Internet Medical Search Results Highlight Prior Bias

The retrospective analysis was conducted on the search and decision behaviors of 75 clinicians (44 doctors, 31 nurses), answering questions for eight clinical scenarios within 80 minutes in a controlled setting. The prospective study was conducted on 227 undergraduate students, who used the same search engine to answer two of six randomly assigned consumer health questions.

The results showed that a person's prior belief (termed anchoring) has a significant impact on their post-search answer. Documents accessed at different positions in a search session, and documents processed for different lengths of time also influenced decision post-search more than expected in the prospective experiment, but not in the retrospective analysis. Reinforcement through repeated exposure to a document did not yield statistical differences in decision outcome post-search. Other factors influencing decisions is where the information appears in the search results; the first or the last document the user sees has a much greater impact on their decisions. The researchers concluded that providing people with the right information on its own may not be enough; even if search engines do find the 'right' information, people may still draw the wrong conclusions, basing decisions on biased opinions. The study was published in the September-October 2007 issue of the Journal of the American Medical Informatics Association (JAMIA).

"We know that the web is increasingly being used by people to help them make healthcare decisions,” said co-author Professor Enrico Coiera, Ph.D., of the center for health informatics. "We know that there can be negative consequences if people find the wrong information, especially as people in some countries can now self-medicate by ordering drugs online. Australians can order complementary medicines online and these can interfere with other medications.”


Related Links:
University of New South Wales