Studies in sub-Saharan African countries have demonstrated a significant reduction of all-causes and Communicable diseases (CDs) related mortality among women[1–4]. Most female mortality is reportedly due to maternal causes and sexual and reproductive health related threats[1, 5, 6].

Ethiopia is the country with high rates of maternal deaths[1, 7]. However, according to 2011 report of Ethiopian health profile, life expectancy of women was higher (61.8 years) than men (59 years)[8]. One of the reasons for this could be high emphasis by the government of Ethiopia given to maternal health resulting in expansion of maternal health care services to prevent pregnancy and childbirth related deaths among women of reproductive age[9]. Despite this effort, little is known about what and in what extent other causes of death are contributing for mortality of females. Some existing studies reported that CDs are still the leading causes of death among females–with notes of increasing burden of Non-Communicable Diseases (NCDs) and injuries in the country[3, 10, 11].

For health interventions to be effective and tailored to the appropriate group, it is important for policy makers and planners to be aware of what causes death among the group[12]. In this regard, mortality data are important indicators of community health and are indispensible inputs in setting priorities for health interventions. Moreover, mortality data are crucial to develop and evaluate effective health interventions and policies[13]. However, many developing countries, including Ethiopia, with the highest burden of diseases continue to lack routine, representative and high quality information on causes of death[12, 14, 15]. Studies have shown that VA is the best available approach in obtaining evidences on causes of death in resource limited setting[16, 17].

In this regard, Health and Demographic Surveillance System (HDSS) using validated VA procedures can be used as an alternative method for ascertaining causes of death out of health facilities[18, 19]. As a result, VA method has been an epidemiological tool for some decades to estimate cause specific mortalities in a community[19, 20]. Thus, Kilite-Awlaelo Health and Demographic Surveillance System (KA-HDSS) in Tigray region has been collecting vital events and VA data since September 2009.

This study investigated causes of death among all age female residents in KA-HDSS. The study will have important policy implications by providing tailored evidences on the most common causes of death among females. It also enables decision makers and planners to allocate appropriate resources to prevent the leading causes of death among this segment of the population. Furthermore, the findings of this study will guide existing maternal health policy and strategies by providing important information on cause specific mortalities among women of reproductive age.

Description of Study setting (including map of the study area), detail VA method and field operation system of KA-HDSS are presented elsewhere[3, 21]. In these two previous publications, mortality and VA data collection procedures and interpretation of VA data are described exhaustively.

Under KA-HDSS, we investigated mortality rates and causes of death in a cohort of female population from 1st of January 2010 to 31st of December 2012. At the baseline, 33,688 females were involved for the prospective follow-up study. Households of the residents were updated every six months by fulltime surveillance data collectors to identify vital events, including deaths.

The enumerators detected deaths in each household during which they record the name of deceased, place of death and unique identification number offered by the surveillance study. All these deaths were reported to the study supervisors for further interview using VA questionnaire. Parents or spouses are identified as respondents for VA interview. VA interviews were performed at least 45 days after the death considering the local mourning period. The surveillance system employed standard data collection tools and procedures adopted from the International Network of demographic Evaluation of people and Their Health (INDEPTH) Network[22].

A total of 398 female deaths were recorded in the follow-up period. Of these, complete VA data were collected for 377 deaths. Due to incomplete nature of VA data, shortage of information or other reasons, causes were not assigned for 21 (5.3%) deaths, which were labeled as “missing” causes of death.

Broad causes of death

Overall mortality among female population in the study period was 3.76 (95% CI: 3.41, 4.15) per 1000 person-years. CDs were the leading causes of death accounting for 148 of 398 deaths (37.2%) or 1.40 deaths per 1000 person-years. NCDs attributed deaths were the second contributing for 104 out of 398 deaths (26.1%) or 0.98 deaths per 1000 person-years. Of all female deaths, pregnancy and childbirth related causes were responsible for few deaths—only 3 of 398 deaths (0.8%) (0.03 deaths per 1000 person-years) (Figure 1).Death rates among old age group females (85 years and above) and infants were 106.84 and 101.16 per 1000 person-years, respectively. The mortality rate decreased among early childhoods, adolescents and late adulthoods which were again increased in older age groups. The leading causes among infants were perinatal causes (51.66 deaths per 1000 person-years) followed by CDs (23.68 deaths per 1000 person-years). Whereas, the leading causes of death among the age group of 85 years and above were CDs attributing for 37.39 deaths per 1000 person-years. In this age group, NCDs were the second leading causes of death causing 21.37 deaths per 1000 person-years; and indeterminate cause of death accounted for 26.71 deaths per 1000 person-years (Figure 2 and Figure 3).

Although CDs were the leading causes of death in early age group, ECs and NCDs were the most common causes among late childhood and adult females. Particularly NCDs were the most common causes of death among females of 50 years and above. Highly linked to lifestyle, poor eating habit and alcohol consumption, these diseases are becoming increasingly common in the developing countries[3, 4, 25].

Our study population depicts epidemiological transition in which NCDs and injuries becoming common and coexisting alongside CDs among females. Studies in Ethiopia reported that CDs are the most common causes of death among females. At the same time, these studies highlights increasing public health problem of NCDs[3, 4, 26]. This can be attributed to nutritional transitions resulting in obesity which has been found to be increasing in some rural areas in developing countries[27, 28]. Moreover, a report indicated that poor and disadvantaged people are more likely to develop NCDs as a result of the exposures to behavioral risks[29]. Despite the continuing interventions against CDs among female population, these findings highlight public health importance of preventing and controlling of NCDs in developing countries.

Tuberculosis, acute lower respiratory tract infection and intestinal infectious diseases were the leading specific causes of death. Particularly, tuberculosis was the leading cause of death among all female residents. We found also that HIV/AIDS and tuberculosis were the major causes of death among women of reproductive age. This implies the need for at least equal attention for other causes of death like HIV/AIDS and tuberculosis among women of reproductive age group in addition to maternal causes. According to the 2013 global tuberculosis report, Africa is one of the regions with slow pace to achieve the mortality elated to tuberculosis. According to the report, tuberculosis remains among the top three killers of women[30]. In line with our findings, different studies have also reported the public health burden of tuberculosis and HIV/AIDS among female population in sub-Saharan African countries including Ethiopia[2, 4, 11, 31–33]. These finds signal the need of more efforts to prevent tuberculosis related mortality among females.

Consistent with the epidemiological trend of mortality in developing countries, higher mortality rates were recorded among old age groups (106.84 per 1000 person-years) and infants (101.16 per 1000 person-years). This is expected as elders and infants have weak and immature immune system to defend diseases. Furthermore, these population groups are prone for variety of diseases like congenital abnormalities and chronic diseases. The leading causes of death among infants were perinatal causes and this is in line with other study in Ethiopia[3]. In concurrence with other studies in Ethiopia[3, 4, 10], pregnancy and childbirth related causes were responsible for few deaths, causing only 3 deaths (4.1%) or 5.34 deaths per 1000 person-years among women of reproductive age.

Our study had some limitations. We had 21 (5.3%) cases whose were their verbal autopsy data were not complete. Hence, they were labeled as “missing” causes of death which could have had important implications on rates of causes of death if they had been correctly collected and assigned. On top of this, the performance related to VA questionnaires and physicians in detecting exact causes of death for a given case could be mentioned as a limitation. It is also important to note that interpreting the VAs and assigning causes of death by physicians have been questioned for its reliability and repeatability[34].

In summary, we got mortality information coupled with VA data which allowed us to estimate cause- and age-specific mortality among female residents in KA-HDSS site. Despite the continuing public health burden of CDs, NCDs and ECs were significantly contributing for female deaths. Tuberculosis, cerebrovascular diseases and accidental falls were the leading specific causes of deaths under the categories of CDs, NCDs and ECs, respectively. CDs are continued to be the leading causes of death among females. HIV/AIDS and tuberculosis were the major causes of death among women of reproductive age.

Despite the attention given and efforts being made to prevent maternal causes of death in Ethiopia, other causes of female deaths are equally important to consider in maternal health related interventions. In rural part of Ethiopia, existing public health and curative interventions on CDs should be strengthened. Alongside with the existing programs and strategies to expand maternal health care services, the Ethiopian Ministry of Health and Tigray regional Health Bureau should scale up the current efforts to reverse the public health burden of other causes of death, like HIV/AIDS and tuberculosis, among females.

Authors are research team members of Kilite-Alaelo Health and Demographic Surveillance System (KA-HDSS) under Mekelle University, College of Health Sciences. All are interested in health and medical researches.