Undernutrition and Associated Factors Among Lactating Women in Dire Dawa Health Facilities, Ethiopia

Background: The nutritional status of lactating women is very important since it also affects the health of their children. However, there was limited information on maternal nutrition status in low-income countries like Ethiopia, especially in the study area. Objectives: Determine magnitude of undernutrition and associated factors among lactating women Methods: Institutional based cross-sectional study was conducted among 422 lactating women in Dire Dawa town health facilities from February 10/02/2019 – March 30/ 03/2019. Result: Prevalence of undernutrition was 22 %. Women who age 15-25 years were four times more likely undernourished than older [AOR=4.04(CI: (1.74, 9.40)]. Unable to read and write Women were almost ve times more likely to be undernourished than formal education [AOR=4.76 CI: (2.31, 9.81)]. Women who have family size >7 were six times more likely to be undernourished than family size < 3 [AOR=5.53 CI :( 1.15, 26.53)]. Women not take additional food during lactating were 4.5 times more likely undernourished than take additional food [AOR=4.56 CI (1.50, 13.9)]. DD score < 5 were four times more likely to be undernourished than (>= 5) DD [AOR= 4 CI: (2.02, 7.90)]. Conclusion: Prevalence of undernutrition in the study area was high: Factors associated with underweight were: Age of lactating women, Education status, Family size, Additional food during lactation and DD score. Thus, multi-sectoral collaboration targeted at improving women’s educational status and increasing food during lactation need to be emphasized.

population being undernourished or have low mineral and vitamin status due to lack of availability or inability to access to food of quality of diet. So providing a nourishing diet for pregnant and lactating mother results in signi cantly better infant health outcomes (WHO, 2014).
Lactating women produce between 700 to 800 ml of milk per day. This needs an extra 500 kilocalories energy provision per day in additional to normal 2200 kilocalories (Patricia, 2014). Breast milk is rich in the vitamins and minerals needed to protect an infant's health and promote growth and development. If the mother's diet is poorly diversi ed, the levels of micronutrients in breast milk may be reduced or the mother's own health may be affected (Allen, 2000).
Adequate nutrition during lactation is not only important for optimum growth, maturation, mental development, physical wellbeing and disease resistance, but also breaks the intergenerational link of

Study Area and Period
The study was conducted in Dire Dawa city administration which is located in the Eastern part of the Ethiopia. According to the Ethiopian central statistics authority's 2008 report, Dire Dawa administrative region has total population of 453,000 of whom almost one to one male to female ratio. It is situated 515Kms from Addis Ababa. Dire Dawa city administration is one of the two City Administration and majority of its population lives in urban area with 233,224 or around 68.22% of the population are urban inhabitants.The public health organizations which are involved in health care delivery include 1public referral hospital, 1 public district hospital and 8 public health centers. All health institutions provide maternal health service in addition to other service. The study was conducted from February 10/02/2019 -March 30/ 03/2019.

Study Design
Institutional based cross sectional study design was used.

Source population
All lactating mothers living in Dire dawa City who were visiting Dire Dawa Health facilities for postnatal care and EPI program was considered as the source of population.

Study population
Selected lactating mothers who visiting Dire dawa Hospitals and Health Centers for postnatal care and EPI program during data collection period were considered as study population.

Inclusion and exclusion criteria Inclusion Criteria
Those lactating mothers living in Dire Dawa City and who was visiting at Dire dawa Hospitals and Health centers for post-natal care and EPI program during the study period.

Exclusion Criteria
Lactating women who were critical ill, pregnant and physical deformity will be excluded from the study to improve quality of anthropometric measurements.

Data collection instrument
Data on socio demographic, dietary diversity and food security variables were collected by using . The adapted data collection tools prepared in English will be translated to Amharic language and again back to English to check consistency.
Food security was assessed using Household Food Insecurity Access Scale (HFIAS), it validated tool use in the Ethiopian context (Seifu et al., 2015). The HFIAS has nine questions asking household's last month experience about three domains of food insecurity: feeling uncertainty of food supply, insu cient quality of food, and insu cient food intake and its physical consequences. (Coates et al., 2007).
The mother's dietary intake pattern will be measured by a qualitative recall of all foods consumed by each woman during the previous 24 hours. Thus, certain food groups was aggregated to calculate Individual (women) dietary diversity score (WDDS) and the mean DDS will be used to classify mothers food intake as adequate or not (Kennedy et al., 2013).
To measure the outcome variable, anthropometric measurement (weight) of lactating women were measured to the nearest 100 g using portable electronic digital scale (Seca, Germany model) and height will be measured to the nearest 0.1 cm using a portable wooden height measuring board with sliding head bar through standard anthropometric measuring technique.

Data collectors and data collection procedures
The data will be collected by 8 nurses' work in the post-natal and EPI service and the data was collect at exit. Data collectors were trained for two days by principal investigator. Two BSC holders in Nursing or health o cer will be recruited and trained for supervising data collectors. Training was given about methods of anthropometric measurement, interviewing technique and lling questionnaires.
To measure weight of mother requested to remove shoe, wear light close and other supportive materials and data collectors were weigh the study participant on calibrated portable digital scale and value will be recorded to the nearest 100 gram or 0.1 kg.
To measure height the study participant was requested to stand erect with their shoulder level, hands was at the side, head, scapulae, buttock and heel were in contact with vertical measuring board with sliding head bar and height value will be recorded to the nearest 0.1 cm. (WHO 2012)

Data quality control
To assure the quality of the data, structured and pretested questionnaire was used. Pretest of the questionnaire was employed prior to actual data collection period among 5% of the study sample on one health center not included in this study. The nal version of the questionnaire which was prepared in English translated into the local language of the respondents (Amharic language) and again translated back to English. The data collectors and supervisors were given two day intensive training by principal investigator (PI) on the instruments, method of data collection, how to take anthropometric measurements and ethical issues.
Relative Technical error Measurement (%TEM) was done to minimize the random anthropometric measurement errors and relative TEMs for intra and inter examiners for weight and height was acceptable if relative technical error Measurement less than 1.5% and 2% respectively (Perini. et al., 2005). Functionality of digital weight scales will be checked using known weight every morning before data collection begin and before every weight measurement the data collectors were assure the scale reading exactly at zero (NHANES, 2007).
Intensive supervision were done by principal investigator and supervisor and they were check the collected data for completeness, accuracy, and consistency throughout the data collection period. The overall supervision was done by the principal investigator. Data double entry was used to make comparisons of two data cells and resolve if there is some difference.

Data processing and analysis
Data was coded and entered on to Epi-data version 3.0 and exported to SPSS Version 22 for analysis.
Missing values checked by conducting simple frequency analysis. Exploratory data analysis was done to check missing values, potential outliers and the normality distribution for those continuous variables.
Body mass index of the mother was calculated through weight in kilogram divided by square of height in meters and based on the result mother was categorized in to underweight with BMI less than 18.5 kg/m2, normal those having BMI 18.5-24.99 kg/m2, overweight with BMI 25-29.99 kg/m2 and obese those having BMI greater than or equal to 30 kg/m2 (WHO 2012). Since the interest was identifying lactating women at risk of undernutrition, the dependent variables are coded as 1 lactating women were undernourished (BMI <18.5 kg/m2) and coded as 0 if not.
Multi-Collinearity effect was checked and variables with SE >2 was removed from analysis and those variables have no collinear effect was included in binary logistic regression model to see the possible relationships with the outcome variables. Covariates with a p-value less than 0.25 in the bivariable logistic regression analysis was candidate for a multivariable logistic regression analysis to control potential confounders and to identify associated factors of undernutrition. The tness of the model was tested by Hosmer-Lemeshow goodness of t test (p-value=0.83). Odd Ratios along with 95% Con dence interval was estimated measure the strength of the association. Level of statistical signi cance was declared at p-value less 0.05. Results were presented using frequencies, summary measures, tables, and gures.

Results
Socio-demographic characteristics of study participants   (Figure 2).

Factors associated with undernutrition among lactating women
To determine the association between undernutrition and explanatory variables, bivariate analysis was performed using logistic regression. The result indicates that there is association between nutritional status of lactating women and some of the explanatory variables under the study such as age of lactating women, educational status of women, Age at rst pregnancy family size, frequency of ANC visit, birth interval for the index child, additional food during lactation, Number of feeding, women dietary diversity and levels of household food insecurity, and ( . Even if all these factor (described above) are not signi cantly associated with outcome variable in the multivariable logistic regression.
In multivariable logistic regression analysis, the covariates: age of lactating women, educational status of women , family size, additional food during lactation and women dietary diversity score were statistically signi cant at 5% level of signi cance and were found to be the predictors of undernutrition among lactating women in the study.

Discussion
The prevalence of undernutrition was 22 %. Age of lactating women, Education status of lactating women (those unable to read and write), Family size, additional food during lactation and women dietary diversity score were signi cantly associated with undernutrition.
The prevalence of undernutrition (BMI < 18.5 kg/m2) among lactating women in this study was 22 %. The present study has a rmed that age of lactating women was signi cantly associated with their nutritional status. Mothers within the age group of 15-25 had four times more likely to be undernourished when compared with those mothers in the age group of 36-49. This is in line with a result from study done in Ambo district indicate that lactating women in age group of 17-25 more exposed to This association could be due to the fact that women from family size higher were in di culties of meeting their dietary requirement, engaged in strenuous works like daily laborer and have minimal health care practice and also food security issue in women with higher family size and related underweight and nutritional depletion of the mother due to successive pregnancies.
Additional food during lactation was another factor associated with undernutrition. Accordingly, women who not got additional food during lactation were 4.56 times more likely to be undernourished compared to those who's got additional food during lactation. This had been also supported by a study done in Tigray (Ismael et al., 2017). This could be because during lactation more energy were required than from normal time even more than from pregnancy time, so if the lactating women did not get adequate food during lactation were more exposed to undernutrition.
Dietary diversity score of studied women was also another factor that affects their nutritional status in this study. Those women whose dietary diversity score below the mean (< 5) were four times more likely to be undernourished than women with the score a greater than or equal the mean (>= 5), implying that eating diverse food groups as a protective to undernutrition. This nding agrees with a study done in Dedo and Seqa-Chekorsa districts, Jimma Zone (Mihiretu et al., 2015). This might be because women with low dietary diversity may not get enough balanced diet which is essential for the human body to be immune and perform metabolic activities and also diversi ed food re ects dietary quality, improves daily nutrient and energy intake. Lactating women who do not get enough energy and nutrients in their diets risk maternal depletion and exacerbates women undernutrition (Allen LH. 2012).
The current study has the following limitations. An anthropometric measurement error was one of the limitations of this study: To minimize this; data collectors were well trained, standardization of anthropometric measures was done and the instrument was calibrated. Since some of the questions are asking the event that occurs 4 weeks back, there is a possibility that some of the responses might suffer from recall bias and this may affect prevalence estimates. This was minimized by probing the respondents about the event. The strength of this study, it has 96.5% response rate.

Conclusion
The result of this study indicated that prevalence of undernutrition (BMI <18.5 Kg/m2) among lactating women in the study area was high. The predictors of undernutrition among lactating women in the study area were: Education status of lactating women (those unable to read and write), Age of lactating women, Family size, additional food during lactation and women dietary diversity score.

Recommendations
Based on the nding of this study, the following recommendations were made: Should strength sustained health and nutrition education to the women, their families and communities regarding increased food intake, proper dietary practices, and dietary diversi cation during their lactation time. Should closely work with health extension workers to increase awareness of lactating women on how to improve their own nutritional status.
They should create strong multisector collaboration targeted at improving women educational status and increasing household food security status.
Should work with bene ciary programs like safety net complemented with nutritional intervention to food insecure households.
For further studies: Furthermore, longitudinal studies should also be conducted for a better understanding of the maternal dietary intake with relating to the nutritional status of lactating women.
Declarations Acknowledgement First of all, I would like to thanks Dire Dawa University for giving me opportunity to develop this research. I extend my appreciation for staffs of Sabian hospitals, and Gende kore, Goro and Legehare Health centers providing important information and for their unlimited co-operation and also I thank the data collectors, supervisors, study participants, and questionnaire translators; without whom the research would not be done.

Ethical considerations
Before starting of the data collection process, Dire dawa University School of medicine and college of health sciences Institutional Health Research Ethics Review Committee (IHRERC) was secured by ethical clearance and Dire dawa University was write O cial letter to hospital and health centers. Informed voluntary written and signed consent was obtained from each participant after explaining the purpose and bene ts of the study. The data collector were trained to respect the culture of the people in these communities throughout the data collection process. Con dentiality of the study participants' information also ensure.