Flowchart for selection of Web sites for analysis.
Proportion of Web sites matching the dietary aspects of the American Diabetes Association (ADA) recommendations for medical nutrition therapy for type 2 diabetes.
Distribution of matching scores among Web sites included in the study.
Post RE, Mainous AG. The Accuracy of Nutrition Information on the Internet for Type 2 Diabetes. Arch Intern Med. 2010;170(16):1504-1506. doi:10.1001/archinternmed.2010.289
A key strategy of diabetes self-management is medical nutrition therapy,1 which has been shown to reduce hemoglobin A1c and low-density lipoprotein cholesterol levels in patients with type 2 diabetes mellitus.2,3 The American Diabetes Association (ADA) recommends a specific diet as part of medical nutrition therapy for secondary and tertiary prevention in patients with diabetes.1
Forty to sixty percent of Americans search for health-related information on the Internet.4,5 Studies regarding health information accuracy on the Internet have shown that health advice varies in quality and accuracy regarding a wide range of topics.6- 8 There is a scarcity of literature regarding the accuracy of self-management strategies for chronic disease on the Internet, especially regarding type 2 diabetes. Therefore, we evaluated the accuracy of nutrition information for type 2 diabetes on the Internet.
A search of “diabetes and nutrition” was performed on both Google and Yahoo on March 18, 2010. The first 100 search results from each search engine were examined. Web sites were excluded from analysis for the following reasons: (1) need to register or pay for an account to access all information or if the Web site (2) was a portal to other Web sites, (3) was a link to a medical journal article, (4) contained information only for type 1 or gestational diabetes, (5) was for the sale of a particular product, eg, a book on diabetic nutrition, (6) was under the same domain as a previously included Web site, (7) was the official Web site of another country's diabetes association, or (8) was the Web site for diabetes education services at a medical center or other setting and did not include specific nutritional information. If a Web site appeared as a search result on both search engines, it was only counted once. Internal links to other pages within the same domain name were followed to try to account for as much information as possible. External links to other Web sites were not followed.
The information collected included recommendations for the 11 dietary aspects included in the established ADA nutrition guidelines for patients with diabetes1 and when the Web site was last updated. The main outcome measures were the proportion of Web sites where information matched that of the ADA recommendations for each of the dietary aspects. Web sites were considered to match the recommendations if the Web sites explicitly stated the same recommendations as the ADA (Table). An overall matching score (maximum score, 11), which was the number of dietary aspects that matched the ADA guidelines, was also assigned to each Web site.
A comparison was also performed of Web sites that updated information before or after the most recent ADA recommendations in 2008. The χ2 test was used to compare the proportions of Web sites that contained information matching each dietary aspect of the ADA recommendations. The independent, 2-sample t test was used to compare the mean matching score assigned to the Websites in each group. Statistical analysis was performed using OpenEpi software.9
Of the first 200 combined search results, 161 Web sites were excluded because of our prespecified exclusion criteria, leaving 39 for analysis (Figure 1). The least frequent recommendation was consumption of 2 or more servings of nonfried fish per week (15.4%) and the most frequent recommendation was to monitor carbohydrates (82.1%) (Figure 2). The mean (SD) matching score for all Web sites was 3.56 (2.20) (Figure 3). The mean matching score for Web sites updated before 2008 was not statistically different from the mean score for those Web sites updated in 2008 or later (3.61 vs 3.50; P = .88).
Nutrition information that is freely available on the Internet for patients with type 2 diabetes is not complete compared with the most recent, established ADA recommendations. Of note, many of these Web sites do not provide misinformation, but rather a lack of information. Examples include many Web sites that mentioned to increase fiber or limit saturated fat intake but did not give a specific range or cutoff for patients to target as a goal. Clear guidance may be more helpful to patients in changing their behaviors. A few Web sites included dietary advice that was close to the ADA recommendations but not accurate. One example advised readers to limit saturated fat intake to less than 10% of daily caloric intake. Though it is true that the ADA recommendation of less than 7% is within this range, this is still incorrect information.
The date of when the Web sites were updated did not affect the accuracy of the information. Web sites that were updated after the most recent ADA recommendations did not include more information matching the recommendations than the older Web sites. Even though the content of these Web sites is being checked on a more frequent basis, the authors of the Web sites are not up to date with the literature.
Web sites that provide nutritional advice for patients with diabetes should reference evidence-based sources for their information. Only 10 of the 39 articles (26%) in this study referenced peer-reviewed sources. Advising diabetic patients to use the Internet to supplement advice given in the office is not recommended at this time and should not be recommended until Web sites improve the accuracy of their information.
Correspondence: Dr Post, Department of Family Medicine, Medical University of South Carolina, 295 Calhoun St, Charleston, SC 29425 (firstname.lastname@example.org).
Author Contributions: Dr Post had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Post and Mainous. Acquisition of data: Post. Analysis and interpretation of data: Post. Drafting of the manuscript: Post. Critical revision of the manuscript for important intellectual content: Post and Mainous. Statistical analysis: Post. Study supervision: Mainous.
Financial Disclosure: None reported.