OBJECTIVE: Published data suggest that the quality of information on diseases accessible on the Internet using non-medical search engines is poor. Such data do not exist for illnesses requiring intensive care. This study investigated the accuracy of health information about head injury pertaining to intensive care on the Internet, and correlated website characteristics with the quality of their content. METHODS: A cross-sectional survey was conducted of the first 20 websites retrieved by the seven most frequently used search engines, with the information evaluated by two independent observers. Inter-observer reliability was evaluated using the kappa statistic. Website information on head injury was compared with "gold standard" guidelines from the Brain Trauma Foundation. Website characteristics were assessed, and their correlation with quality of website content was analysed. RESULTS: 58 websites were assessed. Weighted kappa for interobserver agreement on quality scores was 0.72. The median content score was 2 (interquartile range, 0-4) out of a possible maximum of 23. Logistic regression analysis suggested that medical authors, government sponsors, and being in the second 10 websites retrieved by a search engine were associated with higher website quality scores, while financial incentive and advertisement were associated with lower quality scores. CONCLUSION: This study demonstrated that information retrieved by the public on head injury from non-medical websites may be incomplete and inaccurate. It also identified website characteristics associated with poor content quality.
OBJECTIVE: Published data suggest that the quality of information on diseases accessible on the Internet using non-medical search engines is poor. Such data do not exist for illnesses requiring intensive care. This study investigated the accuracy of health information about head injury pertaining to intensive care on the Internet, and correlated website characteristics with the quality of their content. METHODS: A cross-sectional survey was conducted of the first 20 websites retrieved by the seven most frequently used search engines, with the information evaluated by two independent observers. Inter-observer reliability was evaluated using the kappa statistic. Website information on head injury was compared with "gold standard" guidelines from the Brain Trauma Foundation. Website characteristics were assessed, and their correlation with quality of website content was analysed. RESULTS: 58 websites were assessed. Weighted kappa for interobserver agreement on quality scores was 0.72. The median content score was 2 (interquartile range, 0-4) out of a possible maximum of 23. Logistic regression analysis suggested that medical authors, government sponsors, and being in the second 10 websites retrieved by a search engine were associated with higher website quality scores, while financial incentive and advertisement were associated with lower quality scores. CONCLUSION: This study demonstrated that information retrieved by the public on head injury from non-medical websites may be incomplete and inaccurate. It also identified website characteristics associated with poor content quality.