Is expert opinion reliable when estimating transition probabilities? The case of HCV-related cirrhosis in Egypt
- Equal contributors
1 IAME, UMR 1137, INSERM, F-75018 Paris, France; Univ Paris Diderot, Sorbonne Paris Cité, F-75018 Paris, France
2 EA2694, Université Lille Nord de France, Lille, France
3 Inserm U995, Université Lille 2 – Lille Nord de France, Lille, France
4 Department of Community, Environmental and Occupational Medicine, Faculty of medicine, Ain Shams University, Cairo, Egypt
5 Department of Endemic Medicine and Hepatology, Faculty of Medicine, Cairo University, Cairo, Egypt
6 Service des Maladies de l'Appareil Digestif et de la Nutrition, Hôpital Huriez, CHRU Lille, France
7 National Hepatology and Tropical Medicine Research Institute, Cairo, Egypt
8 Conservatoire National des Arts et Métiers, Chaire Santé et Développement, Paris, France
9 Unité d'Epidémiologie des Maladies Emergentes, Institut Pasteur, Paris, France
10 Service de Maladies infectieuses et tropicales, Hôpital Bichat Claude Bernard, Paris, France
BMC Medical Research Methodology 2014, 14:39 doi:10.1186/1471-2288-14-39Published: 17 March 2014
Data on HCV-related cirrhosis progression are scarce in developing countries in general, and in Egypt in particular. The objective of this study was to estimate the probability of death and transition between different health stages of HCV (compensated cirrhosis, decompensated cirrhosis and hepatocellular carcinoma) for an Egyptian population of patients with HCV-related cirrhosis.
We used the “elicitation of expert opinions” method to obtain collective knowledge from a panel of 23 Egyptian experts (among whom 17 were hepatologists or gastroenterologists and 2 were infectiologists). The questionnaire was based on virtual medical cases and asked the experts to assess probability of death or probability of various cirrhosis complications. The design was a Delphi study: we attempted to obtain a consensus between experts via a series of questionnaires interspersed with group response feedback.
We found substantial disparity between experts’ answers, and no consensus was reached at the end of the process. Moreover, we obtained high death probability and high risk of hepatocellular carcinoma. The annual transition probability to death was estimated at between 10.1% and 61.5% and the annual probability of occurrence of hepatocellular carcinoma was estimated at between 16.8% and 58.9% (depending on age, gender, time spent in cirrhosis and cirrhosis severity).
Our results show that eliciting expert opinions is not suited for determining the natural history of diseases due to practitioners’ difficulties in evaluating quantities. Cognitive bias occurring during this type of study might explain our results.