True North: Building Imaginary Worlds with the Revised Canadian (CADTH) Guidelines for Health Technology Assessment
In March 2017 the Canadian Agency for Drugs and Technologies in Health (CADTH) released the 4th edition of their Guidelines for the Economic Evaluation of Health Technologies: Canada. These guidelines, which were first published and revised for a 3rd edition in 2006 are intended to help decision makers, health systems leaders and policy makers make well-informed decisions. They are designed, apparently, to support best practice in conducting health technology assessments in Canada. The purpose of this commentary is to consider whether or not the evidence standards proposed and the consequent modeled claims for economic effectiveness meet the standards of normal science: are the CADTH standards capable of generating claims for competing products that are credible, evaluable and replicable? The review argues that the standards proposed by CADTH do not meet the standards expected in normal science. Technical sophistication in building reference case imaginary worlds is not a substitute for claims that are experimentally evaluable or capable of assessment through systematic observation. There is no way of judging whether imaginary claims are right or even if they are wrong. CADTH is not alone in setting standards that fail to meet the standards of normal science. Recent commentaries on formulary submission guidelines in a number of other countries, to include Ireland, the Netherlands, France, Australia, the UK and New Zealand conclude that they are subject to the same criticism. If the CADTH guidelines were never intended to support feedback to health system decision makers, then this should be made clear. If not, then consideration should be given to withdrawing the guidelines to ensure they conform to these standards. Hopefully, future versions of the CADTH guidelines will address this issue and focus on a rigorous research program of claims assessment and feedback and not the building of imaginary worlds.