Spatiotemporal mapping and detection of mortality cluster due to cardiovascular disease with Bayesian hierarchical framework using integrated nested Laplace approximation: A discussion of suitable statistic applications in Kersa, Oromia, Ethiopia.

Cardiovascular diseases (CVDs) are the leading cause of death globally and the number one cause of death globally. Over 75% of CVD deaths take place in low- and middle-income countries. Hence, comprehensive information about the spatio-temporal distribution of mortality due to cardio vascular disease is of interest. We fitted different spatio-temporal models within Bayesian hierarchical framework allowing different space-time interaction for mortality mapping with integrated nested Laplace approximations to analyze mortality data extracted from the health and demographic surveillance system in Kersa District in Hararege, Oromia Region, Ethiopia. The result indicates that non-parametric time trends models perform better than linear models. Among proposed models, one with non-parametric trend, type II interaction and second order random walk but without unstructured time effect was found to perform best according to our experience and. simulation study. An application based on real data revealed that, mortality due to CVD increased during the study period, while administrative regions in northern and south-eastern part of the study area showed a significantly elevated risk. The study highlighted distinct spatiotemporal clusters of mortality due to CVD within the study area. The study is a preliminary assessment step in prioritizing areas for further and more comprehensive research raising questions to be addressed by detailed investigation. Underlying contributing factors need to be identified and accurately quantified.

Injury-related gaining momentum as external causes of deaths in Ethiopian health and demographic surveillance sites: evidence from verbal autopsy study.

BACKGROUND: In Ethiopia, though all kinds of mortality due to external causes are an important component of overall mortality often not counted or documented on an individual basis. OBJECTIVE: The aim of this study was to describe the patterns of mortality from external causes using verbal autopsy (VA) method at the Ethiopian HDSS Network sites. METHODS: All deaths at Ethiopian HDSS sites were routinely registered and followed up with VA interviews. The VA forms comprised deaths up to 28 days, between four weeks and 14 years and 15 years and above. The cause of a death was ascertained based on an interview with next of families or other caregivers using a standardized questionnaire that draws information on signs, symptoms, medical history and circumstances preceding death after 45 days mourning period. Two physician assigned probable causes of death as underlying, immediate and contributing factors independently using information in VA forms based on the WHO ICD-10 and VA code system. Disagreed cases sent to third physician for independent review and diagnosis. The final cause of death considered when two of the three physicians assigned underlying cause of death; otherwise, labeled as undetermined. RESULTS: In the period from 2009 to 2013, a total of 9719 deaths were registered. Of the total deaths, 623 (6.4%) were from external causes. Of these, accidental drowning and submersion, 136 (21.8%), accidental fall, 113 (18.1%) and transport-related accidents, 112 (18.0%) were the topmost three leading external causes of deaths. About 436 (70.0%) of deaths were from the age group above 15 years old. Drowning and submersion and transport-related accidents were high in age group between 5 and 14 years old. CONCLUSION: In this study, external causes of death are significant public health problems and require attention as one of prior health agenda.

Trend and causes of adult mortality in Kersa health and demographic surveillance system (Kersa HDSS), eastern Ethiopia: verbal autopsy method.

BACKGROUND: The health problems of adults have been neglected in many developing countries, yet many studies in these countries show high rates of premature mortality in adults. Measuring adult mortality and its cause through verbal autopsy (VA) methods is becoming an important process for mortality estimates and is a good indicator of the overall mortality rates in resource-limited settings. The objective of this analysis is to describe the levels, distribution, and trends of adult mortality over time (2008-2013) and causes of adult deaths using VA in Kersa Health and Demographic Surveillance System (Kersa HDSS). METHODS: Kersa HDSS is a demographic and health surveillance and research center established in 2007 in the eastern part of Ethiopia. This is a community-based longitudinal study where VA methods were used to assign probable cause of death. Two or three physicians independently assigned cause of death based on the completed VA forms in accordance with the World Health Organization's International Classification of Diseases. In this analysis, the VA data considered were of all deaths of adults age 15 years and above, over a period of six years (2008-2013). The mortality fractions were determined and the causes of death analyzed. Analysis was done using STATA and graphs were designed using Microsoft Excel. RESULTS: A total of 1535 adult deaths occurred in the surveillance site during the study period and VA was completed for all these deaths. In general, the adult mortality rate over the six-year period was 8.5 per 1000 adult population, higher for males (9.6) and rural residents (8.6) than females (7.5) and urban residents (8.2). There is a general decrease in the mortality rates over the study period from 9.4 in 2008-2009 to 8.1 in 2012-2013. Out of the total deaths, about one-third (32.4%) occurred due to infectious and parasitic causes, and the second leading cause of death was diseases of circulatory system (11.4%), followed by gastrointestinal disorders (9.2%). Tuberculosis (TB) showed an increasing trend over the years and has been the leading cause of death in 2012 and 2013 for all adult age categories (15-49, 50-64, and 65 years and over). Chronic liver disease (CLD) was indicated as leading cause of death among adults in the age group 15-49 years. CONCLUSION: The increasing TB-related mortality in the study years as well as the relative high mortality due to CLD among adults of age 15-49 years should be further investigated and triangulated with health service data to understand the root cause of death.

Neonatal mortality and causes of death in Kersa Health and Demographic Surveillance System (Kersa HDSS), Ethiopia, 2008-2013.

BACKGROUND: In the world, Neonatal mortality accounts for 40 % of death of children under the age of 5 years. Majority of neonatal deaths occur in developing countries outside of formal health system, among which death in the first hour of first day of their life constitute the huge bulk. This analysis is intended to estimate neonatal mortality rates and identify the leading causes of death based on the surveillance data over 6 years period in Kersa health and demographic surveillance system (Kersa HDSS) site, Eastern Ethiopia. METHODS: Kersa HDSS is an open dynamic cohort of population established in 2007. The surveillance started after conducting a baseline census followed by population update and events registration on house-to-house visits every 6 months. Data were collected using verbal autopsy (VA) questionnaire from close relatives (usually mothers in this case) and causes of deaths were assigned by 2 to 3 physicians. This analysis was done based on 301 neonatal deaths and 10,934 live births occurred during 2008 to 2013. RESULTS: The overall neonatal death rate during the study period was 27.5 per 1000 live births. Nearly all neonatal deaths (94 %) occurred at home. More than four-fifth (82.4 %) of the deaths was occurred in the first week of life. More than 80 % of the deaths were due to perinatal causes. Bacterial sepsis of the newborn accounted for 31.2 % followed by birth asphyxia and perinatal respiratory disorder (28.2 %), and prematurity (17.3 %). Higher number of death was observed in Tolla and Bereka sub-districts located at the southern parts of the study site which are away from the main road network. CONCLUSION: The overall neonatal mortality over 6 years is the same to the national average (27 per 1000 live births). The leading causes of neonatal death were bacterial sepsis of newborn and birth asphyxia. Community-based skilled health care delivery during birth should be emphasized.

Causes of Death among Children Aged 5 to 14 Years Old from 2008 to 2013 in Kersa Health and Demographic Surveillance System (Kersa HDSS), Ethiopia.

BACKGROUND: The global burden of mortality among children is still very huge though its trend has started declining following the improvements in the living standard. It presents serious challenges to the well-being of children in many African countries. Today, Sub-Saharan Africa alone accounts for about 50% of global child mortality. The overall objective of this study was to determine the magnitude and distribution of causes of death among children aged 5 to 14 year olds in the population of Kersa HDSS using verbal autopsy method for the period 2008 to 2013. METHODS: Kersa Health and Demographic Surveillance System(Kersa HDSS) was established in September 2007. The center consists of 10 rural and 2 urban kebeles which were selected randomly from 38 kebeles in the district. Thus this study was conducted in Kersa HDSS and data was taken from Kersa HDSS database. The study population included all children aged 5 to 14 years registered during the period of 2008 to 2013 in Kersa HDSS using age specific VA questionnaires. Data were extracted from SPSS database and analyzed using STATA. RESULTS: A total of 229 deaths were recorded over the period of six years with a crude death rate of 219.6 per 100,000 population of this age group over the study period. This death rate was 217.5 and 221.5 per 100,000 populations for females and males, respectively. 75% of deaths took place at home. The study identified severe malnutrition(33.9%), intestinal infectious diseases(13.8%) and acute lower respiratory infections(9.2%) to be the three most leading causes of death. In broad causes of death classification, injuries have been found to be the second most cause of death next to communicable diseases(56.3%) attributing to 13.1% of the total deaths. CONCLUSION AND RECOMMENDATION: In specific causes of death classification severe malnutrition, intestinal infectious diseases and acute lower respiratory infections were the three leading causes of death where, in broad causes of death communicable diseases and injuries were among the leading causes of death. Hence, concerned bodies should take measures to avert the situation of mortality from these causes of death and further inferential analysis into the prevention and management of infectious diseases should also be taken.

Small area clustering of under-five children's mortality and associated factors using geo-additive Bayesian discrete-time survival model in Kersa HDSS, Ethiopia.

BACKGROUND: Child mortality reflects a country's level of socio-economic development and quality of life. In Ethiopia, limited studies were conducted on under-five mortality and almost none of them tried to identify the spatial effect on mortality. Thus, this study explored the small area clustering of under-five mortality and associated factors in Kersa HDSS, Eastern Ethiopia. METHODS: The study population included all children under the age of five years during the time September, 2008-august 31, 2012 which are registered in Kersa Health and Demographic Surveillance System (Kersa HDSS). A flexible Bayesian geo-additive discrete-time survival mixed model was used. RESULTS: Some of the factors that are significantly associated with under-five mortality, with posterior odds ratio and 95% credible intervals, are maternal educational status 1.31(1.13,-1.49), place of delivery 1.016(1.013-1.12), no of live birth at a delivery 0.35(0.23,1.83), low household wealth index 1.26(1.10 1.43) middle level household wealth index 0.95 (0.84 1.07) pre-term duration of pregnancy 1.95(1.27,2.91), post-term duration of pregnancy 0.74(0.60,0.93) and antenatal visit 1.19(1.06, 1.35). Variation was noted in the risk of under-five mortality by the selected small administrative regions (kebeles). CONCLUSION: This study reveals geographic patterns in rates of under-five mortality in those selected small administrative regions and shows some important determinants of under-five mortality. More importantly, we observed clustering of under-five mortality, which indicates the importance of spatial effects and presentation of this clustering through maps that facilitates visuality and highlights differentials across geographical areas that would, otherwise, be overlooked in traditional data-analytic methods.

HDSS Profile: The Kersa Health and Demographic Surveillance System.

Kersa HDSS was established in 12 sub-districts of Kersa district, Eastern Hararge, Oromia Region, Ethiopia. The site is principally rural with two small towns (Kersa and Weter). The baseline census was conducted in 2007 and since then has been updated every 6 months, with registration of demographic and health events. Data are entered into the HRS-2 relational database. At baseline a total of 10,085 houses, 10,522 households and 50,830 people were registered. The sex ratio and number of persons per household were 1.0 and 5.1, respectively. At the end of 2013, the population was 60,694. Up to the end of 2013, 12,571 births and 3143 deaths were registered, respectively. Over 85% of births and deaths occurred at home. The annual net population growth ranges from 0.06 to 1.6. The majority of the population in Kersa are not working age group; hence the dependency ratio in most of the years is below 1. The total fertility rate ranges from 4.0 to 5.3. A reduction in neonatal, infant and under-five mortalities was observed. For all deaths, verbal autopsies were done. Tuberculosis is the leading cause of death among adults and malnutrition is the leading cause of death among children aged under 5 years. Kersa HDSS is ready to collaborate with interested researchers on health and demographic issues. For further details please visit: [http://www.haramaya.edu.et/research/projects/kds-hrc/].