We believe that the evidence supporting the recommendations by the Canadian Task Force on Preventive Health Care1 has been misinterpreted in several key ways.
First, we disagree that our decision model (commissioned by the Public Health Agency of Canada) was “very low quality evidence.”1 We believe that the model, when judged by the correct standards, is of good quality. For example, in a systematic review by Coward and colleagues, the model received a perfect score using the Consolidated Health Economic Evaluation Reporting Standards checklist.2 The fact that the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system was used to evaluate the model for the guideline represents a misunderstanding of both modelling and GRADE. The GRADE checklist is specifically designed for assessing primary clinical evidence, not secondary synthetic methods such as decision or economic models, which integrate a wide heterogeneous set of epidemiologic, economic and patient preference evidence. The GRADE handbook specifically states that models should not be included in evidence profiles.3 Our model was not only included but considered to be observational evidence, which it is not.
Other limitations for applying GRADE criteria to models are discussed in the review by Rehfuess and colleagues.4 Models have been and are routinely used around the world to inform public health and reimbursement decisions, clinical guidelines and screening policy. In the area of hepatitis C virus screening, where benefits occur decades into the future, modelling is the only practical option we have for fully incorporating all health and cost outcomes.
Second, the evidence cited to highlight the harms of screening seems, on direct inspection, to suggest the opposite. In three of the studies cited, where patients are directly asked their willingness to be screened, the proportion that approve exceeds 90%.5–7 Each paper cited by the task force either concludes with a strong preference toward screening or is designed to highlight implementation issues that need to be overcome to improve screening efficacy. It is never suggested that screening itself is harmful to patients and should not take place. The survey commissioned by the task force concludes: “In considering the benefits of screening and harms of treatment, participants reported a strong preference to [be] screened for hepatitis C (median rating of 8 [interquartile range 6–9] on a 9-point scale).”8 This is a very strong consensus in favour of screening. It is perplexing to us that such a negative conclusion was reached on the basis of this evidence.
We are not screening evangelists, and our concerns involve the interpretation of the evidence presented to the task force. We feel that rational public policy requires a fair and considered evaluation of all available evidence, including evidence from models and from patients.
Footnotes
Competing interests: William Wong and Murray Krahn have received research support from the Canadian Liver Foundation.