IF: 0.644
REUTERS THOMSON

Population-Based Preference Weights for the EQ-5D Health States Using the Visual Analogue Scale (VAS) in Iran

AUTHORS

Reza Goudarzi 1 , Hojjat Zeraati 2 , * , Ali Akbari Sari 1 , 3 , Arash Rashidian 1 , 3 , Kazem Mohammad 2

AUTHORS INFORMATION

1 Department of Health Management and Economics, School of Public Health, Tehran University of Medical Sciences, Tehran, IR Iran

2 Department of Epidemiology and Biostatistics, School of Public Health, Tehran University of Medical Sciences, Tehran, IR Iran

3 Knowledge Utilization Research Center, Tehran University of Medical Sciences, Tehran, IR Iran

How to Cite: Goudarzi R, Zeraati H, Akbari Sari A, Rashidian A, Mohammad K. Population-Based Preference Weights for the EQ-5D Health States Using the Visual Analogue Scale (VAS) in Iran, Iran Red Crescent Med J. 2016 ; 18(2):e21584. doi: 10.5812/ircmj.21584.

ARTICLE INFORMATION

Iranian Red Crescent Medical Journal: 18 (2); e21584
Published Online: February 13, 2016
Article Type: Research Article
Received: July 4, 2014
Revised: August 7, 2014
Accepted: August 30, 2014
Crossmark

Crossmark

CHEKING

READ FULL TEXT
Abstract

Background: Health-related quality of life (HRQoL) is used as a measure to valuate healthcare interventions and guide policy making. The EuroQol EQ-5D is a widely used generic preference-based instrument to measure Health-related quality of life.

Objectives: The objective of this study was to develop a value set of the EQ-5D health states for an Iranian population.

Patients and Methods: This study is a cross-sectional study of Iranian populations. Our sample from Iranian populations consists out of 869 participants, who were selected for this study using a stratified probability sampling method. The sample was taken from individuals living in the city of Tehran and was stratified by age and gender from July to November 2013. Respondents valued 13 health states using the visual analogue scale (VAS) of the EQ-5D. Several fixed effects regression models were tested to predict the full set of health states. We selected the final model based on the logical consistency of the estimates, the sign and magnitude of the regression coefficients, goodness of fit, and parsimony. We also compared predicted values with a value set from similar studies in the UK and other countries.

Results: Our results show that the HRQoL does not vary among socioeconomic groups. Models at the individual level resulted in an additive model with all coefficients being statistically significant, R2 = 0.55, a value of 0.75 for the best health state (11112), and a value of -0.074 for the worst health state (33333). The value set obtained for the study sample remarkably differs from those elicited in developed countries.

Conclusions: This study is the first estimate for the EQ-5D value set based on the VAS in Iran. Given the importance of locally adapted value set the use of this value set can be recommended for future studies in Iran and In the EMRO regions.

Keywords

Visual Analog Scale Quality of life Population Groups EQ-5D Preference-Based Health Measures , Health Related Quality of Life

Copyright © 2016, Iranian Red Crescent Medical Journal. This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/) which permits copy and redistribute the material just in noncommercial usages, provided the original work is properly cited.

1. Background

The quality-adjusted life year (QALY) is a paramount instrument to assess the health outcomes of medical interventions in cost-effectiveness research (1). A preference value set from a sample of populations containing such sets of populations with diverse diseases provides a standard instrument to ensure comparability between outcomes across diseases (2).

Two types of instruments have been developed to capture the HRQoL. Generic instruments measure the general impact of a disease on overall patient life, which allow the benefits of healthcare to be captured. Another type of instrument assesses the specific dimensions of a disease that may not be reflected by a generic instrument (3).

One of most widely used generic instruments for the HRQoL is the EuroQol EQ-5D. It has five generic domains: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression, which are included in the first part of the instrument (2). The second part contains the visual analog scale (VAS) of the EQ-5D. The VAS is a scale from 0 to 100 that evaluates patients’ health state from the best health state (score of 100) to the worst health state (score of 0) (4). This instrument as a rating scale is used to capture values given to “societal preferences” around the world (2, 5-12).

The VAS of the EQ-5D is mostly used for specific patient groups; however, there are studies in which it is used to elicit value sets from general populations in different developed countries (7, 13). The valuation of health state can be performed making use of different methods. These methods are the VAS, the time trade-off (TTO), and the standard gamble (SG) method (6). In this study, we used the VAS method to derive values for the EQ-5D.

2. Objectives

The purpose of this study was to describe the HRQoL, expressed with EQ-5D dimensions and in mean VAS scores and in mean EQ-5D index values, in the general population, by socioeconomic groups, in Iran.

3. Patients and Methods

3.1. The EuroQol EQ-5D

The EQ-5D defines the level of “mobility” (MO), “self-care” (SC), “usual activities” (UA), “pain/discomfort” (PD), and “anxiety/depression” (AD) (14). The response options are no problem with health (score of 1), some problems or moderate problems (score of 2), and severe problems (score of 3). Considering all possible health states, 243 (35) health states can be obtained (15). Response options would abbreviate the health states of participants. An abbreviation 12133, for instance, shows a health state with no problems in walking, some problems with self-care, no problems with performing regular activities, extreme pain, and extreme anxiety.

The VAS evaluates the health state in a visual analogue anchored to 100 for the “best health state” at the top and to 0 for the “worst health state” at the bottom (7). Patients with the same perceived health state by the EQ-5D may assign different values to their health state in this scale.

3.2. Health State Valuation

We make use of Dolan et al.’s (5) approach to select health states in this study. Based on this approach, as a minimum set of health states, 43 states are required to allow the value set to be estimated for a population.

All respondents valued 13 imaginary EQ-5D states. These states belong to 8 EQ-5D health states, which are drawn from a set of 40 states. These sets of health states contain severity and “11111,” “22222,” “33333,” “unconscious,” and “death” states given to all respondents using the VAS from 0 to 100.

Overall, there are 43 health states, which were divided into 5 groups of health states. Five scenarios were common in all groups and therefore were evaluated by all respondents, and 8 other scenarios are specific. A list of value sets is presented in Table 1.

Table 1. Sets of Health States Valued by the Respondents for Each Category
Group AGroup BGroup CGroup DGroup E
Common for all Respondents
11111a11111a11111a11111a11111a
DeadaDeadaDeadaDeadaDeada
2222222222222222222222222
33333a33333a33333a33333a33333a
UnconsciousUnconsciousUnconsciousUnconsciousUnconscious
Very Mild States
1111211121112111211121111
Mild States
1112211131111131131221222
2131212211111332113312121
Moderate States
1321232331133112212212222
2132332211122232233121232
Severe States
3323223232233211333222233
2232332223322323332133323
2331333212

aThese health states are valued twice because they are presented on both pages.

3.3. Study Sample

This survey is the cross-sectional study of the Iranian general adult population (18 years of age and older). We select a sample from this population via a stratified probability sampling method. Assuming that around 170 respondents were required to evaluate each health state (16), a total of 870 people were selected for the interview.

Sampling and data collection has three steps. In step 1, the entire population of Tehran was stratified using 22 regions (districts) of the city. The number of subjects assigned per region was in proportion to the population size. Each district is then divided into several fields. In step 2, the data collection fields were randomly selected per district. In step 3, 10 households per field were randomly invited for the interview. Families who were unwilling to participate in the study were replanced by new ones on a random basis.

The study was approved by the ethics committees in the Tehran University of Medical Sciences (number 17591, dated May 1, 2012). All participants signed a consent form before completing the questionnaire.

3.4. Data Management and Statistical Analyses

The data were collected in the city of Tehran, Iran from July to November 2013. Eight trained interviewers interviewed the study participants. Two training workshops were conducted by Reza to practically tutor interviewers on how to interview the participants. Workshop lasted for 3 hours during which a detail guide for the interview was given to interviewers using audiovisual materials. A guide contains a basic description of methods for capturing health preferences based on theories of VAS and EQ-5D-specific interviewer’s tasks. A number of interviews were simulated to improve interviewers’ skills.

The quality of field work was monitored both through direct supervision and cross-examination of 15% of the entire sample size. As such, 120 respondents were re-interviewed by telephone to evaluate the validation of interviews and double-check some of the demographic characteristics, e.g., age, household dimension, and occupational status. No incentives (monetary or non-monetary) were given to the respondents before or after the interviews. All statistical analyses were conducted using STATA SE 12.0.

3.5. Valuation Method

Respondents received a copy of the EuroQol instrument with a description of health states. The instrument consists of two pages. Per page, there are also 2 columns. Eight health states are listed on page 1 and six on page 2, which will be valued on a 20 cm thermometer similar to the VAS. Two extreme endpoints represent the “best imaginable health” (score of 100) and the “worst imaginable health” (score of 0). Participants rated the health states by drawing a line from the health state box to a point on the VAS that reflects their value for the health state. Having valued 14 health states, all respondents determined their value for death based on the VAS.

Based on a basic assumption set by the EuroQol EQ-5D (7), a value given to a health state by a respondent is a description of the health state for only one year and quality of life afterwards is not known.

Since the cost utility analysis requires values between 0 and 1 from death to full health, values for the VAS were modified at the patient level. The “best health” (100) was modified to 11111 (full health) and “worst health” to 0 (Death). For the modification, the following formula was applied:

Equation 1.

Where Vh is a VAS-adjusted score for a health state h, Sh is a respondent's unadjusted VAS score for state h, Sdead is the respondent-assigned VAS score for the health state “death,” and S11111 is respondent-assigned VAS score for a state 11111. Modified VAS values lower than –1 and larger than +1 have been truncated to –1 and +1, respectively (8).

3.6. Modeling

We modeled data by introducing responses into appropriate regression models. Since respondents may have different patterns of responses, the heteroscedasticity of responses throughout all health states was checked using the Breusch-Pagan test. Subsequently, we compared the simple generalized least-squares (GLS) regressions with random effects (RE) or fixed effects (FE) models. The GLS regression model and FE models were developed based on Hausman’s test.

The number of dependent variables in the regression models was computed as 1 minus the observed rescaled VAS value ranging from 0 to 2, where lower numbers correspond to a higher utility. We investigated several models to find a best fit regression model for our data. A different set of independent variables is therefore investigated where the selection of these variables was based on an evidence of previous research (2, 5, 8-11, 14, 15, 17-22). All models were tested and compared regarding the number of incoherent coefficients, the statistical significance of coefficients, the amount of explained variance (R2), the mean absolute error (MAE), and the Akaike information criterion (AIC). MAE is the absolute difference between the observed and the estimated value in each health state. We examined the assumptions of the model using various tests.

We reported the results of the model that better satisfy all the criteria specified below and compared them the main effects, the UK model (23), and the US model (24). Subsequently, some variables were adopted to account for interactions between different dimensions in an Iranian model, which is further called a final model.

Thus, the value y placed on a health state was as follows:

Y = α + βdl Xdl + ϵ

Where β is the vector of coefficients and X is the vector of dummy variables for dimension d at level l, where Xdl represents ten dummy variables indicating the presence of either level 2 or level 3 in a given health dimension; d stands for dimension and l for either level 2 or level 3.

The dependent variable (y) in the regression analysis was computed as 1 minus the modified VAS value. It represents the measure of disutility by subtracting the value of a given health state from the value of full health.

The model included the following independent variables:

- A dummy variable for level 2 in each dimension (some problems=1; otherwise 0).

- A dummy variable for level 3 in each dimension (extreme problems=1; otherwise 0).

- N2 = 1 if any dimension is at either level 2 or level 3; otherwise 0 (deviation from full health).

- N3 = 1, if any dimension is at level 3; otherwise 0.

- An ordinal variable D1 that represents number of deviations from full health beyond the first (i.e., values ranging from 0 to 4).

- An ordinal variable I3 that represents the number of dimensions at level 3 beyond the first.

- The square of I3 term allowing for non-linearity in an association with the dependent variable.

- The square of I2, an ordinal variable that represents number of dimensions at level 2 beyond the first (5, 11, 17, 24).

3.7. Exclusion Criteria

Respondents were excluded from the dataset when they satisfied the following criteria:

- Completely missing VAS data,

- The same value is given to all health states,

- < 4 health states valued,

- Death ≥ 11111,

- 11111 and / or death that are/is not valued, and

- The number of logical inconsistencies is more than 1 (2, 8, 20-23, 25)

An inconsistency is present when a state that is logically worse is ranked higher than a logically better health state. To clarify more, for a pair of health states, if state 13111 is valued higher than state 12111, respondent expresses a preference for 13111 over 12111. This form of inconsistency is defined as internal inconsistency. However, it is worth noting that a valuation can only be logically inconsistent within the same dimension. If in above mentioned case, respondent has different preferences across different dimensions, ceteris paribus, it cannot be said that 13111 is logically worse than 11121 (8).

4. Results

In total, 869 respondents participated in the survey. Having excluded 16 participants due to meeting at least one of the exclusion criteria, a total of 853 respondents are included in the valuation sample. The interviews lasted 25 min on average.

Table 2 presents the demographic characteristics of respondents. The observed distribution of responses demonstrates an expected distribution in the general public. However, we could directly compare the observed distribution of responses with the Iranian population only per sex and age (stratification variables). Population data are unavailable for the other characteristics to make use for comparison with our data.

Table 2. Sociodemographic Characteristics and Comparison With the General Adult Populationa
VariablesTotal Sample (n = 869)Iranian General Populationb
Sex
Male480 (55.24)42407049 (50.48)
Female389 (44.76)41585166 (49.52)
Age
18 - 24176 (20.25)11275668 (20.94)
25 - 34246 (28.31)15644578 (29.05)
35 - 44166 (19.10)10477767 (19.46)
45 - 54147 (16.92)7557889 (14.03)
55 - 6487 (10.01)4543026 (8.43)
65 >47 (5.41)4343091 (7.98)
mean (SD)38.16 (14.73)29.86 (NA)
Educational status
Low (< 10)186 (21.45)NA
Middle (10 - 13)317 (36.56)NA
High (14 >)364 (41.98)
Smoking behaviorNA
Current smoker121 (13.94)NA
Former smoker44 (5.07)NA
Non-smoker703 (80.99)NA
Average household size
1 - 2 elements148 (17.03)5402339 (25.5)
3 - 4 elements544 (62.60)11313136 (53.4)
5 or more elements177 (20.37)4470172 (21.1)
mean (SD)3.71 (1.31)3.55 (NA)
EQ-5D related health problems
Mobility (MO)94 (10.82)NA
Self-care (SC)11 (1.27)NA
Usual activities (UA)35 (4.03)NA
Pain/discomfort (PD)299 (34.41)NA
Anxiety/depression (AD)
EQ-VAS own health79.49 (16.01)NA
81 - 100405 (46.61)NA
61 - 80340 (39.13)NA
41 - 6098 (11.28)NA
21 - 4020 (2.30)NA
0 - 206 (0.69)NA
mean (SD)290 (33.37)NA

aValues are expressed as No. (%).

b18 years old or more according to the national statistical center, the Islamic Republic of Iran, 2012 (26).

The most common health state problem is pain/discomfort, with 34.41 % of the sample suffering from this. Anxiety/depression is in second place; 33.37 % of the sample are living with anxiety/depression. The average overall health state value in this study sample based on the VAS is 79.58 (SE = 0.54), ranging between 10 and 100.

Each respondent valued 13 health states (including “unconscious” and “death”) using the VAS procedure. The mean values for 42 health states (after the transformation) range from 0.784 for a best observed health state (11121) to -0.07 for the worst observed state (33333). The mean value for the “unconscious” states is -0.167 (SD = 0.82).

The models presented in Table 3 are the best among several models relying on diverse combinations of independent variables. The final model best fits our responses. All parameters made significant contributions to this model. Thus, this model presents a societal tariff for the EQ-5D in Iranian adults. Based on Hausman’s test, RE specifications are not presented since the estimates of the FE models are consistent with the sample.

Table 3. Parameter Estimates and Fit Statistics of Individual’s Level Models Using GLS Regression and FE Modelsa,b,c
GLS/Fix Effect
Main Effects ModelUK Model (N3)US Model (D1)I3 ModelFinal Model (I3 Square)
βS.EβS.EβS.EβS.EβS.E
Statistical parameters
MO20.10610.00640.10120.00640.29740.01000.10120.00640.09900.0064
MO30.12170.00820.11970.00820.48100.01500.24550.01120.15560.0091
SC20.19930.00670.21010.00660.37100.00910.21010.00660.19930.0067
SC30.21530.00810.21270.00800.53560.01410.33850.01100.24640.0088
UA20.16360.00760.11950.00790.31740.01170.11940.00790.14690.0078
UA30.23440.00780.16700.00880.48670.01150.29310.00850.24270.0079
PD20.09970.00590.11520.00600.26950.00860.11520.00600.10380.0059
PD30.18360.00660.15320.00680.46370.01170.27910.00880.20780.0071
AD20.14320.00640.13840.00640.26740.00870.13840.00640.14520.0064
AD30.22800.00690.18820.00720.48700.01100.31420.00860.25450.0074
Constant (N2)0.11110.00390.09600.00400.03080.00470.09600.00400.10590.0040
N30.12590.0078
D1-0.23320.0092
I2square0.01110.0013-0.00890.0010
I3-0.14020.0144-0.12590.0077
I3square0.00540.0018
Goodness-of-fit statistics
R20.550.560.560.560.55
MAE0.200.200.190.200.20
AIC-2523-2801-3547-2802-2607
BIC-2442-2713-3437-2714-2518

aNotes: Dummy dimentions 2 (equal 1 if the Dimensions level = 2, 0 for otherwise); Dummy dimentions 3 (equal 1 if Dimensions level = 3, for otherwise = 0); N2 = 1 if any dimension is at either level 2 or level 3, N3 (1 if any dimension is at level 3; otherwise = 0). D1 (the number of dimensions at level 2 or level 3 beyond the first, ranging from 0 to 4); I3 (the number of dimensions at level 3 beyond the first, ranging from 0 to 4), I3sq (the square of I3), and I2sq (the square of the number of dimensions at level 2 beyond the first).

bAll β are statistically significant at P < 0.05.

cStatistical Results: Breusch-Pagan test, χ2(1) = 578.13, P = 0.0001; Hausman’s test, χ2(1) = 23.96, P = 0.0077.

As an example, we calculate the predicted values for state “21232” based on the above formula for the Iranian model (Table 4).

Table 4. Iranian VAS Tariff for EQ-5D Health State 21232
Iranian VAS TariffValues
Full health (11111)1.0000
Constant term-0.1059
Mobility (MO)-0.0990
Self-care (SC)-0.0000
Usual activities (UA)-0.1469
Pain/discomfort (PD)-0.2078
Anxiety/depression (AD)-0.1452
I3 Square-0.0000
Tariff value for health state 212320.2952

The Iranian model specifications indicate that the differences between no problems and extreme problems (MO3, SC3, UA3, PD3, AD3) are greater than ones representing the difference between no problems and some problems. The respondents ascribe the greatest importance to ‘‘anxiety/depression’’ and ‘‘self-care’’ among the five health dimensions, indicating that disutility of level 3 in one of these two dimensions is greater than that at any level of other dimensions.

Table 5 compares the mean values for the valuations of the 43 health states by the respondents with the predicted value by the final model. As shown, the absolute difference between them is very small. Furthermore, our analysis shows that the sample observed mean values are strongly correlated with the final model value. Given these, such a small difference (0.077) can be disregarded (27).

Table 5. Comparison of Observed and Predicted Values Based on the Fixed Effects Modela,b
StateObservedPredictedDifferenceStateObservedPredictedDifference
111110.97270.89410.0786213230.16210.2030-0.0409
111120.75000.74880.0012221120.35740.4505-0.0931
111130.46420.6396-0.1754221210.37710.4919-0.1148
111210.78370.7903-0.0066221220.22490.3467-0.1218
111220.64610.64500.0010222220.22050.19980.0207
111310.57950.6862-0.1068222330.0746-0.00460.0792
111330.30410.4407-0.1366223230.16070.00360.1570
112110.64520.7472-0.1020223310.14990.1541-0.0041
113120.43150.5061-0.0747232320.27390.05760.2163
121110.61250.6947-0.0823233130.11190.08710.0248
121210.59990.59090.0090233210.25640.21110.0453
122110.49610.5478-0.0517322110.25580.3922-0.1364
122220.21630.2988-0.0825322230.23580.04290.1930
122230.14170.1896-0.0479322320.08000.04800.0320
132120.26600.3556-0.0895323130.16190.07750.0844
133110.27630.4139-0.1376323310.12250.1242-0.0017
133320.14540.08760.0579332120.28620.20890.0773
211110.62610.7950-0.1689332320.03860.02750.0111
211330.23250.3417-0.1092333210.09960.1807-0.0811
212220.34550.3991-0.053633323-0.0593-0.0287-0.0305
212320.25930.2951-0.035833333-0.0742-0.0704-0.0038
213120.39640.4071-0.0106NANANANA

aMean absolute difference: 0.077.

bSpearman’s correlation coefficient: 0.941 (P < 0.0001).

5. Discussion

The present study is the first attempt to elicit population-based value sets for health states in Iran and the Middle East. The research sample consists of the Iranian urban population or households and in respect to Iran’s sociodemographic characteristics (26). Thanks to limited resources for recruiting a larger sample of Iranian adults, our elicited values my not be a perfect reflection of value sets in the Iranian adult population.

At level 3, “anxiety/depression” has the highest coefficient, which is followed by “self-care.” “Usual activity,” “pain/discomfort,” and “mobility” are in further ranks. As a general pattern, our predicted values reflect lower values in comparison with the developed countries, such as Belgium, the UK, Spain, etc. (5, 9, 28). Except MO3, nearly all domains and severity levels highly impact VAS valuations when benchmarked against the UK study (5). All regression coefficients in the final model were statistically significant according to the Wald-type test. The order of these coefficients also differed from those in the UK, Malaysia, and Denmark (5, 8, 22) although the overall value set showed a similar pattern among countries.

In the majority of countries benchmark European counties (such as Germany and Spain) (5, 9, 29) respondents have the least problems in ‘‘mobility’’ among the five domains. In contrast, Iranian people are moderately worried about mobility. This worry stems from the perceived lack of a social security system and negative social norms regarding the use of wheelchairs or staying in nursing homes. On the other hand, in most benchmark countries, “usual activities” has the lowest weight in calculating the social valuation of EQ-5D health states, whereas in Iran this dimension is in third place.

Given the absence of an Iranian tariff, the UK value set has been used in Iran before the early 2020s. We compared values sets for Iran and the UK to understand degree to which the UK tariff differs from the Iranian tariff. We observed that the VAS values estimated in this study deviate from the UK. This implies that previous economic evaluations may have provided inaccurate information for Iranian decision makers.

Figure 1 compares the valuation from different utilities of selected EQ-5D health states between the European countries and Iran. This comparison shows a strong correlation between the Iranian model and the official model for Europe (r = 0.967, P < 0.0001), UK (r = 0.970, P < 0.0001), Denmark (r = 0.976, P < 0.0001), Spain (r = 0.966, P < 0.0001), and Slovenia (r = 0.958, P < 0.0001), respectively, reported by (5, 8, 11, 19, 30). We choose these countries because we used a model specification similar to the one used in the corresponding studies in the countries mentioned above. We also directly evaluated a set of health states that are also evaluated in all of these countries.

Comparison Between the Mean Value Sets of Selected EQ-5D Health States
Figure 1. Comparison Between the Mean Value Sets of Selected EQ-5D Health States

We applied the same exclusion criteria used in other national valuation studies, except for the way in which we dealt with inconsistencies. Inconsistent valuations may occur for two reasons. First, respondents give inconsistent responses. This may occur when they do not understand the valuation task. Second, inconsistencies could also be caused by the difficulty of the VAS practice. Regardless of the reasons, inconsistent values influence the parameter estimates of the regression equation, as they can produce contrary effects to the ones expected. Thus, the exclusion of the inconsistent responses allows a quality dataset and improves the validity of the resulting value set.

The Iranian EQ-5D valuation study differs from the original UK measurement and valuation health (MVH) study, most notably with regard to the sample size and timespan or duration of health states. The UK study has a sample size of 3395 and the health states last for 10 years, whereas our study sample is 869 and lasts only one year. With regard to the number of valued health states, there is a marginal difference. Our study takes 13 scenarios into account compared with 15 scenarios in the UK study.

Health status decreases with age. Women have also reported a worse health status than men, which is supported by studies in other countries (30-32). The EQ-5D instrument distinguishes between the effects of education on health states. Higher education has a positive association with the HRQoL in the study sample. Among the sociodemographic characteristics, smoking status, hospitalization, and marital status did not make a significant contribution to the regression model to explain individuals’ health state.

This study has several strengths and limitations. We are not able to judge the representativeness of the study sample for Iranian adults. The number of health states assigned to each interview might be insufficient to elicit entire preferences. Respondents valued a subset of 13 out of 43 health states. This process was time consuming and patients were reluctant to participate in a second interview. Hence, the sample size is remarkably larger than the sample sizes in studies conducted in France, the Netherlands, Italy etc.

A large number of the participants reported good health, i.e., no problems on any of the EQ-5D dimensions, which might be a valid representation of health status or might be caused by a ceiling effect. A ceiling effect may be caused by the insensitivity of instruments to the severity of health status (33) or culture differences (34), which are worthy of further exploration.

5.1. Conclusions

This study confirmed substantial differences between the Iranian population’s health-related preferences and other countries. These differences between populations warrant the use of locally derived tariffs for preference-based health measures. Researchers can apply the derived value sets to generate QALYs based on Iranian preferences.

Acknowledgements

Footnotes

References

  • 1. Burstrom K, Johannesson M, Diderichsen F. Swedish population health-related quality of life results using the EQ-5D. Qual Life Res. 2001; 10(7) : 621 -35 [PubMed]
  • 2. Cleemput I. A social preference valuations set for EQ-5D health states in Flanders, Belgium. Eur J Health Econ. 2010; 11(2) : 205 -13 [DOI][PubMed]
  • 3. Kontodimopoulos N, Pappa E, Niakas D, Yfantopoulos J, Dimitrakaki C, Tountas Y. Validity of the EuroQoL (EQ-5D) instrument in a Greek general population. Value Health. 2008; 11(7) : 1162 -9 [DOI][PubMed]
  • 4. Shafie AA, Hassali MA, Liau SY. A cross-sectional validation study of EQ-5D among the Malaysian adult population. Qual Life Res. 2011; 20(4) : 593 -600 [DOI][PubMed]
  • 5. Dolan P. Modeling valuations for EuroQol health states. Med Care. 1997; 35(11) : 1095 -108 [PubMed]
  • 6. Greiner W, Claes C, Busschbach JJ, von der Schulenburg JM. Validating the EQ-5D with time trade off for the German population. Eur J Health Econ. 2005; 6(2) : 124 -30 [PubMed]
  • 7. Oppe M, Devlin NJ, Szende A. EQ-5D value sets: inventory, comparative review and user guide. 2007;
  • 8. Wittrup-Jensen KU, Lauridsen J, Pedersen KM. Assessment of the Visual Analogue Scale (VAS) as a Valuation Method for Hypothetical Health States using the EuroQol (EQ-5D). 2008;
  • 9. Xie F, Gaebel K, Perampaladas K, Doble B, Pullenayegum E. Comparing EQ-5D valuation studies: a systematic review and methodological reporting checklist. Med Decis Making. 2014; 34(1) : 8 -20 [DOI][PubMed]
  • 10. Augustovski FA, Irazola VE, Velazquez AP, Gibbons L, Craig BM. Argentine valuation of the EQ-5D health states. Value Health. 2009; 12(4) : 587 -96 [DOI][PubMed]
  • 11. Greiner W, Weijnen T, Nieuwenhuizen M, Oppe S, Badia X, Busschbach J, et al. A single European currency for EQ-5D health states. Results from a six-country study. Eur J Health Econ. 2003; 4(3) : 222 -31 [DOI][PubMed]
  • 12. Little MH, Reitmeir P, Peters A, Leidl R. The impact of differences between patient and general population EQ-5D-3L values on the mean tariff scores of different patient groups. Value Health. 2014; 17(4) : 364 -71 [DOI][PubMed]
  • 13. Mulhern B, Rowen D, Snape D, Jacoby A, Marson T, Hughes D, et al. Valuations of epilepsy-specific health states: a comparison of patients with epilepsy and the general population. Epilepsy Behav. 2014; 36 : 12 -7 [DOI][PubMed]
  • 14. Scalone L, Cortesi PA, Ciampichini R, Belisari A, D'Angiolella LS, Cesana G, et al. Italian population-based values of EQ-5D health states. Value Health. 2013; 16(5) : 814 -22 [DOI][PubMed]
  • 15. Viney R, Norman R, King MT, Cronin P, Street DJ, Knox S, et al. Time trade-off derived EQ-5D weights for Australia. Value Health. 2011; 14(6) : 928 -36 [DOI][PubMed]
  • 16. Ferreira LN, Ferreira PL, Pereira LN, Oppe M. The valuation of the EQ-5D in Portugal. Qual Life Res. 2014; 23(2) : 413 -23 [DOI][PubMed]
  • 17. Lee YK, Nam HS, Chuang LH, Kim KY, Yang HK, Kwon IS, et al. South Korean time trade-off values for EQ-5D health states: modeling with observed values for 101 health states. Value Health. 2009; 12(8) : 1187 -93 [DOI][PubMed]
  • 18. Andrade MV, Noronha K, Kind P, Maia AC, de Menezes RM, Reis CDB, et al. Societal preferences for EQ-5D health states from a Brazilian population survey. Value Health Region Issues. 2013; 2(3) : 405 -12
  • 19. Golicki D, Jakubczyk M, Niewada M, Wrona W, Busschbach JJ. Valuation of EQ-5D health states in Poland: first TTO-based social value set in Central and Eastern Europe. Value Health. 2010; 13(2) : 289 -97 [DOI][PubMed]
  • 20. Devlin NJ, Hansen P, Kind P, Williams A. Logical inconsistencies in survey respondents' health state valuations -- a methodological challenge for estimating social tariffs. Health Econ. 2003; 12(7) : 529 -44 [DOI][PubMed]
  • 21. HitaM JMC. 17th Plenary Meeting of the Euroqol Group. 2001; : 23 -46
  • 22. Yusof FA, Goh A, Azmi S. Estimating an EQ-5D value set for Malaysia using time trade-off and visual analogue scale methods. Value Health. 2012; 15(1 Suppl) -90 [DOI][PubMed]
  • 23. The Measurement and Valuation of Health.Final report on the modeling of valuation tariff.
  • 24. Shaw JW, Johnson JA, Coons SJ. US valuation of the EQ-5D health states: development and testing of the D1 valuation model. Med Care. 2005; 43(3) : 203 -20 [PubMed]
  • 25. Schmidt S, Vilagut G, Garin O, Cunillera O, Tresserras R, Brugulat P, et al. [Reference guidelines for the 12-Item Short-Form Health Survey version 2 based on the Catalan general population]. Med Clin (Barc). 2012; 139(14) : 613 -25 [DOI][PubMed]
  • 26. Vice-Presidency for Strategic Planning and Supervision. 2012;
  • 27. Gudex C. Time Trade-Off User Manual: Props and Self-Completion Methods. 1994;
  • 28. Szende A, Oppe M, Devlin N. EQ-5D Value Sets : Inventory, Comparative Review and User Guide. 2007; [DOI]
  • 29. Garcia-Molina M, Chicaiza L, Rincon CJ, Romano G. PRM36 International Comparisons of EQ-5D Health-States Valuations. Value Health. 2012; 15(4)[DOI]
  • 30. Kind P, Dolan P, Gudex C, Williams A. Variations in population health status: results from a United Kingdom national questionnaire survey. BMJ. 1998; 316(7133) : 736 -41 [PubMed]
  • 31. Fryback DG, Dunham NC, Palta M, Hanmer J, Buechner J, Cherepanov D, et al. US norms for six generic health-related quality-of-life indexes from the National Health Measurement study. Med Care. 2007; 45(12) : 1162 -70 [DOI][PubMed]
  • 32. Sorensen J, Davidsen M, Gudex C, Pedersen KM, Bronnum-Hansen H. Danish EQ-5D population norms. Scand J Public Health. 2009; 37(5) : 467 -74 [DOI][PubMed]
  • 33. Bharmal M, Thomas J. Comparing the EQ-5D and the SF-6D descriptive systems to assess their ceiling effects in the US general population. Value Health. 2006; 9(4) : 262 -71 [DOI][PubMed]
  • 34. Sun S, Chen J, Johannesson M, Kind P, Xu L, Zhang Y, et al. Population health status in China: EQ-5D results, by age, sex and socio-economic status, from the National Health Services Survey 2008. Qual Life Res. 2011; 20(3) : 309 -20 [DOI][PubMed]
  • COMMENTS

    LEAVE A COMMENT HERE: