Quality-related variables at hepatological websites

Background. The amount of hepatology-related information available on the Internet has substantially increased, but little is known about the characteristics and quality of the websites. Aim. The aim of this study was to describe analytically and evaluate critically the information concerning three diseases of hepatological interest: chronic hepatitis, hemochromatosis and Caroli's disease. Methods. In accordance with a validated method, the three search terms were entered into four English language search engines and the first five links of each were considered (a total of 60 sites). The characteristics of the websites were described and their quality was evaluated by three independent reviewers who assigned scores of 1-5 for accuracy, reliability and depth. The relationships between the site characteristics and quality scores were analysed by means of multiple logistic regression. Results. The overall rating score was sufficient (greater than or equal to3) in 51% (95% confidence interval: 38-65%) of cases. The majority of the sites (73%) were aimed at patients rather than at physicians. Commercial sponsorship was significantly more frequent among the chronic hepatitis sites (45%) than among the hemochromatosis (15%) or Caroli's disease sites (0%) (P = 0.002); 61% of the commercial sites did not include a financial disclosure. The only variable that independently related to poor quality was the presence of commercial sponsorship (odds ratio 18.1: 95% confidence interval: 1.7-192.5). Conclusions. Hepatological websites are characterised by poor quality and are mainly aimed at patients. Quality is negatively affected by commercial interests, which are often undeclared. Guidelines for the certification and surveillance of websites relating to liver diseases are highly advisable