JMIR Publications

ABSTRACT

Background:

Critical evaluation of online health information has always been central to consumer health informatics. However, with the emergence of new Web media platforms and the ubiquity of social media, the issue has taken on a new dimension and urgency. At the same time, many established existing information quality evaluation guidelines address information characteristics other than the content (e.g., authority, currency), target information creators rather than users as their main audience, or do not address information presented via novel Web technologies.

Objective:

The objectives of this formative study are to 1) develop a methodological approach for analyzing health-related webpages and 2) apply it to a set of relevant webpages.

Methods:

The study analyzed twenty-four type 2 diabetes pages which were the results of a Google search with the keywords “diabetes, reversal, natural.” The coding scheme, developed via a combination of theory-driven and data-driven approaches, includes five categories from existing guidelines (resource type, information authority, validity of background information sources, objectivity, currency) and seven novel categories (treatment/reversal method, promises and certainty, criticisms of establishment, emotional appeal, vocabulary, rhetoric and presentation, use of science in argumentation). The coding involves both categorical judgement and in-depth narrative characterization. Upon establishing satisfactory level of agreement on the narrative coding, the team coded the complete dataset of twenty-four pages.

Results:

The results set included “traditional” static pages, videos, and digitized versions of printed newspapers or magazine articles. Treatments proposed by the pages included a mixture of conventional evidence-based treatments (e.g., healthy balanced diet, exercise) and unconventional treatments (e.g., dietary supplements, optimizing gut flora). Most pages either promised or strongly implied high likelihood of complete recovery. Pages varied greatly with respect to the authors’ stated background and credentials, as well as the information sources they referenced or mentioned. The majority included criticisms of the traditional healthcare establishment. Many sold commercial products ranging from dietary supplements to books. The pages frequently used colloquial language. A significant number included emotional personal anecdotes, made positive mentions of the word “cure,” and included references to nature as a positive healing force. Most pages presented some biological explanations of their proposed treatments. Some of the explanations involved the level of complexity well beyond the level of an educated lay person.

Conclusions:

Both traditional and data-driven categories of codes used in this work yielded insights about the resources and highlighted challenges faced by their users. This exploratory study underscores the challenges of consumer health information seeking and the importance of developing support tools that would help users seek, evaluate, and analyze information in the changing digital ecosystem.