A framework for multi-indication evidence synthesis in oncology Health Technology Assessment
Lead Research Organisation:
University of York
Department Name: Centre for Health Economics
Abstract
An increasing number of drugs are used to treat different types of cancer. These present a common mechanism of action, such as blocking cancer blood vessel growth (anti-angiogenics such as bevacizumab). These drugs are typically first licensed by regulatory authorities for use in a specific indication (for a particular cancer histology in a particular line of treatment), and over time their licences are extended to include additional indications. Bevacizumab, for example, is currently approved by regulators for use in 11 indications.
Closely after the regulatory license is awarded for a particular indication, the drug is appraised (for clinical and economic value) by health systems to determine whether it is made available (funded) to patients in that indication. This is called Health Technology Assessment (HTA). The evidence supporting HTA typically comprises of an indication-specific study that can be powered on intermediate (surrogate) outcomes, such as progression-free survival, rather than the final patient-relevant endpoints required by HTA agencies (e.g. overall survival).
The focus on single indications means that every appraisal is subject to a high level of evidential uncertainty on final endpoints, even where considerable evidence on the drug exists for other indications. We propose exploring whether better use of evidence can be made by looking across, as well as within, indications. Such analyses would share information across indications, reducing decision uncertainty across the existing indications and providing more realistic predictions of the value of the drug in future indications. The overall aim of this project is, therefore, to develop a framework for multi-indication evidence synthesis to support oncology HTA. The framework will outline the judgements and analyses required to specify, conduct, and interpret the synthesis of multi-indication evidence, and will incorporate a package of tools.
Three complementary work packages of methodological work will support the development of the framework:
WP1: Presenting and summarising evidence
In this work package, we will identify the relative effectiveness evidence for bevacizumab across its 11 indications, the real case study that will be used as an exemplar in this programme of work. To support multi-indication analyses, it is important that the evidence base is summarised appropriately for consideration by clinical and policy decision makers. Therefore, we will extend existing methods for the mapping of evidence to the multi-indication oncology setting (across indications, with multiple outcomes, and displaying the level of immaturity of evidence).
WP2: Methods for multi-indication evidence synthesis
In this work package, we will develop methods for multi-indication evidence synthesis in oncology. We will consider models for sharing of information across indications on a single endpoint and, alternatively, for sharing of information across indications on surrogate relationships. We will also develop methods for making explicit the level of sharing across indications imposed by the methods, ensuring that the appropriateness of the underlying assumptions can be explicitly considered by decision makers and clinical experts.
WP3: Additional work to support a policy framework
We will perform a simulation study to generalise from the bevacizumab case study to a broader set of multi-indication drugs and support the framework's recommendations and guiding principles relating to when multi-indication analyses may, or may not, be worthwhile. We will also develop methods for the formal elicitation of the judgment of clinical experts to support multi-indication analyses. Finally, we will demonstrate how the fuller use of multi-indication evidence can contribute to HTA decisions, by evaluating the impact on cost-effectiveness of integrating multi-indication evidence synthesis using the bevacizumab case study.
Closely after the regulatory license is awarded for a particular indication, the drug is appraised (for clinical and economic value) by health systems to determine whether it is made available (funded) to patients in that indication. This is called Health Technology Assessment (HTA). The evidence supporting HTA typically comprises of an indication-specific study that can be powered on intermediate (surrogate) outcomes, such as progression-free survival, rather than the final patient-relevant endpoints required by HTA agencies (e.g. overall survival).
The focus on single indications means that every appraisal is subject to a high level of evidential uncertainty on final endpoints, even where considerable evidence on the drug exists for other indications. We propose exploring whether better use of evidence can be made by looking across, as well as within, indications. Such analyses would share information across indications, reducing decision uncertainty across the existing indications and providing more realistic predictions of the value of the drug in future indications. The overall aim of this project is, therefore, to develop a framework for multi-indication evidence synthesis to support oncology HTA. The framework will outline the judgements and analyses required to specify, conduct, and interpret the synthesis of multi-indication evidence, and will incorporate a package of tools.
Three complementary work packages of methodological work will support the development of the framework:
WP1: Presenting and summarising evidence
In this work package, we will identify the relative effectiveness evidence for bevacizumab across its 11 indications, the real case study that will be used as an exemplar in this programme of work. To support multi-indication analyses, it is important that the evidence base is summarised appropriately for consideration by clinical and policy decision makers. Therefore, we will extend existing methods for the mapping of evidence to the multi-indication oncology setting (across indications, with multiple outcomes, and displaying the level of immaturity of evidence).
WP2: Methods for multi-indication evidence synthesis
In this work package, we will develop methods for multi-indication evidence synthesis in oncology. We will consider models for sharing of information across indications on a single endpoint and, alternatively, for sharing of information across indications on surrogate relationships. We will also develop methods for making explicit the level of sharing across indications imposed by the methods, ensuring that the appropriateness of the underlying assumptions can be explicitly considered by decision makers and clinical experts.
WP3: Additional work to support a policy framework
We will perform a simulation study to generalise from the bevacizumab case study to a broader set of multi-indication drugs and support the framework's recommendations and guiding principles relating to when multi-indication analyses may, or may not, be worthwhile. We will also develop methods for the formal elicitation of the judgment of clinical experts to support multi-indication analyses. Finally, we will demonstrate how the fuller use of multi-indication evidence can contribute to HTA decisions, by evaluating the impact on cost-effectiveness of integrating multi-indication evidence synthesis using the bevacizumab case study.
Technical Summary
Health Technology Assessment (HTA) evaluates the health and cost consequences of competing health care interventions. Typically, this evaluation is focused on evidence of the effects of a therapy in a specific indication. However, an increasing number of health technologies, particularly in oncology, are multi-indication gradually attaining license extensions to other indications as trial evidence emerges.
Indication-by-indication HTA typically ignores evidence for other indications on the same treatment and therefore commonly presents high levels of evidential uncertainty. The aim of this project is to make more efficient use of the entirety of evidence, within and across indications, thus reducing decision uncertainty and improving decisions. We plan a methodological programme of work, grounded on a case study in oncology. We will extend methodology for evidence maps to allow summarising evidence in multi-indication oncology HTA. We will extend and adapt two existing methodologies (mixture prior and mixture hierarchical models), which enable different ways for information to be shared across indications, to multi-indication network meta-analysis. We will conduct a simulation study to generalise the case study findings and help us make appropriate recommendations for information sharing in this context. We will develop methods for eliciting expert judgements on the appropriateness of information sharing. Finally, we will identify the implications of multi-indication evidence synthesis to decisions, aiming to demonstrate how the fuller use of evidence contributes to the understanding of the value of the treatment across indications. We will engage with relevant stakeholders to develop a policy framework to facilitate rapid translation into the methods and processes of HTA agencies.
Indication-by-indication HTA typically ignores evidence for other indications on the same treatment and therefore commonly presents high levels of evidential uncertainty. The aim of this project is to make more efficient use of the entirety of evidence, within and across indications, thus reducing decision uncertainty and improving decisions. We plan a methodological programme of work, grounded on a case study in oncology. We will extend methodology for evidence maps to allow summarising evidence in multi-indication oncology HTA. We will extend and adapt two existing methodologies (mixture prior and mixture hierarchical models), which enable different ways for information to be shared across indications, to multi-indication network meta-analysis. We will conduct a simulation study to generalise the case study findings and help us make appropriate recommendations for information sharing in this context. We will develop methods for eliciting expert judgements on the appropriateness of information sharing. Finally, we will identify the implications of multi-indication evidence synthesis to decisions, aiming to demonstrate how the fuller use of evidence contributes to the understanding of the value of the treatment across indications. We will engage with relevant stakeholders to develop a policy framework to facilitate rapid translation into the methods and processes of HTA agencies.