Documental Quality of Websites Concerning Physical Activity, Lifestyle and Sedentarism Available on the Internet

Objectives: To analyze the documental quality of 389 websites in Portuguese about physical activity, healthy lifestyles and sedentary lifestyles found on the Brazilian version of the general search engine Google. Methods: The documental quality of the 389 websites was estimated based upon the following parameters: 1) a combination of quality criteria from the Health Information Locator (LIS—OPS/BIREME) and those from Chile’s Pontifical Catholic University, organized into 17 variables; 2) uniformity of reference criteria (Vancou-ver); 3) association between the presence of authorship and a higher number of the quality criteria being fulfilled. We also studied the ranking of the results presented by Google in addition to attributes connected to the websites’ target audience, the types of content, their sponsors and country of origin. Results: Of the 389 websites studied, 111 links were not active (28.53% CI 95% [24.05 - 33.02]) and none of the websites in the sample met all of the 17 quality variables. Authored websites displayed remarkable differences in quality when compared to those which did not identify their authors. Conclusions: Faced with the issue of the proliferation of websites with questionable quality content, and the fact that the ranking of results interferes directly in the internal evaluation of content relevance, we propose that public-health research institutions cooperate with web-searching developers to improve the web-site-positioning formula, in which the “identified authorship” criterion should play a major role in the ranking system.

Health field of therapy and disease diagnosis. On the other hand, while "efficacy" discourses underlie the production of scientific content, there is also a profusion of content directed at the mass consumption of health-related products [1]. Various studies have related that the Internet fills a huge gap as the most popular place of reference about these issues [2] [3]. For the producers of the information sites, the "Web" does not pose restrictions on the suitability of any person or body who wants to add content about health promotion, whether the motivation is preventive or commercial. This adds to the difficulties in controlling the quality of health information put on the Internet [4] and makes assessing this issue of the Internet highly interesting [5] [6].
The rapid growth in websites dedicated to the health sciences causes difficulties when it comes to ensuring and judging the quality of the information they offer. There are many proposals by Health Organizations that seek criteria to identify high-quality information on the Internet. Because of the disparity in technical knowledge between experts and the general public, accurate information on these subjects is not always clear and easily understandable. It becomes increasingly difficult to discern the information that is updated and produced with technical rigor from the inaccurate, or obsolete information that may, eventually, mislead the public [5].
Nowadays it is not uncommon for patients to turn to their doctors with a great volume of information obtained on the Internet regarding their diseases. In about 15% of cases, the diagnosis and treatment offered by the health professional are questioned based upon information from dubious sources [7]. It is becoming increasingly frequent for patients to resort to the Web in order to find a "virtual diagnosis" or "elixir" to alleviate symptoms of illness. As there are no widely accepted certification filters, the user cannot evaluate the quality of health information content and tends to make the mistake of thinking that, because it is on the Internet, all the information there has unquestionable scientific value. The questionable content of the specialized information on websites is also growing in the field of health promotion. The rhetoric ranges from acclaiming physical activity as a simple measure for solving complex health problems to the legitimization of products on the beauty market. The paradox is that the imperative to lead a "healthy lifestyle" seems more related to physical appearance and moral values than to health itself [8] [9].
Whether it focuses on self-treatment or on health promotion, the ease of finding a website on the Internet (given that this is connected to its greater visibility) is directly related to its credibility. This fact boosts the belief that "friendly (understood as familiar) is also dependable". The website positioning algorithms (page rank) and Google's "sample fallacies", described below, directly interfere with the positioning of the results that connect to the websites, deciding the sequence in which they appear, which directly influences the user's judgment of their pertinence and relevance.
This study sought to analyze (through criteria postulated by widely accepted institutions in this field) the documental quality of websites whose content deals Health with combating a sedentary lifestyle and promoting healthy lifestyles (through physical activity), found on Google. We also checked the sites' main attributes, their visibility, as well as the positive association between the presence of authorship and compliance with quality indicators.

Methods
Three hundred and eighty-nine (389) Portuguese language websites found on the Brazilian version of the general search engine Google (http://www.google.com.br/) were evaluated (Search date: Nov 15, 2012), through "advanced search" and the option "include all words", using as metadata (search terms) the phrases: "physical activity", "lifestyle" and "sedentarism".
It is well-known that the Google search engine tends to change the results displayed from one search to the next or on different computers [10] [11]. In order to avoid mistakes caused by changes in results and retain the "positioning" (original order of the results), the results were recorded in pdf format (option: "100 results per page") with a hyperlink. Exclusion criterion: pages that requested payment before consultation or that limited access to registered users.
An Excel 2003® spreadsheet was used for storing the data and hyperlinks, mathematical calculations and creating tables and figures. The statistical analysis procedures were performed using the program SPSS ® version 15.0 for Windows.
In the documental quality evaluation we used a check-list of 17 quality indicators for websites (shown in Table 1 The following features were also analyzed: inactive links; target audience; and objectives (academic, commercial, governmental, etc.).
We also studied the correlation (Pearson method) between "authorship presence" and "greater compliance with these information quality criteria".  Table 1. Check-list-indicators related to documental quality. 1) Authorship-people responsible for the displayed contents.

Description of the Websites
2) Promoting body-institution that supports the website.
3) Endorsement-relevant professional or academic institutions that give guarantees and endorse the site.

Documental Quality
No website fulfilled all the 17 indicator criteria in Table 1, and none of them indicated the date of update or declaration of conflicts of interest (Table 2). A maximum score of 14 indicators fulfilled (1 case, or 0.3% CI 95% [0.0 -1.1]) and a minimum of 0 (3 cases 1.1% CI 95% [0.0 -2.3]) were observed. A mean of 9.2 (± 0.1) and a median of 10 variables were met. In relation to the fulfillment of the six Vancouver uniformity criteria, no site met all these variables-the mean was 3.5 ± 0.1 and the median was 4, with a maximum of 5 (29 cases 10.4% CI 95% [6.8 -14.0]) and minimum of 0 (4 cases 1.4% CI 95% [0.0 -2.8]).

Authorship Presence
Among the pages studied, the presence of authorship may be positively correlated to six quality indicators ( Table 3)

Discussion
From the point of view of documental quality, it is possible to state that the websites analyzed displayed unsatisfactory results. Additionally, the number of inactive links on homepages was higher than described in other works [16] [17] [18], which may indicate the presence of a significant number of unstable and, perhaps, untrustworthy sources. The documentation uniformity criterion may contribute to the evaluation of this picture, due to the need to identify the origin of the sources and when they were updated: failures in websites references make it difficult to verify their trustworthiness [19] [20] [21] [22].
In order to find out whether the information disseminated by a periodical and a general or technical book is up-to-date, it is essential to know the date of its creation or publication as well as, in some cases, the date on which it was last updated. Equally essential to the credibility of texts, especially in specialized technical literature, is the possibility of contact with the person responsible for the published document, which may not be specified in about a third of the sample analyzed in our investigation. It is also worrying to note that among the homepages studied, more than three quarters did not record the date when they were first published and none recorded the dates they were updated. A technical book would be useless without this information.
Considering such results and in the face of the high number of websites dedicated to the health sciences, the question arises: how can we ensure and judge the quality of so much information that deals with such complex subjects and need specialized and up-to-date technical validation? There are similar works in the literature that use indicators comparable to the ones in this study [23].

Conclusion
We concluded that the documental quality of information found in Brazilian websites is very poor. However, it should be admitted that even fulfillment of all the variables presented here, would not by itself totally guarantee the quality of the contents of the analyzed websites. Authored websites offer some degree of accountability in the sense that their authors may be held partially responsible for their contents. Faced with the issue of the proliferation of websites with questionable quality content, and the fact that the ranking of results interferes directly in the internal evaluation of content relevance, we propose that public-health research institutions cooperate with web-searching developers to improve the website-positioning formula, in which the "identified authorship" criterion should play a major role in the ranking system. This concept should be the object of further analysis by specialists in scientific documentation, who would be responsible for recommending selection criteria, above all the existence of website authorship, for the ranking of results. This would facilitate the internal review and evaluation (those that the users themselves use in order to select their texts) in terms of pertinence and relevance.

Collaborators' Participation
The authors developed and organized all parts of the text in conjunction.