Toward a Smaller Design for EQ-5D-5L Valuation Studies

A Methodological Study to Compare Alternative Modes of Administration with Value EQ-5D Using Preference-Elicitation Techniques

OBJECTIVES

Time trade-off (TTO) and discrete choice experiment (DCE) preference-elicitation techniques can be administered using face-to-face interviews (F2F), unassisted online (UO) surveys, or remote-assisted (RA) interviews. The objective of this study was to explore how the mode of administration affects the quality and reliability of preference-elicitation data.

METHODS

EQ-5D-5L health states were valued using composite TTO (cTTO) and DCE approaches by the UK general population. Participants were allocated to 1 of 2 study groups. Group A completed both F2F and UO surveys (n = 271), and group B completed both RA and UO surveys (n = 223). The feasibility of survey completion and the reliability and face-validity of data collected were compared across all modes of administration.

RESULTS

Fewer participants reported receiving sufficient guidance on the cTTO tasks during the UO survey compared with the 2 assisted modes. Participants across all modes typically reported receiving sufficient guidance on the DCE tasks. cTTO data were less reliable from the UO survey compared with both assisted modes, but there were no differences in DCE data reliability. cTTO data from all modes demonstrated face-validity; however, the UO survey produced higher utilities for moderate and severe health states than both assisted modes. Both F2F and RA modes provided comparably reliable data.

CONCLUSIONS

The reliability of DCE data is not affected by the mode of administration. Interviewer-assisted modes of administration (F2F or RA) yield more reliable cTTO data than unassisted surveys. Both F2F and RA surveys produced similar-quality data.

Exploring non-iterative time trade-off methods for valuation of EQ-5D-5L health states

INTRODUCTION

The composite time trade-off (cTTO) method is used as the primary method for valuing EQ-5D-5L health states, but it requires intensive interviewer training and stringent quality control, which increases the burden of conducting cTTO studies. In this study, two non-iterative variants of the TTO method, non-stopping TTO (nTTO) and open-ended TTO (oTTO), were tested head-to-head with the cTTO method aiming to reduce the administration burden.

METHODS

31 EQ-5D-5L health states from an orthogonal array was selected and valued by a general public sample in China. Data were collected by 7 interviewers with all interviewers performed an equal number of interviews using all three TTO methods. We compared the value distribution, logical consistency, administration burden, and modeling performance of these three TTO methods.

RESULTS

In total, 422 participants participated in the valuation interviews, with 139 using the nTTO method, 140 using the oTTO method, and 143 using the cTTO method. Both oTTO and nTTO methods saved around 10 min for conducting an interview. The mean values of three methods were similar with each method showed different characteristics in their value distributions. cTTO outperformed the other two methods in terms of modeling performance.

DISCUSSION

Both non-iterative TTO methods showed potential for valuing EQ-5D health states, although their data distributions and modeling performance were inferior to the cTTO method. The results of this study showed the potential of these two alternative non-iterative TTO methods.

Valuation of EQ-5D-5L health states from cancer patients' perspective: a feasibility study

OBJECTIVES

To assess the feasibility of estimating an EQ-5D-5L value set using a small study design in cancer patients and to compare the EQ-5D-5L values based on the preferences of cancer patients with those of the general public.

METHODS

Patients with clinically diagnosed cancers were recruited from two hospitals in Shanghai, China. In face-to-face interviews using the EQ-PVT survey, health states were valued by cancer patients using both cTTO and DCE methods. cTTO data was modelled alone or jointly with DCE data. Forty-eight models using different model specifications (cross-attribute level effect [CALE] and additive models), random/fixed effects model assumptions, data heteroscedasticity and censoring were estimated. The best performed model was identified in terms of monotonicity of estimated model coefficients and out-of-sample prediction accuracy.

RESULTS

Data collected from 221 cancer patients who participated in the study were included. The hybrid CALE model using both TTO and DCE data performed best in terms of prediction accuracy (Lin's concordance coefficient = 0.989; root mean squared error = 0.058) and suggested that pain/discomfort and anxiety/depression were the most undesirable health problems. Compared to values based on general Chinese public's health preferences, the values based on cancer patients' preferences were much higher and lower for health states characterized by extreme mobility problems and severe/extreme pain or discomfort, respectively.

CONCLUSION

This study demonstrated the feasibility of using a small design to develop EQ-5D-5L value sets based on cancer patients' health preferences. Since there were signs of differences between preferences of patients and general population, it may be valuable to develop patient-specific value sets and use them in clinical decision making and economic evaluations.

Evaluation of EuroQol Valuation Technology (EQ-VT) Designs to Generate National Value Sets: Learnings from the Development of an EQ-5D Value Set for India Using an Extended Design (DEVINE) Study

INTRODUCTION

Countries develop their EQ-5D-5L value sets using the EuroQol Valuation Technology (EQ-VT) protocol. This study aims to assess if extension in the conventional EQ-VT design can lead to development of value sets with improved precision.

METHODS

A cross-sectional survey was undertaken in a representative sample of 3,548 adult respondents, selected from 5 different states of India using a multistage stratified random sampling technique. A novel extended EQ-VT design was created that included 18 blocks of 10 health states, comprising 150 unique health states and 135 observations per health state. In addition to the standard EQ-VT design, which is based on 86 health states and 100 observations per health state, 3 extended designs were assessed for their predictive performance. The extended designs were created by 1) increasing the number of observations per health state in the design, 2) increasing the number of health states in the design, and 3) implementing both 1) and 2) at the same time. Subsamples of the data set were created for separate designs. The root mean squared error (RMSE) and mean absolute error (MAE) were used to measure the predictive accuracy of the conventional and extended designs.

RESULTS

The average RMSE and MAE for the standard EQ-VT design were 0.055 and 0.041, respectively, for the 150 health states. All 3 types of design extensions showed lower RMSE and MAE values as compared with the standard design and hence yielded better predictive performance. RMSE and MAE were lowest (0.051 and 0.039, respectively) for the designs that use a greater number of health states. Extending the design with inclusion of more health states was shown to improve the predictive performance even when the sample size was fixed at 1,000.

CONCLUSION

Although the standard EQ-VT design performs well, its prediction accuracy can be further improved by extending its design. The addition of more health states in EQ-VT is more beneficial than increasing the number of observations per health state.

HIGHLIGHTS

The EQ-5D-5L value sets are developed using the standardized EuroQol Valuation Technology (EQ-VT) protocol. This is the first study to empirically assess how much can be gained from extending the standard EQ-VT design in terms of sample size and/or health states. It not only presents useful insights into the performance of the standard design of the EQ-VT but also tests the potential extensions in the standard EQ-VT design in terms of increasing the health states to be directly valued as well as the number of observations recorded to predict the utility value of each of these health states.The study demonstrates that the standard EQ-VT design performs good, and an extension in the design of the standard EQ-VT can lead to further improvement in its performance. The addition of more health states in EQ-VT is more beneficial than increasing the number of observations per health state. Extending the design with inclusion of more health states marginally improves the predictive performance even when the sample size was fixed at 1,000.The findings of the study will streamline the systematic process for generating precise EQ-5D-5L value sets, thus facilitating the conduct of credible, transparent, and robust outcome valuation in health technology assessments.

Using Time Trade-Off Values to Estimate EQ-5D-Y Value Sets: An Exploratory Study

OBJECTIVES

The published international EQ-5D-Y-3L valuation protocol does not recommend the composite time trade-off (cTTO) method as the primary valuation method because of 2 major concerns. First, the cTTO method was shown to generate high values. Second, the cTTO method is not as feasible for valuing children's health as other established methods. This study aimed to explore the feasibility of using cTTO values alone to estimate EQ-5D-Y-3L value sets.

METHODS

We analyzed the cTTO data derived from the recently completed Chinese EQ-5D-Y-3L valuation study in which a total of 28 health states were valued. We assessed the feasibility of the cTTO tasks in terms of survey completion time and participant-reported difficulty of understanding the task, differentiating the health states, and deciding the answer. We also examined the data distribution characteristics and modeled the data using different models.

RESULTS

In total, 418 participants completed the cTTO interview. On average, participants took approximately 35.70 minutes (SD 12.42) to complete the interview and made 13.21 moves (SD 9.00) in the cTTO tasks. There were 74.16%, 59.33%, and 11.48% of participants indicated that it was easy to understand the cTTO tasks, easy to differentiate between the health states, and difficult to decide on an indifference point, respectively. The data distribution was smooth, and a random-effects model performed the best in terms of coefficient significance, monotonicity, and predictive accuracy.

CONCLUSION

Our finding suggests that estimating EQ-5D-Y-3L value sets using cTTO data alone is feasible and therefore could be considered as an option in future valuation studies for EQ-5D-Y-3L.

Cross-Attribute Level Effects Models for Modeling Modified 5-Level Version of EQ-5D Health State Values: Is Less Still More?

OBJECTIVES

Cross-attribute level effects (CALE) model has demonstrated better predictive accuracy for out-of-sample health states than the conventional additive main-effects model in cross-validation analysis of the 5-level version of EQ-5D (EQ-5D-5L) composite time trade-off (cTTO) datasets. In this study, we aimed to further test the performance of CALE model using a different design and modified EQ-5D-5L states.

METHODS

A total of 29 EQ-5D-5L self-care bolt-off states, 30 EQ-5D-5L states, and 31 EQ-5D-5L vision bolt-on states were selected from the same orthogonal array. A total of 600 university students were interviewed face-to-face to value a subset of these health states using the cTTO method. For each type of health state, we fitted both the conventional main-effects model and the CALE model. Predictive accuracy was assessed in a series of cross-validation analysis using the leave-one-state-out method.

RESULTS

Overall, the CALE model outperformed the conventional model for each of the 3 types of health states in predicting the cTTO values of out-of-sample health states. The prediction accuracy of using the CALE model improved with the number of dimensions in health states, for example, the MAE decreased about 24%, 67%, and 77% for the EQ-5D-5L self-care bolt-off, EQ-5D-5L, and EQ-5D-5L vision bolt-on states, respectively, when using CALE models.

CONCLUSION

Our study supported the strengths of the CALE model for modelling the utility values of both original and modified EQ-5D-5L health states. Investigators with limited resources may consider using the CALE model to lower the costs for their valuation studies for EQ-5D-5L or similar health state descriptive systems.

Censoring Time Trade-off Values at 0 Versus at -1: How Does the Assumption for Worse-Than-Dead Time Trade-off Values Affect the Modeling of EQ-5D-5L Valuation Data?

OBJECTIVES

A recent study found that negative utility values elicited using composite time trade-off (TTO) were barely associated with the severity of EQ-5D-5L health states, suggesting poor discriminative ability. Assuming negative values provide limited information, this study aimed to explore the usefulness of censoring negative TTO values at 0 in modeling EQ-5D-5L valuation data.

METHODS

We analyzed EQ-5D-5L valuation data from China, The Netherlands, Canada, Singapore, and Thailand. For each data set, we estimated value sets using 2 Tobit models, one left-censored at -1 (current practice) and one left-censored at 0 (our proposed method), and compared the model performances. We hypothesized that censoring at 0 and censoring at -1 would produce similar values, though on slightly different scales.

RESULTS

When censoring at 0, logical inconsistencies and statistical significance were improved but the value range was compressed. In the cross-attribute level effects model, the 3-level parameters were similar between the models censored at 0 and -1, but the rank order of some dimension parameters was altered. Health state values predicted by the 2 censoring models approximated a perfect agreement after rescaling.

CONCLUSIONS

Censoring TTO values at 0 improved model estimation and fit but produced higher utility values than models censoring at -1. Investigators of future EQ-5D value set studies using the composite TTO method are advised to examine the validity of negative TTO values before choosing modeling strategies.

Interaction Effects in Health State Valuation Studies: An Optimal Scaling Approach

OBJECTIVES

This study aimed to introduce a parsimonious modeling approach that enables the estimation of interaction effects in health state valuation studies.

METHODS

Instead of supplementing a main-effects model with interactions between each and every level, a more parsimonious optimal scaling approach is proposed. This approach is based on the mapping of health state levels onto domain-specific continuous scales. The attractiveness of health states is then determined by the importance-weighted optimal scales (ie, main effects) and the interactions between these domain-specific scales (ie, interaction effects). The number of interaction terms only depends on the number of health domains. Therefore, interactions between dimensions can be included with only a few additional parameters. The proposed models with and without interactions are fitted on 3 valuation data sets from 2 different countries, that is, a Dutch latent-scale discrete choice experiment (DCE) data set with 3699 respondents, an Australian time trade-off data set with 400 respondents, and a Dutch DCE with duration data set with 788 respondents.

RESULTS

Important interactions between health domains were found in all 3 applications. The results confirm that the accumulation of health problems within health states has a decreasing marginal effect on health state values. A similar effect is obtained when so-called N3 or N5 terms are included in the model specification, but the inclusion of 2-way interactions provides superior model fits.

CONCLUSIONS

The proposed interaction model is parsimonious, produces estimates that are straightforward to interpret, and accommodates the estimation of interaction effects in health state valuation studies with realistic sample size requirements. Not accounting for interactions is shown to result in biased value sets, particularly in stand-alone DCE with duration studies.

The remarkably frequent use of EQ-5D in non-economic research

INTRODUCTION

EQ-5D is an instrument which has been utilized for a variety of purposes, including in health-economic appraisals as an input into quality-adjusted life year (QALY) calculations. Indeed, it is the most-widely applied instrument for health-economic appraisal worldwide, and is recommended for use in QALY calculations by many national Health Technology Assessment (HTA) agencies. There is also a growing body of evidence for its usefulness in a variety of settings other than economic appraisals, but such use has not been well-documented. This study addresses this issue and documents how EQ-5D has been applied in both the non-economic and economic contexts.

METHODS

The PubMed database was searched using the terms 'EQ-5D', 'EQ-5D AND cost', and 'EQ-5D AND cost AND QALY' from 1 January 1980 to 31 December 2019. We concentrated on 2019 publications for more detailed analyses. All the data collected for 2019 were downloaded and collected in EndNote. For 2019 only, we classified economic and non-economic use based on the inclusion of 'cost'. We also checked by manual inspection whether the search terms were suitable in correctly identifying economic and non-economic use. Variants of the non-economic use of EQ-5D were classified as follows: (a) as a quality of life outcome measure; (b) as a tool for methodological research; (c) methodological issues of EQ-5D itself; (d) comparisons with other quality of life questionnaires; (e) mapping studies; (f) value sets; (g) alongside costs but no QALY calculated; and (h) other.

RESULTS

The first publication found was from 1990. Up to and including 2019, 10,817 publications were identified, of which more than two in three did not contain any reference to costs or QALYs. In 2019, a total of 1409 manuscripts were identified, of which 239 were specifically for EQ-5D-5L.  Four hundred and seven (28.9%) included some form of 'costs' and 157 (11.1%) both 'costs' AND 'QALYs' terms. For EQ-5D-5L, the corresponding numbers were 104 (43.5%) and 29 (12.1%), respectively. After manually checking all the 1409 papers, three were duplicated records, which were omitted. In the remaining 1406 papers, only 40 (2.8%) contained the term 'cost', but not 'cost per QALY', and only 117 (8.3%) were identifiable as economic evaluations using the term 'cost per QALY'. Most non-economic use of EQ-5D was as a quality-of-life outcome measure (72.8%). Other applications were: as a tool for methodological research (6.7%); comparison studies (3.7%); EQ-5D methodological issues (3.5%); containing costs but not QALYs (2.8%); mapping (1.3%); value sets (0.4%); and other papers (0.4%).

CONCLUSIONS

The majority of the studies retrieved, covering a wide variety of research areas, reported upon the non-economic use of EQ-5D. Despite being the most-used instrument worldwide for QALY calculations, economic appraisal accounted for only a small, but important, part of published use.

Use of a Non-parametric Bayesian Method to Model Health State Preferences: An Application to Polish and Irish EQ-5D-5L Valuations

Valuations of preference-based measures for health are conducted in different countries. There is scope to use results from existing countries' valuations to generate better valuation estimates than analyzing the data from each country separately. We analyse data from two smaller design EQ-5D-5L valuation studies where a sample of 119 Polish migrants and 123 native Irish valued 30 common health states using similar composite time trade-off protocols. We apply a non-parametric Bayesian method to provide better predictions of the Polish (Irish) population utility function when the existing Irish (Polish) results were used as informative priors. The resultant new estimates were then compared to those obtained by analyzing the data from each country by itself via different prediction criterions. The results suggest that existing countries' valuations could be used as potential informative priors to produce better valuation estimates under all prediction criterions used. The implications of these results will be hugely important in countries where valuation studies are expensive and hard to conduct. Future application to other countries and to other preference-based health measures are encouraged.

Examining the Association between Polish Migrant Status and Health Preferences Using a Novel Application of a Smaller Design EQ-5D-5L Valuation Study

INTRODUCTION

Migrants have different utilisation of healthcare services and health-related behaviours than host populations. A potential factor that may contribute to the notable differences in healthcare use and health-related behaviours between migrants and host populations is how these groups value health. Those who place a high value on health have greater healthcare-seeking practices than those who do not.

OBJECTIVE

The aim of this study was to examine how Polish migrants and native Irish differ in health state utility valuations using a novel application of a smaller design EQ-5D-5L valuation study.

METHODS

This study uses health preferences as a predictor of how one values health. We examined the EQ-5D-5L health preferences of 119 Polish migrants and 123 native Irish, both residing full-time in Ireland. To do so, we used a novel application of a smaller design EQ-5D-5L valuation study that consisted of 30 health states and a targeted sampling strategy coupled with a Bayesian statistical nonparametric model. We collected data from June 2018 to September 2019.

RESULTS

Our results highlight that Polish migrants and native Irish differ in their health preferences for and valuation of severe health states. Polish migrants place meaningfully higher utility valuations of 0.1 or more on the three most severe health states compared with the native Irish.

CONCLUSION

This study can provide an understanding of a potential new factor underpinning some of the disparities in healthcare utilisation and health-related behaviours among migrants and host populations in Europe. This study also provides proof of principle for using a smaller design EQ-5D-5L valuation study to explore differences in health preferences among other minority subgroups, which can otherwise be hard to uncover when using the secondary analysis of national EQ-5D-5L valuation studies.

Developing the EQ-5D-5L Value Set for Uganda Using the 'Lite' Protocol

OBJECTIVE

A 'lite' version of the EQ-5D-5L valuation protocol, which requires a smaller sample by collecting more data from each participant, was proposed and used to develop an EQ-5D-5L value set for Uganda.

METHODS

Adult respondents from the general Ugandan population were quota sampled based on age and sex. Eligible participants were asked to complete 20 composite time trade-off tasks in the tablet-assisted personal interviews using the offline EuroQol Portable Valuation Technology software under routine quality control. No discrete choice experiment task was administered. The composite time trade-off data were modelled using four additive and two multiplicative regression models. Model performance was evaluated based on face validity, prediction accuracy in cross-validation and in predicting mild health states. The final value set was generated using the best-performing model.

RESULTS

A representative sample (N = 545) participated in this study. Responses to composite time trade-off tasks from 492 participants were included in the primary analysis. All models showed face validity and generated comparable prediction accuracy. The Tobit model with constrained intercepts and corrected for heteroscedasticity was considered the preferred model for the value set on the basis of better performance. The value set ranges from - 1.116 (state 55555) to 1 (state 11111) with 'pain/discomfort' as the most important dimension.

CONCLUSIONS

This is the first EQ-5D-5L valuation study using a 'lite' protocol involving composite time trade-off data only. Our results suggest its feasibility in resource-constrained settings. The established EQ-5D-5L value set for Uganda is expected to be used for economic evaluations and decision making in Uganda and the East Africa region.

Do health preferences differ among Asian populations? A comparison of EQ-5D-5L discrete choice experiments data from 11 Asian studies

Introduction

Many countries have established their own EQ-5D value sets proceeding on the basis that health preferences differ among countries/populations. So far, published studies focused on comparing value set using TTO data. This study aims to compare the health preferences among 11 Asian populations using the DCE data collected in their EQ-5D-5L valuation studies.

Methods

In the EQ-VT protocol, 196 pairs of EQ-5D-5L health states were valued by a general population sample using DCE method for all studies. DCE data were obtained from the study PI. To understand how the health preferences are different/similar with each other, the following analyses were done: (1) the statistical difference between the coefficients; (2) the relative importance of the five EQ-5D dimensions; (3) the relative importance of the response levels.

Results

The number of statistically differed coefficients between two studies ranged from 2 to 16 (mean: 9.3), out of 20 main effects coefficients. For the relative importance, there is not a universal preference pattern that fits all studies, but with some common characteristics, e.g. mobility is considered the most important; the relative importance of levels are approximately 20% for level 2, 30% for level 3, 70% for level 4 for all studies.

Discussion

Following a standardized study protocol, there are still considerable differences in the modeling and relative importance results in the EQ-5D-5L DCE data among 11 Asian studies. These findings advocate the use of local value set for calculating health state utility.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11136-021-03075-x.

Sample Size and Model Prediction Accuracy in EQ-5D-5L Valuations Studies: Expected Out-of-Sample Accuracy Based on Resampling with Different Sample Sizes and Alternative Model Specifications

Background. National valuation studies are costly, with ∼1000 face-to-face interviews recommended, and some countries may deem such studies infeasible. Building on previous studies exploring sample size, we determined the effect of sample size and alternative model specifications on prediction accuracy of modeled coefficients in EQ-5D-5L value set generating regression analyses. Methods. Data sets (n = 50 to ∼1000) were simulated from 3 valuation studies, resampled at the respondent level and randomly drawn 1000 times with replacement. We estimated utilities for each subsample with leave-one-out at the block level using regression models (8 or 20 parameter; with or without a random intercept; time tradeoff [TTO] data only or TTO + discrete choice experiment [DCE] data). Prediction accuracy, root mean square error (RMSE), was calculated by comparing to censored mean predicted values to the left-out block in the full data set. Linear regression was used to estimate the relative effect of changes in sample size and each model specification. Results. Results showed that doubling the sample size decreased RMSE by on average 0.012. Effects of other model specifications were smaller but can when combined compensate for loss in prediction accuracy from a small sample size. For models using TTO data only, 8-parameter models clearly outperformed 20-parameter models. Adding a random intercept, or including DCE responses, also improved mean RMSE, most prominently for variants of the 20-parameter models. Conclusions. The prediction accuracy impact of further increases in sample size after 300 to 500 were smaller than the impact of combining alternative modeling choices. Hybrid modeling, use of constrained models, and inclusion of random intercepts all substantially improve the expected prediction accuracy. Beyond a minimum of 300 to 500 respondents, the sample size may be better informed by other considerations, such as legitimacy and representativeness, than by the technical prediction accuracy achievable.

Methodological similarities and variations among EQ-5D-5L value set studies: a systematic review

OBJECTIVES

(1) To descriptively compare the selected elements of valuation methods for EQ-5D-5L value set studies, (2) to compare the characteristics of the value sets, and (3) to examine the associations between the selected elements of valuation methods and the EQ-5D-5L value sets.

METHODS

A systematic literature search of EQ-5D-5L valuation studies from 1 January 2009 to 22 April 2021 was conducted in selected databases. Following the initial search, we also explored additional studies published during the completion of the final version of the manuscript. Similarities and variations for selected elements of valuation methods were descriptively compared. The relative importance of dimensions, utility decrements between the levels, and distribution of the utility scores were used to compare value sets. A meta-regression analysis examined the associations between the selected methodological elements and the utility scores and dimension levels of EQ-5D-5L.

RESULTS

A total of 31 studies were included in this review. Methodological similarities centered around data collection and preference elicitation method. On the other hand, variations include sampling technique, sample size, and value set modeling. The variations in value sets based on the relative importance of dimension, decrement in utility score, and distribution of utility score across countries were observed. Although the distribution of the utility scores differed across countries, higher levels of each dimension tended to have a larger decrement in the utility scores. Mean utility scores for the experience-based value sets were higher than those estimated using stated choice methods. The selected methodological elements were not significantly associated with the mean predicted utility scores or most dimension-level coefficients.

CONCLUSIONS

EQ-5D-5L health state valuation methods and characteristics of value sets differed across studies. The impact of the variation of methodological elements on the value sets should be further investigated.

Health preferences and preventive care utilisation: How EQ-5D-5L health preferences may affect uptake

Despite the economic and health benefits of preventive care being well established, the uptake of many cost-effective preventive services remains lower than desired in many cases, especially among specific sub-populations. The value an individual places on health can influence their uptake of preventive care. One way to capture the value an individual places on health and future health status is to examine their health preferences. This study used a novel use of EQ-5D-5L health preferences to determine if health preferences are associated with the uptake of a range of preventive care services, including a cancer screening, blood pressure check, cholesterol check, blood test and urine test. We collected EQ-5D-5L composite time trade-off data in 2018/2019 on 242 respondents residing in Ireland. We estimated an initial tobit model to predict an individual’s health preference to capture health preferences as a regressor. We then estimated a bivariate probit model to examine the uptake of each preventive service and GP use. Each model controlled for health preferences, education, sex, type of health coverage, self-reported health, employment status, age and marital status. Health preferences are a significant determinant of all five preventive services while controlling for other covariates. The results shows that the higher an individual values good health, the more likely they are to avail of preventive care. Health preferences can be noted as a potential determinant of preventive care use that could guide policy responses seeking to increase demand-side factors for preventive care uptake.