Greater access to web-based health information may empower women by giving them necessary facts or advice about health-care that their providers do not have time or depth of knowledge to communicate. Greater participation in decisions regarding care and enhanced self-care may result. If the indicated actions produce benefits, then greater access to information will lead to better health outcomes. Research has not conclusively established these benefits, however; consequently, the potential for harm exists and should be carefully considered. Although women routinely use the Internet to seek answers to their health questions (23), it is not at all clear whether the answers they find are based on the best available scientific evidence most likely to lead to optimal health outcomes.
With steadily increasing rates of Internet use among health consumers, there is accelerating potential for as much misinformation as valid information to be widely diffused. In a general survey of Canadians that found 22 percent adult web use in the general population, women were found to be more likely than men to search for health information (21). A recent survey has shown even higher rates in other Western countries, with a third of Europeans and almost half of Americans seeking health information online (6). Rates among current patients may be higher than the general population: 60 percent of patients in a U.S. primary-care practice were found to use the Internet for health-related information, but without necessarily divulging Internet use or the information they found with their physicians (5).
Increasing concern with the quality of health information on the Internet parallels increased consumer use. In response to such concerns, numerous initiatives have sought to develop instruments, certification, and rating systems to help consumers gauge the quality of a health information site. These tools have themselves been challenged as lacking a basis of quality. In a repeat of a seminal study of instruments used to assess website quality, Gagliardi and Jadad found that none of the 98 tools identified reported on the reliability and validity of the measure or provided instructions on how to obtain the ratings (12). The authors' original study concluded: “It is unclear, however, whether [the quality rating instruments] should exist in the first place, whether they measure what they claim to measure or whether they lead to more good than harm” (18). By looking for transparent and pervasive infrastructures, new initiatives aim to surmount the considerable logistical and methodologic problems in the task of helping consumers identifying quality health information. Nevertheless, the potential benefits have to date not been established through rigorous scientific evaluation (10;12).
In the meantime, reports of inaccurate, misleading, and therefore, potentially harmful web-based information are plentiful (26). Illustrative is the classic study by Impicciatore et al. (17), which found that only four of forty-one websites offering information on home management to parents of children with fevers adhered closely to the main recommendations in published guidelines. Although a recent re-evaluation of the same topic and many of the same sites found that the quality of health information on the Internet had improved, the results were not entirely reassuring—still only 45 percent of the new pages (compared with 8 percent of the original pages) adhered to accepted clinical guidelines (24). This finding led the authors to conclude “monitoring health information on the Internet for accuracy, completeness, and consistency is still fundamental” (24).
The standard against which information on the web is evaluated is often provided by a clinical practice guideline (14), or an individual clinical expert recruited for the purpose of the study (12). Discrepancies within and between guidelines and expert opinion are not uncommon, which raises questions about the appropriate comparator (4;14;19). We posit that health technology assessments (HTA; as topic-specific reports) have a distinct advantage as a comparator, for several reasons. HTA reports are produced by independent organizations that are not in a position of conflict of interest with provider, consumer, or industry groups, because they do not stand to gain financially from the provision of services, sale of products, or diversion of costs. More importantly, HTA reports are conducted with a specific mandate to examine and report the scientific research underpinning emerging technologies. HTA reviews, therefore, should be scientifically rigorous, which is the benchmark of quality. Lastly, HTA reports are most frequently conducted for the purpose of providing input into public policy decisions likely to have widespread social impact. Comprehensive HTA reports will provide an analysis from the perspective of population groups making explicit all the expected consequences of widespread application of the technology. While policy-makers rather than consumers are the intended audience, this type of information may be of interest to consumers as well.
Two of the present authors previously had conducted HTA studies on bone mineral density (BMD) testing in well women (13;20). Discrepancies with the study findings emerged in subsequent discussions with women who had obtained health information on the Internet. This awareness of inconsistencies motivated the present study, whose goal is to evaluate the quality of the information content posted on consumer health websites (CHWS), compared with HTA reports that have undergone rigorous systematic and peer review and, therefore, are held to a scientific standard.
In light of the foregoing, the authors decided to use information on BMD as the topic area. Not only was the level of discrepancy identified as appropriate for exploration, the magnitude of the implications regarding information availability is considerable. According to some guidelines aimed at reducing future fragility fractures, women will clinically qualify for bone-mineral density testing based on their age alone (1).
Numerous studies have quantified the differences between CHWS information and clinical standards, thereby establishing consistent and significant discrepancies. This qualitative analysis was undertaken to gain greater insight into discrepancies in the type of information and messages that were presented on CHWS compared with HTA reports than are possible using quantitative designs. While the goals of presentation for the authors of CHWS and HTA reports are divergent and the two media are aimed at different audiences and for different purposes, an alignment could, and indeed should, materialize in relation to scientific knowledge. It is the scientific dimension of information that this study defines as quality and seeks specifically to illuminate. Content analysis provides an appropriate method of discovering messages in texts, comparing conceptual themes, and gaining insight into complex communication patterns using a theoretical and empirical basis such as that provided by the clinical sciences (35).
METHODS
The study research objective was to evaluate the quality of information content regarding BMD testing posted on CHWS, as compared with HTA reports. HTA reports provided a scientific standard for information quality as they are undertaken to examine the available scientific evidence in support of technologies. Fifteen websites with relevant content identified by popular search engines and six HTA reports on BMD identified from the HTA Database were selected according to predefined search strategies and criteria. Relevant content in both CHWS and HTA reports were then identified and compared by two independent reviewers.
Identification of CHWS
Search engine selection
Based on the May 2001 Jupiter Media Metrix ratings, five search engines were identified as the five most frequently used, each having a market share of greater than 20 percent (with a range of 21.7 to 61.4 percent). Jupiter Media Metrix ratings were selected because the user-based rating system is one of the longest established, and its monthly ratings are widely cited as a measure of popularity. One limitation is that the ratings are based on the total amount of traffic to a specific site, and, therefore, includes nonsearch use, such as use of the portal for email.
Search strategy
By using the selected search engines, the search for websites was undertaken in September 2001, using the key words “bone scan,” “osteoporosis,” and “bone density test.” The Boolean operator “AND” was a default search parameter in all five search engines to improve the precision of the results. Identical key-word searches were used across all five search engines.
Inclusion criteria
The top fifteen websites were selected that (i) pertained to bone mineral density testing; (ii) were written in a style directed at patients/consumers, including a lack of technical language more appropriate to providers or purchasers; and (iii) provided information on BMD testing, including information on benefits or risks and related information, and on which its quality could be judged.
Exclusion criteria
Excluded were sites that provided information directed at providers, equipment purchasers, academic course material, news items, book reviews, access to database holdings, and sites without content. Evaluation of content was limited to that appearing under the initially identified domain name, excluding content that might be obtained by following links to external sites.
Selection process
Inclusion and exclusion criteria were applied by each reviewer independently to the first twelve websites identified by each of the five search engines. The order of evaluation was as follows: the first website returned by each of the five search engines in order of popularity were reviewed before going on to second website listed by each search engine and so on until fifteen websites had been identified. Differences were resolved by discussion. Content was identified and extracted from each website that could be used by a patient to answer the following questions: Should I get a bone density test? Should I get a bone density test to detect osteoporosis? What are the benefits of taking this test, and are there any risks? What are the health benefits and risks of this test? Is the information given in this website based on proven, reputable studies? Have the health benefits been scientifically proven? What is the scientific basis of this information?
Identification of HTA reports
The Health Technology Assessment Database at the U.K. York University portal provided by the National Health Service Centre for Reviews and Dissemination (http://agatha.york.ac.uk/htahp.htm) was used to identify HTA reports. Included were reports that met the following criteria: (i) they pertained to BMD testing and (ii) were published with the 5 years immediately preceding September 2001 when this analysis was conducted. Six relevant reports were located, and research evidence on indications, clinical effectiveness and safety of bone mineral density testing was identified independently by two reviewers from the report summaries where available.
Qualitative content analysis
Major messages on individual CHWS sites and in individual HTA reports were first isolated and then systematic comparisons were made within and between groups to identify overlap, omissions and discrepancies in the type of information and messages presented. The scientific findings of the HTA reports were taken as the standard for scientific content quality. The types of websites and reports identified were analyzed descriptively to provide context.
Findings
The included CHWS are described and listed in Table 1 with Uniform Resource Locators (URLs) and sponsors. The organization which produced the included HTA reports follow along with a description of potential conflict of interest. Discrepancies between the content of the two textually based media are then discussed under following themes: limited disclosure on test accuracy; disclosure of informational basis; self screening questionnaires; use of statistics, storytelling; and presentation style.
Included CHWS
Represented among the top fifteen CHWS returned by the five most popular search engines that met inclusion criteria are websites sponsored by media companies (Netdoctor; SpineUniverse.com, W3COMMERCE inc; iVillage.com: The Women's Network), public organizations (US National Osteoporosis Foundation; Osteoporosis Society of Canada), a pharmaceutical company (Merck & Co. Inc.), for-profit clinics (Muskogee Women's Clinic and Bone Densitometry Center, Missouri Osteoporosis Health and Education Center, Cedarhurst Osteoporosis Center of the Five Towns, Oregon Osteoporosis Center), governments (the British Columbia [Canada] Ministry of Health Services Protocol Steering Committee, the U.S. Department of National of Health and Human Services, National Institutes on Aging), and medical device suppliers (South East Medical Equipment Co., Distributors of Norland Medical Systems, Inc.).
Included HTA Reports
The HTA reports were produced by the following organizations: Alberta Heritage Foundation for Medical Research (AHFMR; 16), The British Columbia Office of Health Technology Assessment (BCOHTA; 13), The Catalan Agency for Health Technology Assessment (CAHTA; 8), The International Network of Agencies for Health Technology Assessment (INAHTA; 15), the Swedish Council of Health Technology Assessment (SBU; 34), and Centre for Health Economics at York University (31).
Information on the HTA reports is available publicly through websites such as the UK York University portal where the Health Technology Assessment Database can be accessed and searched (http://agatha.york.ac.uk/htahp.htm). The organizations that produced the HTA reports also have specific websites that can be accessed by the public, although they are not designed for specific consumer access nor are they among the sites with favorable placement in lists of sites retrieved by the most popular search engines.
Conflict of Interest
The organizations producing HTA reports do not have financial interests in the purchase of equipment, therapeutic products, or clinical service provision. The gains accrued to media companies contributing CHWS to this study is in general not transparent to those accessing the site. Some had advertisement banners, pop-up windows, and links to commercial sites selling products, such as calcium supplements, directly related to the topic of interest. Such promotional material is indicative of at least collateral financial interests but equally indicates that the content may need to be kept in accordance with advertisers' interests. The funding of communication companies is not sufficiently transparent to preclude content appearing as paid “infomercials.” There is no regulation that would require the disclosure of such relationships.
Among the CHWS sponsors, the single pharmaceutical company identified stands to make the greatest financial return from increased pharmaceutical sales that would be expected after increased BMD testing (11). Pharmaceutical sales occur over a wide geographical area; therefore, the market for their products could expect the highest volume. In this study, each website was included only once for analysis though different search engines may have retrieved them and they may have appeared under different domain names. Even though Merck contributed only two sites to this study—one site with an osteoporosis focus and the other with a product focus—these two sites appeared with the greatest frequency among the links returned and reviewed. Multiple strategies for gaining favorable placement with search engines had evidently been used. Three URLs led to a single website (www.bmdtesting.com/=bmdtesting.com/=www.bonedensitytesting.com). These URL reveal a preferred strategy of having the topic included as a domain name within the URL. Many consumers might be expected to enter the topic of interest for example BMD testing followed by “.com” preceded by “www” (or this may be added by the search engine). Domain names are obtained by either applying for the name first or purchasing the name from the original owner. Several other techniques that would not be obvious to the consumer accessing the sites may have been used to garner such favorable placement amongst search engine retrievals. Merck is also an acknowledged sponsor of the U.S. National Osteoporosis Foundation site.
The single distributor of the BMD testing equipment included in this study stands to make financial gain through equipment sales, although appreciably less than the pharmaceutical company might be expected to make through drug sales.
The clinics that provide BMD testing are in evident conflict of interest, as they are likely to profit directly through payments related to BMD testing and with increased follow-up monitoring. The geographical reach of for-profit clinics, however, is limited to their catchment area, with consequent limits on potential profits. In addition to earnings for clinical services, the Oregon Osteoporosis Center is also privately funded by the pharmaceutical industry to undertake drug trials.
Government sponsored CHWS are not in as clear a position of conflict of interest with regards to providing information on BMD testing as other CHWS sponsors. If BMD were effective then governments would have a moral dilemma as to how to allocate funds to BMD under general cost-control conditions in health care. A conflict of interest would arise if government were getting a benefit (monetary, political, or any other type) while acting against the public interest.
Limited Disclosure on Test Accuracy
The predominant discrepancy between the CHWS and HTA reports emerged over the characterization of the accuracy of BMD testing. The majority of CHWS identified by popular search engines do not disclose the limited capacity of BMD to discriminate between low-risk individuals who will suffer future fractures and those who will not (with one exception). Table 2 provides a sample of statements that demonstrate this discrepancy. Whereas the findings of HTA reports are consistent across reports in revealing the limitations of BMD test parameters, the findings are not reflected in the CHWS content. The CHWS are uniformly consistent within their group (with one exception, a government CHWS) in presenting to consumers a positive message about BMD with information on its limitations. If consumers access more than one CHWS using the most popular search engines, the effect of this consistency is likely to reinforce the confidence they have in the message.
Disclosure of Informational Basis
There is not enough information provided on the CHWS to determine the extent to which the messages conveyed are based on research evidence; therefore, the extent to which the advice is valid and reliable. References to scientific research are largely absent from CHWS sites. More often, indirect association was implied. The affiliation of the site with physicians associated with the clinic or with an editorial board may be interpreted by some users of the information as suggesting that the information has a scientific basis.
The predominant CHWS messages are that BMD testing is quick, painless, and noninvasive. Information on osteoporosis is prominently paired on CHWS sites with BMD testing information, with an emphasis on “silent disease” and devastating consequences of advanced disease. Information is presented in the format of a structured argument. Typically the argument in support of BMD was that it is a cheap and harmless test for a silent disease having devastating future consequences unless treated and that effective treatment is available.
In contrast, the HTA reports contain a method section making transparent the process used for reviewing scientific evidence and making statements supported by research while acknowledging the considerable uncertainty and lack of good research on health outcomes. With the exception of the government CHWS, the process by which the information is worked up is not apparent. In contrast, the review methods used by HTA organizations are written in the report for public scrutiny and peer review. The function of peer review is to ensure that scientifically valid methods are used, resulting in interpretations based on evidence. There are no such safeguards on CHWS. None of the returned sites had accompanying seals of approval indicating that the site had met even minimal scientific standards.
Self-Administered Screening Questionnaires
Testing is frequently recommended on CHWS on the basis of risk factors for fragility fractures or low bone density, and identifying risk factors that are widespread among the general population. From a scientific perspective, the evidence for the validity of these self-administered screening tools is suspect. No information on which to judge the validity of any of the tools or risk factor advice was provided. Although research has quantified the differences in the contribution-to-risk the presence of each factor signifies and the level of evidence for each, these issues or distinctions were not presented in a way that would be easily assessable by visitors to the website.
On some sites, on-line questionnaires could be filled in, submitted, and automatically graded to provide an individual risk score. This is an example of a two-step type screening process, in which the evaluation of “risk factors” is used as a screening test to support the use of a second screening tool such as BMD testing. For example, an extensive on-line survey tool was provided on the South East Medical Clinic website. It provided automatic calculation of a score and advice, by means of a submission button labeled “Do I need a bone density test?” The scoring system embedded in the software is not explicit; therefore, it is not possible to accurately evaluate the validity of the tool.
The Osteoporosis Risk Questionnaire for Women provided on-line by the Muskogee Clinic demonstrates the ease with which perimenopausal, well women can, on the basis of age, race, and non-use of estrogen, elicit advice to contact the clinic or a personal physician. The questionnaire has only six questions the answers of which may garner points. A Caucasian or Asian woman (five points) not taking estrogen (one point) would reach the threshold score of six points, which elicits the advice to consult either the clinic or a physician, unless they further qualify for point subtraction. To qualify for point subtraction, they would have to fall above weight and age guidelines, that is, 130 lbs for age 40, 160 lbs for age 50, and 190 lbs at age 60.
From a scientific perspective, body weight without a consideration of height is difficult to interpret as a risk factor; therefore, a body mass index value is a preferred measure. Rheumatoid arthritis and fractures also garner points in this survey, but these diagnoses presumably warrant medical care regardless of the other factors. In the absence of evaluative research to validate self-administered risk assessment questionnaires, the information provided is of uncertain benefit. Furthermore, this type of tool will likely motivate consumers to contact the clinic, therefore, fulfill a marketing function.
Use of Statistics
The use of statistics was not uncommon on CHWS but was presented in such a way as to motivate clients to book a BMD test, rather than provide a balanced appraisal of risk. The following example was from a for-profit medical clinic: “A woman's life time risk of hip fracture alone is equal to the combined risk of developing breast, uterine, AND ovarian cancer. A 50-year-old woman is just as likely to die from the complications of osteoporosis as from breast cancer. And osteoporotic fractures are four times more common than strokes. Osteoporosis is still undiagnosed and untreated in over fifteen million individuals” (www.muskogeewomansclinic.com). From an HTA report, we find the statistical data that put such marketing fear in perspective: for example, the Years of Potential Life Lost (PYLL) are estimated to be approximately 29 (per 1,000 persons) for stroke, 20 for from breast cancer, versus 9.2 for hip fracture (7).
Storytelling
HTA reports discount anecdote as the lowest level of scientific evidence in a hierarchy that has randomized controlled (RCT) trial at the pinnacle (9). Authors of CHWS use anecdote journalistically, with power to engage human emotions, sway opinion, and influence decisions. The following example comes from a U.S. government site. “Helen grew up on a dairy farm in the Midwest. She drank 3 glasses of milk a day as a child. After high school she began work as a secretary in a local law office where she spent her entire career. Helen never jogged, walked as exercise, or played tennis. She went through menopause at age 47. Shortly before retirement at age 61, she slipped on a small rug in her kitchen and broke her hip. After Helen recovered, she needed a cane to walk. Helen had osteoporosis, but she didn't know it” (www.nih.gov).
The multiple messages conveyed by this story are subtle but not difficult to decipher. Helen is like the majority of North American women—unexceptional in her life history. If the reader can identify with Helen, they are learning that a minor accident could lead to a permanent disability before retirement without their knowing that their bones are fragile—clearly a frightening suggestion. If a reader thought that milk consumption in childhood would protect them then they are being led to question that belief. If the reader did not undertake vigorous weight-bearing exercise then they are being led to believe that they have put themselves at risk. If, after reading the story, the consumer feels concerned that they may unknowingly have osteoporosis, they will want a bone-density test. Even readers who are not the perimenopausal women identified in this type of story are likely to recognize that many women are like Helen and be inclined to support BMD testing.
Presentation Style
The HTA reports are written in a technical report writing style that is aimed at nonexpert policy-makers—consequently they seek to make scientific knowledge reasonably accessible. Although the language is not as technical as would be found in the research journals on the topic, it is based on the same scientific tradition that aims to eliminate the subjective (personal and emotional) and to maintain objectivity (unbias) in describing scientific experiments and their interpretation.
DISCUSSION/CONCLUSIONS
Content analysis demonstrates the omission and divergence of information on BMD testing available on CHWS as compared with HTA reports, which are based on systematic review of clinical research evidence. Fourteen of the fifteen CHWS identified by popular search engines fail to disclose the limited capacity of BMD to identify individuals who will suffer future fractures accurately. It is a research standard in clinical epidemiology to evaluate tests like BMD on the basis of how well individuals are correctly or incorrectly identified as having the condition of interest (29). Pre- and posttest predictive values of correct assignment are then calculated on this basis.
Clinical leaders and osteoporosis researchers have resisted evaluating the BMD test this way and instead use a correlation between the test and outcomes of interest (often surrogate) as proof of predictive value. In so doing, they side step the issue of false-negative and false-positive test results. HTA reports have shown how such omissions undermine the validity of such evaluations when applied to clinical populations. In Canada, for example, a national charitable organization with the acknowledged sponsorship of Avantis, Lilly, Merck Frost, and P&G pharmaceuticals (all pharmaceutical manufacturers) produced BMD testing guidelines that do not deal with issues of the predictive value of the test (1) and recommend testing for everyone at age 65. The guideline authors claim the recommendations are evidence based because they provide a systematic review of the literature, cite levels of evidence and have the endorsement of a scientific advisory council made up predominantly of Canadian medical and academic leaders. These guidelines gain further credibility through publication in the national medical association journal. These national guidelines were then taken up by a provincial research organization as a gold standard against which the appropriateness of provincial utilization of BMD testing was evaluated (27). A group of provincial clinical and scientific leaders convened by the provincially funded research organization then also promoted both a recommendation to increase capacity to offer all citizens BMD testing at age 65 and to promote the national guidelines through local educational initiatives (27) even though the report concurrently acknowledged: “The effectiveness of screening to reduce the incidence of fractures has not been evaluated. Cost effectiveness studies on the use of BMD testing and osteoporosis drugs are also not available. Finally, there are no generally accepted criteria for when treatment should be started” (27).
As a consequence of the failure to adequately evaluate BMD, there is limited evidence from clinical sources that could provide consumers with useful and balanced information, in particular, the importance of knowing how this information is arrived at and what the expected benefits are. Consumers might find it highly useful to know how many people in their reference age and sex category would have a fragility fracture over a given time period and how much these outcomes could realistically be altered by the proposed testing and treatment interventions.
Given the expanded capacity of information and communication technology, the capability to collect and analyze this type of information is highly feasible. Pharmaceutical trials and database analysis are being analyzed in as short a time frame as 1 year, using fracture data. This strategy effectively eliminates any suggestion that the time frame is prohibitive as a rationale for not conducting this type of important evaluative research.
The accepted gold standard research design to evaluate tests to be applied to large populations of well individuals with or without known risk factors is the RCT. The RCT has not been used to evaluate BMD. Instead, from the National Osteoporosis Risk Assessment (NORA) database (32), the finding that there is a linear and inverse relationship between BMD and fragility fractures is being actively used to promote BMD testing. It comes from an analysis. This observational study was conducted with a database containing information on BMD, risk factors and self-reported fracture 1 year post-BMD measurement (32).
In an accompanying editorial from the pages of the Journal of the American Medical Association, this finding is being heralded as “generally supporting the validity of using [BMD at peripheral sites] in screening of large populations for osteoporotic risk” (emphasis ours; 2). Results from this study on previously identified risk factors have also been taken up by the National Dairy Council of Canada in a press release that announces “more evidence that women need to build up their bone bank” (http://www.dairycouncilofca.org/media/medi_pres39.htm).
There is no evidence from RCT directly linking BMD tests of well populations to favorable changes in fracture outcomes. Peto cautions that nonrandomized designs “cannot discriminate reliably between moderate differences and negligible differences in outcome, and the mistaken clinical conclusions that they engender could well result in the undertreatment, overtreatment, or other mistreatment of millions of future patients (25). This result clearly argues for a higher standard of evidence than is currently being applied to BMD test as an intervention proposed for screening large populations through general practitioner practices.
The importance of the RCT gold standard in research design was recently highlighted in another study of relevance to women's health. Hormone replacement therapy (HRT) is a companion technology to BMD, as it is one of the commonly prescribed therapies after a finding of low bone density through a BMD test. The Women's Health Initiative (WHI) study on HRT was stopped early, after a 26 percent increase in breast cancer was found in the HRT group—an unexpected finding from a trial that sought to discern the effects of HRT on chronic diseases such as heart disease (28). The trialists had initiated the study because “Despite decades of accumulated observational evidence, the balance of risks and benefits for hormone use in healthy postmenopausal women remains uncertain” (28). This assessment of the state of the science had been consistently reiterated in numerous HTA reports, many of which are included in this study, because of its therapeutic connection to BMD testing. The results are relevant to the assessment of BMD testing because of the study's finding that, for every 10,000 person-years treatment, the benefits of five fewer hip fractures and six fewer colorectal cancers were offset by the excess in harmful outcomes of seven more coronary heart disease events, eight more strokes, eight more pulmonary embolism, and eight more invasive breast cancers (28). The evidence of these harms previously were unavailable and would contribute to even less favorable estimates of the harm to benefit ratio after BMD testing contained in prior HTA reports.
It is noteworthy that the messages on the CHWS obtained using the most popular search engines did not acknowledge important and relevant debates on the posttest treatment implications. Generally BMD testing is presented as quick, painless, noninvasive, and widely advisable, based on risk factors widely present in the general well population. Information on HRT is offered as one of the “effective treatments” available but the expected outcomes after testing are not explicitly discussed.
BMD testing information is prominently paired on CHWS sites with information on osteoporosis, as the “silent disease” having devastating consequences. Research based statistics, if presented at all, frequently pertain to fragility fractures. Consumers are, therefore, led to focus on the harmful outcomes associated with the disease rather than the benefit of the testing and treatment. The absence of a balanced or critical perspective was found to be uniform with one exception. A government CHWS that carried more critical information was picked up in out study because one of the search engines uses the strategy of returning sites of local interest. This government website is in the same geographical location as the HTA report on the topic and this current study.
There are CHWS with messages that diverge from the ones appearing in this study. For example, information from the BC Office of Health Technology Assessment report on BMD testing (among other sources) was taken up by Jill Sandson, a New Zealand women's health educator and consumer advocate, who has presented a critical perspective in formats accessible to consumers: these include a website, a paperback book (30), and television coverage. Her main message was formulated as follows: “If you are a healthy woman of menopausal age you may be the unwitting target of a vast and successful marketing campaign. Every year, millions of women are told that they have low bone density. Hormone replacement therapy or other drugs are frequently recommended to prevent onset of the debilitating disease, osteoporosis. Women are also bombarded with information about calcium and dairy products and their ability to prevent bone loss. But how accurate is the diagnosis of ‘low bone density’ and what is the probability of low bone density progressing to fracture later in life?” (www.bonestory.com). This website is not among the top sites selected by the popular search engines. This finding raises a fundamental issue in the availability of information on the Internet—how to get placed highly in a search engine's URL returns.
Like the web pages of the HTA organizations, CHWS unwilling or unable to pay to ensure favorable placement with the most popular search engines are increasingly unlikely to appear among in the coveted first ten sites listed, and, therefore, increasingly inaccessible to the majority of women looking for health information. To ensure favorable placement with search engines, sponsors of websites must use increasingly sophisticated strategies, often with the help of paid consultants, all beyond the means of not-for-profit enterprises. Many of these strategies are technically based, such as ensuring that the meta-tag keywords used to index the site are an effective means to ensure a given relevant site is placed appropriately by the search engine. More recently though, search engine providers have started to accept payment from websites in return for priority placement—a practice called “paid placement.”
Consumer complaints about “paid placement” led the U.S. Federal Trade Commission (FTC) to issue a warning letter to seven search engine providers (including some included in the present study) “outlining the need for clear and conspicuous disclosures of paid placement,” acknowledging that, while such disclosure is attempted, it is frequently insufficiently clear (33). The FTC action was in response to a formal deceptive advertising complaint made to it by Commercial Alert, a nonprofit organization with a mission to “keep the commercial culture within its proper sphere” (www.commercialalert.org). The rationale for this complaint was that the listings “look like information from an objective database selected by an objective algorithm. But really they are paid ads in disguise” (3). This complaint was launched in the month before the search for this study and is supported by the clear conflict of interest among the CHWS demonstrated by the sample in this study. The sponsors of CHWS sites were frequently either providers of BMD test or companion drugs and, consequently, in a position of conflict of interest with regard to decisions to undergo BMD testing.
The prominence of one pharmaceutical company among the sample of CHWS in this study, is the latest in a long series of extensive corporate involvement in “changing the way populations think about bone loss” (22). Moynihan and colleagues have drawn attention to the ‘disease-mongering’ of osteoporosis, stating: “Drug companies have sponsored meetings where the disease was being defined, funded studies of therapies, and developed extensive financial ties with leading researchers. They have funded patient groups, disease foundations, and advertising campaigns (on both drugs and disease) targeted at doctors and have sponsored osteoporosis media awards offering lucrative prizes to journalists” (22).
By contrast, HTA organizations have no documented conflict of interest nor do they invoke emotional arguments. HTA reports are based on comprehensive evaluations of scientific evidence and strive to present an unbiased perspective. Their approach is to emphasize the effects of testing on populations on the basis of referenced research findings. HTA reports could potentially provide consumers with an unbiased, high-quality source of information that would help consumers make an informed health decision. The strength of HTA reports lies in the critical appraisal methods and systematic approach that is transparent and reproducible. To date, the focus of HTA has been to provide health policy decision-makers with appropriate information on a specific topic.
This present study indicates that the type of extensively reviewed scientific information contained in HTA reports is not easily accessible to women by means of the Internet—a source to which increasing numbers look to for health information. The challenges of obtaining sufficiently prominent placement in the returns of popular search engines and their increasing commercialization would appear to put this medium out of the range of not-for-profit HTA information providers and consequently beyond the reach of the average consumer of health information who lacks sophisticated search skills.
A further barrier to consumer access is the writing style and presentation of the technical, scientific, and policy oriented HTA report. Although the latter are generally more accessible than scientific journal publications, they are similarly constrained by traditions of scientific communication. This does not usually use stylistic devices to make information accessible to the nonexpert consumers. Consumer-friendly versions of HTA results could be produced and disseminated. Surmounting the barriers to access through creative partnerships, with informaticians, librarians, writers, and visual artists for example, would make the presentation of valuable HTA information more accessible to consumers.
Women increasingly look for information from sources other than their health-care professionals to make independent health-care decisions. Valuable resources are available to supply appropriate, unbiased, and accurate information. The challenge to ensure they are the preferred focus for the general public should be identified as an important priority in knowledge translation research.
