Birth weight and delivery practice in a Vietnamese rural district during 12 year of rapid economic development

During the last few decades, there has been a dramatic change in the economic conditions in Vietnam. Yearly income per capita has increased from $130 in the early 1990s to $900 in 2008. Poverty according to the World Bank definition (less than 1.25 US$ per person and day) has been reduced from 58% of the population in 1993 to 13% in 2008 [1]. The total health expenditure in Vietnam increased from 4.9% of GDP in 1999 to 5.9% in 2005. Expenditures for public health care and services decreased while expenditures for care by private providers increased [2]. Some health care and health effects have been noted. The percentages of women, who used at least one antenatal care visit, increased from 71% in 1997 [3] to over 80% in 2005 [4]. The proportion of low birth weight infants decreased from 9% in 2000 [5] to 7% in 2007 [6]. The infant mortality rate (IMR) decreased from 39 to 18 per 1,000 live born births and the maternal mortality ratio (MMR) decreased from 92 to 80 per 100,000 live births between 1997 and 2005 [4].

The Health and Demographic Surveillance Site (HDSS) FilaBavi in the rural district Bavi started in 1999. An open cohort of about 12,000 households has since then been followed regularly. From 1999 to 2009, the median self-reported yearly income per household in FilaBavi increased from 5,990,000 Vietnamese dong (VND) (18,932 VND = 1 US$ in 2010) to 15,000,000 VND, taking inflation, (118% according to International Monetary Fund) over the period, into account.

In a situation of economic growth but with possibly widening gaps in socioeconomic conditions within populations there is an urgent need for surveillance of health and well-being. A vulnerable group is children, particularly the newborn and the very young. Birth weight is an indicator of the mother’s health and nutritional status. It is also an indicator of a newborn’s chances for survival, growth, long-term health and psychosocial development [7]. Low birth weight (LBW, birth weight less than 2,500 grams) raises the risk of infant death and adverse health outcomes later in life such as asthma, coronary heart diseases, type 2 diabetes and hypertension [8‐10]. LBW is a public health problem in most developing countries, where an estimated 15 per cent of births result in LBW babies. Reducing the incidence of LBW is one of the major goals in ‘A World Fit for Children,’ the Declaration and Plan of Action adopted at the United Nations General Assembly Special Session on Children in 2002 [5].

Several genetic and environmental factors influence the intrauterine growth of a fetus and hence the birth weight. The different strengths and prevalences of those factors lead to differences in birth weight between countries, ethnic groups and different social and economic contexts. At the household and individual levels variations occur due to a manifold of influences. For example demographic, social and economic conditions are known as determinants of health in general, including birth weight. It is so far not known if the economic changes in Vietnam have been associated with birth weight changes and changes in the proportion of low birth newborn. Therefore the primary aim of this paper is to study trends in birth weight together with the proportion of LBW newborn over the time period 1999–2010 in FilaBavi. There is also an aim to study trends in delivery place, presence of medical attendance and delivery type, particularly the use of Cesarean section (CS). Finally we want to study how birth weight and delivery characteristics are associated with economic and educational characteristics of the mother and the household. This will be done at the macro level, over time, as well as at the level of individual households.

The main finding in this study of the comparatively long time series is that the overall mean birth weight has been stable over the years in spite of a rather drastic economic and technological development. Most previous studies in Vietnam, and other countries, have found increases in the mean birth weight and reductions in the incidence of low birth weight over time periods with improved economy [10, 14‐16]. One study even showed that the low birth weight proportion decreased in spite of an increased rate of multiple births [17]. Some possible explanations have been mentioned. Improvement of socioeconomic conditions can lead to better nutrition of mothers during pregnancy and increased mother’s weight gain during pregnancy [10, 14‐16]. Increased economic resources can facilitate access to antenatal care. An interesting contradictory result is given in a study of term singleton births in the United States where the mean birth weight decreased from 1990 to 2005. The finding could not be explained by maternal and neonatal characteristics or by trends in induced labor or CS. Neither did the observed decline in gestational age at birth, which is normally strongly correlated to birth weight, explain the decreasing birth weight [18]. The present results show no systematic changes over time in the gestational age proxy indicator used.

In the yearly analyses at the level of individual children we find associations between birth weight and indicators of economic resources as well as education. This is not a contradiction to the finding that the mean birth weight did not increase over years. First a certain time lag from economic changes to birth weight effects should be expected. Further, the general economic improvement does not necessarily influence on the general birth weight level, particularly if the economic and nutritional inequalities remain or increase. Over the years there can also be changes in household priorities for the use of economic resources.

UNICEF has suggested that economic development alone does not automatically lead to improvement in nutrition. Good nutrition is an outcome of many factors both biological and social [19]. Most pregnant women in rural areas are not covered by any social insurance scheme [20]. The present study shows that household health expenditures roughly doubled during the period and were not much different between income quintiles. This means that poor households spend a much higher share of their total income on health than the higher income groups. Rural mothers have to pay themselves for antenatal care if they are not covered by any social insurance and even if they are covered, may still have to pay themselves for delivery or antenatal care in hospital, particularly if they come directly, without referral, to a hospital.

The present study shows that the percentages of women who delivered in hospitals as well as women who delivered with CS, increased during the time period. This tendency might be an indication of increased confidence in technology. Along the same line, the use of ultrasound has become increasingly common. Possible overuse of ultrasound has been seen in another study in Vietnam in 2007 [21].

Compared to results from urban Hanoi, the pregnant women in FilaBavi attended antenatal care later, had fewer visits and much fewer medical services [22]. The health expenditure of pregnant women may to a large extent be payment for health technology rather than for nutrition and good antenatal health care. E.g. the prevalence of anemia of pregnant women in Vietnam did not change noticeable between 2001 (32.2%) [23] and 2008 (31.4%) [24].

A finding is that the percentages of women, who gave birth in hospitals as well as delivered with trained attendance, increased over the twelve years. Vietnam has a strict two child policy. From a recent study in FilaBavi, we know that the proportion of delivery in hospitals is higher among women with higher education and among mothers without previous sons in FilaBavi [25]. Most likely all women want to have good delivery care for themselves and their newborns and believe that the quality of delivery care is better in the hospitals than in the commune health centers. There is no formal barrier for women to go directly to any hospital to give birth, if they can pay for themselves. According to the Health Strategy and Policy Institute, about 90-95% of all patients in the Central Obstetric Hospital in Hanoi, including mothers with low economic condition, went directly to the hospital without transferring from local hospital [26]. The main reason for this care-seeking behavior can be that people have higher confidence in the professional staff, medical equipment, and infrastructure of the central hospitals [26].

We found that the use of CS increased over the 12 years in FilaBavi along with the increase of economic resources and mother’s education. This is in accordance with the result from a recent study in FilaBavi, in which CS use increased with education level and among women living at lowland areas or in richer households [25]. CS has increased also in other countries. The highest percentages of CS in Asia were in China (46.2%). The second country was Vietnam (35.6% in 2007 in three obstetric hospitals) [27]. A reason for CS without medical indication can be that families want to choose a good time and day for operation, that the operation is believed to be safe and that the child will have a good future life as a result. There are beliefs that CS will prevent asphyxia for newborns and that it is a safe and painless technique for both mother and baby. A study in nine Asian countries showed that CS is associated with improved perinatal outcome, but it also pointed to increased risk of neonatal intensive care treatment, maternal death and maternal morbidity [27]. The rate of CS has increased in FilaBavi since 1999 to 2010 but we do not have information about the distribution of medical and non-medical indications.

During the last 12 years, the male to female sex ratio at birth (SRB) has increased in Filabavi. The sex ratio at birth in 2010 was quite far from the normal. In Vietnam as a whole, SRB increased from 1.05 in 2001 to 1.12 in 2006 [28]. This is similar to results from other studies in Vietnam [28] and other Asian countries, particularly India and China. It has been estimated that around 80 million females in China and India are “missing” [29]. Sons are considered more valuable to parents than daughters. The first son will maintain the family line and be responsible for carrying out the cult of the ancestors. Almost all parents want to give their house and other wealth only to their sons, especially the first son [30]. Another reason for the high SRB is that stricter population regulations were also introduced in Vietnam at the end of 1980s saying that families should not have more than two children [28]. Parents therefore may try harder to have at least one son. It is not difficult for women to know the sex of the foetus very early through ultrasound scanning and women can decide freely about early abortion. One can suspect that the beliefs and traditions behind the elevated SRB also influence the choices of delivery place and method as mentioned above.

This study gives a picture of birth weight trends in FilaBavi during twelve years from 1999 and the incidence of LBW, delivery place, presence of medical attendance at delivery and type of delivery, particularly the use CS. We also found an association between birth weight and delivery characteristics on the one hand and economic condition of the household as well as educational characteristics of the mother on the other.

The births studied are assumed to be all or close to all live births that actually took place in FilaBavi 1999–2010. Different small studies have been done comparing the numbers of reported births and deaths in FilaBavi with the official information given by local authorities. The latter reports roughly 10% fewer births. The deficit is higher regarding deaths [31].

The quality of the birth weight information used must be judged knowing how it was obtained. Mothers have reported the information they received in hospitals or community health centers immediately after delivery. In the Commune Health Centers and the District hospital, the newborn children were weighed and measured directly after delivery wearing light clothes. The birth weight, recorded in the clinical records of the district hospital or in the delivery notes of the commune health station, was adjusted for the weight of clothes.

In Vietnam, it is considered important to know and remember the birth weight of children. Three months after delivery, the field workers interviewed mother about birth weight, why recall bias should be low. We have also compared information of birth weight between the different kinds of interview, quarterly interviews in FilaBavi 1999–2010 and the quarterly interviews in FilaBavi and DodaLab 2008–2010 [25]. The results are that the estimated means of birth weight match very well for both sexes.

Birth weight was measured by different health staffs working differently and using different equipment. Poor standardization can create bias and excess random variation. This source of error is shared with many studies and difficult to eliminate in any context. The health staff may also report the birth weight to the mother incorrectly, e.g. with a tendency to “increase” for low weight babies and “decrease” for high weight babies as a wish to avoid extremes and thereby “please” the mother. This could lead to underestimation of the proportions of low birth weight children and to a biased estimate of the standard deviation (underestimated). Gifts to the health care staff, often monetary, are important parts of the Vietnamese health care system.

To explore possible bias a small study was done. For each year, 1999–2010, we collected the birth weight information for 30 randomly selected children from the records in the Bavi district hospital and compared the estimated means of birth weights with those obtained from the interviews with mothers. There were no trends over the 12 years in the birth weight means estimated using this particular data. The overall means were 3,110 g for boys and 3,047 g for girls. The corresponding estimates using the mother reports were 3,136 and 3,057 g, i.e. differences are 26 and 10 g. We also collected birth weight data for 40 children in each of two health stations for two years, 2005 and 2009. There was no large difference between the health station and the hospital data.

During the studied twelve years there were improvement of economic conditions and education as well as increased expenditure for health care and use of health technology in a rural area of Vietnam. Surprisingly, no positive secular trend in birth weight was observed in this period. Some reasons can be that the increased economic resources are not distributed in an equitable way or not used in the best possible way. The research was carried out in one rural area of Vietnam and the results may not be representative for other rural areas of Vietnam. To better understand the mechanisms behind this result, similar studies in other areas of Vietnam would be needed. In future studies birth weight information should ideally be obtained from validated delivery registers to avoid biases.