Institute of Health Systems

     

Training
Burden Of Disease
AP Burden of Disease Study
AP Health State Valuation Study
Cause of Death
BOD Training Programmes
DFHS
Indoor Air Pollution
Burden of Disease and Socioeconomic Impact of HIV/AIDS
Air Pollution and Cause of Deaths in Hyderabad
Technical Support for Verbal Autopsy
Cause of Death Coding for AP Rural Health Initiative
Health Effects of Air Pollution in Hyderabad
Socioeconomic Impact of Asthma
Health System Performance Assessment
Health Sector Reform 
Health Informatics
Public Health Databases
Health System-Fact Sheets
Publications
Library
About IHS
Home

 

IHS Mission & Goals:
Groom Skills,
Gather Evidence and
Generate Knowledge for people's health.

To Improve the Efficacy,
Quality & Equity
of Health Systems.

  

AP Health State Valuation Study    

 

Review of literature Methods of the HSV Study
 
Future research needs Summary findings

Review of literature
Theoretical considerations about use of health state valuations:

Theoretical interpretations about the object of measurement have some impact on the methodological path leading up to health status weights. Three different interpretations have been made in the literature, namely: (a) individual preference, (b) descriptive measure of health state and (c) social preference weight attached to the health state. Culyer (1989) distinguished between "welfarist" and "extra-welfarist" approaches to health status measurement among economist. The "Welfarist" approach consists of viewing individual preferences (utility) as the source of all social welfare. Health state or disability weight is viewed to represent individual preference for different health states. Since health outcomes at a personal level are characterised by uncertainty, Von Neuman and Morgenstern’s (VNM) expected utility theory is applied. Thus health state weights are viewed as VNM utility. Viewing health state weight as a measure of personal utility means that an equivalence with utility of other goods and services is straightforward. Advocates of personal utility interpretations may not, however, emphasise this equivalence in deference to strong emotional responses against valuation of human life in money terms. "Extra-welfarists" view health as the principal outcome of health services. The health state weight describes this output i.e. the health -related quality of life (Bleichrodt, 1997). The third interpretation, that quality adjustment weights are values under a social welfare function ,is proposed by Nord (1994). Patrick et al (1973) also recognised that quality adjustment weights derived from the equivalence (person tradeoff) method would represent valuation under some social welfare function.

 

Another analytic issue is concerned with how to treat the large amount of illnesses that exist but for which people do not seek treatment. It has been observed that aggregate measures like the sickness rate (number of persons falling ill per time period) or illness episodes per time period based on simple count and the implied equal weightage to illnesses of all severity, are too stable and non-responsive to variations of incidence of more severe illnesses (Logan and Brooke 1957). This is in fact the primary motivation behind the search for a set of unequal health state weights. The relative weight to illness of different severity will depend on the current concept of health. At a time characterised by survival as the dominant concept of health, weights attached to all forms of morbidity are nearly zero. As the concept of health evolves to include absence of disease, more severe forms of disability begin to receive higher weightage, both in the minds of the patients and public. Further evolution of the concept of health to include quality of life would naturally enhance the weightage received by illnesses considered minor and trivial in an earlier era. Rosser (1983) notes that even though Logan and Brooke, in 1957, sought to increase the sensitivity of aggregate indicators of morbidity by splitting down some categories (thereby assigning zero weightage to excluded conditions), such an approach would not be relevant in view of increasing concerns about these so-called minor and trivial illnesses.

 

Health status measures differ in sensitivity to minor and trivial illnesses. For example, the Quality of Well Being scale (QWB) is known to be more sensitive to small departures from perfect health (McDowell and Newell, 1996 p.483-491). This is because the QWB construct includes a symptom complex dimension. For example, Erickson et al (1989) found that the QWB scale classified 95% of the 45-64 year -old population in less than perfect health compared to 75% when activity limitation was used as the criterion. On the other hand EuroQol is less sensitive to small departures from perfect health (McDowell and Newell, 1996 p480-483). This instrument differentiates between more severe forms of morbidity but lumps all morbidities at the healthy end of the scale.

 

Thus, the choice of instrument in assigning health state weight will have an impact on the kind of policy application to which the resultant aggregate measure of disease burden can be assigned. For example, generic health status measures insensitive to small variations in functional status in a particular dimensions or sub-dimension, would not be suitable to demonstrate the efficacy of a new therapy acting to improve the quality of life in that dimension or sub -dimension alone. This is the rationale for many condition-specific health status measurement scales. Here the policy application of health status measurement is in approval of new formulations and procedures, etc. A different kind of policy application involves the allocation of resources at the macro economic level. For this purpose, a generic health status measurement allowing for comparison of a large number of health care interventions would be useful. Within the class of generic health status measurement, scales may vary in their sensitiveness to different levels of disability. For example, if a health state valuation instrument is not sensitive to minor and trivial illnesses, the resultant disease burden estimates would not be useful for planning of ambulatory services.

 
How to describe health states for valuation:

When asked to value time spent in a health state without information about the key domains of health, a respondent must guess what the level in those dimensions will be. This will inevitably introduce measurement error and a potential for bias in the results. Another consideration is how to convey relevant information about a hypothetical health state to the individual undertaking the valuation, who might not have personally encountered the state. Disease labels are short and parsimonious, but do not convey adequate information about functional status. Moreover, disease labels are vulnerable to different interpretations based on cultural and personal settings. Issues relevant to development of a health state description systems have been described by Boyle and Torrance (1984), and Froberg and Kane I-IV (1989). Briefly, four important considerations guide us in defining the description space (number of dimensions) and the inclusion of specific attributes of human health.

 
1. Conceptual definitions of health and deduced description systems.
2. Empirically gathered health-related attributes, and description systems induced by them.
3. Attention span and cognitive capacity of the human mind to process multi-dimensional information.
4. Statistical analysis of multi-attribute measurements.
 
How to convey health state descriptions effectively to an individual undertaking the valuations:

Effective communication of the description to individuals acting as valuers has many difficulties. The descriptive system must be comprehensible to the young, middle-aged and older adults with widely varying levels of educational attainment, socio-economic and cultural backgrounds. For example, differences have been found between using paragraphs written in the first person in describing a health state, and using straight lists of levels in each domain of health (Llewellyn-Thomas et al. 1982). The descriptive system should be meaningful across cultures. Translation of instruments should produce equivalence in terms of word meanings and idioms i.e. semantic and idiomatic equivalence; equivalence in terms of situations and concepts evoked in the descriptions i.e. experiential and conceptual equivalence, respectively (Guillemin et al, 1993). The description system should enable communication with semi-literate as well as illiterate persons. The description systems used so far have been developed for literate societies like North America and Europe. Even here, studies have experienced communication difficulties due to language barriers. For example, in the Canadian study by Sackett and Torrance (1978), about 12% of the randomly selected sample had to be excluded, because the interviewees could not communicate in English. One way to deal with this problem is to supplement written descriptions with appropriate graphical representations. Some researchers have used multimedia methods for valuation exercises (Lennert and Hornberger 1996, Lennert and Soetikno 1997). One problem with multimedia solutions is that the computer may be a source of distraction, particularly where the general community has limited experience with multimedia. In any case, multimedia solutions need a graphical description system to start with. So description systems for partially literate and multi- lingual communities should ideally include a graphical description sub system.

 

Next  

  

| AP State Profile  | Public Health Database  | IHS Resource  |  Careers  | Contact us |