David Hadorn1, Giorgi Kvizhinadze1, Lucie Collinson1, Tony Blakely1. 1. Burden of Disease Epidemiology,Equity and Cost Effectiveness Programme,Department of Public Health,University of Otago,Wellington,PO Box 7343,Wellington,New Zealandtony.blakely@otago.ac.nz.
Abstract
OBJECTIVES: The aim of this study was to determine the prevalence and methods of expert knowledge elicitation (EKE) for specifying input parameters in health economic decision models (HEDM). METHODS: We created two samples using the National Health System Economic Evaluations Database: (1) 100 randomly selected HEDM studies to determine prevalence of EKE and (2) sixty studies using a formal EKE process to determine methods used. RESULTS: Fifty-seven (57 percent) of the random sample included at least one EKE-derived parameter. Of these, six (10 percent) used a formal expert process. Thirty-four studies from our second sample of sixty studies (57 percent) described at least one aspect of the process (e.g., elicitation method) with reasonable clarity. In approximately two-thirds of studies the external experts estimated parameters de novo; the remainder confirmed or modified initial estimates provided by authors, or the method was unclear. The majority of elicitations obtained point estimates only, although a few studies asked experts to estimate ranges of parameter values. CONCLUSIONS: The use of EKE for parameter estimation is common in HEDMs, although there is room for improvement in the methods used.
OBJECTIVES: The aim of this study was to determine the prevalence and methods of expert knowledge elicitation (EKE) for specifying input parameters in health economic decision models (HEDM). METHODS: We created two samples using the National Health System Economic Evaluations Database: (1) 100 randomly selected HEDM studies to determine prevalence of EKE and (2) sixty studies using a formal EKE process to determine methods used. RESULTS: Fifty-seven (57 percent) of the random sample included at least one EKE-derived parameter. Of these, six (10 percent) used a formal expert process. Thirty-four studies from our second sample of sixty studies (57 percent) described at least one aspect of the process (e.g., elicitation method) with reasonable clarity. In approximately two-thirds of studies the external experts estimated parameters de novo; the remainder confirmed or modified initial estimates provided by authors, or the method was unclear. The majority of elicitations obtained point estimates only, although a few studies asked experts to estimate ranges of parameter values. CONCLUSIONS: The use of EKE for parameter estimation is common in HEDMs, although there is room for improvement in the methods used.
Authors: Laura Bojke; Bogdan Grigore; Dina Jankovic; Jaime Peters; Marta Soares; Ken Stein Journal: Pharmacoeconomics Date: 2017-09 Impact factor: 4.981
Authors: Duncan O S Gillespie; Kirk Allen; Maria Guzman-Castillo; Piotr Bandosz; Patricia Moreira; Rory McGill; Elspeth Anwar; Ffion Lloyd-Williams; Helen Bromley; Peter J Diggle; Simon Capewell; Martin O'Flaherty Journal: PLoS One Date: 2015-07-01 Impact factor: 3.240