Underweight and associated factors among lactating women in Uganda: Evidence from the Uganda demographic health survey 2016

Abstract Background Lactating mothers are at increased risk of being underweight because of the physiological changes that lead to disproportionately higher energy and nutrient requirements compared to their non‐pregnant and non‐lactating counterparts. Objective We aimed to determine the prevalence and factors associated with being underweight among lactating women in Uganda. Methods We used the Uganda Demographic and Health Survey (UDHS) 2016 data of 1356 women aged 20 to 49 years. Multistage stratified sampling was used to select study participants. The data were collected using validated questionnaires. We used multivariable logistic regression to determine factors associated with underweight among 20 to 49‐year‐old lactating women in Uganda. Results The prevalence of underweight was 8.2% (111/1356) (95% confidence interval, [CI]: 7.0‐10.0). Women who had no education were 10.21 (adjusted odds ratio, [AOR] = 10.21; 95% CI: 1.61‐64.74) times as likely to be underweight as those who had higher (post‐secondary) education levels. Women who were not working were 50% (AOR = 0.50; 95% CI: 0.26‐0.94) less likely to be underweight compared to those who were working. Women in the Western (AOR = 0.15; 95% CI: 0.07‐0.32), Eastern (AOR = 0.34; 95% CI: 0.18‐0.66), and Central (AOR = 0.30; 95% CI: 0.12‐0.74) regions were 85%, 66% and 70% respectively less likely to be underweight compared to those in the Northern region. Conclusion Based on the findings of this and other studies, it is important for the different stakeholders to design targeted nutrition programs for lactating women particularly those with low levels of education and those from the Northern region.


| INTRODUCTION
Maternal and child nutrition are good indicators of a society's overall wellbeing. 1 Globally, about 10% of women aged 20 to 49 years are underweight 2 with the greatest burden observed in low-income countries. 3 Underweight is considered an indicator of undernutrition in an adult with no underlying comorbidities and is defined as body mass index (BMI) below 18.5 kg/m 2 . 4,5 Pregnancy and lactation are nutritionally demanding periods due to increased calorie and nutrient requirements. This can be detrimental to a woman's health if food intake is not commensurate. 1,6 The World Health Organization (WHO) recommends up to 2 years of breastfeeding, with exclusive breastfeeding for the first 6 months of a baby's life. 7 The production of breast milk largely depends on the nutritional status of the mother, hence the need for adequate nutrition during lactation. 7 In addition, this period is enough to ensure full recovery from a preceding pregnancy and reversal of the associated physiological changes, the WHO recommends a birth-spacing interval of at least 24 months. 8 Maternal underweight negatively affects the quality and quantity of nutrients in breast milk leading to increased risk of child morbidity, mortality, and adverse long-term effects on the child's health. 9,10 Undernutrition among lactating mothers induces nutrition-related metabolic disturbances in early infancy and irreversible physiologic alterations in infants. 11 In low-income countries, women are at a high risk of unmet nutrient requirements because of inadequate food supply mainly attributed to financial constraints. 12 Most women in Uganda are economically disadvantaged; they have less control of resources such as land, are much more involved in unpaid care work, their work hours are longer compared to men, and they have less access to credit services. 13 16,17 Uganda's health system has six levels ranging from the highest level of national referral hospitals to the lowest level at the community level. 18 Agriculture contributes about 24% of the country's gross domestic product (GDP), providing half of the export earnings, and is the primary source of income for the majority of Ugandans. 17

| Study sampling and participants
Samples were collected using a stratified two-stage cluster sampling design with census enumeration areas as the primary sampling units. 15 The first stage of sampling involved selecting 697 enumeration areas (EAs), including 162 urban and 535 rural enumeration areas selected from the list of the 2014 population and housing census sample frame. 15 One enumeration area in the Acholi region was excluded due to land disputes hence ending up with 696 EAs.
Enumeration areas with over 300 households were segmented and only one segment was selected with probability proportional to the segment size as this helped minimize the burden of the household listing. 15 The enumeration areas that were involved in the survey were chosen independently from each stratum with probability proportional to size. The second stage of sampling involved the selection of households through equal probability systematic sampling. A list containing all households and maps in the selected enumeration area was made available, and households that were in institutional living arrangements were excluded. 15 Women aged 15 to 49 years who were either permanent residents or slept in the selected household the night before were eligible for inclusion in Uganda's demographic health survey 2016. 15 During the survey, anthropometric measurements were taken by trained technicians in about a third of the sampled women. 15 Our secondary analysis only considered 20 to 49-year-old lactating women and excluded 15 to 19-year-old women (adolescents) because the recommended anthropometric indicators for assessing underweight for those above 20 years (BMI and Height) are different from those of adolescents (BMI for age and Height for age) and cannot be directly compared.
Of the 18 506 women who consented and filled in the questionnaires, 14 242 were aged 20 to 49 years and of these, 4731 were selected for anthropometric assessment of which 4640 had their measures taken. From this group, 1356 were lactating women (Table 1), and 289 of them were excluded from logistic regression analysis because of being overweight or obese, leaving a final weighted sample of 1067 women. The sample selection process is summarized by the consort flow diagram in Data S1. To maintain the representativeness of the sample and account for possible differences in response rates across regions, sampling weights were used.

| Outcome variables
In this study, underweight was the primary outcome variable of interest, defined as body mass index (BMI) less than 18.50 kg/m 2 . 4 Other BMI categories were defined as follows; normal = between 18.50 and 24.99 kg/m 2 , overweight = between 25.0 and 29.99 kg/m 2 and obesity = above 29.99 kg/m 2 . 5 In the logistic regression analysis, we only considered underweight women and those with normal BMI and excluded those who were overweight and obese.

| Explanatory variables
This study included determinants of underweight based on evidence from the available literature. These factors were divided into individual (age, marital status, working status, frequency of antenatal care [ANC] utilization and education level), household (wealth index, household size and sex of household head), and community (region and residence) levels. The "Wealth index" is a measure of relative household economic status and was calculated by DHS from information on household asset ownership using Principal Component Analysis. 15,19 Different household assets were used to calculate separate wealth indices for rural and urban areas which were combined to form a national wealth index and then divided into quintiles namely; the poorest, the poorer, the middle, the richer, and the richest. 15,19 Place of Residence was aggregated as urban or rural.
The national region was categorized as Northern, Central, Eastern, and Western. Level of Education was categorized into; no education, primary education, secondary, and higher education. Age was categorized into three groups; 20 to 29, 30 to 39, and 40 to 49 years. Household Size was categorized as less than six members and six and above members based on the national and dataset average of six members per household. The sex of Household Head was categorized as male or female. Working status was categorized as: not working and working. Marital Status was categorized into married, and this included those in formal and informal unions, and not married. ANC utilization frequency was divided into less than four visits and four and above visits.    15.8% (95% CI: 13.7-17.5) and obesity was 5.6% (95% CI: 4.4-6.9). Table 2 shows the distribution of underweight by sociodemographic characteristics.

| Factors associated with underweight
In the final logistic regression model, factors associated with underweight were region, education level, and working status as shown in Table 3. Women who had no education were 10.21 times more likely to be underweight compared to those who had higher (post-secondary) education level. Women who were not working were 50% less likely to be underweight compared to those who were working.
Women in the Western, Eastern, and Central regions were 85%, 66%, and 70% less likely respectively to be underweight compared to those in the Northern region.

| DISCUSSION
This study investigated the prevalence and factors associated with underweight among lactating Ugandan women aged between 20 and 49 years. Based on the anthropometric assessment, 8.2% (95% CI: 7.0-10.0) of the lactating women were underweight, a prevalence that is within the standard acceptable malnutrition rate of less than 10%. 22 Of the 8.2%, 3.6% were severely underweight, 20.5% were moderately underweight and 75.9% were mildly underweight. This prevalence is lower than studies conducted among lactating women in Ethiopia 6,9,12 and Vietnam 14 but higher than a study done in Nigeria. 21 The difference in the prevalence of underweight between the current study and the other studies above could be attributed to the differences in the sociodemographic and economic characteristics between these study areas.
Underweight was significantly associated with working status, education level, and region. Women who belonged to the western, eastern, and central regions of the country were less likely to be underweight compared to those in the Northern region. Region of residence has also been shown to be associated with undernutrition in similar low-income African settings. 4 Women who were not working were 49% less likely to be underweight compared to those who were working. Working status has been found to be associated with being underweight in Ethiopia, Senegal, and Cambodia. 12,28,29 The workload due to laborintensive activities increase energy expenditure which further predisposes these women to underweight. Similarly, a study done among lactating Senegalese women showed weight loss due to negative energy balance associated with agricultural labor, 28 which agricultural activities are the commonest source of income in Uganda. 17 Another possible reason could be tight working schedules that affect the dietary patterns of lactating mothers, which coupled with the increased nutrition needs, makes lactating mothers prone to underweight. 30,31 Similar to studies done in Ethiopia, 12 Nepal, 2 Tanzania, 4

| Strengths
Standardized procedures are a requirement of DHS surveys in data collection and validated questionnaires were used which ensured the internal and external validity of the results.
Secondly, we used a nationally representative sample and weighed the data for analysis and therefore our results are generalized to all Ugandan women aged 20 to 49 years.

| Limitations
The cross-sectional design is limited by lack of temporality hence causal inferences cannot be made. Most data on the predictors were based on self-reporting and could not be verified through records, which could have led to socially acceptable answers hence information bias. Other important predictors of underweight such as dietary diversity score, nutritional knowledge, co-morbidities, and food security were not collected.

| CONCLUSION
Our study established that the factors associated with being underweight among Ugandan lactating women were level of education, working status, and region. Based on the findings of this and other studies, it is important for the different stakeholders to design targeted nutrition programs for lactating women with special emphasis on working, women from the Northern region, and those with low levels of education.

ACKNOWLEDGMENTS
We thank the MEASURE DHS program for availing us with the data.

FUNDING
No funding was obtained for this study.

CONFLICT OF INTEREST
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

TRANSPARENCY STATEMENT
The manuscript is an honest, accurate, and transparent account of the study being reported; and no important aspects of the study have been omitted.