcomes of Mexican Americans. American Journal of Public Health, 86, 303–305.

123