The purpose of this research is to investigate the causes of the high infant mortality rate in Ethiopia, and determine if travel time to healthcare facilities and an inadequate amount of funding and resources plays a major role in the shockingly high statistics. By taking a deeper look at the “why” behind infant mortality and by using several different data science methods, it will be clear to see that the problem isn’t random. Ethiopia lacks a strong healthcare system and sufficient resources, and the infants are paying the price. This research includes data collected using Kaplan Meier curves, multilogistic regression lines, and J48 and JRip classification algorithms, in order to discover the true reasoning behind the deaths of infants. This research will build upon previous research on infant mortality and travel time to healthcare facilities with mothers, in order to determine the best data science methods necessary for pinpointing the exact cause of these deaths, and figure out the most efficient solution.
Every woman deserves the resources and healthcare facilities necessary for performing a safe birth, both for the mother and child’s benefit. This takes on Amartya Sen’s “friendly” approach to human development. There are unfreedoms done to entire groups of people, when there is an economic gap that divides those that are able to give birth safely and comfortably, verus those who must walk hours or even days just to get to an underfunded, overcrowded healthcare facility that isn’t even neonatal focused. The economic gap divides groups of women, and essentially says that some women don’t deserve the safety for them and their children in the same way that the women above the economic gap do. To put things in perspective, the state of Virginia has a population of around 8.5 million, and Virginia has around 1,243 healthcare facilities. Ethiopia has a population of around 105.4 million, and they have only 228 health care facilities. More than 500,000 women and girls from Ethiopia suffer from disabilities caused by complications caused by childbirth each year. In Ethiopia, 80% of newborn deaths are caused by treatable cases. The neonatal mortality rate accounts for 41% of under-five-years old deaths. Compared to the United States, a mother is twenty five times more likely to die during childbirth, and an infant is nine times more likely to die before reaching the age of one. This is unjust, and this issue needs more people utilizing the newest data science platforms and technologies, in order to figure out how to bridge the gap. This problem could be hugely benefited by an increase of healthcare facilities and resources going to those women who need it most.
There are several reasons that Ethiopia has yet to implement a system that would decrease child mortality and increase attendance in postnatal care, and better resources and conditions for mothers. In all of Ethiopia, there are few healthcare centers. Some women have to travel through very tough terrain to get to these healthcare facilities. As one can imagine, traveling four hours to reach a healthcare facility, which is understaffed and underfunded, is not exactly ideal for women in labor or those who are close to childbirth. Because of this, many women opt to give birth at home with a midwife, or decide to make the long trek to a hospital, but will not go back for postnatal health care. This decision is understandable considering many healthcare facilities do not even have properly trained professionals to aid women in the birthing process. With the use of several different technologies, the areas in Ethiopia that require more resources can be determined, and we can figure out which places most need new healthcare providers and more resources. In order to pinpoint exactly where we could allocate resources, we need to use data science methods that can take in multiple parameters. By using multiple covariates, we will be able to form a more accurate map of where women of child-bearing age reside, and how we can aid them, so that they aren’t walking across the country to get what they need and deserve. Personally, I don’t believe this problem is a complex adaptive system. I believe that we have the tools and data out there to determine and understand what Ethiopia, and countries like Ethiopia need. It is more a problem of resources. Recently, Ethiopia has a more stable political and economic situation. If they could figure out, with the help of other stable countries, how to allocate resources and loans effectively, then more money and funding could be put into healthcare resources in Ethiopia. There are many problems with the current quantity and quality of their healthcare facilities, and women are arguably the most impacted. Using bayesian models and geospatial data, we can figure out exactly where it would be most efficient to build more hospitals and funnel in resources.
The main focus of the research in neonatal and maternal health and safety improvement, focuses on pinpointing where women of child-bearing age reside, and how close they are, or how easy it is for these women to get the help they need. When a woman is pregnant, she will need easy access to several things, namely vitamins and supplements, a certified healthcare worker, and facilities that are adequate enough to take care of women who are giving birth. These women need to be able to get to facilities and have access to health care workers in an adequate amount of time, and right now there are not nearly enough hospitals or workers to supply these women with necessary resources. In order to help improve upon the lives of these women, we need to figure out where the population resides, and which areas would be most efficient to place more healthcare facilities, and which areas need more resources than others. We can use data science to pinpoint directly these areas in which funding and contributing to the process of building and providing extra healthcare facilities for Ethiopia would be best spent. We can use machine learning to find these areas. In Maternal and newborn health services utilization in Jimma Zone, Southwest Ethiopia: a community based cross-sectional study, the authors used multivariable logistics models. These are models used when there are more than two variables, in order to more accurately define the outcome. In this case, the authors used a multivariable logistics model to input several factors, such as demographic and socioeconomic factors and other, more arbitrary factors, such as waiting time by clients. These independent variables provided possible outcomes for several different dependent variables, such as healthcare utilization and postnatal care. By using a multivariable logistics model, the authors were able to predict more accurate results, because they were able to provide more inputs.
In Effect of Geographical Access to Health Facilities on Child Mortality in Rural Ethiopia: A Community Based Cross Sectional Study, the authors use a Kaplan Meier estimator to determine the risk of death for children under five, and cross reference that with travel time, in order to determine if travel time is a significant enough factor into why the children are dying. This data science method uses lifetime data to help narrow in on the true risk of a certain factor. This model shows that as the children age, there is a higher probability of death. This probability increases the farther away a child lives from a healthcare center. The Kaplan Meier estimator could also be used to determine a child’s survival rates based on demographics and socioeconomics relating to the mother.
A helpful data set was one given in Effect of Geographical Access to Health Facilities on Child Mortality in Rural Ethiopia: A Community Based Cross Sectional Study. This data set used GIS to determine where women reside, and how difficult it was for each woman to get the resources necessary. This was done by first collecting the coordinates of healthcare facilities and selected participants’ households. Then the authors used ArcGIS10 to calculate straight line distances between households and health care facilities, which were then used in a “Cost Analysis” module, which contains two layers of data. The first of which is the location of the health care centers. The second of which is travel time associated with moving through different geographical nuances, such as water, mountains. By using this module, the authors were able to come up with a more accurate travel time. This study looked more specifically at the number of newborn deaths in children under age five, and compared specific variables that aligned with an inadequate access to resources and facilities. Some of the variables included travel time, household wealth, and mothers’ education. It is clear to see that the longer the travel time and the greater the travel distance, the number of newborn and children under five death rate increases. Of the children that passed, 33.7% of them were three and a half to six and a half hours away. This data is extremely interesting because figuring out how far away children who passed away were from health care centers could be crucial in determining where it is best to place critical health care resources. This data set also gave odd ratios to determine how more likely one child was to die than another, based on mothers income and education, but above all, the difficulty of travel to the nearest healthcare facility, and the percentages increase as the distance moves further away.
Another useful data set was the Ethiopia Demographic and Health Survey (EDHS) 2011 dataset used in Ethiopic maternal care data mining: discovering the factors that affect postnatal care visit in Ethiopia. This data set covers family planning, fertility levels, mother, child and infant mortality rates, nutrition, and more. The authors used this data, and used the J48 and JRip classification algorithms to update and more accurately describe the Ethiopian female population. They started by first creating bar graphs to represent the number and demographics of the women that attend postnatal care visits, and those who don’t, ultimately trying to uncover factors behind the women who do not attend postnatal care. The data is summarized at the end, and is formatted to give the most important and determining factors in the reasons behind why women do not attend postnatal care. The most important reasons were delivery place, prenatal health professional, and age. Overall, if a woman had inadequate resources, either pertaining to the healthcare professional or to getting to the health care facility, there is a much higher chance that she will not attend postnatal care. When this happens, the child mortality rate increases. By using J48 and JRip classification algorithms, the authors have determined a set of rules in order to determine in each scenario, what a woman will do given each factor. These rules can really come in handy when setting up resources, in order to know what can be supplemented into Ethiopia to help educate women on the importance of postnatal care, but also by making it attainable to all.
There was also census data taken. Specifically in Cause of neonatal deaths in Northern Ethiopia: a prospective cohort study, there were surveys taken of mothers by midwives dealing with risk factors and infant survival rates. This was in order to compile a dataset with the information of demographics of mothers on infant mortality, in order to see if there was a connection. The authors surveyed 1,152 live births, and of those, 68 were deaths. This was in order to create a dataset that would be prepared to map out the causes of death and figure out where more resources needed to go. In urban areas, there were no deaths by infection. However in rural areas, 21% of infants died by infection. Had there been a hospital closer to the mothers in these areas, maybe they would have been able to bring their child to a healthcare center, or even have prevented the problem all together, by attending postnatal care. Nearly half of the infants who were born premature, were born to mothers who were in the lowest income class. These are women who do not have the means to make it to some health care facilities. If they live far away, it can be an expensive endeavor to take a three hour trek as a pregnant woman. They may also have other children to take care of, and it quickly becomes a nearly impossible feat to get these mothers to travel a far distance to attend pre and postnatal care. Two of the most prominent characteristics found in this study were that residence and travel time is extremely important. Asphyxia proved to be the leading cause of death in infants. This is due to poorly trained healthcare professionals, and an inadequate amount of funding and resources. If there were more resources for not only Ethiopia as a whole, but also those women who do not have the same higher economic status as others, there is a chance that the overall infant mortality could decrease.
Through the use of these articles and through data science methods, I am looking at whether infant mortality could be lessened by a decrease in travel time to healthcare facilities by mothers. There is clear data showing that the further away a mother lives from a healthcare facility, the more likely she is to not attend postnatal care, which in turn heightens the child’s chance of death. The articles mainly used technology that took in parameters, in order to narrow in on the reasons behind mortality in children under five. By using multi logistic regression, and ARCGis technology, we can improve upon predictions of where exactly women of child bearing age reside, and where children younger than five reside. By getting a more accurate count of exactly where these people live, we will be able to put more healthcare resources in more efficient parts of the country. Right now, I believe we are missing the data that is needed in order to make an educated assumption about where it would be most efficient to place resources. I was also unable to find information with bayesian models. However, right now, many women are suffering, and unable to get the care that they need and deserve. These women are walking miles simply to give birth, and many of these women have other children. This is a lot of stress to put on a pregnant woman. A decrease in travel time and an increase in adequate resources and properly trained healthcare professionals, would likely yield a higher standard of living for many Ethiopian women, as well as decrease the shockingly high infant mortality rate for Ethiopia.
Alegana, V. A., Atkinson, P. M., Pezzulo, C., Sorichetta, A., Weiss, D., Bird, T., … Tatem, A. J. (2015). Fine resolution mapping of population age-structures for health and development applications. Journal of The Royal Society Interface, 12(105), 20150073. doi: 10.1098/rsif.2015.0073
Bailey, P. E., Keyes, E. B., Parker, C., Abdullah, M., Kebede, H., & Freedman, L. (2011). Using a GIS to model interventions to strengthen the emergency referral system for maternal and newborn health in Ethiopia. International Journal of Gynecology & Obstetrics, 115(3), 300–309. doi: 10.1016/j.ijgo.2011.09.004
Dadi, Lelisa Sena, et al. “Maternal and Newborn Health Services Utilization in Jimma Zone, Southwest Ethiopia: a Community Based Cross-Sectional Study.” BMC Pregnancy and Childbirth, vol. 19, no. 1, 2019, doi:10.1186/s12884-019-2335-2. Ebener, S., Guerra-Arias, M., Campbell, J., Tatem, A. J., Moran, A. C., Johnson, F. A., … Matthews, Z. (2015). The geography of maternal and newborn health: the state of the art. International Journal of Health Geographics, 14(1). doi: 10.1186/s12942-015-0012-x
Mcpadden, J., Durant, T. J., Bunch, D. R., Coppi, A., Price, N., Rodgerson, K., … Schulz, W. L. (2019). Health Care and Precision Medicine Research: Analysis of a Scalable Data Science Platform. Journal of Medical Internet Research, 21(4). doi: 10.2196/13043
Mengesha, Hayelom Gebrekirstos, and Berhe W. Sahle. “Cause of Neonatal Deaths in Northern Ethiopia: a Prospective Cohort Study.” BMC Public Health, vol. 17, no. 1, 2017, doi:10.1186/s12889-016-3979-8.
Okwaraji, Y, et al. “PS53 Effect Of Geographical Access To Health Facilities On Child Mortality In Rural Ethiopia: A Community Based Cross Sectional Study.” Journal of Epidemiology and Community Health, vol. 66, no. Suppl 1, 12 Mar. 2012, doi:10.1136/jech-2012-201753.152. Sahle, Geletaw. “Ethiopic Maternal Care Data Mining: Discovering the Factors That Affect Postnatal Care Visit in Ethiopia.” Health Information Science and Systems, vol. 4, no. 1, 2016, doi:10.1186/s13755-016-0017-2.
Tatem, A. J., Campbell, J., Guerra-Arias, M., Bernis, L. D., Moran, A., & Matthews, Z. (2014). Mapping for maternal and newborn health: the distributions of women of childbearing age, pregnancies and births. International Journal of Health Geographics, 13(1), 2. doi: 10.1186/1476-072x-13-2
Tekelab, Tesfalidet, et al. “Neonatal Mortality in Ethiopia: a Protocol for Systematic Review and Meta-Analysis.” Systematic Reviews, vol. 8, no. 1, 2019, doi:10.1186/s13643-019-1012-x.