FE43DF93-13CD-493E-BEF2-A77A9E7D6926 Copenhagen Consensus Center Logo
Copenhagen Consensus Center

Post-2015 Consensus: Health Assessment, Jha et al.

Assessment Paper

Summary of targets from the paper

Health
Target Costs in 2030 ($B) Benefits in 2030 ($B) Benefit for every dollar spent
By 2030, reduce premature mortality by 40% in low income countries (LICs) $42 $584 $13
By 2030, reduce premature mortality by 40% in lower-middle income countries (LMIs) $402 $1,080 $3
By 2030, reduce premature mortality by 40% in LICs and LMIs $444 $1,664 $4

Summary 

A variety of priorities have been proposed for 2030 as “Sustainable Development Goals (SDGs)” to replace the highly-influential 2015 Millennium Development Goals (MDG). These proposed goals cover a wide range of development objectives and include SDG3, “Ensure healthy lives and promote well-being for all at all ages”. Within this overarching health goal, many sub-goals have been proposed, some of which are realistic and others that are not.

Benefit-cost analysis is one of the methods which can be used to evaluate the goals. Although this allows a comparison to be made between different interventions, individual ones are seldom provided without an extensive delivery system comprising community outreach services, first referral and specialty hospitals, as well as support services for quality, patient safety, monitoring and evaluation and other services, and some reduce deaths beyond the specific diseases they cover. Traditional BCA, applied to individual interventions, fails to fully capture the cumulative and synergistic benefits or costs of implementation within a health system, often done in tandem with other health-promoting activities.

Thus, overall goals of reducing child and adult mortality are required as an overarching framework target. However, it should also be emphasized that within this framework, careful consideration should be given to the specific sub-population needs for each major age group (0-4, 5-49 and 50-69 years), as they bear different burdens of disease.

In this paper, we attempt to provide a method to provide a BCR for a broad bottom-line outcome of reducing premature mortality in low-income and lower-middle income countries by 40% by 2030 (40x30). We propose this outcome as a new sub-goal of SDG3. At the same time, we try to indicate the affordability of achieving this goal by estimating the cost per capita. A secondary objective is to identify interventions that contribute positively to a reduction in premature mortality within the framework of the SGD3 sub-goals and frame these within the context of the overall goal.

We focus on mortality, but nevertheless appreciate the importance of sub-goals to reduce disability and suffering. However, in low- and middle-income countries, the burden of this is smaller than for mortality. Although we focus on premature death (prior to age 70), this is not at the exclusion of strategies to reduce deaths among older people. Reducing premature death and disability can make the time period between onset of disease and death in later age shorter and less painful by reducing key risk factors in middle age.

Methodology

Section One:

In this section we describe the method for estimating the BCR for the 40% reduction of premature mortality in low and lower-middle income countries. The focus of this analysis is on the 0.9 billion people living in 34 low-income countries (GNI per capita <$1045 in 2013 as defined by the World Bank) and 2.6 billion people living in 50 lower middle-income countries (GNI per capita $1046 to $4125 in 2013). The benefit of meeting the goal is measured as Disability Adjusted Life Years (DALYs) averted by the 40% reduction in premature mortality and the cost as the incremental healthcare expenditures required to achieve the reduction.

To determine an over-arching, achievable and quantitative goal that covers a full range of ages and diseases, we conducted an analysis of the trends in mortality from 1970 to the present, as well as the current projections to 2030. From this data, we can extract the risk of dying in each of the age ranges 0-4, 5-49 and 50-69 for particular calendar years. WHO Global Health Observatory figures give cause-specific mortality rates for 2000 and 2010. Application of these rates to the UNPD medium-fertility projection of the 2030 population yielded two numbers of deaths, comparison of which gave the change (% per decade); this is also the change in the death rate from 2000 to 2010, standardized to the projected 2030 population. Such age-standardized comparisons avoid issues of competing risks.

To estimate the BCR for the over-arching goal (for low- and middle-income countries only), we converted the mortality reductions to DALYs. We calculated these for each region for two cases: first assuming they achieved the mortality projections estimated by the UN in 2030 and second, assuming the regions achieved a higher premature mortality reduction articulated in the proposed goal. To convert the DALYs into a monetary value, we assumed each DALY was valued at 1) $1000; 2) $5000 or 3) by multiply the DALYs by two times the GDP per capita income in that region for that year.

To determine the costs of achieving the goal, we estimated the incremental government health expenditures required to reduce child and adult mortality to the proposed level. This is because the majority of the funds required would be derived from national health spending and it would be difficult to identify the full suite of interventions for each individual country or region and further estimate the cost to achieve a specific level of intervention coverage. We assume that this expenditure would lead to higher access to basic health services such as health promotion and disease prevention and treatment services.

WHO estimates that current public spending on health is about 2% of GDP in low income countries, and slightly lower for lower-middle income countries, but that a higher share of GDP to health (crudely at ~5%) would allow countries to achieve a “grand convergence” in health, which we take as the expenditure needed to achieve the mortality reduction target.

To conduct a crude sensitivity analysis on costs we use another study that shows a reduction of about 8 deaths of children under 5 years old per 1000 for a 10% increase in government health expenditures per capita; this reduction increases to about 12 deaths per 1000 when considering only low- and middle-income countries. For adults in low- and middle-income countries, the impact is less dramatic, at about 2 per 1000 for the same increase in expenditure. This allows us to extrapolate the benefits of government-funded health expenditure to estimate the funding needed to achieve the mortality reduction target. 

Section Two:

Our second objective was to comment on the health-related proposed Sustainable Development Goal and sub-goals, particularly where they are likely to be poor in terms of BCR. We reviewed existing BCRs for the sub-goals, in all examining nine key interventions in terms of their cost-effectiveness and the burden of disease they address. The cost effectiveness results are expressed as the cost of buying a DALY (Disability-Adjusted Life Year). These generate a measure of the disease burden resulting from premature mortality by integrating a discounted, potentially age-weighted, disability-adjusted stream of life years from the age of incidence of the condition to infinity using a survival curve based on the otherwise expected age of death.

The literature shows that there is a broad range of reasonable estimates for the cost-effectiveness of most interventions. Most CE studies have limited generalizability and caution is advised in extrapolating results to other jurisdictions and countries. Given the broad ranges, we have taken the approach of identifying major opportunities for investment in interventions that address a large disease burden highly cost-effectively.

Results

The broad target of a 40% reduction in premature mortality translates to reductions of: 2/3 of child and maternal deaths; 2/3 of TB, HIV and malaria deaths; 1/3 of premature deaths from non-communicable diseases; and 1/3 among remaining causes.

In low-income countries, a 40% reduction in premature death would yield a gain of 2.2m averted deaths or 195m DALYs in 2030. For lower middle-income countries, this would translate to 4.5m deaths and 352m DALYs averted in 2030. For these income groups, it is estimated that an additional $42 million and $402 million in government health expenditures would be required respectively for these two cases. Together, this suggests a benefit to cost ratio of 5 to 25 for low-income countries and 1 to 5 for lower middle-income countries. A population weighted BCR across both income groups ranges from 2 to 9, while the combined BCR for the sample is 1 to 6.We categorize the proposed SDG targets on the basis of their BCR. A BCR of 15 or above was considered “excellent”, a BCR between 5-15 was considered “good”, a BCR between 1-5 was considered “fair”, a BCR of less than 1 (or where the target was internally inconsistent, poorly specified, or unrealistic) was considered “poor”, and an “uncertain” rating was given when the economic evidence was unclear or the policy response to reach the goal was uncertain.

The first two targets in the proposed SDG3 relate to maternal and child health. Target 1, to reduce the global MMR (maternal mortality ratio) to less than 70 per 100,000 live births by 2030, was rated “unrealistic”. An alternative of essential surgery to address difficult childbirth, trauma, and other complications has a BCR of 10 (“good”). Target 2, to end preventable deaths of newborns and children under five by 2030, was also assessed as “unrealistic”.  Interventions aimed at childhood diseases and expanded immunization coverage have an “excellent” BCR of 20.

Target 3, to end the epidemics of AIDS, tuberculosis, malaria, and neglected tropical diseases and combat hepatitis, water-borne diseases, and other communicable diseases by 2030, was rated “unrealistic”. Alternative, more realistic targets, include HIV “combination prevention” (found to have a “good” BCR of 12), appropriate case finding and treatment of TB (including dealing with MDR TB), which has an “excellent” BCR of 15, and a subsidy for appropriate malaria treatment via AMFm (Affordable Medical Facility malaria) which has an “excellent” BCR of 35.

The fourth target, to reduce by one-third premature mortality from non-communicable diseases (NCDs) through prevention and treatment and promote mental health and wellbeing by 2030, was rated “uncertain” due to the lack of economic evidence. However, some alternatives are much better: tobacco taxation to reduce cancer, heart disease and other disease has an “excellent” BCR of 40, acute management with low-cost drugs to reduce heart attacks has an excellent BCR of 25, and salt reduction to address heart disease and strokes also has an excellent BCR of 20. The fifth target, to strengthen prevention and treatment of substance abuse including narcotic drug abuse and harmful use of alcohol was also rated “uncertain” with low priority. The sixth and final target within this category, to halve global deaths and injuries from road traffic accidents was decided to be “fair”.

Target seven, to ensure universal access to sexual and reproductive health care services, including family planning information and education, and the integration of reproductive health into national strategies and programs by 2030, was rated “excellent”. Target eight, to achieve universal health coverage (UHC), including financial risk protection, access to quality essential health care services, and access to safe, effective, quality and affordable essential medicines and vaccinations for all, though an excellent goal in principle, is rated “uncertain” due to the lack of data.

Target nine aims to substantially reduce the number of deaths and illnesses from hazardous chemicals and air, water, and soil pollution and contamination by 2030. This target was rated “uncertain” due to the lack of cost-effectiveness data.

Discussion

Our analysis suggests that an overarching goal of 40% reduction in premature mortality is highly cost-effective under a range of assumptions. The overall BCRs for broader goals of mortality reduction of 40% are more attractive in terms of BCR than some of the specific sub-targets, such as maternal delivery services. This might reflect the overall efficiencies of funding and supporting broad health system goals. The 40% reduction is also relatively easy to monitor, as global systems from the United Nations provide reasonable estimates of age-specific mortality. 

Our analysis has the advantage of introducing quantification to the target-setting process, based on rigorous analysis of mortality trends by age as well as by disease category. The proposed targets focus on premature mortality before age 70 and avoid more complex metrics that are much harder to measure and track over time, such as changes in DALYs or QALYs.

The level of increase in health expenditures remains to be a topic of research. Current public spending on health is about 2% of GDP in low income countries, and slightly lower for lower-middle income countries, but the opportunity (as income grows) to devote a higher share of GDP to health exists—with some countries able to achieve a 4% allocation completely from domestic resources and others supplemented via development assistance for health. It has been estimated that health system performance increased greatly with expenditure up to about $100 per capita a year (in 2010 adjusted dollars).

Technological progress can of course reduce the costs of these interventions and thereby also reduce the public spending necessary. Our previous research shows that the cost to save a child’s life has fallen by half from 1970 to 2010, with the greatest decline achieved after 1990, coinciding with actions following the UN's World Summit for Children. Unfortunately the spread of HIV/AIDS, an increase in smoking in LMICs and less attention devoted to adult mortality has allowed the costs to rise for older populations. Indeed, a goal of reducing premature mortality would not necessarily give preferential treatment by age or disease and could stimulate research that can reduce the adult mortality cost curve already achieved in children.