Health state preference scores for Canadians

In past newsletters, we have described the Population Health Impact of Disease in Canada (PHI) research program, a collaboration of Statistics Canada, Health Canada, and researchers from McGill University, the University of Ottawa, the University of Manitoba, the Institute for Clinical Evaluative Sciences (ICES) and the Montérégie Regional Board of Health and Social Services.

This program is building a framework to examine the relative impact of diseases and health conditions relevant to Canadians and the risk factors that contribute to them. Summary measures are used to combine the impact of mortality and morbidity in a single indicator. An important component of this program is the development of preference scores that help quantify the impact of a health condition in terms of health-related quality of life, or more specifically, limitations to functional capacity.

This article describes the process and preliminary results of eliciting preference scores from panels of lay Canadians. A preference score represents an individual’s relative preference for a health state compared with full health.

Health state descriptions
Preference exercises
Next steps
In summary

Health state descriptions

We used 239 health state descriptions selected to provide a broad coverage of the impact of diseases relevant to Canadians. Each health state, which could represent a specific stage in the progression or treatment of a disease, had been assigned a level for each of the 11 attributes of the Classification and Measurement System of Functional Health (CLAMES). (See Table 1.)

Top of Page

Standardized descriptions were developed for use in preference exercises to assure a common understanding of the health state. For instance, the card ML describes severe chronic asthma (Figure 1). The core attributes were listed on every card; to simplify the presentation blank space was used to indicate no limitations. Supplementary attributes were listed only if they applied to that health state.

The development of CLAMES and health state descriptions are detailed in our September 2003 newsletter.

Figure 1
Description card used in preference measurement

Top of Page

Preference exercises

A subset of these health state descriptions was considered by groups of Canadians in 9 Canadian communities: Vancouver, Edmonton, Saskatoon, Toronto, Ottawa, Montréal, Québec, Moncton, and Halifax. Full-day sessions were led by a trained facilitator from Statistics Canada assisted by a member of the study team. Each group of 8 to 11 individuals included a mix of gender, age group and socio-demographic profile. Individuals with activity limitations due to health conditions were included in each group. One panel comprised only rural dwellers, and four groups were conducted in French.

We used a thermometer-like scale (visual analogue scale) to orient participants to the descriptions and the concept of preference measurement. Then we used the Standard Gamble technique to elicit preference scores on a scale where dead is 0 and full health is 1.

Participants were asked to imagine that they were living in the health state represented on the card and to consider how living with the health state would affect their own life. They were asked to assume that they would be in that state of reduced health for the rest of their life.

They were then offered a hypothetical procedure that had a certain probability of restoring them to full health with a corresponding probability of ending their life instantly (for instance, 80% probability of full health and 20% probability of death).

They had to choose whether they wished to take the procedure (with the specified probability of success) or to remain in the state of reduced functional capacity for the remainder of their life. The probabilities were then varied until the point at which participants found it most difficult to make a decision whether to take the procedure or not.

For 12 anchor states considered by all groups, individuals recorded their scores before and after group discussion. Group discussion helped participants clarify information about their decisions and hear other opinions, but they were not encouraged to reach consensus.

Another 227 health states were randomly assigned as individual exercises (14 to each participant). Each health state was thus considered by at least six individuals. Although the health states were presented without disease labels, most represented a disease under study.

Table 2 shows the average or mean of the scores provided in the group exercises for the anchor states. As one could expect, there was some variation in the scores between individuals. This variation could reflect personal situations and values, for instance, a parent might view functional limitations differently from a young adult without children. Individuals with strong family supports systems might consider severe functional limitations differently from those who live by themselves without these supports.

These exercises demonstrated that lay Canadians were willing and able to use the Standard Gamble in a group setting. Participants expressed enthusiasm about contributing to measurements for health state preferences and were eager to be informed about the results of these exercises.

Next steps

The mean scores for 239 health states are being used to develop a statistical function that establishes a relative weighting for each level of each attribute. With this function, we can estimate preference scores for any combination of levels for the eleven attributes—a theoretical possibility of over 10 million health states!

Preference scores are an important component of our summary measures, health-adjusted life years lost to a disease (HALYs), because they contribute a measure of severity for each health state. We first calculate year-equivalents lost to reduced functioning (YERFs) which are the product of incidence (number of cases), duration, and a weight for severity (1 minus the preference score). These are calculated for each stage in the progression and treatment of the disease across every age group and for both genders. YERFs are then combined with years of life lost to the disease (YLL) to obtain HALYs lost to the disease.

To date, we have calculated HALYs for 26 cancer types and five risk factors that contribute to them. Our focus in the next year will be on the contribution of obesity to several major disease groups including diabetes and cardiovascular diseases. The ultimate goal of this research program is to obtain a relative ranking of diseases and to determine the overall impact of health determinants that affect Canadians.

In summary

As part of the Population Health Impact of Disease in Canada (PHI) research program, we elicited preference scores from small groups of the Canadian lay population.

We used standardized descriptions based on literature review and medical expertise to classify and measure the impact of functional limitations for 239 health states. Standardized descriptions helped make sure that individuals had a common understanding of health states.

The observed scores are being used to establish a statistical function that contributes to measures of the relative impact of disease and health determinants in the Canadian population.

Health preference measurement makes an important contribution to assessing the relative impact of diseases on health-related quality of life because it quantifies the severity of functional limitation.

Julie Bernier and Kathy White

Julie Bernier came to Statistics Canada in 1995, after working as a statistical consultant and teaching statistics at Université Laval. She joined the Health Analysis and Measurement Group in 2000 as Chief of the Health Measurement Section. Her research interests are generic measures of population health, health-related quality of life indicators, and health state preference measurement.