Implications of forecasted disease burden on envisioning health strategies in Ethiopia; Findings from Global Burden of Disease 2017 Forecasting Study

Background Universal health coverage is the main goal of the health sector in the coming decade for Ethiopia, in the county’s transitions to a middle-income state. We used GBD 2017 forecasting 2017–2040 results to support Ethiopia’s envisioning framework with baseline scenario and calibration targets Methods We used GBD 2017 forecasting estimates for Ethiopia which modelled 250 causes and cause of death from 2017–2040. The data sources for Ethiopia include surveys, surveillance, case notications, facility reports, sibling history, verbal autopsy and police records to estimate mortality and causes of death. We reported Life Expectancy (LE), death and premature mortality rates using GBD broader and detail categories with 95% uncertainty Intervals (UI). Results Ethiopians average life expectancy will increase from 66.0 years (64.0-68.5) in 2017 to 73.8 years (70.3–77.3) in 2040. In 2040, the all-cause age-standardized death rate is 854 per 100,000 people of which NCDs caused 567.1 and CMNNDs caused 211.2 of the deaths. NCDs and injuries show a smaller reduction of 26% and 23% respectively between 2017 and 2040 compared to the 53% decrease for CMNNDs. Cardiovascular and neoplasm causes 224.7, 137.8 deaths/100,000 respectively. Diarrheal, lower respiratory infections, and other common infectious diseases combined caused 144.5 age-standardized deaths/100,000 in 2040. The combined age-standardized death rate for HIV/AIDs and tuberculosis is 35.2 deaths/100,000. The age-standardize premature mortality percentage contribution of CMNNDs declined from 62.4% in 1990, 45% in 2018 and 34% in 2040, whereas NCDs contribution increased from 25.2% in 1990, 46% in 2019 and 54% in 2040. further eradicating drivers. For each independent drivers-65 risk factors, selected interventions, income per person, educational attainment, and total fertility rate under 25 years- reference forecasts through 2040 and two alternative scenarios; better health and worse health were developed. These scenarios corresponded with the relative effect of these drivers on health outcomes. A hypothetical future scenarios were constructed using annualized rates of changes observed across all years in the past for the better and worse scenarios to show what would happen if Ethiopia had that level of change in the future. The data sources for Ethiopia include surveys, surveillance, case notications, facility reports, sibling history, verbal autopsy and police records to estimate mortality and causes of death. We reported Life Expectancy (LE) compared with countries considered benchmark in Ethiopia’s envisioning strategy, death and premature mortality rates using GBD broader and detail categories with 95% uncertainty Intervals (UI). Years of life lost (YLLs) calculated as a measure of premature mortality by summing up the remaining life expectancy for people dying in each age group. inequalities across the different regional states in Ethiopia and to implement relevant strategies.

considered as the core of Ethiopia's envisioning strategy to move to UHC for which UHC is assumed to guarantee access to essential services while providing protection against nancial risk (7)(8)(9). However, lack of baseline evidence for speci c strategic areas and targets is presented as a challenge to monitor progress and predict gains and resource need. Besides, understanding future disease burden and the nature of the epidemiologic transitions with its major drivers and determinates would be essential for priority setting and implementing high impact interventions to achieve UHC targets. Therefore, to ll this evidence gap this study used GBD 2017 forecasting results to support Ethiopia's envisioning framework with baseline scenario and calibration targets for intervention that intends to guide the overall investment direction of the country in health. Health planning, investments in health requires considerations of possible future trends in health and the corresponding drivers.

Methods
The methods used were described elsewhere (10) For each independent drivers-65 risk factors, selected interventions, income per person, educational attainment, and total fertility rate under 25 years-reference forecasts through 2040 and two alternative scenarios; better health and worse health were developed. These scenarios corresponded with the relative effect of these drivers on health outcomes. A hypothetical future scenarios were constructed using annualized rates of changes observed across all years in the past for the better and worse scenarios to show what would happen if Ethiopia had that level of change in the future.
The data sources for Ethiopia include surveys, surveillance, case noti cations, facility reports, sibling history, verbal autopsy and police records to estimate mortality and causes of death. We reported Life Expectancy (LE) compared with countries considered benchmark in Ethiopia's envisioning strategy, death and premature mortality rates using GBD broader and detail categories with 95% uncertainty Intervals (UI). Years of life lost (YLLs) calculated as a measure of premature mortality by summing up the remaining life expectancy for people dying in each age group.
To identify LMIC and UMIC bench marking countries from low and middle income countries and upper and middle income countries, their health status; health care system resource and health care system performance and health status of the population were considered. There were 48 LMIC and 55 UMIC using world bank classi cation of GNI per capita. The team got GDP by years from 1960-2011 and their health pro le from WHO's World Health Statistics report (WHO). LMIC were countries with the best health pro le, population of 10 million+ and were low income country in 1970's; with MMR, UMR, Age-standardized mortality rates by cause per 100,000 populations for CD, NCD, Injuries), cause-speci c mortality rate per 100, 000 populations for malaria, TB and HIV and life expectancy at birth. The average of health status and health care performance achievements of best countries selected by the above criteria were considered as a bench mark target for Ethiopia as a best case scenario (Average achievements of best LMIC for Ethiopia's 2025 target as a best case scenario and average of achievements of best UMIC for Ethiopia's 2035 targets as a best case scenario). The average of health status and health care performance achievements of all LMIC and UMIC countries are considered as a bench mark target for Ethiopia as a base case scenario (Average achievements of all LMIC for Ethiopia's 2025 target as a base case scenario and average achievements of all UMIC for Ethiopia's 2035 targets as a base case scenario).

Life expectancy
In Ethiopian the average life expectancy will increase from 66.0 years (64.0-68.5) in 2017 to 73.8 years (70.  In 2017, 311.7 deaths per 100,000 were due to CMMNN disease while 283.1 deaths per 100,000 were due to NCDs, and 55.5 deaths per 100,000 were due to injuries. In 2040, non-communicable disease accounted 277.9 deaths per 100,000, 129.6 deaths per 100,000 were due to CMNN disease and injuries accounted 50.4 deaths per 100,000. All-cause death will decline by 30%; from 650.3 death per 100,000 in 2017 to 457.9 deaths per 100,000 in 2040. The transition from high CMNN disease burden to high NCDs burden occurs in 2020. From 2038, deaths due to NCDs will be two times higher than CMNN disease. Mortality due to CMNN causes will show a 58.4% signi cant decline, while NCDs have a 1.8% decline and injuries a 9% decline between 2017 and 2040. people. Transport injury, unintentional injuries, self-harm and interpersonal violence causes 29.3 (95% UI; 21.5-39.9), 31 (95% UI; 26.6-36) and 25.8 (95% UI; 18.2-36.9) deaths per 100,000 people respectively. Between 2017 and 2040, diarrhea, lower respiratory infections and other common infectious disease causes collectively decline by 47%, HIV/AIDS and tuberculosis collectively decline by 56%, neonatal disorder causes decrease by 80% while cardiovascular disease fall only 15%, unintentional injuries 28% and self-harm and interpersonal violence decrease 6%. Whereas mortality due to neoplasm and diabetes causes increase by 22% and 14% respectively and transport injury increased by 21%, between 2017 and 2040.
Age-standardized death rates and the Epidemiologic transition The CMNNDs were leading causes of age-standardized death rate since 1990, whereas non-communicable disease become leading causes of age-standardized death rate after 2007 through 2040 ( Table 2). The epidemiologic transition from CMNNDs to NCDs and the burden of injuries in terms of age-standardized death rates for both sexes and all age groups, has shown in Figure 3. Overall, total mortality and age-standardized death rates for each of the broad groups of causes decreases between 2017 and 2040. In 2040, the allcause age-standardized death rate is 854 (95% UI; 635.3-1168.7) per 100,000 people of which NCDs caused 567.1 (95% UI; 429. 4-753.3) and CMNNDs caused 211.2 (95% UI; 132.8-331.9) of the deaths. NCDs and injuries show a smaller reduction of 26% and 23% respectively between 2017 and 2040 compared to the 53% decrease for CMNNDs ( Between 2017 and 2040, except diarrheal, lower respiratory infections, and other common infectious diseases which decline by 42%, all CMNND level two categories of causes of death showed a 57% and above decline. The age-standardized death rates from cardiovascular diseases and neoplasms will have a 37% and 8% reduction, respectively, between 2017 and 2040. Age-standardized death rates due to unintentional injuries will also decline by 33% and due to transport injuries by 2% ( Table 2).

Discussions
Ethiopians average life expectancy is expected to increase signi cantly with females' gaining more than male between 2017 and 2040. Mortality due to NCD and injuries are expected to show smaller reduction compared to CMNNDs. NCDs continue being the leading causes of age-standardized death rate from 2017 through 2040. Of the CMNNDs, the age-standardized deaths due to diarrheal disease, tuberculosis, meningitis and protein-energy malnutrition are expected to decline by half or more, whereas lower respiratory infections and HIV/AIDS showing one third or less decline from the rates in 2017 to 2040. During the same period, cardiovascular diseases declines by 37%, neoplasms by 8%, diabetes by 3%. In the contrary, prostate cancer is expected to increase by 18%. In 2040, twenty leading causes account for 72% of the total age-standardized death rates and six of 10 leading causes are NCDs. The ve leading causes of age-standardized death rates are ischemic heart disease, lower respiratory infections, diarrheal disease, stroke and diabetes. In terms of premature mortality, NCDs are expected to be the leading causes of age-standardized rate from 2017 through 2040.
Percentage contribution of CMNNDs to the total premature deaths will decline from 47% in 2017 to 34% in 2040. By contrast, the contribution of NCDs is expected to increase from 44% in 1990 and 54% in 2040. The top twenty leading causes accounts for two thirds of the total age-standardized premature mortality in 2040. In 2040, 10 of the 20 leading causes of premature mortality rate expected to be NCDs. The top ve leading causes of premature mortality are lower respiratory infections, Ischemic heart disease, diarrheal diseases, stroke, and road injuries.

Policy Implications
Improve life expectancy at birth In Ethiopia life expectancy at birth is expected to increase by 8 years from 66 years in 2017 to 74 years by 2040 for which the country could target to achieve during the envisioning period. According to previous studies, major gains in life expectancy for Ethiopia that happened between 1990 and 2015 was attributed to reductions in under-ve child mortality and burden of major communicable diseases and also linked to extensive efforts and intensive investment on the Primary Health Care system, which improved access and coverage of health care services (11,12). More gains could happen from reduction in mortality due to NCDs and injuries, improved health care availability and access for NCDs treatment, prevention and control services. The average life expectancy gain for Ethiopia is expected to be higher than many lower middle income countries (LMIC) by 2040. This increase will be one-year in Philippine, 4 years in Egypt, 5 years in Indonesia and 6 years in India. During the same period, life expectancy is expected to increase by 3 years in Brazil and by 6 years in China, benchmark upper middle income countries (UMIC) for Ethiopia. In terms of years gain, Ethiopia still be better than china and brazil but the baseline affects to join UMIC and gaining more years of life expectancy would be challenging for UMIC. The ndings showed that Ethiopia is more likely to join LMIC in terms of life expectancy than to join UMIC by 2040.
The rate of unemployment and income inequalities is quite high in Ethiopia, which contributes signi cantly to health inequalities, access to quality health services is limited and there is lack of sustainable health care resources, which negatively impact Ethiopia's life expectancy gains (13). The current move to use universal health coverage (UHC) approach to guarantee access to essential health services for all and ensuring nancial risk protection through strengthening primary health care system is commendable for Ethiopia (14). However, it is important to examine effective coverage of primary health care to address non-communicable diseases and expand the role of health extension program and community involvement. The country is implementing health care nancing, and there is a need to ensure nancial risk protection, improve the contribution of domestic nancing and increase overall health budget for the sustainability of the progress (15). Furthermore, sub-national analysis is important to understand life expectancy inequalities across the different regional states in Ethiopia and to implement relevant strategies.

Reduce neonatal, child and maternal mortality
Ethiopia is expected to reduce neonatal mortality from 23 in 2017 to 14 deaths per 1000 live births by end of 2030. This success is comparable to India that expect neonatal mortality to decline from 24 in 2017 to 14 deaths per 1000 live births in 2030 (16). Neonatal mortality in Ethiopia is expected to be higher than other lower-middle income countries such as Philippine, Egypt and Indonesia in the coming 11 years. More success with neonatal mortality has been longstanding challenge for Ethiopia in the MDG (17) and is expected to continue in the envisioning period of Ethiopia. Risk factors such as low birth weight and short gestation appears to be highly prevalent in this period. In 2030, Ethiopia is less likely to achieve neonatal mortality rate success of current upper-middle income countries such as Brazil and China (16). There is a need to strengthen quality care at birth and high impact survival interventions to address leading causes of neonatal deaths that includes neonatal encephalopathy, neonatal sepsis and neonatal preterm birth complications for Ethiopia (16).  (10). In 2040, Ethiopia could target malaria, tuberculosis and HIV/AIDs cause speci c age-standardized death rate 2017 achievement of some LMICs such as India, Philippine and Indonesia but higher than Egypt and much higher than UMIC such as Brazil and China (10). India is expected to reduce malaria, tuberculosis and HIV/AIDS cause speci c age-standardized death rate by 56% (4 to 2 death per 100,000), 64% (42 to 15 deaths per 100,000), 29% (5 to 4 deaths per 100,000) respectively from 2017 to 2040. Philippine is expected to reduce malaria and tuberculosis cause speci c age-standardized death rate by 43% (0.2 to 0.1 death per 100,000), 52% (44 to 21 deaths per 100,000) respectively but HIV/AIDS increased from 2017 to 2040. Indonesia reduce malaria, tuberculosis and HIV/AIDS cause speci c age-standardized death rate by 37% (3 to 2 death per 100,000), 54% (49 to 23 deaths per 100,000), 8% (2.6 to 2.4 deaths per 100,000) respectively from 2017 to 2040. In Egypt cause speci c age-standardized death rate for malaria and HIV/AIDs is expected to be close zero from 2017 to 2040. In Brazil cause speci c age-standardized death rate for malaria is close to zero in 2017 and 2040, tuberculosis and HIV/AIDS is expected to decline by 49% (2.6 to 1.3 death per 100,000), 37% (7 to 4 deaths per 100,000) respectively from 2017 to 2040. Ethiopia need to strengthen malaria elimination, end tuberculosis and HIV/AIDS prevention and control programs to account more success close to upper middle income countries by 2040.
Reduce the burden of non-communicable diseases Ethiopia has considered NCDs as a national priority since 2015 (5). Progress has been made to reduce population based risk factors including intervention that the government has passed bill restricting smoking in public places and banning alcohol advertisements on billboards and limiting the time where promotion is broadcasted on media. The law bans smoking within 100 meters of public and work places, health institutions and youth recreational centers. The rati ed bill also bans anyone from selling alcoholic drinks to people under 21 (18). Of the top ten risk factors (high blood pressure, high body-mass index, and high fasting plasma glucose) stood out as having a wide range of potential effects on future Ethiopian health, 23% of the total age-standardized YLL in 2040 compared to 10% in 1990.
There is a need to empower the community to play a signi cant role in the health sector to prevent non-communicable diseases and strengthen health service delivery with emphasis on primary health care units (PHCU) within the wider health sector context in Ethiopia.

Limitation of the study
A key strength of GBD forecasting study is the innovative method for combining projections from multiple models to more completely capture the uncertainty about future trends in life expectancy. The forecasted mortality rates do not tell us what interventions will achieve what health gain, for whom, which is crucial for modelling effects on health inequalities, and the cost effects (19). The key limitation of this study, shared by all projections of the future, is the inability to account for completely unexpected events and changes in the social, technological, and health systems determinants of health. Moreover, limitations of the data have been indicated with wide uncertainty interval that challenges public health decisions.

Conclusions
Ethiopians average life expectancy is expected to increase. This major gain in life expectancy is expected to be attributed to further reductions in under-ve child mortality and decline in burden of major communicable, maternal and nutritional diseases. However, more gains could be expected postponing death from non-communicable disease and injuries. Ethiopia is more likely to achieve the success of lower middle income countries in terms of life expectancy; neonatal, child and maternal mortality rates; eradicating malaria, tuberculosis and HIV/AIDS causes of mortality however less likely to achieve the success of UMICs by 2040. Non-communicable disease and injuries are expected to show smaller reduction and expected to be leading causes of age-standardized death rate from 2007 through 2040. Ethiopia is using UHC approach to address health inequalities with different segments of the population and addressing the largest share of unmet health-care needs due to nancial hardship. The cornerstone of these strategy would be having an equitable and effective primary health care system that provides free access to high-quality primary and secondary care for prevention and treatment, and uses regulation and economic tools for substantially reducing tobacco and harmful alcohol use. The health-care need to go beyond simply increasing the number of facilities and personnel, and should consider how and where care is delivered including considering a more integrated care provided in the community setting to reduce the double burden of communicable and non-communicable diseases.      Trends of age-standardized death rate per 100,000 by major causes for both sex and all age groups in Ethiopia, 1990-2040 Page 19/20 Figure 4 Percentage contribution to total age-standardized YLL rates per 100,000 by years for major causes for both sex and all age groups in Ethiopia, 1990Ethiopia, -2040 Page 20/20 Figure 5 Ranks of age-standardized death rate per 100,000 people for both sex and all age group in 1990, 2017 and 2040 in Ethiopia Ranks of causes of age-standardized YLL rate per 100,000 people for both sex and all age group between 1990, 2017, and 2040 in Ethiopia.