Nearly 160 million Americans search the Internet for health-related information every month. Health-focused Web sites have sprung-up everywhere in response to this extraordinary demand. The problem is that the information on many health-related Web sites is neither accurate nor complete, and to the extent that people rely on misinformation, dangerous health consequences become more likely.
How serious is the problem? To find out, orthopedic surgeon James Starman and colleagues from the Carolinas Medical Center in Charlotte selected 10 common conditions in the field of sports medicine (e.g. anterior cruciate ligament rupture, rotator cuff tear) and developed a scheme to assess the quality of on line information for each one.
The team’s assessments focused on the type of Web site, the information content on that Web site, and transparency of the information.
Type of Web Site-Starman’s group divided the universe of health-related Web sites into 5 categories: academic, commercial, news-oriented, personal, physician and nonprofit. Academic Web sites were affiliated with a medical journal, a university or a medical society. Commercial sites displayed ads, sold products or received industry funding (this group included WebMD and eMedicine). News-oriented sites were nonmedical ones that contained articles and anecdotes about the sports medicine conditions Starman’s group had selected. Personal Web sites were authored by physical trainers and laymen. Nonprofit sites were supported by the government and by donations (e.g. the NIH web site and Wikipedia).
Information Content-The scientists parsed the health information content on each site into 5 categories: disease summary, pathogenesis, diagnostics, treatment and complications, and outcomes and prognosis. Then, for each category they graded the sites on a 0-100 scale, with 100 representing the most outstanding levels of accuracy and completeness.
Transparency of Information-The scientists determined whether each Web site displayed a banner indicating its intention to comply with Health on the Net Foundation (HON) protocols, which assure transparency of information and purpose. They also assessed compliance with HON principles using a 16-point scale.
The scientists then searched for their 10 sports medicine conditions on Google and Yahoo, and asked 3 reviewers to analyze the top 10 sites returned for each search.
Using these methods, 200 potential web sites could have been generated for the review (10 diagnoses, 2 search engines, 10 results for each search). Tellingly, 154 Web sites were selected for review. That means only 46, or less than half of the search findings were returned by both Google and Yahoo.
Of these, commercial sites were the most common (74, or nearly half of the total). Next up were academic (32).
The average information content scores were 56.8, 56.0, and 54.8 on a 100 point scale, with values for treatment options consistently receiving higher scores than those for pathogenesis and outcomes. The authors noted in particular, that there was a consistent lack of information regarding potential complications of treatment. Non-profit sites had the highest information content score, while news-related sites had the lowest.
Across the board, the average HON scores on a 16-point scale were 9.8, 9.5, and 8.5 (out of 16) for the 3 reviewers, respectively. Personal web sites returned the lowest HON scores, while non-profit sites had the highest scores. Academic and commercial site scores were observed to have the most variation, although WebMD and eMedicine came out quite well across the board.
Only 37 of the 154 evaluated sites displayed the HONcode seal; 117 did not. The average HON and information content scores for sites with the HONcode seal were 12.0 and 65.4, respectively. For those without the seal, the numbers were 8.4 and 52.9. These differences were significant.
In their write-up, which appears in the Journal of Bone and Joint Surgery, the authors concluded that the transparency and information content of health information on the most searched-for Web sites is quite variable for sports medicine conditions. They were particularly concerned about the presentation of biased information and lack of accountability on certain commercial sites.
They encouraged seekers of health information to avoid commercial web sites other than “reputable” ones like “WebMD and eMedicine,” and to look for the HONcode seal of compliance for transparency and accountability.
Sounds like a plan to me.
Glenn Laffel, MD, PhD
Sr. VP Clinical Affairs
Practice Fusion EMR
