Article type
Year
Abstract
Background: Treatment sequences relate to the order in which interventions are administered within treatment pathways. For many conditions several alternative treatments are available should patients respond poorly. The potential effectiveness of each may differ according to its position in the treatment pathway. When a new drug is introduced it is necessary to compare its value and determine its optimum point of delivery. Subgroups of patients may benefit more, or experience greater benefit if the new treatment is used earlier; new drugs are generally evaluated at later stages.
Objectives: To identify quantitativemethods developed to estimate the treatment effect of interventions conditional on previous treatments in the sequence being ineffective, or only partially effective.
Methods: A comprehensive literature review to identify studies describing methods assessing treatment sequences, and Health Technology Appraisals implementing them. Due to scarce relevant indexing terms and no clear methodological taxonomies, a conventional systematic search was insufficient. A more pragmatic and iterative process based on ‘pearl growing’ was used. Included analyses were summarised as a narrative and appraised using predefined criteria.
Results: Most commonly, treatment sequences were evaluated in a limited way as part of an economic evaluation. These evaluations used simplistic assumptions for sequence specific treatment effects. Importantly, they tended to consider isolated decision points, ignoring potential impact of upstream events or treatment decisions. Main challenges include lack of directly matching evidence and poor reporting of previous treatments. Recent developments such as multi-parameter evidence synthesis, individual patient data meta-analysis, and whole disease modelling could potentially be extended to address theses and incorporate treatment sequences.
Conclusions: There is a growing need for policy and practice decision-making to consider evidence on treatment sequences for chronic conditions. No current guidance or empirically tested methods exist for conducting quantitative evidence synthesis to develop effect sizes conditional on previous treatment used.
Objectives: To identify quantitativemethods developed to estimate the treatment effect of interventions conditional on previous treatments in the sequence being ineffective, or only partially effective.
Methods: A comprehensive literature review to identify studies describing methods assessing treatment sequences, and Health Technology Appraisals implementing them. Due to scarce relevant indexing terms and no clear methodological taxonomies, a conventional systematic search was insufficient. A more pragmatic and iterative process based on ‘pearl growing’ was used. Included analyses were summarised as a narrative and appraised using predefined criteria.
Results: Most commonly, treatment sequences were evaluated in a limited way as part of an economic evaluation. These evaluations used simplistic assumptions for sequence specific treatment effects. Importantly, they tended to consider isolated decision points, ignoring potential impact of upstream events or treatment decisions. Main challenges include lack of directly matching evidence and poor reporting of previous treatments. Recent developments such as multi-parameter evidence synthesis, individual patient data meta-analysis, and whole disease modelling could potentially be extended to address theses and incorporate treatment sequences.
Conclusions: There is a growing need for policy and practice decision-making to consider evidence on treatment sequences for chronic conditions. No current guidance or empirically tested methods exist for conducting quantitative evidence synthesis to develop effect sizes conditional on previous treatment used.