Cost-Effectiveness and Priority Setting

The principal analytic tool throughout this volume is CEA, which compares the cost of an activity, called an intervention, with the known or expected health gain. The result is summarized in a cost-effectiveness ratio (CER), as explained more fully below. This ratio corresponds to the concept of (health) value for money. Favoring activities that are more cost-effective over those that are less so is consistent with the ethical view that "limited resources for health should be allocated to maximize the health benefits for the population served" (chapter 14).

Cost-effectiveness provides the clearest simple way to promote value for money in health: hence, the emphasis on it here. CEA allows comparisons throughout the health sector and not only for the same health outcome. It does not allow comparison to nonhealth outcomes unless these outcomes can be incorporated into costs, and the calculation of the CER by itself makes no pretense of monetizing the intrinsic value of health. To use CERs for choosing what to buy and what not to, decision makers must determine a maximum willingness to pay for units of health gain, unless other criteria are considered to justify buying something with relatively poor cost-effectiveness.

For risk factors, CEA requires estimating the gain in health consequent on introducing an intervention to reduce the risk of acquiring or transmitting a condition. For packages of interventions or elements of the health system, such as hospitals, effectiveness is estimated by judging how much mortality and morbidity would be reduced by providing the whole package or set of services or by operating the facility. With some exceptions, the analyses may describe but do not quantify the non-health benefits of an intervention. Apart from the difficulty of obtaining enough data, such quantification requires attaching values to nonhealth outcomes, which is problematic when comparisons are made over large cultural and income differences.

All comparisons are relative, with no absolute distinction between being and not being cost-effective. In assigning priorities among interventions for public funding or for other policy actions, one must also consider the magnitude of health problems to which interventions apply because that affects what is affordable. Calculations of the effect of spending US$1 million or the total cost and health gain in a population of 1 million people offer ways of looking at such choices. Equity, poverty, and risk of impoverishment from ill health may also influence priorities; so do the budgets available—and the decisions of how much to make available—for buying interventions. Finally, the effectiveness of an intervention and, therefore, the degree to which it deserves priority depend on how far it is culturally appropriate or acceptable for the population it is intended to benefit. The identical intervention, technically speaking, may lead to different degrees of use or compliance in different population groups, and information and incentives may be needed to achieve the full potential outcomes.

Cost-effectiveness is only one of at least nine criteria relevant for priority setting in health if the object is to decide how to spend public funds (Musgrove 1999). Cost matters by itself, as do the capacities of potential beneficiaries to pay for an intervention. The other criteria that may affect priorities include horizontal equity (equal treatment for people in equal circumstances); vertical equity (priority for people with worse problems); adequacy of demand; and public attitudes and wants. Two criteria—whether an intervention is a public good and whether it yields substantial externalities—are classic justifications for public intervention, because private markets could not supply them efficiently, just as in other sectors. As noted in chapter 1, the interventions analyzed in this volume are not limited to public or semipublic goods. The emphasis is on value for money—that is, whether an intervention is worth buying, not who pays for it. Nonetheless, when one is choosing which public goods to buy, several criteria become irrelevant, and cost-effectiveness can be used as the chief or even the only consideration. Cost-effectiveness can similarly determine what to include in a mandatory universal public package of health care alongside competitive voluntary insurance (Smith 2005).

Cost-effectiveness can conflict with both kinds of equity—that is, the more cost-effective of two interventions may also lead to a less equitable distribution of health benefits. Equity and cost-effectiveness are compatible when a cost-effective intervention is provided to only part of the population that would benefit from it because everyone in the group suffers from the same problem. Then expanding coverage will generally also promote horizontal equity. These equity effects are reinforced when those who are better off already benefit while the poorer and sicker population does not. Choices about vertical equity—doing more for those in greatest need—are more complicated. Doing very little for people with severe health problems—because the available interventions for those problems are not very effective at reducing their suffering—is not necessarily preferable to doing more for people with less severe problems that are more amenable to intervention. When an intervention is reaching only part of a potential beneficiary population and those not benefiting tend to have more severe illness, then expanding coverage can improve both horizontal and vertical equity. Where possible, chapters consider the equity effects of expanding or changing interventions.

Definition and Characteristics of Interventions

The object of a CEA—the thing to which it is applied, the costs and outcomes of which are to be analyzed—is an intervention. An intervention is an activity using human, physical, and financial resources in a deliberate attempt to improve health by reducing the risk, duration, or severity of a health problem (Jamison 2002, table 2). The term usually refers to an activity undertaken by a health system rather than by an individual. The emphasis on a deliberate, systemic effort means that an intervention is not simply anything that improves health; for example, if more rainfall leads to higher crop yields and better nutritional status, the rain does not count as an intervention. Similarly, although breastfeeding protects infants' health, it is not itself an intervention as the word is used in this volume. In contrast, a program to encourage new mothers to breastfeed is an intervention (as described in chapter 27). How effective such a program is, of course, depends on how many mothers it persuades to adopt the practice when they are neither currently breastfeeding nor planning to do so.

Interventions can be directed against an injury or disease (such as trachoma), a condition associated with or deriving from a disease (such as blindness), or a risk factor that makes the disease or condition more likely (such as the lack of hygiene that leads to trachoma). An intervention may pursue primary prevention at the population level—promoting personal behavior change, controlling environmental hazards, or delivering a medical intervention such as immunization to a large population—or individual action for primary prevention, cure, acute management, chronic management, secondary prevention, rehabilitation, or palliation. Box 15.1 defines these terms, and the figure in the box illustrates how interventions may prevent ill health events or deal with their consequences. Characterizing an intervention fully also means distinguishing the level at which it is delivered (home, primary care facility, district hospital, or referral hospital); indicating whether it involves drugs, immune enhancement, surgery, or physical or psychological therapy; and determining whether it requires a physician or uses diagnostic, laboratory, or imaging procedures. Such procedures are most often evaluated relative to the interventions they screen for or lead to, because they produce no health gain by themselves (although the information they provide can be valuable for reassurance or for promoting behavioral changes).

Box 15.1

Intervention Categories, with Examples. The following figure illustrates how interventions are related to a health event; the definitions of these categories are given below. Interventions Related to the Occurrence of an Adverse Health Event

An intervention in the everyday sense includes such activities as immunizing a child, performing a surgical procedure, or treating an infection with antibiotics. The authors of some chapters use the term only in this sense—for example, in discussing interventions that contribute to meeting the Millennium Development Goals (chapter 9). Authors of other chapters use the term in several other senses as well. It can mean modifying an existing intervention—for example, adding Haemophilus influenzae type B (Hib) antigen to the Expanded Program on Immunizations (EPI). Immunization against Hib is itself an intervention, but instead of analyzing it separately, one can use CEA to evaluate the additional cost of incorporating that antigen and the additional health gain that is expected to result (see chapter 20). The intervention studied is then not Hib immunization as such but the change in the full vaccination procedure. A change in the scale of an existing activity can also be considered an intervention, even if the activity itself is unchanged: that is, one can analyze the change in costs and in outcomes associated with expanding or contracting the coverage of the activity—for example, extending antiretroviral treatment for HIV and AIDS to a larger population (chapter 18) or screening more newborns for sickle cell anemia (chapter 34). In most chapters, the authors assume that expansion affects costs and outcomes linearly, so that the CER does not change. The chapters on vaccine-preventable diseases (chapter 20) and malaria (chapter 21) provide explicit estimates of the differential costs of expanding coverage.

Adding one intervention to another to deal with the same disease or condition is also an intervention, and combinations of interventions can be analyzed to determine which is most cost-effective or how the cost-effectiveness of one intervention depends on the other activities with which it is combined. Examples include successively adding drugs for treatment of epilepsy (chapter 32) or secondary prevention of cardiovascular disease (chapter 33) or combining several quite different interventions to control tobacco addiction or alcohol (chapters 46 and 47, respectively). The analyses of community health and nutrition programs (chapter 56) and integrated management of infant and childhood illness (IMCI; chapter 63) define "the intervention" as a whole program incorporating several different activities. Generally, even less empirical evidence exists concerning combinations of interventions than for individual activities, but IMCI is an exception; it has been evaluated more thoroughly than most single interventions.

Box 15.2 includes a more detailed discussion, using a hypothetical example of three different ways to deliver immunization, of how CEA can be applied to four of the meanings of intervention used here: an existing intervention at its current coverage, changes in the scale of that intervention, the addition of one intervention to another when expanding coverage, and the complete shift from one intervention to a different (and more cost-effective) one.

Box 15.2

Average and Incremental Cost-Effectiveness and Intervention Choices. In the figure below, which compares three ways of delivering immunization, point X describes the status quo of a current intervention, delivering immunization by means of fixed facilities. (more...)

Depending on the comparison undertaken, the result may be an average cost-effectiveness ratio (ACER) or an incremental cost-effectiveness ratio (ICER). The former compares total costs and total results, starting from zero, whereas the latter compares additional costs and additional results, starting from the current or some other level of coverage of an intervention.

Either shifting completely from one intervention to another or partially replacing one with another may reduce costs while producing more health gain. For example, if spending is high on hospitalization for acute myocardial infarction, a program using a "polypill" (several medications in a single pill) would reduce expenditures by lowering incidence (chapter 33) and would be cost saving, because less hospitalization would be needed. If the status quo is no hospitalization (as is typical at low incomes), a polypill program increases costs but may more than correspondingly increase health gains and therefore be more cost-effective. If the polypill both reduces costs and improves outcomes compared with hospitalization, it is said to dominate a hospital-only strategy. The second figure in box 15.2 illustrates the concept of dominance; table 45.4 and box 45.1 of chapter 45 provide examples of interventions that are dominated by others.

Unfortunately, reliable information on current intervention coverage, costs, and results is not always available even in high-income countries (iMTA 2005) and is extremely scarce in low-and middle-income countries. Studies showing whether an intervention is effective or cost-effective seldom cover the entire potential beneficiary population, and service provision in the private sector is often not recorded. Many chapter authors describe only the ACER of an existing or potential intervention, whereas others explicitly compare alternatives to current practice (for an example, see chapter 16).

Many of the activities analyzed here aim at promoting changes in personal behavior, by informing and persuading individuals to eat differently, to avoid smoking and excessive alcohol, to reduce the risks of sexually transmitted infections, or to practice better hygiene. Such efforts can be considered interventions in themselves, and as such are crucial for controlling HIV and AIDS (chapter 18), promoting better infant and child care (chapters 20 and 27), preventing inherited disorders (chapter 34), encouraging healthful diets and exercise (chapters 44 and 45), and avoiding addiction (chapters 46–48). But they can also be used to improve the effectiveness of other interventions by increasing awareness and demand, combating mistaken beliefs about diseases and risks, or reducing anxiety and stigma. In that sense, information, education, and communication delivered to consumers or providers or both are examples of policy instruments. They can facilitate or promote the use of such interventions as condom distribution, screening for diseases or congenital disorders, prenatal care, or immunization.

Other activities that can be classified either as interventions or as policy instruments include the following:

• Measures to increase the quality of care, such as some kinds of staff training or the introduction of better recordkeeping. These activities may simultaneously affect a large number of specific interventions in a health facility (chapter 70).
• Legislation and regulation to impose an intervention (for example, limiting the salt content of foods, chapter 45, or requiring that salt be iodized, chapter 28); to limit or prohibit an intervention that is ineffective or dangerous or to reduce unhealthful behavior such as smoking and excessive drinking (chapters 46–47); or to codify how an intervention should be delivered and determine who may provide it, as by licensing doctors, nurses, and health facilities (chapter 71).
• Economic incentives, which can take the form of subsidies or taxes (chapter 11) for particular items of consumption other than health goods or services, such as tobacco and alcohol (chapters 46–47) or condoms to reduce HIV transmission (chapter 18), or can be provided through protection of property rights, as for patented drugs (chapter 72).

These activities of informing, mandating, legislating, regulating, and taxing or subsidizing, which are at one remove or more from medical interventions, are also often called functions of the health system (WHO 2000, chapter 2; see also chapter 9 in this volume). Several of these instruments may be used together, such as increased taxes on tobacco or alcohol along with measures to educate consumers and to restrict the times, places, or quantities of consumption. Sometimes the instrument is needed before introducing or expanding an intervention to overcome barriers to its use or to make it cost-effective enough to be worth pursuing. Educating the affected population, for example, is crucial to screening and treatment of cancers and hemoglobin disorders. The need for a particular instrument may vary from country to country even if the intervention that it facilitates is identical, because the legal, regulatory, or financial environment differs.

Estimating Effectiveness in Health

Using cost-effectiveness for resource allocation requires health effects to be represented in common units in order to facilitate comparison across interventions, diseases, or conditions. All analyses start with some natural unit: cases of disease or injury, deaths, or numbers of people who quit smoking or adopt some other health-improving behavior. All interventions that avert death are alike in that regard. Preventing a child's death at a particular age, independent of the cause, means that the averted death alone is an adequate measure of outcome. However, when lives are saved at different ages—averting death from malaria at age 2 versus death from beta thalassemia at age 10–15—the outcome is no longer identical, and some measure must account for the difference in years of life saved. These cases provide another natural unit, subject to estimating how much longer a person spared death might live. The choice of life expectancy to assume for such calculations is discussed later.

The unit of time becomes a less natural and more synthetic measure if the future is discounted, as in all these analyses. Discounting means reducing the value of each variable in each future year by an amount that increases the further in the future that year is. The discounting procedure reflects inherent uncertainty about the future and preferences for timing of consumption, and it avoids two problems. First, outcomes that potentially generate benefits forever, such as smallpox eradication, appear to have infinite benefits if the future is not discounted and therefore seem to justify any finite cost at all. Second, it makes little sense to postpone interventions forever simply because funds to finance them could be invested today and be worth more tomorrow. Even discounting the future at the low rate of 3 percent annually has a substantial effect—that is, dividing the values for future years by successive powers of 1.03. That means dividing values for year 1 by 1.03; those for year 2 by 1.03 squared, or 1.0609; and so on. At that rate, averting an infant death saves not all the 80 calendar years of life expectancy at birth (or fewer in low-income countries) but at most 30 discounted years.

For interventions that avert mortality, analysis starts by estimating the deaths prevented, uses age at death to yield numbers of life years saved, and then discounts those years as described above. When interventions improve health by averting or reducing nonfatal disability, different disabilities must be compared in severity. As with mortality, age at the time of intervention matters for long-lasting conditions, and so does discounting. In contrast, age is irrelevant for episodes of illness or injury that are self-limited or quickly resolved by intervention, because the duration of ill health does not depend on age, and all ages are treated alike in this analysis. Discounting also makes little difference over short intervals.

Determining Costs for Interventions

Whatever outcome measures are used to evaluate an intervention, its costs must be estimated. This need raises several questions about which costs to attribute to the intervention and how some of them should be valued.

Direct and indirect costs should be distinguished, and choices should be made about which, if any, of the latter to include. In addition to the direct costs to the health system of producing an intervention, the U.S. Public Health Service guidelines (Gold and others 1996) recommend including the indirect costs to patients and their families of consuming it. This recommendation means, in particular, the value of time needed for travel, waiting, and undergoing medical tests and procedures, or the value of time used in caregiving, as well as any income forgone during treatment. Externalities, or costs imposed on third parties, such as on the school system or the environment, should also be included. The analyses in this volume generally exclude such costs and report only the direct costs of delivering interventions, partly because published analyses seldom include the various indirect costs, and they are harder to estimate. Walker and Fox-Rushby (2000) found that only 20 of 101 studies included some element of indirect costing. Valuing time according to local wages or income, for example, may underestimate how valuable time actually is to poor people. Estimating such costs, even if time is not valued in money, may show whether time or monetary costs or both account for a relatively low level of use and therefore impede expanding coverage. Applying one or more of the policy instruments discussed earlier, along with the intervention, may then be important in order for it to be cost-effective.

Including such costs also raises a question of interpretation. If an intervention appears low in cost-effectiveness because it requires much travel or waiting time, the fault may lie not with the intervention itself but with health facilities that are located too far from the beneficiary population, are understaffed, or are inefficiently managed. For this reason, cost-effectiveness is estimated assuming a functional health system that does not impose prohibitive time costs on users.

Not only the characteristics of the interventions themselves, but also the capacity to deliver interventions greatly affect cost-effectiveness across many activities. In a complete analysis, each intervention is characterized by how demanding it is of managerial or institutional capacity. This element is difficult to measure directly, but authors often provide at least an intuitive description of how easy or hard delivery of an intervention is or what factors facilitate or impede its implementation. Where capacity to deliver several interventions together is important, authors deal explicitly with the issue, as in the chapters on health facilities (chapters 64–66), resources (chapters 71–72), service management (chapter 73), and whole packages of interventions (chapters 56 and 63).

Dollar values of unit costs need to be calculated for international comparisons. The inputs used to produce an intervention—the time (and training) of human resources; drugs and supplies; and depreciation or rental value of equipment, vehicles, and buildings—are either produced in the country or imported. If the latter, they already have prices in U.S. dollars; if the former, prices in local currency must be converted to U.S. dollars for comparison with other interventions and other countries. The usual distinction between tradable and nontradable goods is that tradables move from producing to importing countries at relatively constant "world" prices. In fact, the same good may be imported at different prices to different countries or may be imported to one country but locally produced in another, so that it has both an international and a local price. This situation is increasingly true of drugs and supplies, which middle-income countries (Brazil, for example) and some low-income countries (India) now produce and sometimes export.

Prices in local currency can be converted to U.S. dollars by exchange rates or by purchasing-power parity rates (as estimated in World Bank 2003). The former may reflect under- or overvaluation of the local currency, making goods systematically cheaper or more expensive than at world prices, and they may change quickly and substantially in response to changes in a country's trade balance, indebtedness, or capital flows. Nonetheless, they represent what is actually paid for locally produced inputs at any given moment. Purchasing-power parity rates, in contrast, attempt to say what local currency is worth in purchasing power, correcting for systematic price differences. Such rates can be calculated for the country as a whole, for the health sector, or for specific inputs or combinations thereof (Wordsworth and Ludbrook 2004). This calculation means valuing local inputs at external prices, assuming they are equally productive or of equal quality in the particular country as in the countries from which purchasing-power parity rates are derived. A doctor in South Asia or Sub-Saharan Africa is treated as costing just as much as a physician in high-income countries. This approach approximates measuring the real resource cost of intervention by comparing quantities of inputs among countries, eliminating price differences as a source of cost variation. Estimates of real national income are derived this way, making poor countries usually look less poor in dollar terms than if income in local currency were valued at exchange rates.

Granted that purchasing-power parity rates are reasonable for comparing large aggregates such as income across countries, but they bear little relation to the allocation of resources and budgetary choices within a country. The cost calculations in this volume are, therefore, all based on exchange rates. Exchange rates more accurately reflect what a domestic buyer—or a foreign donor or investor—has to pay for imported versus domestic inputs and, therefore, are more relevant for choices between interventions with high or low imported content. (If exchange rates are artificially fixed, the country pays a cost for that distortion that affects all interventions to the extent that they require foreign exchange.) In general, the more an intervention is produced with local inputs, the more cost-effective it will appear when priced using exchange rates, compared with its cost at purchasing-power parity rates. For decision makers and purchasers in the country, efficiency means choosing interventions according to what they actually cost, not according to what they would cost if prices were more nearly uniform among countries. If, in local currency, physicians are paid little more than nurses are, it may make sense to employ more doctors per nurse—even if at international prices doctors would cost much more and should be replaced by nurses when possible. Of course, the staffing decision turns on the competencies of the two groups as well as on their costs; for certain health problems, more nurses might be the better choice even if they cost more.

Two other reasons besides that of efficiency in buying interventions support basing cost-effectiveness on exchange rate prices. First, authors who have used published costs (which usually involve exchange rates) rather than building up estimates from individual prices and quantities seldom break down costs into imported and domestic components. Local inputs cannot be repriced at purchasing-power parity rates or can only be repriced very approximately. Second, for readers accustomed to dealing with prices converted using exchange rates, real resource estimates may simply appear to penalize the use of local inputs by valuing them at unrealistic prices. The problem with exchange rate prices, in contrast, is that when rates change, so may the relative cost-effectiveness of interventions, as imported inputs become relatively more or less expensive. Cost-effectiveness is not static or intrinsic but depends on prices as well as on quantities and on the results of an intervention—and prices can change individually or generally, through exchange movements. Priorities sometimes need to shift because of such price changes, as well as because of technological changes that make interventions more effective, and so analyses should be kept up to date.

More and Less Comprehensive Data and Analysis

Several authors (Drummond and others 1997; Gold and others 1996; Sloan 1995) provide similar guidance and recommendations for relatively comprehensive economic evaluation in general or for specific medical procedures. This volume aims at estimating cost-effectiveness for interventions against many different problems in all low- and middle-income regions, for which varying amounts and quality of information are available. It has therefore not always been possible to conduct as complete an analysis as would be desirable. Some degree of modeling is usually inescapable (Buxton and others 1997).

More complete analysis starts by characterizing, in each regional setting where an intervention is relevant (where the health problem causes some measurable burden and the intervention appears feasible), estimates of the quantities of inputs required (Q), the unit costs of those inputs (P), and the effectiveness or health gain (E). Authors were provided regional estimates of unit costs for the major inputs—salaries, facility costs, fuel and vehicle operation, drugs, representative equipment, diagnostic tests, and buildings (Mulligan and others 2003). The total cost of delivery is the sum of the input costs PQ, which is compared with effectiveness E and the CER calculated from the total costs and total effects of the proposed intervention or from the changes in those costs and outcomes compared with current practice.

The data on unit costs, quantities, and outcomes may all derive from published literature; what is original is how that information is combined to calculate cost-effectiveness rather than taking the ratios from existing studies. Estimates are built up using prices and physical inputs in the chapters on tuberculosis (chapter 16), vaccine-preventable diseases (chapter 20), malaria (chapter 21), cancers (chapter 29), psychiatric disorders (chapter 31), neurological disorders (chapter 32), cardiovascular disease (chapter 33), hemoglobin disorders (chapter 34), water and sanitation (chapter 41), indoor air pollution (chapter 42), tobacco (chapter 46), alcohol (chapter 47), community programs (chapter 56), family planning (chapter 57), surgery (chapter 67), emergency care (chapter 68), and complementary medicine (chapter 69). Several chapters analyze some interventions more fully and others less fully, depending on the available information.

As indicated in box 15.2, expanding or contracting the scale of an intervention may change the CER because of difficulty in reaching more of the population. The ratio may also vary because of the cost of identifying who would benefit most from the intervention—for example, whether to screen all newborns for sickle cell disease or only those of African origin (chapter 34). And expansion may change the cost-effectiveness because it would require considerable fixed investment. The costs of expanding capacity to deliver an intervention, including physical capital and training of human resources, should be amortized over a reasonable interval (10 years is the standard in this volume) and included in the total costs. Ideally, one would know the complete production function of the intervention, including the possibilities of substituting one input for another to minimize costs in response to differences in prices. However, analysis of this level of complexity is difficult to achieve, so most chapters assume fixed input proportions. Q, then, does not depend on P, and the CER varies (at most) only with coverage, prices, and outcomes. This result could be an underestimate of the true cost-effectiveness if much substitution is possible (see chapter 16 on tuberculosis).

Approximations are required when the average and incremental CERs have to be taken directly from the literature and when key parameter values are not easily available. Existing estimates of total cost or effectiveness may or may not incorporate the standard assumptions about discounting, disability weights, and life expectancy. Authors then need to judge how to adjust the available estimates for a more consistent analysis.

Local cost and outcome estimates that have not been constructed transparently from inputs and prices provide a less complete basis for secondary analysis. Where such estimates are used, information about how costs are constructed or how results vary with the scale of the intervention is usually not available, but the data may explicitly show regional differences in one or both elements, thereby permitting regionally differentiated recommendations (or may show differences so small that recommendations need not differ regionally). If costs and results refer to only one moment or are specified year by year, they can be discounted at 3 percent. Published analyses often use higher constant rates of 5, 6, or even 10 percent and may specify only total costs and outcomes rather than the respective streams through time. In that case, both costs and health gains occurring in the future are valued less, but conversion to a CER based on 3 percent discounting may be impossible and is at best only approximate. Some published analyses discount costs but not health outcomes, which makes interventions look more cost-effective when costs are spread over long intervals (for examples, see chapter 29). Imported estimates of cost and effect—that is, estimates from other regions, commonly from high-income countries—are often all that is available. Sometimes data on costs and outcomes derive from the same source; in other cases, they come from different sources and even different regions and are difficult to compare directly. More appropriate adjustments to total costs are possible with information on quantities of inputs. In the absence of data on quantities of resources used, differences in average cost can sometimes be calculated using estimates of input proportions. Such an approximation characterizes the analysis for diabetes (chapter 30), in which proportions were known in one region and assumed to be the same elsewhere and costs were estimated from regional cost ratios.

Cost-Effectiveness and Population Impact

An intervention CER, whether average or incremental, is based on assumptions about introduction, expansion, contraction, or modification of the activity compared with current (or sometimes "best") practice. Comparison of ratios indicates whether one intervention offers better or worse value for money than another at the individual level but says nothing about how either one affects the whole population. The analysis, therefore, includes, wherever possible, two ways of describing the latter effect. One is to consider a population of 1 million, with a typical regional age and sex structure, and to suppose that the intervention were delivered to all the potential beneficiaries. That number of people is just the prevalence or incidence of the condition times 1 million. The total cost would then be the unit cost times that number (or the cost of reaching that many people if the unit cost varies with coverage). The total health gain would be the individual effectiveness times that same number (or the overall outcome if that depends on externalities, such as the transmission of communicable disease, that are sensitive to coverage). Standardizing on a population of 1 million allows comparisons among regions and interventions in which the incidence or prevalence may vary greatly.

A second approach standardizes not on population but on expenditure: if an additional US$1 million were devoted to the resources needed for an intervention, how many people could benefit from it and how large would the health gain be? The coverage of the intervention would be US$1 million divided by the average cost, and the total gain in DALYs would be that number of people times the average effectiveness. This approach is applied in relatively few chapters because of the information requirements; its advantage is to facilitate judgments as to where increased spending would be most justified—where it would yield the largest improvement in health, reach the most people, or account for the largest share of burden from a condition. Table 1.3 in chapter 1 provides examples for some interventions to reduce child mortality, prevent or treat HIV and AIDS, reduce smoking prevalence, treat heart attack and stroke, detect and treat cervical cancer, and operate a basic surgical ward. The estimates of DALYs gained per US$1 million vary from less than 100 to more than 100,000—a thousandfold difference in value for money. Annex 26.A of chapter 26 provides both kinds of calculations, per million population, to compare the cost-effectiveness of interventions for improved maternal health in South Asia and Sub-Saharan Africa.

Improvements and Further Applications

What would improve the kind of estimates and conclusions reported in this volume? Most crucially, more and better data are needed in low- and middle-income countries to reduce reliance on extrapolation from high-income countries and on expert judgments. The need for information starts, in some cases, with better estimates of incidence and prevalence, but even where the epidemiology is well known, data on coverage and outcomes of existing interventions are scarce. Evidence of what it would cost to change coverage of existing interventions or add new interventions, and with what results, is particularly scarce and depends heavily on assumptions. This situation is sometimes true even for activities that have been conducted widely for many years and have been extensively analyzed, notably the EPI (chapter 20). Analyses should when possible be conducted at the level of a country or even smaller units, to take full account of all the reasons cost-effectiveness varies from place to place and to develop priorities on the basis of analyses appropriate to local circumstances. The methods used here are intended to help guide such efforts, and they can and should be refined through research to provide more robust help to policy.

Finally, a more concerted approach is needed for clarifying the options facing different decision makers and incorporating the results from systematic literature reviews into analytic models that compare the costs and effects of alternative interventions (Buxton and others 1997; Kuntz and Weinstein 2001). Modeling encourages explicit decision making and can deal comprehensively with the inputs and outcomes of decision options, which allows a range of uncertainties to be reflected. Thus, hypotheses about interventions can be formulated and tested statistically. Specifying models explicitly (as in chapter 16, for example) can also help identify gaps in current evidence and can capture details specific to particular populations and settings.

Acknowledgments

The authors are grateful to Jo Mulligan for the estimation and explanation of input costs used in this volume; to Sonbol A. Shahid-Salles for help in drawing methodological examples from numerous chapters; to George Alleyne, Joel G. Breman, Mariam Claeson, Anthony R. Measham, and Elinor Schwartz for helpful comments; and to Dean T. Jamison and Anne Mills for overall guidance.

References

1. Brooks, R., R. Rabin, and F. de Charro, eds. 2003. The Measurement and Valuation of Health Status Using the EQ-5D: European Perspective. (Evidence from the EuroQol BIOMED research program.) Dordrecht, Netherlands: Kluwer Academic Publishers.
2. Buxton M. J., Drummond M. F., Van Hout B. A., Prince R. L., Sheldon T. A., Szucs T., Vray M. Modeling in Economic Evaluation: An Unavoidable Fact of Life. Health Economics. 1997;6(3):217–27. [PubMed: 9226140]
3. Claxton K., Sculpher M., Drummond M. F. A Rational Framework for Decision Making by the National Institute for Clinical Excellence (NICE). Lancet. 2002;360(9334):711–15. [PubMed: 12241891]
4. Dolan P., Jones-Lee M. The Time Trade-Off: A Note on the Effect of Lifetime Reallocation of Consumption and Discounting. Journal of Health Economics. 1997;16:731–39. [PubMed: 10176781]
5. Drummond, M. F., B. O'Brien, G. L. Stoddart, and G. W. Torrance. 1997. Methods for the Economic Evaluation of Health Care Programmes. Oxford, U.K.: Oxford Medical Publications.
6. Fox-Rushby, J. A. 2002. Disability-Adjusted Life Years (DALYs) for Decision-Making? London: Office of Health Economics.
7. Gabbay J., le May A. Evidence Based Guidelines or Collectively Constructed `Mindlines?' Ethnographic Study of Knowledge Management in Primary Care. British Medical Journal. 2004;329(7473):1013. [PMC free article: PMC524553] [PubMed: 15514347]
8. Glick, H., D. Polsky, and K. Schulman. 2001. "Trial-Based Economic Evaluations: An Overview of Design and Analysis." In Economic Evaluation in Health Care: Merging Theory with Practice, ed. M. Drummond and A. McGuire, 113–40. Oxford, U.K.: Oxford University Press.
9. Gold, M. R., J. E. Siegel, L. B. Russell, and M. C. Weinstein, eds. 1996. Cost-Effectiveness in Health and Medicine. New York: Oxford University Press.
10. Hoffmann C., Stoykova B. A., Nixon J., Glanville J. M., Misso K., Drummond M. F. Do Health-Care Decision Makers Find Economic Evaluations Useful? The Findings of Focus Group Research in U.K. Health Authorities. Value Health. 2002;5(2):71–78. [PubMed: 11918822]
11. iMTA (Institute for Medical Technology Assessment). Newsletter. 2005;3(1):1–3.
12. Jamison, D. T. 2002. "Cost-Effectiveness Analysis: Concepts and Applications." In Oxford Textbook of Public Health, 4th ed., ed. R. G. Detels, J. McEwen, R. Beaglehole, and H. Tanaka, 903–19. Oxford, U.K.: Oxford University Press. Also published as Disease Control Priorities Project Reprint 3.
13. Jamison, D. T., and J. S. Jamison. 2003. "Discounting." Disease Control Priorities Project Working Paper 4, World Bank, Washington, DC.
14. Jamison, D. T., W. H. Mosley, A. R. Measham, and J. L. Bobadilla, eds. 1993. Disease Control Priorities in Developing Countries. New York: Oxford University Press.
15. Karlsson G., Johannesson M. The Decision Rules of Cost-Effectiveness Analysis. Pharmacoeconomics. 1996;9(2):113–20. [PubMed: 10160090]
16. Kuntz, K. M., and M. C. Weinstein. 2001. "Modelling in Economic Evaluation." In Economic Evaluation in Health Care: Merging Theory with Practice, ed. M. Drummond and A. McGuire, 141–71. Oxford, U.K.: Oxford University Press.
17. Mathers, C. D., A. Lopez, C. Stein, D. Ma Fat, C. Rao, M. Inoue, and others. 2005. "Deaths and Disease Burden by Cause: Global Burden of Disease Estimates for 2001 by World Bank Country Groups." Disease Control Priorities Project Working Paper 18, World Bank, Washington, DC.
18. McDaid D., Cookson R. ASTEC Group. . Evaluating Health Care Interventions in the European Union. Health Policy. 2003;63(2):133–39. [PubMed: 12543526]
19. Morgenstern, O. 1963. On the Accuracy of Economic Observations. Princeton, N.J.: Princeton University Press.
20. Mulligan, J., J. A. Fox-Rushby, T. Adam, B. Johns, and A. Mills. 2003. "Unit Costs of Health Care Inputs in Low and Middle Income Regions." Disease Control Priorities Project Working Paper 9, World Bank, Washington, DC.
21. Murray, C. J. L. 1996."Rethinking DALYs." In The Global Burden of Disease, ed. C. J. L. Murray and A. D. Lopez, 1–98. Cambridge, MA: Harvard University Press.
22. Murray, C. J. L., and A. D. Lopez, eds. 1996. The Global Burden of Disease. Cambridge, MA: Harvard University Press.
23. Musgrove P. Public Spending on Health Care: How Are Different Criteria Related? Health Policy. 1999;47:207–23. [PubMed: 10538919]
24. ———2000A Critical Review of `A Critical Review': The Methodology of the 1993 World Development Report, Investing in Health Health Policy and Planning 15(1):110–15. [PubMed: 10731242]
25. NICE (National Institute for Clinical Excellence). 2004. Guide to the Methods of Technology Appraisal. London: NICE. http://www​.nice.org.uk. [PubMed: 27905712]
26. Sheldon T. A., Cullum N., Dawson D., Lankshear A., Lowson K., Watt I. et al. What's the Evidence That NICE Guidance Has Been Implemented? Results from a National Evaluation Using Time Series Analysis, Audit of Patients' Notes, and Interviews. British Medical Journal. 2004;329(7473):999. [PMC free article: PMC524545] [PubMed: 15514342]
27. Sloan, F. A. 1995. "Introduction." In Valuing Health Care: Costs, Benefits, and Effectiveness of Pharmaceuticals and Other Technologies, ed. F. A. Sloan. New York: Cambridge University Press.
28. Smith, P. C. 2005. "Statutory Packages of Health Care Alongside Voluntary Insurance: What Treatments Should Be Covered?" York, U.K.: Centre for Health Economics, University of York.
29. Taylor R. NICE (National Institute for Clinical Excellence). . HTA Rhyme and Reason? International Journal of Technology Assessment in Health Care. 2002;18(2):166–70. [PubMed: 12053416]
30. Walker D., Fox-Rushby J. A. Critical Review of Economic Evaluations of Communicable Disease Interventions in Developing Countries. Health Economics. 2000;9(8):681–98. [PubMed: 11137950]
31. WHO (World Health Organization). 2000. World Health Report 2000. "Health Systems: Improving Performance." Geneva: WHO.
32. Wordsworth S., Ludbrook A. Comparing Costing Results in Across Country Economic Evaluations: The Use of Technology Specific Purchasing Power Parities. Health Economics. 2004;14(1):93–96. [PubMed: 15386663]
33. World Bank. 1993. World Development Report: Investing in Health. New York: Oxford University Press.
34. ———. 2003. World Development Indicators. Washington, DC: World Bank.