|Health Economics for Developing Countries: A Survival Kit (London School of Hygiene and Tropical Medicine, 1998, 134 p.)|
1. Introduction: The Value of Health Indicators
In Chapter 6, the need for measures of health outcome both in physical and in value terms was noted. These are required because it is necessary to compare the magnitude and value of inputs and outputs in order to evaluate economic efficiency. Why though should this present a problem in the health sector, when many other fields of human activity seem to require no such explicit measures?
The answer lies partly in the fact that in most other areas of production or service provision, it is possible to measure and value output directly in money terms by market prices. It is also possible to infer some measure of efficiency through profits earned and the behaviour of the industry. However, in many of the 'social services', and in the health sector in particular, prices have often been reduced or eliminated, so that profit has ceased to be a useful measure of efficiency or of effectiveness. Furthermore, certain characteristics of the market for health and health services and the institutional characteristics of providers may reduce pressures for efficiency.
Perhaps a more obvious problem is the relative difficulty of identifying just what the output of a health system is. By comparison it is easy to say what the output of, say, a car factory is. These difficulties have meant that historically, the evaluation of health systems has focused on aspects of performance that are relatively easy to measure and define. Measures of service provision (e.g. beds), or activity (e.g. number of patients), or other inputs have frequently been used as if they were measures of outcome.
If it is accepted that the primary purpose of health services is to produce health, it is necessary that one clearly defines what one means by 'health'. This question has been receiving a great deal of research attention in recent years. Amidst the wide variety of approaches, it is possible to discern a number of central themes. Most approaches to the definition of health recognize that it is not merely the absence of disease or infirmity but a generic term seeking to encompass all dimensions of an individual's, group's or population's level of well-being. Likewise most definitions accept that health is not an either/or state, but rather a continuum: at the lower boundary one might place the state of death; and at the upper boundary a state of perfect health, or, perhaps more realistically, some statistically derived norm. All definitions accept that health is a multi-dimensional concept and not merely some state of biological dysfunction, although this acceptance is not necessarily reflected in measures of health used in practice.
However, from this broad level of consensus, opinions differ on a number of crucial points. In particular, they differ over what elements of well-being should be included in the definition, how they should be measured, and how and whether or not they should be valued and incorporated into a single unitary index of health status.
A number of techniques have been developed that attempt to use some of these conceptual approaches to health status measurement. A necessarily brief overview of these is given below.
2. Measures of Mortality and Morbidity
These have been the most frequently used measures of health. The first, mortality, is simply an attempt to measure the quantity of life; whereas morbidity is an attempt to measure the quality or healthfulness of life. Both indicators suffer from serious deficiencies.
Mortality, despite its relative lack of ambiguity, is a highly insensitive measure of health and gives little indication of health dimensions (such as disability) that do not result in death. Morbidity, although a potentially more sensitive measure also has its deficiencies. Whilst it is a measure (if very imperfect) of the quality of life, morbidity studies have often relied on individuals recognizing a health problem and seeking treatment. In practice much illness is not reported and is not therefore recorded in morbidity statistics. In addition it is difficult to make comparisons between a day of illness for different types of morbidity.
3. Measures of Function and Dysfunction
This approach to health definition and measurement recognizes that there is a social role to health and illness as well as a biological one. Health is therefore measured indirectly by its effects on the ability of individuals to perform their normal social roles. For example, different health problems have been studied with regard to the number of days of work that are lost because of them. This provides a useful way of measuring actual health impact, and is a considerable advance on simple measures of resource or service provision. Alternative measures look at the ability to perform everyday activities such as climbing stairs, dressing oneself or shopping.
4. Health profiles
The general feature of the health profile approach is the identification of a number of dimensions of health with a subsequent scaling of an individual or population along each dimension. As the name implies, the approach provides a profile of certain selected dimensions of health at one time. Repeated measurements make it possible to compare health profiles over time; and arguably, to assess the relative efficacy of particular interventions. A characteristic of the health profile is that it is possible to tailor the dimensions of health measured to suit the particular health problem under study.
5. Global Health Indicators
A common characteristic of all the approaches previously discussed is that they attempt to describe and measure health along one dimension of health at a time without providing any basis for the comparison or ranking of one dimension of health with another. This has been a conscious and intended decision of some writers who argue that any relative valuation or ranking of different dimensions of health is necessarily highly subjective and best avoided.
This view has been challenged by those who argue that it is necessary to rank different dimensions of health relative to each other, and thus to be able to value different levels of health in a single index. They argue that the health profile approach is useful when studying changes in one dimension of health at a time; or to a lesser degree, where more than one dimension changes at the same time, but where the changes are all in the same direction. It is of no use they argue, where one is required to assess changes where one dimension of health is traded-off at the expense of another.
A number of different approaches to the global health indicator have been developed. Although they differ in some respects the common features can be identified as:
- the selection of the dimensions of health to be included in the index
- a classification of health stales within these descriptive categories
- the derivation of a scale allowing the relative valuation of the above stales.
Typically, different health states will be described in terms of such dimensions of health as pain and distress, or discomfort and disability. These dimensions will then be incorporated together to provide one index along which all feasible health outcomes can be valued relative to each other.
The value of the global indicator approach is that, in theory at least, it allows direct comparison and ranking of options that produce different types of health improvement, or involve different types of interventions. It is the ideal type of indicator for cost-utility or cost-benefit evaluation (see Chapter 9).
6. The Quality Adjusted Life Year (QALY)
A problem common to all of these approaches is that they focus on a measure of health status at a particular time. Health is not a static concept and in practical terms one is also concerned with how health changes over lime and with life expectancy. One method that attempts to include the duration as well as the quality of health output is the QALY, or Quality Adjusted Life Year, where the number of years of life gained from an intervention is adjusted by a measure of their quality. This is particularly relevant to the evaluation of treatment for chronic diseases (such as kidney dialysis) where treatment enables survival for a certain period of time at a less than perfect state of health.
Central to this approach is the assumption that it is possible, even necessary, to make comparisons and 'trade-offs' between the quality and the quantity of life. The first part of the QALY approach is essentially the same as that for the global health indicator approach. However, having derived a single index, it is then necessary to provide a weighting for the health index relative to the duration of the health state. This allows assessment of the relative desirability of choices between being relatively healthy for a short period of time, or being in poorer health for a longer period. As this is the sort of choice that needs to be made in the real world the value of the approach is clear. At present, studies treat each additional year of life gained as being of equal value regardless of the age at which it is gained, although this weighting could of course be changed.
Recently the QALY has been used to compare the desirability of different health intervention strategies. The resource costs of different types of intervention are calculated and are then expressed in terms of how many quality adjusted life years they would produce. This makes it possible to identify the relative outputs of different strategies and to determine which procedures represent good value for money in terms of cost per QALY.
One rough form of QALY is the measure Healthy Days of Life Lost, which has been used with Ghanaian data to compare the relative impact of different diseases. The measure combines information on the incidence rate, case-fatality rate and the extent and duration of disability produced by a disease. It has also been modified to allow for time preference (discounted years of healthy life) and for age preference (discounted productive years of healthy life).