If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
School of Public Health, University of Montreal, Montreal, QC, CanadaCentre de recherche de l’IUSMM, Montreal, QC, CanadaCRCHUS, CIUSSS de l’Estrie–CHUS, Sherbrooke, QC, Canada
Population norms for the EQ-5D-5L were published in Canada but only for Alberta province. The purpose of this study was to derive Quebec population norms from the EQ-5D-5L.
Methods
The data came from a larger study conducted between September 2016 and March 2018 using elicitation techniques for a quality-adjusted life-year project. The online survey was distributed randomly in the province of Quebec. To best describe the entire population, data were stratified by various sociodemographic characteristics such as age, gender, urban and rural populations, whether disadvantaged or not, immigrant or nonimmigrant, and health problems.
Results
A total of 2704 (53.8%) respondents completed the EQ-5D-5L. Mean (95% confidence interval) and median (interquartile range) utility scores were 0.824 (0.818-0.829) and 0.867 (0.802-0.911), respectively. The EQ-VAS scores were estimated at 75.9 (75.2-76.6) and 80 (69-90). Subjects with lower scores were those who had a low or high body mass index; were smokers; were single, divorced, or widowed; had no children; were unemployed or sick; had lower education or lower annual income; and had a family or personal history of serious illness. Immigrants had higher scores. There was no difference in gender and urban or rural population. The score logically decreased with worsening health status, from a mean score of 0.896 (0.884-0.908) to 0.443 (0.384-0.501; P < .0001. Similar results were observed for subjects’ satisfaction with their health or life. Subjects with lower scores were less willing to take risks. Subjects who declared they were affected by health problems presented significant lower utility scores, ranging from 0.554 (nervous problem) to 0.750 (cancer), compared with those without health problems (0.871; confidence interval: 0.867-0.876).
Conclusion
This is the first study to present utility score norms for EQ-5D-5L for the Quebec population. These results will be useful for comparison with quality-adjusted life-year studies to better interpret their results. Moreover, utility norms were provided for 21 health problems, which was rarely done.
For several years, economic studies have used cost-utility analysis with outcomes expressed as quality-adjusted life-year (QALY), which combines quantity of life with a measure of quality of life into a single score, namely, a utility score.
Utility score norms estimated for a population may serve as a reference group, especially for studies that do not have a control group and want to estimate the effect of their intervention.
Using the Internet to collect EQ-5D norm scores: a valid alternative?.
in: Busschbach J. Rabin R. De Charro F. Proceedings of the 24th Scientific Plenary Meeting of the EuroQol Group; Kijkduin-The Hague. The EuroQOL Group,
The Netherlands2009
These data can also be used to document the burden of disease experienced by these individuals, particularly in comparison with other population groups and the general population, to help establish priority strategies in the provision of care for the population.
The original version, the EQ-5D-3L, comprised 5 dimensions with 3 levels each. A new version, the EQ-5D-5L, was recently developed, keeping the original 5 dimensions but expanding the response options from 3 to 5 levels to reflect no, slight, moderate, severe, and unable/extreme problems.
This allowed obtaining a wider range of health state possibilities. Population norms for the EQ-5D-5L were published in Canada but only for Alberta province. Otherwise, value sets for the EQ-5D instrument were published in Canada, a first for the short version EQ-5D-3L
Nevertheless, no population norms were derived from these studies, and for the EQ-5D-5L, the authors indicated that the sample used cannot be considered as representative for the province of Quebec. Considering other health-related quality-of-life (HRQoL) measures, Canadian normative data were established for the SF-36 health survey, including the Quebec province.
recently provided utility score norms for several age, gender, and jurisdiction strata in Canada, using the Health Utilities Index Mark 3.
Quebec’s population is different from other Canadian provinces. Indeed, Quebec is the only Canadian province that has a predominantly French-speaking population,
It is therefore important to have reference values for Quebec and not only for Canada for studies in this province.
The purpose of this study was to derive Quebec population norms for the EQ-5D-5L. The specific objective was to describe the entire population, stratified for various sociodemographic characteristics of individuals, particularly for age, gender, urban and rural populations, whether disadvantaged or not, immigrant or nonimmigrant, and health problems.
Methods
Data
The data come from a larger study conducted between September 2016 and March 2018 using elicitation techniques for a QALY project. Specifically, the aim of this larger study was to compare different elicitation methods for QALY (i.e., standard gamble versus discrete choice experiment). The company Research Now SSI (Survey Sampling International) distributed the online survey randomly among its members in the province of Quebec. Members decided whether they were interested in participating by clicking within a list of surveys provided by Research Now SSI. The first page of the survey presented the survey content, usefulness of this study, advantages and disadvantages for the participant, and a contact person for questions or complaints. The company made sure there could not be multiple completions by using a unique ID by respondent. As the EQ-5D-5L was the last issue of the survey, only those who completed it entirely were selected for analysis. Respondents did not receive an amount directly for completing the survey. The survey company gave them an amount depending on the number of questionnaires answered.
Variables
The population HRQoL was measured with the EQ-5D-5L questionnaire, which included 5 dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) with 5 levels each, ranging from no problem to extreme problem. The EQ-5D-5L questionnaire also included a vertical visual analog scale (VAS), with 100 on the top representing the “best imaginable health state” and 0 at the bottom representing the “worst imaginable health state.” Health utility scores were calculated with the value set developed by Xie et al
and using their recommended model. They used a composite time trade-off composed of a traditional and a lead time trade-off. Their sampling was from the general population of 4 cities across Canada, including Montreal, a city in the Quebec province. The health utilities elicited ranged from –0.148 for the worst (55555) to 0.949 for the best (11111) EQ-5D-5L health states. Because health utility scores were calculated with a linear model, worst and best states were different from 0 and 1. The ceiling effect was the proportion of respondents reporting “no problems” for all dimensions.
The questionnaire included a large number of covariates to describe the sample, including sociodemographic variables (age, gender, occupation, education, marital status, annual income, smoking), if the respondent lives with another adult, is responsible for children younger than 18 years, lives in a rural or urban area, has home ownership (yes/no), has residency status in Canada, has a personal or family history of health problem, has a self-reported health problem affecting quality of life, and their general health status and satisfaction with general health. Health status was measured with a 5-point Likert scale ranging from excellent to poor. Health and life satisfaction scores were measured with a Likert scale ranging from 0 to 10, from not at all satisfied to fully satisfied. Willingness to take risks was measured using a VAS, where 0 represented hate taking risks and 10 represented love taking risks.
Statistical Analysis
The health utility score was estimated with the mean (95% confidence interval [CI]) and median (interquartile range [IQR]). To evaluate whether the sample was representative of the general population, descriptive statistics of the Quebec population were added when available by Statistique Canada or the Institut de la Statistique du Québec. To describe utility scores among subpopulations, EQ-5D-5L scores were compared with covariables using Kruskal-Wallis and Wilcoxon tests. A multivariate ordinary least squares regression model was computed on inverse logarithm utility score transformation to determine which demographic characteristics were important indicators. Inverse logarithm transformation was used to obtain residuals normally distributed. Only demographic variables with P < .05 were included in the multivariate model (ie, age, gender, body mass index [BMI], smoking, married/living with a partner, employed, education in years, annual income, living with an adult, have a child, history of serious illness, family history of serious illness, home ownership, Canada as country of origin, health status, health satisfaction, life satisfaction, willingness to take risks). The White correction was used to control heteroscedasticity. As the EQ-5D-5L was requested at the end of the questionnaire, there were no missing data for the variables included in the model, so no imputation was needed. The ceiling effect was calculated for each subpopulation. The health utility score was described for age categories, BMI, and health status for men and women separately. The health utility score was calculated for each HRQoL. Finally, the percentage of patients for each level of each dimension was presented and compared by gender and age. All analyses were also computed for EQ-VAS and presented in a supplementary analysis. Analyses were computed with SAS software (version 9.4; SAS Institute Inc, Cary, NC) and graphed with GraphPad Prism version 7.00 for Windows. A P value less than .05 was considered significant.
Ethical Approval
This project was approved by the institutional ethics committee (Comité d’éthique de la recherche of the CIUSSS de l'Estrie–CHUS #2016-1350). The subject’s consent was obtained by clicking on the start button at the end of the explanatory letter. All questionnaires were anonymous, and the data were stored on a protected network at our institution to ensure confidentiality and protection of the respondents.
Results
Sample Characteristics
Of the 5028 subjects who were directed to the explanatory letter, 366 (7.3%) did not go further; 1958 (38.9%) started it, but did not complete until the end; and 2704 (53.8%) completed the EQ-5D-5L, but 9 did not complete the EQ-VAS. The 2324 subjects who did not complete the questionnaire were mostly older, were alone, had no children, were retired, were less educated, had lower income, and had more health problems (Supplementary Table 1 found at https://doi.org/10.1016/j.jval.2019.08.008). The 2704 completers were globally representative of the general population but had a little more education, were less urban, and fewer were immigrants (Table 1). People older than 75 years were less well represented.
Table 1Study sample characteristics: comparison with Quebec general population aged 18 years or older and EQ-5D-5L index.
General population for education was 25 to 64 years. For history of serious illness, Canadian citizenship, Canada as origins country, and health status, the population was 15 years and older.
For health status, 89.7% considered themselves to be in good health (excellent to good).
0.896 (0.884-0.908)
0.949 (0.905-0.949)
<.0001
50.3%
Very good
37.1%
0.879 (0.873-0.884)
0.905 (0.867-0.949)
27.7%
Good
37.9%
0.815 (0.807-0.823)
0.854 (0.785-0.905)
10.4%
Fair
9.7%
10.3%
0.664 (0.640-0.687)
0.724 (0.558-0.808)
2.3%
Poor
2.8%
0.443 (0.384-0.501)
0.423 (0.227-0.660)
1.3%
Satisfaction with health
0-3
8.5%
0.582 (0.551-0.613)
0.645 (0.404-0.771)
<.0001
1.7%
4-6
26.3%
0.777 (0.765-0.789)
0.828 (0.726-0.867)
8.4%
7-10
65.1%
0.874 (0.870-0.879)
0.905 (0.860-0.949)
28.3%
Satisfaction with life
0-3
5.7%
0.608 (0.568-0.647)
0.645 (0.447-0.810)
<.0001
3.9%
4-6
24.6%
0.775 (0.763-0.788)
0.829 (0.725-0.874)
7.8%
7-10
69.7%
0.858 (0.853-0.864)
0.905 (0.829-0.949)
26.7%
Willingness to take risks
0-3
29.5%
0.786 (0.772-0.799)
0.866 (0.742-0.905)
<.0001
15.0%
4-6
41.4%
0.836 (0.828-0.845)
0.867 (0.808-0.911)
22.0%
7-10
29.0%
0.843 (0.834-0.853)
0.867 (0.810-0.911)
25.0%
CI indicates confidence interval; IQR, interquartile range (25th–75th percentiles).
∗ General population for education was 25 to 64 years. For history of serious illness, Canadian citizenship, Canada as origins country, and health status, the population was 15 years and older.
† Last categories for body mass index (BMI) included BMI ≥30.
‡ For occupational status, 27.4% were “other” (probably included retired and at home).
§ For education, 25.5% included baccalaureate, masters, and PhD degrees.
‖ For health status, 89.7% considered themselves to be in good health (excellent to good).
A histogram of the utility score distribution and EQ-VAS is presented in Figure 1. As expected, the distribution was not normal. The upper bound for utility scores was 0.949, and the lower bound was –0.148. Only 6 subjects had negative utility scores. Of the subjects, 20.8% had the highest EQ-5D-5L utility score (EQ-5D-5L utility score = 0.949, 11111 health state) and only 4.3% for the EQ-VAS (VAS = 100). The other most common health states were 11121, 11122, 11112, and 11123 in 14.5%, 11.4%, 6.7%, and 3.5% subjects, respectively. The mean (95% CI) and median (IQR) utility scores were estimated at 0.824 (0.818-0.829) and 0.867 (0.802-0.911), respectively (Table 1).
Figure 1Distribution of utility score and VAS for EQ-5D-5L. The values on the x-axis correspond to the central value with an interval of 0.05 for EQ-5D-5L score and 5 points for EQ-VAS. The first band/value corresponds to –0.148 to –0.10 for EQ-5D-5L score and 0 to 2.5 for EQ-VAS. The last band/value corresponds to the ceiling effect, 0.949 for utility score (n = 562, 20.8%) and 100% for VAS (n = 116, 4.3%). VAS indicates visual analogue scale.
EQ-5D-5L Utility Scores and EQ-VAS According to Sample Characteristics
Subjects with lower scores were those who had low or high BMI; were smokers; were single, divorced, or widowed; had no children; were unemployed or sick; had lower education or lower annual income; and had a family or personal history of serious illness (Table 1). Immigrants had higher scores. There was no difference in gender and urban or rural population. Logically, the score decreased with worsening health status, with a mean score of 0.896 (0.884-0.908) to 0.443 (0.384-0.501; P < .0001). Similar results were observed for subjects’ satisfaction with their health or life. Subjects with lower scores were less willing to take risks. The mean (95% CI) and median (IQR) EQ-VAS were estimated at 75.9 (75.2-76.6) and 80 (69-90), respectively (Supplementary Table 1). Similar results were observed for the EQ-VAS and are presented in a supplementary file. In the multivariate regression model, because EQ-5D-5L values were transformed with an inverse logarithm to obtain a normal distribution of residues, the estimates must be interpreted in the opposite sense. Women of older age with higher BMI, history of serious illness, and family history of serious illness independently and significantly reduced the health utility score. On the contrary, employment, higher education, living with an adult, higher health status, higher health, and life satisfaction independently and significantly increased the health utility score (Supplementary Table 5).
Norm Scores for Age, BMI, Health Status, and Comorbidity According to Gender
The EQ-5D-5L utility score was presented by age, BMI, and health status categories and divided by gender (Table 2). Results were almost similar for men and women. The distribution was different for age between genders. For men, the utility score decreased with age, and the lowest score was for those 75 years and older. For women, the lowest mean score was between 45 and 59 years, and the lowest median score was for those between 50 and 54 years. A description of the EQ-5D-5L utility score according to problems affecting quality of life is presented in Table 3. For men, problems that affected them the most were the nervous system, injury, and musculoskeletal problems other than osteoarthritis or arthritis. For women, it was principally genitourinary problems, but these were represented in only 13 subjects. In addition to genitourinary issues, nervous system and musculoskeletal problems other than osteoarthritis or arthritis were most common, as in the men.
Table 2EQ-5D-5L index score according to gender.
Men
Women
n
Mean (95% CI)
Median (IQR)
n
Mean (95% CI)
Median (IQR)
Age, y
18-24
62
0.835 (0.797-0.873)
0.905 (0.828-0.949)
130
0.847 (0.824-0.869)
0.874 (0.829-0.911)
25-29
68
0.861 (0.826-0.896)
0.911 (0.829-0.949)
125
0.867 (0.850-0.884)
0.874 (0.829-0.949)
30-34
101
0.827 (0.795-0.858)
0.905 (0.785-0.949)
152
0.844 (0.824-0.864)
0.867 (0.829-0.911)
35-39
93
0.864 (0.839-0.890)
0.905 (0.847-0.949)
147
0.819 (0.794-0.844)
0.867 (0.771-0.911)
40-44
87
0.848 (0.814-0.881)
0.905 (0.829-0.949)
133
0.808 (0.781-0.835)
0.867 (0.784-0.905)
45-49
112
0.822 (0.788-0.855)
0.879 (0.806-0.920)
124
0.789 (0.752-0.826)
0.867 (0.765-0.905)
50-54
153
0.803 (0.773-0.833)
0.867 (0.784-0.905)
200
0.793 (0.770-0.816)
0.847 (0.751-0.905)
55-59
160
0.812 (0.785-0.840)
0.867 (0.784-0.908)
169
0.799 (0.773-0.825)
0.866 (0.784-0.905)
60-64
157
0.822 (0.798-0.845)
0.867 (0.790-0.905)
154
0.822 (0.796-0.847)
0.867 (0.807-0.905)
65-69
126
0.842 (0.819-0.866)
0.867 (0.810-0.905)
107
0.820 (0.792-0.848)
0.867 (0.771-0.905)
70-74
58
0.836 (0.795-0.876)
0.872 (0.803-0.911)
57
0.818 (0.780-0.855)
0.866 (0.802-0.905)
≥75
18
0.779 (0.680-0.877)
0.856 (0.743-0.905)
11
0.842 (0.756-0.929)
0.867 (0.828-0.949)
Body mass index, kg/m2
<18.5
26
0.755 (0.672-0.837)
0.820 (0.704-0.911)
57
0.813 (0.767-0.860)
0.867 (0.725-0.949)
18.5-24.9
360
0.847 (0.831-0.862)
0.905 (0.829-0.949)
572
0.848 (0.839-0.858)
0.867 (0.829-0.911)
25-29.9
483
0.849 (0.838-0.861)
0.905 (0.828-0.911)
415
0.832 (0.818-0.846)
0.867 (0.803-0.911)
30-34.9
200
0.812 (0.786-0.838)
0.867 (0.795-0.911)
249
0.787 (0.765-0.810)
0.854 (0.764-0.905)
35-39.9
77
0.740 (0.693-0.787)
0.821 (0.679-0.885)
125
0.773 (0.740-0.806)
0.828 (0.725-0.905)
≥40
49
0.744 (0.687-0.801)
0.800 (0.708-0.905)
91
0.734 (0.688-0.780)
0.829 (0.673-0.867)
Self-reported health status
Excellent
164
0.894 (0.875-0.914)
0.949 (0.905-0.949)
174
0.897 (0.882-0.912)
0.911 (0.867-0.949)
Very good
456
0.885 (0.877-0.893)
0.905 (0.867-0.949)
547
0.873 (0.866-0.880)
0.905 (0.866-0.911)
Good
430
0.818 (0.805-0.830)
0.859 (0.795-0.905)
595
0.813 (0.803-0.824)
0.847 (0.784-0.905)
Fair
107
0.669 (0.632-0.705)
0.726 (0.560-0.802)
154
0.660 (0.630-0.691)
0.717 (0.558-0.810)
Poor
38
0.447 (0.360-0.533)
0.442 (0.255-0.660)
39
0.439 (0.357-0.521)
0.411 (0.224-0.616)
CI indicates confidence interval; IQR, interquartile range (25th-75th percentiles).
When the 5 dimensions of the EQ-5D-5L were considered separately (Table 4), 27.1% considered having mobility problems, 8.4% self-care problems, 29.1% problems for usual activity, 67.9% pain or discomfort, and 53.3% anxiety or depression. Women had significantly fewer self-care problems (P = .03), more pain or discomfort (P = .01), and more anxiety or depression (P < .001). Distributions by age categories and gender are presented in Supplementary Table 6. The difference between genders was principally between 25 and 35 years for self-care problems, between 25 and 50 years for pain or discomfort, and between 35 and 39 years for anxiety or depression.
Table 4Reported problems by 5 dimensions of the EQ-5D-5L (%).
The objective was to establish population norms for Quebec, a province of Canada. To our knowledge, this is the first Quebec study to report EQ-5D-5L results for a randomly selected large sample. This will help better compare utility values in a different cultural setting and provide reference values for clinical studies conducted in Quebec.
Some studies have established population standards for Canada, but some do not include Quebec,
Results for the EQ-5D-5L from Alberta province clearly showed a decrease in the age scores (0.88 to 0.82) with slightly higher values in men, but these were not significant. In our case, the EQ-5D-5L is lower in men 75 years and older, but this decrease is not as dramatic as for Alberta.
For women, it is rather the categories 45 to 59 years that are the worst, suggesting that older participants were in better health and possibly less representative of their cohort, unless this means the first years of retirement are a period during which people may benefit more from their life, thus experiencing a better HRQoL.
In Alberta, the median (IQR) EQ-5D-5L index scores for men and women were 0.90 (0.83, 0.95) and 0.90 (0.82, 0.95), respectively, which was slightly higher than our results at 0.87 (0.81-0.91) and 0.87 (0.79-0.91), but their histogram distribution was similar to ours.
have fewer problems than our patients, but they were also younger and less representative of the general population.
This study also confirmed some assumptions made in the literature regarding the association between sociodemographic variables and HRQoL. For example, we found lower utility scores for low and high BMI
A number of people living alone may suffer from social isolation, and isolated people may experience financial stress, which could lower their well-being.
A strength of our study is that the sample size is considerable and generally representative of the population. Nevertheless, those aged 75 and older are not as well represented. This may be a limitation of an online sample because older people are generally less connected to the Internet.
Considering the aging population, it would have been interesting to have more subjects aged 75 and older. Another limitation was that respondents were more educated, and rural people slightly more represented. In addition, the utility score increased with education, as demonstrated in other studies.
This may limit the generalization of the results. As it was an online survey, it would have been difficult to change the response rates by categories, unlike a telephone study, as in Ferreira et al,
who contacted 6 times more subjects aged 70 years and older to have a better representation.
Another limitation is that there is no statistical model for calculating EQ-5D-5L scores based on the preferences of the Quebec population (i.e., no value set), but at least the equation used
is a Canadian one and includes data from Quebec. Another strength of our study was that we presented health utility scores according to a wide range of variables, including but not limited to several age categories, BMI, and health status for men and women separately. This makes it easier for other studies to compare their data to our data. In addition, our study is one of the few to provide utility scores for various health problems. As in van den Berg,
we found a high heterogeneity in the utility scores associated with specific health problems, which could be useful for healthcare decision makers in their priority exercises. On the other hand because numbers are quite small for some health problems, these results may not be fully representative of the population and should be used with caution.
Finally, as compared with other studies describing population norms with the EQ-5D-5L, our study is the only one showing a ceiling effect as low as 20.8%, whereas it is generally greater than 40%.
(ie, 19% and 4%, respectively), thus indicating that our study is probably more representative in terms of health problems affecting the HRQoL of the population. Also, the results in Figure 1 showed more linearity than in previous studies,
which is more in line with what is expected from a representative sample.
Conclusion
This study provides utility score norms for the EQ-5D-5L for the Quebec population, a French-speaking province of Canada. These results could be used for other studies that do not have a control group or to compare populations. In addition, these results will be useful for comparing with QALY studies to better interpret their results and assist researchers in interpreting their findings. Moreover, utility norms are provided for 21 health problems, which has rarely been done.
Acknowledgments
This work was funded through a research fund of the Social Sciences and Humanities Research Council of Canada, Knowledge Development Grants Program 2015-2017, file number 430-2015-00712.
Using the Internet to collect EQ-5D norm scores: a valid alternative?.
in: Busschbach J. Rabin R. De Charro F. Proceedings of the 24th Scientific Plenary Meeting of the EuroQol Group; Kijkduin-The Hague. The EuroQOL Group,
The Netherlands2009
32352-6&id=gr1.jpg){kind=link}