Productive Safety Net Program and Dietary Practice Among Pregnant Women in Kacha Birra District, Southwest Ethiopia: Comparative Cross-Sectional Study

Background: Morbidity and mortality associated with poor dietary practice among pregnant women are a common public health burdens especially in low income countries like Ethiopia, for which, household food insecurity can be a determinant factor. Productive safety net program (PSNP) is implemented in Ethiopia to alleviate this problem with emphasis on pregnant and lactating mothers even though little is known about the dietary practice among pregnant women in the program user and non-user households. Therefore, this study is aimed to compare the proportion of good dietary practice and associated factors among pregnant women in productive safety net program user and non-user households in Kacha Birra District. Method: A Community based comparative cross sectional study was conducted in Kacha Birra District from March 12 to April 20, 2019. A total of 715 pregnant women were included by simple random sampling technique. A pretested and structured interviewer administered questionnaire was used to collect the data. Data was entered into EpiData version 3.1 and transferred to Stata version 14 for analysis. Binary logistic regression model was tted to identify factors associated with dietary practice. Crude odds ratio and adjusted odds ratios with 95% condence intervals were calculated to assess the strength of associations and signicance of the identied factors with dietary practice. Result: The overall proportion of good dietary practice among pregnant women was 14.97% (95%CI: 12.3-17.6), while it was 11.17% (95%CI: 6.5-15.8) and 16.23% (95%CI: 13.09-19.36) among pregnant women in PSNP user and non PSNP user households respectively. Dietary knowledge (AOR=2.64, 95%CI:1.67-4.18), nutrition information (AOR=2.36, 95%CI:1.41-3.95) and husband occupation (AOR=3.51, 95%CI:1.02-12.08) were signicantly associated with dietary practice of the pregnant women. Conclusion: The proportion of good dietary practice of the pregnant women in the study area was low with no signicant difference among the comparative groups. Dietary knowledge, nutrition information and husband occupation were signicantly associated with the dietary practice. So, giving due attention to


Background
Pregnancy is a state associated with many physiological changes like: increased plasma volume, formation of new fetal tissue, growth of existing maternal tissues and deposit of energy in different tissues of the body. It is associated with increased demand of both macro and micronutrients that makes pregnant women highly vulnerable for malnutrition [1][2][3].
Globally, the burden of maternal and child under nutrition is very high, affecting hundreds of millions of pregnant women and young children [4]. More speci cally, micronutrient de ciency among pregnant women is common across regions and countries of the world. Anemia and vitamin A de ciency affect approximately 32 million and 19 million pregnant women worldwide respectively [5]. Almost 40% of pregnant women in developing countries are anemic [6]. This high burden of maternal under nutrition is more signi cant in developing countries of south Asia and Sub-Saharan Africa [7]. Maternal under nutrition in Africa also ranges from 5%-20% [8].
In Ethiopia, the rate of chronic energy de ciency among women is a signi cant burden [9], whereas, anemia is detected on 33.6% of the pregnant women [10]. After all, by a systematic review of studies from the developing countries, inadequate dietary practice during pregnancy has been identi ed as one of the causes of maternal under nutrition and death during pregnancy [11].
Maternal dietary practice during pregnancy is thought to be one of the factors in uencing the health of both the mother and her growing fetus, posing a great burden of morbidity and mortality for both the mother and the fetus [12], abnormal fetal programming so that increasing the future risk of chronic diseases [13], poor birth outcomes [14,15] and perpetuation of intergenerational cycle of malnutrition [16].
Previous community based studies conducted in Ethiopia showed that, the proportion of good dietary practice ranges from 26.9-40.1% [2,15,17,18], for which, sociodemographic and economic factors [17], dietary knowledge [15], attitude of the pregnant women for diet [19], exposure to nutrition information [3,12,15] and pregnancy related factors [2,20] were explained as a determinant factors even though their signi cance shows variation from setting to setting. Also, as explained by WHO report and ndings from studies, food insecurity is one of the major determining factors for dietary practice of pregnant women, increasing the burden of fetal and maternal risk [21,22]. To alleviate this problem, several nutrition speci c interventions like: micronutrient, most commonly Iron folate supplementation and supplementary feeding for malnourished pregnant women are undertaken in Ethiopia [23] and showed some improvement in the nutritional status and prevalence of anemia among the pregnant women through time [9]. Along with the nutrition speci c interventions, one of the nutrition sensitive interventions, Productive Safety Net Program (PSNP), was started to be implemented to support food insecure HHs living in chronically food insecure areas through public works employment and unconditional transfers with the aim of improving food security status of the households and avoid depleting their productive assets to meet their basic food requirements [24,25]. The 2015 revised PSNP started to provide temporary direct support and linkage to different social services for pregnant and lactating women in food insecure HHs [23] to improve their dietary practice and nutritional status as well as their new born child.
However, little is known about the proportion of good dietary practice and factors affecting it among pregnant women in PSNP bene ciary areas. Also, there is information gap, regarding difference in the proportion of good dietary practice of pregnant women in PSNP user and non PSNP user HHs in the country as well as in the district. Therefore, this study was aimed to compare the proportion of the dietary practice of pregnant women in PSNP user and non PSNP user HHs in Kacha Birra District and to identify factors affecting it for further intervention.

Study design and setting
A community based comparative cross sectional study design was conducted from March 12 to April 20, 2019 in Kacha Birra District, one of the administrative districts found in Kembata Tembaro zone, South nation nationalities and peoples region (SNNPR), Southwest Ethiopia. The district covers an area of 1666.724 sq. km and has 21 kebeles. The capital town of the district is Shinshicho town, which is located 293 km from Addis Ababa, the capital of Ethiopia. The district has 1 primary hospital, 4 health centers and 23 health posts providing services including maternal and child care. The total number of population in the district was 153,677, of which, 5178 were pregnant women [26].
Households in the area depend both on livestock and crop production. As the crop production is completely rain-fed, the area suffers of a food de cit every year. The district comprises both highland and lowland areas, where the lowland areas are more prone to drought and food insecurity. PSNP was started to be implemented in the district in 2005. All kebeles of the district are bene ciary since the very beginning of the program. Currently, the PSNP in the district is delivered in two modalities; cash is given to households with able bodied members who can repay it by engaging in any developmental activities, like water and soil conservation, road construction, etc. On the other hand, food aid is given to households with no able-bodied members. According to the 2011 district disaster and risk management o ce early warning and response cluster report,16,563 people of the district were PSNP users [27].
Study population and sampling procedure All pregnant women in PSNP user and non PSNP user HHs who had lived at least for six months in the study area were included in the study. As the study was a comparative type, the minimum sample size was determined by using the double population proportion formula that the two groups were classi ed based on their PSNP bene ciary status. Group one were pregnant mothers from PSNP user HHs, considered as exposed and group two were pregnant mothers from non PSNP user HHs, considered as not exposed. To estimate the minimum sample size, the proportion of good dietary practice 26.9% was used, which was obtained from previous study conducted in non PSNP user HHs [17]. Since there was no previous study conducted among pregnant women in PSNP user HHs, the proportion of good dietary practice among pregnant women in PSNP user HHs was assumed to be decreased by 15% from that of non PSNP user HHs, in fact that PSNP user HHs are food insecure and have low access to resources, so that p1 was 11.9%. The nal sample size was calculated by using Epi Info software by using 95% con dence level, 80% power, ratio of unexposed to exposed (r) = 3, design effect of 2, and non-response rate of 10%. The nal minimum sample size was 739. As the ratio of exposed to unexposed used in the study was 1:3, the sample size for the exposed group (n1) was 185 and for the unexposed group (n2) was 554.
Of the total 21 kebeles found in the District, eight kebeles were selected by using simple random sampling (SRS). Then, HHs with pregnant mothers in each selected kebele were strati ed based on PSNP bene ciary status. Since there was sampling frame of the pregnant women in the health post of each kebele, pregnant women from both PSNP user and non PSNP user HHs of each selected kebele were selected by simple random sampling after proportional allocation of the sample sizes calculated to the two groups to the total number of pregnant mothers in each stratum (users and nonusers) of each selected kebele. Attitude: It is the pregnant women's emotion, motivation, perception and thought which in uence her eating behavior [19].

Variables
Favorable attitude: is when the respondents' attitude score > median [19].
Have nutrition information: if the pregnant mother heard or was advised to have a healthy diet, to eat more and to eat different vegetables and fruits at any time in current pregnancy [29].
Wealth index: It is analyzed by using Principal Component Analysis(PCA) and classi ed as low, middle and high [30].
Data collection tool and quality control Data from the pregnant women were collected through home to home visit using a structured interviewer administered questionnaire. The questionnaire was designed to capture socio-demographic characteristics, health service utilization, pregnancy related characteristics, nutrition information during the current pregnancy, dietary knowledge of pregnant women and attitude about diet of the pregnant women.
The questionnaire was rst prepared in English, translated to local language (Kembatgna) and then retranslated back to English to keep consistency. The data was collected by six diploma nurses, who uently speak both Amharic and the local language (Kembatgna) and two health o cers were recruited for the supervision purpose. To maintain the data quality, two days training was given to the data collectors and supervisors and the questionnaire was pre-tested on 5% of the actual sample size in nonselected Kebele. Frequent and timely supervision of the data collectors was done by the supervisors. Data was monitored and checked for completeness on daily basis by the data collectors as well as by the supervisors and by the Principal investigator (PI) during the data collection. To minimize the social desirability bias, the pregnant women were informed about the objectives of the study in detail and the information they give is handled con dential.

Methods of assessment
For assessing the dietary practice, ten questions were used and each question was given one mark if the answer was correct and zero score was given if the response was wrong [2,15]. The score was obtained by summation of responses of each question and converted to percentage, then if the respondent answered less than 75% of the ten questions classi ed as poor dietary practice. Also for assessing dietary knowledge of the respondents, ten questions were used and each question was given one mark if the answer was correct and zero score was given if the response was wrong [2]. The score was obtained by summation of responses of each question and converted to percentage and then classi ed as poor dietary knowledge if the respondent answered less than 70%. Dietary attitude was assessed using nine questions obtained from previous study and converted to percentage. Since the data was skewed, median was calculated to consider it as favorable or unfavorable [19] .
Data processing and analysis   practice. Relative to the pregnant women whose husband was daily worker, the proportion of good dietary practice was 3.5 [AOR = 3.51(95% CI:1.02-12.08)] times higher among pregnant women whose husband was employed (Table 3)

Discussion
Mothers' sub optimal dietary practice during pregnancy negatively affects the growth of the fetus [3] and therefore predisposes to low birth weight, premature birth and prenatal mortality. It also adversely affects the health of the pregnant women [31].
This study showed that only 14.97% of the pregnant women had good dietary practice during pregnancy, which was lower than the results from Guto Gida district(33.9%)[18], Bahir Dar city(39.3%) [2], Gondar town(40.1%) [15], Addis Ababa city(34.5%) [19] and Ambo district(26.9%) [17], Ethiopia. Also, it is lower than the results from different African and world countries: Nigeria(56.6%) [3], Kenya(25.9%) [20], Malaysia(74%) [32] and Swaziland(51%) [33]. The possible reason for this disparity may be low coverage of knowledge about maternal nutrition in the study area relative to that of from the above listed studies(for example, those studies in Gondar town [15], Bahir Dar [2], Malaysia [32] and Swaziland [33] have relatively high coverage of knowledge about maternal nutrition than that of the study area), variation in the residence and study setting (for example studies conducted in Addis Ababa [19], Kenya [20], Swaziland [33] and Nigeria [3]), while this study was conducted in rural setting and was community based) and dissimilarity in the sample size will also be an important determinant for the difference. Furthermore, the difference may also be due to the fact that the study area is chronically food insecure, while the above listed study areas are food secured.
This study also revealed the proportion of good dietary practice among pregnant mothers in PSNP user and non PSNP user households. Accordingly, 11.17% and 16.23% of the pregnant women residing in PSNP and non PSNP user households had good dietary practice respectively. Even though the study indicated that the proportion of good dietary practice among pregnant women residing in non PSNP user households was higher, the difference was not statistically signi cant (p = 0.1). The probable reason for the absence of statistically signi cant difference in the two groups may be due to subjectivity in inclusion of the HHs to the program so that those who should be included in the program may be left out. Moreover, the staple diet of the comparative groups was cereal based monotonous diet, resulting in low consumption of other food items on daily basis, rain dependent production of food items in all the selected kebeles and vulnerability to drought at the time of climate change can also be listed as factors that affect the presence of difference in the two groups. Furthermore, provision of targeted supplementary feeding (TSF) for pregnant mothers from the two groups whose MUAC is less than 23 cm at the stage of third trimester, not having big difference in the knowledge of the pregnant mothers about diet during pregnancy among the two groups and the small sample size from PSNP user households may mask the difference in the proportion of dietary practice in the two groups.
Concerning the determinant factors of dietary practice among the pregnant women, husband occupation was identi ed as a signi cant factor for the dietary practice of the pregnant mothers, which states that pregnant mothers whose husbands were employed had good dietary practice than those whose husbands were daily laborer, which is supported by evidences from Addis Ababa [19] and Ambo district [17]. It may be because of those employed husbands have relatively better educational status and knowledge level than the daily laborers, which may contribute a lot in giving information about the importance of having good dietary practice during pregnancy and increase a tendency to practice it. Also, those employed husbands have good income status so that able to purchase different kinds of food for the household and also may have relatively better support to their pregnant wives to have a healthy dietary practices than daily laborers.
In addition, nutrition information during pregnancy showed signi cant statistical association with the dietary practice among pregnant mothers, explaining that mothers having nutrition information had better dietary practice than those who had no exposure to nutrition information during their current pregnancy.

Declarations
Ethical approval and consent to participate Ethical clearance was obtained from the Ethical Review Committee of the Institute of Public Health, the University of Gondar. Also, o cial permission letter was obtained from Kacha Birra District administrative and health o ces after detail explanation of the purpose of the study. The objective of the study was brie y clari ed and informed consent was obtained from the pregnant women as all of them participated in the study were aged above eighteen. Any risk on participants during data collection was minimized and name or any other identifying information of the pregnant women was not recorded on the questionnaire and all information was kept strictly con dential. The information retrieved was only used for the study purpose. It was informed that participants who are unwilling to participate on the study have the right to quit at any time.

Consent for publication
Not applicable Availability of data and material