Cost-utility Threshold For China From The Demand-Side Perspective

BackgroundWith the further acceleration of aging process in China,health-related expenditures are increasing every year.The balance of medical insurance funds is severe.Pharmacoeconomic evaluation,especially cost-utility analysis,is a vital tool for the efficient allocation of limited resources in healthcare,which has been rapidly developed and widely used in China.The cost-utility threshold is always considered as criterion to determine whether a project is economic.Hence,a proper threshold is an important factor for the reliability of cost-utility analysis.The 1-3 times per capita GDP as cost-utility threshold,which was recommended by World Health Organization,is widely used in China.However,this threshold has been criticized for its limitations.International studies on cost-utility thresholds focus on both theoretical and empirical aspects.When determining whether interventions are economical,some scholars believe that the opportunity cost should be considered,while many scholars believe that consumption value of health should be adopted.The latter view can also be considered as the demand-side threshold,which represents the maximum value that the society is willing to pay for additional health benefits.It is often expressed as the maximum willingness to pay for a quality-adjusted life year(WTP/Q).Empirical research on cost-utility thresholds is rare in China.If any,the samples are often not representative,or the measurement method of health outcome is not the quality-adjusted life year(QALY),which is commonly used in pharmacoeconomic evaluation.Those studies provide limited evidence of the cost-utility threshold in China.Aims and contentsOverall,the aim of this research is to explore the cost-utility threshold in China with QALY as health-related outcome from the demand-side perspective,which we hope,could be used in the cost-utility analysis in China.The detailed research contents are as follows:first,using systematic method,we describe the research background and status,related concepts and theories,methods of eliciting cost-utility threshold and willingness to pay surveys.Second,we conduct an updated systematic review of studies estimating WTP/Q in the general population to summarize the methods and contexts of how the WTP/Q was elicited and to examine whether the WTP/Q estimate varies in a systematic way in relation to important aspects such as the size of QALY,types of health gain,certainty of health outcome and the study population,with the intention of informing future research.Third,we carry out a large sample interview to investigate WTP/Q with quota sampling and snow-ball sampling,two-part regression models are used for handling the heterogeneity and test how different factors influence the WTP/Q.MethodsFirst,based on conceptual analysis and theoretical methods,we summarize the shortcomings of using 1?3 times GDP per capita as the cost-utility threshold,then we discuss the best method of setting cost-utility threshold and investigate the willingness-to-pay method in detail.Second,we use systematic review and meta-regression technique to summarize studies on the cost-utility threshold from the demand perspective.A random-effect meta-regression was applied to integrate outcomes from different studies in the presence of heterogeneity.We adopted the Method of Moments(DerSimonian and Laird)to evaluate true between-study variance(?2)with Knapp-Hartung modification.Furthermore,sensitivity analyses were conducted by replacing the lowest and highest 5%and 2.5%of the values of WTP/Q by percentiles,in order to reduce the impact of outliers.Third,based on the outcomes of theoretical research and systematic review,we composed our survey questionnaires,which contain five parts including introduction,quality of life,hypothetical scenario,WTP,and demographic items.Quota sampling with snow-ball sampling was used in the final survey with quotas based on sex,age,income,location,and education.Quotas were calculated according to the China Statistical Yearbook 2020.The interviewers in this survey are all undergraduate or master students.All interviewers are required to participate in two training sessions and pass the final test.Finally,the data analysis is carried out in the following way:for continuous variables,the mean,standard deviation(SD),median,interquartile range(IQR)and value range are reported;for categorical variables,we compute the number and proportion in each group.Moreover,we distinguish actual zero value,protest value and irrational value.Protest value and irrational value are excluded for further data analysis.We adjust the probability value of health outcome using prospect theory.Subgroup WTP/Q is also reported.In accordance with demand theory of medical service,threshold model of multiple determinants is developed with two-part regression model,which is composed of probit model and generalized linear function.OutcomesOutcomes of conceptual analysis and theoretical methods:The main limitations of using 1-3 times per capita GDP as cost-utility threshold are:the health outcome indicator is DALY instead of QALY,which is much more common used as health-related outcome measure in health economics.Furthermore,the threshold recommended by the WHO is the threshold of average cost-effectiveness ratio rather than the threshold of the incremental cost-utility ratio,which are quite different concepts.The optimal method of cost-utility threshold should combine both supply-side and demand-side perspective.Adopting threshold from only one perspective would eventually lead to a sub-optimal distribution of social resources.Currently,studies about willingness-to-pay method mainly focus on two aspects:questionnaire design and data analyzing.It was recommended that for the representative of survey,various of questionnaires would be necessary.For a valid WTP answer,the payment card values should cover respondents true WTP.As for data analyzing of WTP,more detailed methods were developed to distinguish the protest responses with zero response.Outcomes of systematic review and meta regression:33 studies with 102 WTP/Q estimates were included.The QALY gain was due to improvement of quality of life in 20 studies,whereas 4 studies applied change of life length,and in 30%of the studies,the QALY gain was combined with both change on the quality of life and life length.More than half research(18 studies)only asked a hypothetical question of health gain with certainty,8 studies involved risk in the health gain,and 7 studies incorporated both health gain with certainty and health gain with uncertainty in their scenarios.More than half of the studies used payment card as survey technique.The meta-regressions demonstrated that types of health gain(quality of life or life length)and certainty of health outcomes are statistically significant factors.Moreover,the declining trend of QALY gains and positive effect with statistical significance of the sample age were also noticed.For valid and representative values of WTPQ,future researchers should therefore take into consideration various scenarios and investigate both health gain with certainty and uncertainty,health gain from both life length and quality of life,and different size of QALY gains.Outcomes of WTP survey:overall,we include 2008 valid samples in this survey.Sample characteristics regarding gender,age,education,location and income are quite similar with quotas,which means our respondents can be considered as representative of national population.The survey contains 4,016 WTP responses,of which 270 were protested and 298 were irrational.Excluding the protest value and irrational value,a total of 3058 WTP responses were included in the subsequent data analysis.The average value of WTP/Q was 113,987 yuan,and the median was 36,236 yuan.Sensitivity analysis results show that the average value of WTP/Q is 96,760 yuan,and the 95%confidence interval is(92,612,101,062).For the WTP/Q that improves the quality of life,the average value is 82,204 yuan,and for the extended life scenario,the average WTP/Q is 174,838 yuan.Outcomes of two-part regression model:we use WTP/Q as the dependent variable,quality of life,sociodemographic variables as well as hypothetical variables as independent variables.The results indicate that the older people are,the lower the probability of people choosing to spend money on medical interventions.However,age has no significant effect on the size of WTP/Q.Education and income have positive impact on the size of WTP/Q.At the same time,the subject's feelings such as pain and depression will also affect the WTP/Q results.The size of QALY is negatively related with the WTP/Q.Moreover,the WTP/Q value of terminally illness is significantly higher than that of curable diseases.Compared with 50%of health outcome,the WTP/Q is smaller when it is 100%chance of health outcome.Future researchFirst,future research on the cost-utility threshold of the supply perspective with QALY as health-related outcome would be needed;second,given that the WTP/Q for terminal illness is much higher than that of ordinary curable diseases,the threshold of cancer and rare diseases should get extra attention;finally,the perspective of supply and demand should be integrated to establish a long-term research perspective of cost-utility threshold.