Abstract
Aim
To explore the prevalence of anemia in three cohorts of women, namely, married yet to be mothers, married and are mothers, and currently pregnant, to ascertain the patterns in anemia in women.
Methods
We analyzed a sample of 130,965 married women from four Demographic Health Surveys: 2000, 2005, 2009 and 2015. The primary focus for the analysis was married women aged 15 to 49 years. In the absence of a longitudinal data that followed the same women over the periods, a synthetic cohort of the women of that age-group was constructed to get women aged 15 to 64 years over the four surveys. Women who were aged 15 to 19 years in 2000 were the same as those 30 to 34 years in 2015, while those aged 45 to 49 years in 2000 were the same as 60 to 64 years in 2015.
Results
Logistic regression revealed that young mothers were significantly more infected (p < .001). Pregnancy affected anemia in the women (p < .001). Being younger and richer were associated with odds ratios of 0.599 (95% confidence interval, CI: [0.560, 0.640]) and 0.765 (95% CI: [0.726, 0.807]) for anemia, respectively. Being pregnant had odds ratio of 1.642 (95% CI: [1.439, 1.872]) for anemia.
Conclusion
Public health strategies should target social deprivation at the household level while addressing maternal health issues. An analysis of data on unmarried women and their children is recommended.
Anemia is basically defined as a red blood cell count, which is below the accepted lower level of the normal range. 1 It is a major global health problem, especially in developing countries, and affects the health, quality of life, and working capacity in billions of people all over the world. 2 The World Health Organization estimates that anemia affects nearly 2 billion people globally. 3 These figures indicate that in many countries, anemia constitutes a serious health problem with prevalence rates of ≥40. Although a complex of factors (nutritional deficiencies, different and disease condition), working in symphony have been identified to be responsible for anemia, nutritional deficiencies have been isolated as the single most common cause of anemia, the nutritional iron deficiency anemia being the commonest. 3 Iron deficiency anemia is considered to cause about 75% to 80% of the total burden of anemias.2–4 Preschool age children (47%), pregnant women (42%), and nonpregnant women (30%) are considered the worst victims of anemia, whereas prevalence in school-age children, men, and the older adults is put at 25%, 13%, and 245%, respectively. 4 It is thus surmised that in both developing and developed countries, the prevalence of anemia has a social imbalance, and more in the lower socioeconomic strata of societies.1,3
In Cambodia, anemia is blamed for the worsening maternal mortality ratios despite the strategic place of the eradication of anemia during pregnancy in safe motherhood initiative globally. 5–11 It is estimated that between 45% and 58% of women in Cambodia have a hemoglobin level (<11.0 g/dl), indicative of anemia. 12 Anemia-related maternal and childhood mortalities and morbidities could be prevented with timely and tested interventions if the patterns are known and promptly targeted. 13
However, the point at which anemia often set in among women is still unclear. Information on pattern and trajectory of anemia is necessary for presumptive interventions to protect women and children against anemia. Many studies published on pattern of anemia over time among married women do so within a cross-sectional context. Such studies merely present snapshots of anemia at a point in time in the life of a woman, usually during pregnancy, rather than focusing on trend over time and fail to tell whether changes that occur are due to the uptake of interventions or due to the characteristics that are known to vary across cohorts.
This lack of information is attributed, in part, to gaps in data. In assessing relative changes among cohorts overtime, there is a need for control in the different cohorts to remove biases resulting from time trends and unobserved factors. 14 According to Campbell and Hudson, studies of the determinants and effects of life events such as anemia and motherhood suffer from the fact that these are often events with some levels of peculiarity, and thus demand research designs that require unavoidable compromises. Some of these compromises introduce biases and one possible way of avoiding some of these biases is to form synthetic cohorts of women who have experienced anemia and motherhood by pooling observations from quasi panel surveys. 15 The synthetic cohort approach yields a multiwave data focused on anemia with fixed amounts of pre, current, and postpregnant women. This type of study of anemia among married women has never been published before, at least not within our literature search. This article presents the outcome of a synthetic cohort analysis of women aged 15 to 49 years between 1999 and 2014 to establish a trajectory of anemia patterns among married women, to see at what point, pre, during, or postpregnancy, they are susceptible to anemia.
Data and Methods
Data Source
There are many possible sources of data on anemia, which could be employed. However, constructing synthetic cohorts requires a source with historical data that provide information on changes in anemia. With the help of the Demographic Health Survey (DHS) stat compiler, a dashboard of anemia incidence in countries covered in the DHS was constructed. Establishing a trend required three or more data points taken at regular intervals. A number of countries showed multiple data points. There include Cambodia, Jordan, and Rwanda with four data points, respectively, and Peru with six data points. However, the intervals between data points for these countries were very irregular except for Cambodia. Table 1 shows Cambodia DHS as presenting the most useful historical data set with anemia covered in four DHS. The Cambodia DHS presented four data points with regular 5 years intervals in 20 years.
Dashboard of Demographic Health Surveys of Countries containing Data on Anemia Incidences in Married Women of Reproductive Ages.
Abbreviation: DHS, Demographic Health Survey.
To this end, the Cambodia DHS is the main source of data on anemia in mothers and children. The DHS contains detailed information on household health and has been conducted consistently for many years and will consequently constitute the source of data for this study. The DHS has been conducted four times, with regular 5 years interval, in Cambodia. It is one of the few DHS with data on anemia and with multiple data points for more than 10 years. Data will be extracted on women falling into the age groups 15 to 49 in DHS of 2000 to 2015.
Table 2 shows that those who were 15 years old in 2000 became 20 years old, 5 years later and 30 years old by 2015. By implication women who were 30 years old in 2015 represented those that were 15 years in 2000. Similarly, women who were aged 60 years in 2015 represented those that were aged 45 years in 2000. Three cohorts of women, namely, married yet to be mothers, married and are mothers, and currently pregnant constitute the study population.
Extraction of Data by Age and Year.
Data Tool
The uniform set of questionnaire employed for the DHS was used for data extraction. Focus was on:
Demographic information of mothers eligible for the study. Their child bearing experiences. Experiences with anemia. Practices on anemia.
Data Analysis
All data were analyzed with SPSS. Simple descriptive statistics were employed in characterizing the respondents. The Pearson’s R illustrated the relationship between certain sociodemographic variables and the anemia status of the women in the DHS. Logistic regression was conducted to test the predictability of the socioeconomic characteristics including the maternal and pregnancy statuses of the respondent on anemia. One-way and multiple analysis of variance (ANOVA) were done to establish differences in occurrence of anemia in women with diverse characteristics
Results
Demographic Characteristics of Respondents
A total of 37,965 married women were included in the analysis distributed in four overlapping cohorts shown in Table 2. More than three quarters (76.3%) of the sample were from rural area. Table 3 shows that 69.7% had less than secondary education. A majority (95.1%) practiced birth spacing of more than 3 years. A small proportion (2.1%) has never delivered any baby while 5.8% was pregnant at the time of the survey. There was a normal distribution around the third quintile with 19.1% of the sample. More of the older cohorts were found in the richest (fifth) wealth quintiles. Conversely, more of the younger cohorts fell within the poorest quintile.
Distribution of Respondents by Age Cohort and Personal Characteristics (% in Parenthesis).
Almost half (47%) of the women had anemia of any type. In addition, 10% had severe anemia and moderate anemia while 37% had mild anemia. Fifty-three percent was not anemic at the time of the survey. Furthermore, Figure 1 shows that incidence of anemia was inversely correlated with the age of the women. Maternal anemia rate declined with rise in the age cohort. Similarly, Figure 2 showed that childhood anemia declined with the age cohort of the mothers.
Trend in Maternal Anemia by Demographic Cohort.
Trend in Childhood Anemia by Demographic Cohort.
Factors Associated With Maternal Anemia
Table 4 presents the correlation of maternal anemia with some personal characteristics of the married women in the sample. For instance, it shows significant correlation with the different age cohort (p < .001). Similarly, more of the rural dwellers (20.7%) than the urban dwellers (17.8%) suffered anemia (p < .001). It also showed that those with lower levels of education than those with higher levels of education were anemic (p = .011). The married women who have never delivered a baby (30.5%) were more anemic than their counterparts who have delivered babies (19.8%) and this was significant at p<.001 level. The correlation of incidence of any form of anemia with birth spacing practices, wealth index, and pregnancy status were all significant at p<.001 level. There was also a significant correlation between maternal anemia and childhood anemia at p<.001 level of significance.
Distribution of Respondents by Their Experience of Any Anemia by Their Personal Characteristics.
Abbreviation: DHS, Demographic Health Survey.
Variables, which correlated significantly with any form of anemia, were included in a multivariate binary logistic regression model. Before running the regression, the variables were transformed into binary variables. For instance, the age cohorts assumed the value of “1” if age cohort is 15 to 54 years else age cohort was “0.” Residence was “1” if the respondent lived in urban area else residence was “0.” Wealth index was “1” if richer or richest, else “0.” Education was “1” if less than secondary schooling else “0.” Birth spacing was “1” if <2 years “0.” Delivery experience was “1” if the respondent is a mother, else “0.”. Similarly, pregnancy status was “1” if the respondent was pregnant at the time of the survey, else “0.”
Unlike the multiple bivariate analysis in Table 4, education and urban residence did not influence the anemia experiences of the mothers. On the other hand, being a poor currently pregnant young mother was significantly associated with maternal anemia. Put more clearly, being a young mother was significantly associated with being infected with any form of anemia (p < .001). The richer the mother the less the incidence of anemia (p < .001) and being currently pregnant was directed associated with being anemic (p < .001; see Table 5).
Logistic Regression: Binary Logistic Regression on Personal Characteristic of Respondents Associated With Any Maternal Anemia.
Abbreviations: CI, confidence interval; OR, odds ratio.
Further analysis employing a one-way ANOVA revealed that there is a significant difference between mothers who were currently pregnant and those who were not (F = 242.904; p < .001). Similarly, there was a significant difference in anemia incidence among the different cohorts of mothers (F = 91.260; p < .001). In terms of child-bearing experience, a one-way ANOVA revealed that there was no significant difference between married women who are mothers and those yet to be mothers (F = 0.485; p = .486). This corroborates the results from the multivariate logistic regression.
Discussion and Conclusion
There has been stress on the point that anemia is a major challenge to maternal and child health globally.9,10,15–19 However, another school of thought has argued that rather than pregnancy, other factors may play a role because of social assortative mating due to phenotype and cultural factors 20 and similarity in socioeconomic backgrounds. This school concludes that beyond the pregnancy-related pathways to childhood nutrition and anemia, there are other pathways such as the environment, household income, and parental nutritional practices, including the father’s. All of these must be considered for an effective intervention. Given this debate, this study set out to examine the nature of the relationship between pregnancy and anemia in married women. It proposed to answer the following questions: what is the trend in anemia in different cohorts of married women and their children?; is there any difference in anemia rates as women grow older?, and identify any discernable trajectory of anemia with age in the different cohorts of women. It also set to establish any association between pregnancy and childbirth with anemia in the women and their children and ascertain if there is any difference in anemia trends between women and their children.
The results show that as married women grew older, the rate of anemia tended to drop in both the mothers and their children. Severe anemia in the married women dropped from 1.6 in the youngest cohort of 15 to 49 years to 0.9, 0.3, and 0.2 in the older cohorts of 20 to 54 years, 25 to 59 years, and 30 to 64 years, respectively. Similarly, moderate anemia rates in the women dropped from 14.6 in the youngest cohort of 15 to 49 years to 12.2, 7.5, and 6.4 in the older cohorts of 20 to 54 years, 25 to 59 years, and 30 to 64 years, respectively. A one-way ANOVA test of difference in means revealed that this observation was statistically significant with an F-statistics of 91.260 (p < .001). This finding revealed that anemia is more in the younger age cohorts where child-bearing activities were more, thus suggesting that anemia is exacerbated by the pregnancy experiences of the mothers. Previous studies have also arrived at the same conclusions demonstrating that the processes and complications of pregnancy expose women to anemia. These studies argue therefore that high proportions of pregnant women are susceptible to anemia.3,9,10,16,21,22
In this study, maternal anemia was equally associated with childhood anemia. A similar trend in anemia was reported in the mothers and their children. For instance, severe anemia was highest (2.2) to the children born to the mothers in the youngest age cohort of 15 to 49 years. It dropped to 0.9, 0.6, and 0.3 as the ages of the mothers advanced to 20 to 54 years, 25 to 59 years, and 30 to 64 years, respectively. However, the rates of moderate anemia in the children of the women initially rose from 27.2 recorded among the youngest (15–49 years) age cohort of mothers to 29.0 in the children of women in the second age cohort (20–54 years) but later dropped to 21.8 and 21.1 in the older age cohorts of 25 to 59 years and 30 to 64 years, respectively. The difference was statistically significant (F = 27.977; p < .001).
This result corroborates earlier findings on intergenerational studies that suggested direct relationships between maternal and childhood anemia. They stress the impact of anemia during pregnancy on the development of anemia in the newborn, because maternal iron is positively correlated with ferritin from umbilical cord blood suggesting a causal relationship in maternal anemia and childhood anemia. 23 Some have also argued the direct links between antenatal anemia in mothers and low birth weight and prematurity, both of which increase the risk of childhood anemia. 23 Furthermore, it is argued that the failure of the mother to adequately breastfeed the child increases the risk of offspring anemia, which is exacerbated by the reality of their shared socioeconomic environment, and dietary quality between mother and child.23–26 There is also the evidence for the genetic transmission of anemia, which derives from heritability studies. 24 For instance, the study on the genetic influences on F cells and other hematologic variables revealed that additive genetic effects account for 37% and 42% of hemoglobin and red blood cell count, respectively. Other studies demonstrated that normal Caucasians have higher hemoglobin levels than Black individuals matched for age and sex. 20 There is thus a mechanism to link maternal hemoglobin/anemia with childhood anemia, which has been supported with the findings from this study, where a Pearson correlation also revealed direct association between anemia in mothers and those in their children (p < .001).
Furthermore, the study also revealed significant difference between the pregnant and nonpregnant women with respect to anemia. Women who were pregnant at the time of the survey reported higher prevalence of anemia (57.0%) compared with their counterparts who were not pregnant (44.6%) with a p-value of <.001. Further test of difference using one-way ANOVA gave an F-value of 242.904 and p value of <.001, both suggesting significant difference.
All the same, there was no significant association between being a mother and the prevalence of anemia. A test of difference in the rate of anemia in the married women who were mothers and those yet to be mothers proved that there is no significant difference. Instead, other factors were found to mediate the prevalence of anemia in the women. Some of such factors included the socioeconomic standing of the women (wealth quintile) and birth spacing practices among the women influenced the prevalence of anemia in the women. Thus the multiple binary logistic regression analysis showed that being young, poor, practising short birth spacing, being a mother and currently pregnant are the major drivers of anemia in women of reproductive ages. In other words, younger poor mothers practising short birth spacing and currently pregnant are more susceptible to anemia than their counterparts from richer home.
Other studies have equally highlighted the important associations of household wealth and maternal cum offspring anemia.23,27 These results point to a direct link between broader socioeconomic conditions and hemoglobin levels in children and their mothers, attributable to shared malnutrition, 28 deficiencies in other micronutrients, 29 exposure to biofuel smoke, 30 and other possible mechanisms associated with lower socioeconomic status. 20
It could be argued that household financial crises may threaten the health status of children in low- and middle-income countries 31 and may play a role in maternal and offspring anemia through its effect on food insecurity. This will worsen if the stressors listed earlier undermine socioeconomic advancement or worsen food insecurity in Cambodia. Interventions that support nutrition and address socioeconomic conditions may help mitigate these phenomena. Our data provide major insights into household, and economic factors associated with anemia in mothers and their children.
The results should, however, be taken with the following limitations in mind. First, this is a synthetic cohort analysis of multiple cross-sectional study, for which we report association rather than causation. It was impossible to do longitudinal studies. All the same, the synthetic cohort analysis provided the opportunity of following proxy women who have progressed through the different age ranges. Similar to cohort analysis, but instead of using successive observations of the same group of people, the age distribution of the married women is treated as if it were a cohort passing through time. This yields different result from true cohort analysis in periods of rapid change. However, given the dearth of such longitudinal data, synthetic cohorts have provided the basis for most commonly used measures of demographic behavior such as life expectancy, total fertility rate, and median age at marriage. 32
In conclusion, this study has successfully answered the questions it set out to address. It showed that pregnancy is a major driver of anemia in women of reproductive health, and special efforts should be made to protect the women against anemia during pregnancy. It also showed that younger women of reproductive ages were more susceptible to anemia. This is understandable because as the women grew older their reproductive activities witness sustained decline. It also showed some links between maternal and offspring anemia and at the same time demonstrating the influence of socioeconomic realities of the mothers and anemia. This called for an expansion of any intervention to address anemia in women cover the household and the socioeconomic environments.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
