Levels of mother-to-child HIV transmission knowledge and associated factors among reproductive-age women in Ethiopia: Analysis of 2016 Ethiopian Demographic and Health Survey Data

Background The world community has committed to eliminating the mother-to-child transmission of human immunodeficiency virus. Even though different studies have been done in Ethiopia, to the knowledge of the investigators, the Ethiopian women’s level of knowledge on the mother-to-child transmission of human immunodeficiency virus is not well studied and the existing evidence is inconclusive. The current study is aimed to study the Ethiopian women’s level of knowledge on the mother-to-child transmission of human immunodeficiency virus and its associated factors using the 2016 Ethiopian Demographic and Health Survey Data. Methods Data of 15,683 women were extracted from the 2016 Ethiopia Demographic and Health Survey. Descriptive statistics and multilevel ordinal logistic regression were respectively used for the descriptive and analytical studies. Results 41.1% [95% CI: 39.5%, 42.7%] of the Ethiopian reproductive-age women have adequate knowledge of the mother-to-child transmission of human immunodeficiency virus. 77%, 84% and 87.8% of the women respectively know that human immunodeficiency virus can be transmitted during pregnancy, delivery, and breastfeeding. There are wider regional variations in the women’s level of knowledge of the mother-to-child transmission of human immunodeficiency virus. Being an urban resident, having better educational status, being from a wealthy household, owning of mobile phone, frequency of listening to the radio, frequency of watching television, and being visited with field workers were significantly associated with having adequate knowledge of the mother-to-child transmission of human immunodeficiency virus. Conclusion Despite all collective measures put in a place by different stakeholders to prevent the mother-to-child transmission of HIV in Ethiopia, a large proportion of the Ethiopian women do not know about the mother-to-child transmission of the disease. Stakeholders working on HIV prevention and control should give due emphasis to promoting mobile phone technology and other media like radio and television by giving due focus to rural residents and poor women to promote the current low level of the knowledge. Emphasis should also be given to the information, education, and communication of the mother-to-child transmission of the disease through community-based educations.

112 million, Ethiopia is the second most populous nation in Africa following Nigeria. The country is among the fast-growing economy in the region; however, it is also one of the poorest, with a per capita income of $850. Administratively, the country is divided into nine regions and two city administrations [32,33]. Even though different small-scale studies have been done in Ethiopia to study women's level of knowledge on MTCT of HIV and its associated factors, to the best knowledge of the investigators, the problem is not well studied using nationally representative data, and the existing evidence is inconclusive. Hence, the current study is aimed to study the level of mother-to-child HIV transmission knowledge and its associated factors among reproductive-age women in Ethiopia using the nationally representative 2016 Ethiopian Demographic and Health Survey (EDHS) Data.

Data sources
The 2016 EDHS data were collected by the Central Statistical Agency (CSA) and other stakeholders both in Ethiopia and abroad. The authors accessed the processed and organized data from open datasets of the MEASUREDHS by permission. Variables anticipated to be associated with women's knowledge about mother-to-child HIV transmission were extracted from the 'women dataset' based on the reviewed literature and then processed for further analyses. For collecting the EDHS data, standard protocols and three types of tools; the Household Questionnaire, the Woman's Questionnaire, and the Man's Questionnaire were used. Further contextualization and standardization of the questionnaires were also done by governmental and non-governmental shareholders to maintain the validity of the tools [33].

Study population and sampling procedures for the 2016 EDHS
The 9 regions and 2 two city administrations in Ethiopia were considered based on the 2007 census that divided each kebele, the lowest governmental administration unit, to a subdivision called census enumeration areas (EAs). The survey followed a two-stage sampling design with stratification into urban and rural. At the first stage, 645 EAs, 202 from urban, and 443 from a rural were selected according to probability proportionate to the size of the EAs. At the second stage, approximately 28 households from each EA were selected by systematic random sampling and then all women whose ages were from 15-49 who live in the selected households were included in the study [34]. By this procedure, 15,683 eligible women were identified and interviewed during the parent study and so were considered for the current study.

Measurements
The dependent variable of the study was knowledge of mother-to-child HIV transmission among reproductive-age women and generated from the scoring of four questions each woman was asked. The questions were "the virus that causes AIDS can be transmitted from a mother to her baby during pregnancy"; "The virus that causes AIDS can be transmitted from a mother to her baby during delivery"; "The virus that causes AIDS can be transmitted from a mother to her baby by breastfeeding"; and "There are special drugs that a doctor or a nurse can give to a woman infected with the AIDS virus to reduce the risk of transmission to the baby" [33,35]. The scoring which ranges from '0' to '4' was done from the responses of the 4 questions and then grouped into 3 levels of knowledge of MTCT of HIV.

Operational definitions
No knowledge of mother-to-child transmission. If the score of the four measurement questions of MTCT of HIV summed to '0'.
Inadequate knowledge of mother-to-child transmission. If the scores of the four measurement questions of MTCT of HIV sum ranges from 1 to 3.
Adequate knowledge of mother-to-child transmission. If the scores of the four measurement questions of MTCT of HIV summed to '4' i.e. if they answered the four questions correctly.
Agricultural workers. Refers to those females who were market-oriented skilled agricultural workers, market-oriented skilled forestry, fishery and hunting, and agricultural, forestry, and fishery laborers.
Professional workers. Include chief executives, senior officials, and legislators, administrative and commercial managers, production and specialized services managers, science and engineering professionals, health professionals, teaching professionals, and other professionals were included.
Trade or sales workers. Encloses sales workers, building and related trades workers, excluding electricians, metal, machinery and related traders, handicraft and printing workers and, electrical and electronic traders.
Elementary occupation. Covers cleaners and helpers, laborers in mining, construction, manufacturing and transport, food preparation assistants, street and related sales and service workers and, refuse workers and other elementary workers.
Others workers. Consists of hospitality, retail, and other services managers, general and keyboard clerks, numerical and material recording clerks, other clerical support workers, personal service workers, personal care workers, protective services workers, handicraft and printing workers, food processing, woodworking, garment, and other craft and related trades workers.

Data analysis
Data processing, management, and analyses were performed using Stata 14.2 statistical software. The multilevel ordinal logistic regression model was fitted to assess regional variation of knowledge of women about MTCT of HIV and to identify factors associated with the outcome of interest for the target population of reproductive-age women in Ethiopian. The appropriate statistical method that can capture inflation of variability due to the application of staged sampling is multilevel analysis. Models used for the analysis of hierarchical data structure must account for associations among observations within clusters to make efficient and valid inferences. When the variance of the residual errors is correlated between individual observations as a result of these nested structures, single ordinal logistic regression is inappropriate [36]. Consequently, in this study, multilevel ordinal logistic regression was used to assess the relationship between levels of knowledge of HIV and associated factors using the 2016 EDHS data.
In the analysis of multilevel regression, the clustering effect plays a great role in the estimation of the parameters and this clustering effect can be quantified by intraclass correlation (ICC). ICC is the proportion of total variation in the response variable that is accounted for by between-group variation [36]. In this study, the effect of the clustering variable (region) where the subjects were residing during the study period was given an emphasis and all other predictors were considered at the first level.
All the outputs for descriptive as well as fitting multilevel ordinal logistic regression analyses were carried out by weighting provided by the MEASUREDHS program. The weights from DHS were used to carry out multilevel analysis but adjusted as per the recommendation by Adam [37]. Moreover, we have checked the goodness of fit after weighting the dataset by both the DHS and Adam's. Compared to the multilevel ordinal logistic regression fitted by using the unadjusted weights (AIC = 30,307.7, BIC = 30,392.0), the model fitted using the adjusted weights (AIC = 25,826.1, BIC = 25,902.7) had lower AIC and BIC. Besides the choice of weights, the model with fewer numbers of variables in the model was considered due to the principle of parsimony. The final model with significant variables is presented in Table 3.

Ethics statement
The EDHS 2016 survey protocol was reviewed and approved by the Federal Democratic Republic of Ethiopia, Ministry of Science and Technology, and the Institutional Review Board of ICF International. Additionally, written consent was obtained from each respondent. All participant identifiers were removed during data entry of the parent study, earlier of doing data management and any analyses. For the current study, the authors received permission from the public domain MEASUREDHS website and re-analyzed the data.

Characteristics of the participants
Fully, 15,683 participants were included in this study and the average age of the participants was 28.17 (± 9.16) year. On average, each woman had 0.7 (±0.84) number of births in the last five years and had 2.37(±2.34) mean number of antenatal visits during pregnancy. As presented in Table 1, regarding religion, 43.3%, 23.4%, and 31.2% of the participants were Orthodox, Protestant, and Muslim religious followers respectively. Most of the study participants (65.2%) were in a union by the time the survey was conducted, whereas about one in four were never been in a union. Half (50%) of the women were not engaged in a paid type of work and 20.8% were engaged in agricultural works. The majority of the participants (77.8%) were rural residents. Of all women included in this study, 47.8% never attended school; 35.0%, 11.6%, and 5.6% attended primary, secondary and higher education, respectively. More than 50% of the participants belong to the poor to the middle class of wealth status. Concerning the partners of the respondents, about 84% of them attended primary education; whereas most of the partners close to 63% were engaged in agricultural works.

Antenatal care-related and individual characteristics of reproductive-age women in Ethiopia, EDHS 2016
About 46.3% of women had the opportunity to talk about the transmission of HIV from mother-to-child during the antenatal visit; whereas 47% of them also discussed how to prevent HIV. More than half (59.0%) of the participants in this study were tested for HIV as a part of the antenatal care during the visits. Government health centers were the dominant (70.7%) places where HIV tests were given as part of the antenatal visit. Nearly half (48.4%) of the women had at least one birth in the past five years before the survey. Almost four in nine of the women were tested for HIV; in contrast one in four of them do not know where to get tested for HIV. On the other hand, only 36% of the women discussed with health workers about family planning. Often (66%) decisions about health care were made by both respondent and husband/partner. About two in five of the women visited a health facility in the last 12 months before the survey. Only 7.2% of the women were pregnant by the time the survey was conducted ( Table 2).  21.5%] of the women have inadequate and no knowledge of the MTCT of HIV respectively. The study also revealed that 77%, 84%, and 87.8% of the women respectively know that HIV can be transmitted from a mother to her child during pregnancy, delivery, and breastfeeding (Fig 1). There are wider regional disparities in the level of the knowledge of the MTCT of HIV among the women residing in the different regions of the country. Nearly two-thirds (66.8%) of the women residing in the Addis Ababa region have an adequate knowledge of the MTCT of HIV followed by the women

PLOS ONE
Levels of mother-to-child HIV transmission knowledge and associated factors among Ethiopian women residing in the Tigray (50.7%) and Harari (47.6%) regions. The women residing in the Somali region are the least to know MTCT of HIV where only 11.3% of the women have adequate knowledge and 60.5% of the women do not have any knowledge of the MTCT of HIV (Fig 2). The multilevel ordinal logistic regression also revealed that about 13.3% of the variation in the level of women's knowledge of the MTCT of HIV was explained by the variations among the regions.

Factors associated with mother-to-child HIV transmission knowledge among reproductive-age women in Ethiopia, EDHS 2016
In all the forthcoming interpretations, by odds ratio, we mean adjusted odds ratio (AOR) and while interpreting AOR for a selected variable we further assume that all the other variables in the model are held constant. The residence of the respondents has a statistically significant association with the level of knowledge of mother to child HIV transmission. The odds of having adequate knowledge of

Discussions
The results of this study showed that nearly one-fifth (19.9%) of the reproductive-age women in Ethiopia do not have any knowledge of how HIV is transmitted from mother-to-child, whereas nearly two-fifth (39.0%) of them have inadequate knowledge, and only 41.1% have adequate knowledge. The study also revealed that about 13.3% of the variation in the level of women's knowledge of the MTCT of HIV was explained by the variations among the regions; 66.8% of the reproductive-age women residing in the Addis Ababa region have adequate knowledge of the mother-to-child transmission of HIV, whereas, only 11.3% of the reproductive-age women residing in the Somali region do have the adequate knowledge of the motherto-child transmission of HIV The current study indicates that Ethiopian women's knowledge of MTCT of HIV has shown little improvement as compared to the findings from the secondary data analysis of the previous EDHS (the 2011 EDHS) during which only 34.9% of the women had adequate knowledge of MTCT of HIV [28]. The finding from the current study is higher than the finding from the study done in Cameroon where only 37% of women had adequate knowledge of MTCT of HIV [38] and the finding from the study done in Northwest Ethiopia where only 19% of the women who had been included in the study knew the MTCT of HIV [26]. However, the finding from the current study is lower than that of the studies done in Zimbabwe and Tanzania where 70.5% and 46% of the reproductive-age women respectively had a comprehensive knowledge of the MTCT of HIV [35,39] and the two other

PLOS ONE
Levels of mother-to-child HIV transmission knowledge and associated factors among Ethiopian women studies done in Southwest Ethiopia and Northern Ethiopia [29,31] where 65.9% and 52% of the women respectively had a comprehensive knowledge of the MTCT of HIV. The difference might be explained by the differences in uptake of maternal health care services among the reproductive-age women residing in the different regions. The 2016 EDHS report showed that the maternal health care services utilization highly varies across the difference regions of Ethiopia. According to the same report, for example, the ANC coverage from a skilled provider varies highly across the different regions of the country being highest in Addis Ababa (97%) and lowest in Somali (44%) [33]. According to the guideline for prevention of mother-to-child transmission of HIV in Ethiopia, one of the integral components of antenatal care service in the country is routine offer of HIV counselling and testing which might promote the women's knowledge of the mother-to-child transmission of HIV [23,24].
The finding of the current study also showed that rural-resident women are less likely to have adequate knowledge of MTCT of HIV instead of no knowledge or inadequate knowledge. This finding is in line with the findings from the different small-scale studies done in Ethiopia where urban-resident women had better knowledge of MTCT of HIV [26][27][28]. This might be because urban-resident women might have better access to maternal health care services and mass media than rural-resident women. The 2016 EDHS report showed that urban women are more likely than rural women to receive any ANC from a skilled provider; 90% of urban women received any ANC service from a skilled provider as compared to 58% rural women [33]. The study done in Amhara regional state, Ethiopia showed that ANC service utilization has positive significant association with knowledge of the mother-to-child transmission of HIV [40]. The current study and the previous study done in Ethiopia [28] showed that exposure to mass media has a positive and significant association with MTCT of HIV-related knowledge of reproductive-age women. On the other hand, urban-resident women might have better access to education than rural-residents, which might boost their knowledge of MTCT of HIV. The 2016 Ethiopian Demographic and Health Survey report showed that 57% of rural women have no formal education as compared with 16% of urban women [33]. The current study and many other studies [26][27][28][29][30]38] have shown that a better educational level is positively associated with better MTCT of HIV knowledge among reproductive-age women.
The odds of having adequate knowledge of MTCT of HIV instead of not having this knowledge or having inadequate knowledge among the reproductive-age women who attended primary, secondary and higher education as compared to those who have never attended formal school were 1.47, 1.71, and 1.81 respectively. This finding is supported by the findings from many other studies done in Ethiopia [26][27][28][29][30]. This might also be explained by different factors. Firstly, women may get MTCT of HIV-related knowledge through their formal academic process. On the other hand, women with better academic status might have the ability to gain more MTCT of HIV-related knowledge through their day-to-day life experience as they might have better communication skills. Educated women might also have better access to mass media which has a positive impact on MTCT of HIV-related knowledge among women.
The wealth status of the households in which the women live was significantly associated with knowledge of MTCT of HIV. Those women who were from the rich and the middleincome households were 33% and 52% more likely to have adequate knowledge of MTCT of HIV instead of having no or inadequate knowledge respectively. This finding is concordant with the finding from the study done in Tanzania using a nationally representative sample [35] and the finding from the secondary data analysis of the 2011 EDHS [28] where women from higher wealth quantile households had higher knowledge of MTCT of HIV as compared to women from the lowest wealth quantile households. This could be explained by the inequalities in accessing educational services, health care services, and social media between women from rich households and those from poor households which might have significant impacts on MTCT of HIV-related knowledge among the women. The 2016 Ethiopian Demographic and Health Survey report showed that educational attainment highly varies by wealth quintile; 74% of women in the lowest wealth quintile have no formal education, as compared with 19% of women in the highest wealth quintile. The same report also showed that there is high disparity between rich and poor women in accessing mass media; only 1% of women in the lowest wealth quintile read a newspaper at least once a week, compared with 10% of women in the highest quintile [33]. Different studies had also witnessed that wealth quantile has positive significant association with maternal health care services utilization [41,42]. The study from Ethiopia using data from the three-round EDHSs (the year 2000, 2005, and 2011) showed that socioeconomic inequality among the reproductive-age women had highly disadvantages the poor women in the uptake of maternal health care services. The same report has shown that inequalities in education and media access significantly contribute to inequalities in maternal health service utilization favoring the non-poor [43].
Owning mobile telephones was also significantly associated with MTCT of HIV-related knowledge among reproductive-age women in Ethiopia. The odds of having adequate knowledge of MTCT of HIV instead of having no or inadequate knowledge among those who have a mobile telephone was 1.43 times more likely than those who don't have a mobile telephone. The qualitative study done in Nyanza, Kenya showed that using mobile phone technology enhances linking with health workers, protecting confidentiality, and receiving information and reminders. The same study also concluded that the mobile communications platform holds considerable potential in preventing the MTCT of HIV [44].
Media exposure is expected to be associated with having MTCT of HIV knowledge. Precisely speaking, those women who listen to the radio at most 1 time a week and those who listen to the radio at least 1 time a week were by 24% and 27% more likely to have adequate knowledge of MTCT of HIV instead of not having or having inadequate knowledge of MTCT of HIV respectively. Similarly, those women who were watching television at most 1 time a week and those who were watching television at least 1 time a week were by 28% and 61% more likely to have adequate knowledge of MTCT of HIV instead of not having or having inadequate knowledge of the transmission respectively. This is concordant with the result of the study done in Ethiopia where exposure to mass media was significantly associated with the knowledge of the MTCT of HIV [28] and the cross-sectional study done in the SSA where exposure to mass media had shown a potential effect on HIV-related knowledge [45]. The study from Nigeria also showed that radio and television are the main sources of information to have knowledge of the MTCT of HIV among pregnant women in Nigeria [46]. This might directly be related to the HIV-related educations transmitted through radio and television broadcasting.
Those women who were visited by fieldworkers in the last 12 months before the survey were by 38% more likely to have adequate knowledge of MTCT of HIV instead of not having or having inadequate knowledge of the transmission as compared to those women who were not visited by the fieldworkers in the same period. This finding is consistent with the findings from the cross-sectional study done in Northeast Ethiopia and Tanzania where women who reported receiving information on HIV from health care providers had adequate knowledge of the MTCT of HIV [26,35]. This finding witness that community-based information, education, and communication of HIV-related information lifts women's knowledge of MTCT of HIV.

Strength of the study
As the EDHS sampling techniques, data collection techniques, and data processing and management are very strong, the pieces of evidence from the current study are more valid and dependable than the pieces of evidence yielded from the prior small-scall studies done in the country. On the other hand, the current study has sufficient power than other prior small-scale studies done in Ethiopia as the EDHS included a large sample size in the study. The weighting of the data was also done before the analyses to minimizes biases which could have been introduced due to the clustering effect. Moreover, multilevel order logistic modeling was also applied to account for the variation of the level of women's knowledge of the MTCT of HIV across the regions in the country.

Limitation of the study
The EDHS data, which were extracted and used for the current study were collected using a single-time survey; therefore, the temporal relationship between women's knowledge of the MTCT of HIV and the independent factors identified cannot be ascertained and the yielded evidence should be utilized with cautions. Besides, due to the absence of qualitative data on EDHS, the authors failed to investigate the association between pertinent qualitative variables like socio-cultural factors and women's knowledge of the MTCT of HIV.