Some Contributions to Aggregate and Individualized Cost-Effectiveness Analysis
Loading...
Links to Files
Permanent Link
Author/Creator
Author/Creator ORCID
Date
2019-01-01
Type of Work
Department
Mathematics and Statistics
Program
Statistics
Citation of Original Publication
Rights
Distribution Rights granted to UMBC by the author.
Access limited to the UMBC community. Item may possibly be obtained via Interlibrary Loan thorugh a local library, pending author/copyright holder's permission.
This item is likely protected under Title 17 of the U.S. Copyright Law. Unless on a Creative Commons license, for uses protected by Copyright Law, contact the copyright holder or the author.
Access limited to the UMBC community. Item may possibly be obtained via Interlibrary Loan thorugh a local library, pending author/copyright holder's permission.
This item is likely protected under Title 17 of the U.S. Copyright Law. Unless on a Creative Commons license, for uses protected by Copyright Law, contact the copyright holder or the author.
Abstract
The field of cost-effectiveness analysis (CEA) addresses the question of whether a new treatment provides value for the money compared to a standard treatment. The relevant statistical methodologies are based on outcomes of cost and effectiveness, typically obtained from a clinical trial. These methodologies assist health policy-makers in deciding if a new treatment should be assigned to patients, and what amount of investment can provide cost-effectiveness for the new treatment. There are several commonly used metrics to quantify a new treatment's cost-effectiveness. The most widely used metric is the incremental cost-effectiveness ratio (ICER), which is the ratio of the average incremental costs to the average incremental effectiveness between two competing treatments. Another popular metric for CEA is the incremental net benefit (INB), which is the difference between the incremental effectiveness and the incremental cost, after multiplying the former with a willingness-to-pay parameter, which is the maximum amount a policy-maker is willing to pay for a unit of effectiveness gained under the new treatment. Yet another CEA metric is the cost-effectiveness proportion (CEP), which is the proportion of patients for whom the new treatment is less costly and more effective, up to specified margins of cost and effectiveness. Sometimes net monetary benefits (NMBs) are also compared for CEA; the NMB for a treatment is the difference between the effectiveness and the cost, after multiplying the former with the willingness-to-pay parameter. This thesis develops statistical methodologies for CEA that enable decision making at the population and patient levels. Aggregate analysis carried out at the population level aims to provide information of a new treatment's value for an entire population of patients. Such an analyses often omits the heterogeneity amongst patients. The realization that between-patient variability can affect cost-effectiveness has underscored the need for individualized CEA, new metrics and methods are necessary for such an investigation. In addition, there has also been interest in developing cost-effectiveness metrics when there are multiple effectiveness measures. The research reported herein deals with individualized criteria for CEA, as well as aggregate criteria when there are multiple effectiveness measures. For individualized cost-effectiveness analysis, the CEA literature currently recommends subgroup analysis based on a stratification approach for constructing the subgroups. However, these stratification methods are somewhat arbitrary, and there is no clear way of constructing the subgroups in a well-defined fashion. In our work, we have considered a multivariate regression model for incorporating the patient-level covariates, and covariate specific CEA metrics are then defined and investigated. This appears to be a natural approach for individualized CEA, and avoids the need to construct subgroups. The individualized criteria that we have investigated include the INB, CEP and NMB. In terms of comparing the NMBs at the population-level, we have explored the stochastic comparison of the NMB distributions of the new treatment and the standard treatment. In the presence of multiple effectiveness measures, the traditional approach, labelled multi-criteria decision analysis, has been to combine the different effectiveness measures into a single quantity by taking a weighted linear combination; however, the weighting is clearly subjective. Our research on this topic has focused on the CEP metric, modified to take into account the availability of multiple effectiveness measures. This work deals with the case of only continuous cost and effectiveness random variables. For the CEP metrics investigated, the major focus is interval estimation. Both parametric and non-parametric approaches are investigated. The parametric set up assumes a joint normal distribution for the cost and effectiveness, where the normality may hold only after a monotone-transformation, notably a log-transformation for costs. In the parametric set up, a fiducial approach, the delta method, and the bootstrap are all investigated and compared for the interval estimation of the relevant CEA metrics. The non-parametric method that has been investigated is based on U-statistics. Extensive numerical results are reported in order to assess the accuracy of the confidence intervals, and the results are all illustrated with examples.