The Internet has become increasingly popular among consumers as a source of health care information. A poll in 2001 showed that almost 100 million American adults regularly go online for health care information [1].

Although there are quality-standards guidelines for medical publishing on the Internet [2,3] there is currently no governing body acting as a gatekeeper of Web page publications. Anybody can create a Web site and they are free to write whatever they wish. This has led to some serious concern as to the accuracy of the information of some of these sites [4,5]. Even those sites that offer to evaluate health Web sites are often incomplete or fail to reveal how they perform the evaluation [6,7].

LASIK (Laser in Situ Keratomileusis) is a very popular [8] combined surgical and laser procedure, which is used to correct myopia (shortsightedness) and hyperopia (farsightedness). As the procedure is not available within the United Kingdom's National Health System, but is available privately, many commercial companies have taken a keen interest in promoting its uptake. Although LASIK is predominantly a safe procedure there are many potential complications [9- 16]. Most of these are mild and/or transient but some are severe, permanent, and may require corneal grafts to correct [10,11]. LASIK was first performed only 11 years ago and United States Food and Drug Administration approval was only granted 6 years ago; therefore there is concern as to what may happen to these corneas in the future.

Based on a Medline search there have been no previous studies that have specifically examined the quality of ophthalmic Web sites. This study examines information a member of the public might read when searching for information on LASIK on the Internet. The sites are evaluated regarding the quality of the content, authorship, and technical aspects of the information on the complications of LASIK. that have specifically examined the quality of ophthalmic Web sites. This study examines information a member of the public might read when searching for information on LASIK on the Internet. The sites are evaluated regarding the quality of the content, authorship, and technical aspects of the information on the complications of LASIK.

Between July 16 and July 22, 2002, 21 Web sites that described LASIK were evaluated. Included in the evaluation were the first 17 English-language sites that appeared when LASIK was searched using the Google search engine [17] and the first 4 UK (United Kingdom) sites selected using the same keyword but using the Yahoo [18] search engine and limiting the selection to UK sites. Although the Google search returned 150000 hits and the Yahoo search returned 1010 hits, the evaluation was restricted to the first 21 English-language sites that appeared, as it was felt that most English-speaking potential LASIK customers would not extend their search beyond this number of sites. There was no consumer involvement when devising the search strategy. Of the Web sites included in the study, 16 were North American, 4 were United Kingdom, and 1 was Indian. Web sites that just provided information on addresses of LASIK surgeons/surgeries, were not included.

Each site was assessed by one rater (DF) who was blinded to the source. Objective measurement was performed of the following: source. Objective measurement was performed of the following:

Most of the information on complications was assessed and marked subjectively based on authorship, content, and technical quality. The evaluation form used (Table 1) for the Web sites visited was created based on appraisal criteria from a number of sources [2,4,5,19- 21].

General information about the Web sites visited is summarized in Figure 1 and Figure 2. 16 of the 21 sites visited were from the United States and 13 of the sites were devoted to LASIK only.

Figure 3 shows that 16 of the 21 the sites mentioned the complications of LASIK and this was more common in the American sites compared with UK sites (75% vs 50%), although only 4 of the sites visited were UK. This is not a statistically significant difference ( P= .285, 95% CI). Figure 4 illustrates that 12 out of the 16 US (United States) sites visited were commercial compared with all 4 of the UK sites. This is not a statistically significant difference ( P= .264, 95% CI). Twelve out of the 16 commercial sites compared with 4 out of the 5 individual/government sites gave information on complications of LASIK, which is also not a statistically significant difference ( P= .819, 95% CI).

Table 2 shows that out of the 16 sites that had details on complications, 2 showed when they were last updated, authorship of page could only be clearly identified in 3, and referenced information was only found in 6. Of the 16 sites visited only 2 used links to enhance the quality of their complication information and only 6 waivered any responsibility for the information given.

Figure 3 showed that 5 of the 21 Web sites visited had no information on the possible complications of LASIK. In those sites that have information on complications, Table 2 and Table 3 illustrate that it is often inadequate because of:

Table 3 shows that of the 2 Web sites that provided information on the complications the references were generally accurate (score 8 out of 10) and were mostly were up-to-date (score 6 out of 10). One of these Web sites scored 2 out of 10 for balanced information because there was only 1 quoted reference for all its information; this was from a commercial site.

Although this study presents a mostly-negative picture of the complication information on LASIK Web sites there were some positive findings. Table 3 shows that most sites scored well on the quality of headings, subheadings and footers.

This study assessed the quality of information on the Internet on LASIK complications and found it to be poor. Our findings support recent work in the College of Optometry in Southern California [22], which looked at 96 Web sites containing LASIK information. They rated 26% of sites as "markedly informative," 28% were rated "moderately informative," and 46% were rated as "minimally informative." The poor quality of the information represents a negligent omission as the public are being misled into believing that LASIK is without risk. This may lead to liability cases by patients with complications whose decision to have LASIK was based on the information they read on the Web site. Of the sites that gave information on the complications of LASIK, only 38% (Table 2) included a waiver of responsibility for the information. There is an argument that the authors of the Web site could be held liable unless they have a waiver. On the other hand perhaps health Web sites should be prevented from using a waiver to ensure greater accuracy of the information provided.

Detailed, well-referenced, up-to-date, good quality information from recognized authorities should be found on all medical Web sites. This is particularly important for procedures such as LASIK that are often primarily cosmetic but whose complications can be devastating. The risk of irreparable sight-threatening complications such as corneal ectasia [11] particularly needs to be explained. The combination of graphics, video, and hyperlinks should be fully utilized to explain difficult concepts and allow patients to make fully-informed decisions regarding their medical treatment. Comprehensive consent forms available on the Internet should be mandatory for elective procedures such as LASIK. This was only seen in 1 of the Web sites analyzed [23].

In a systematic review of studies that assess the quality of health information for consumers on the World Wide Web, Eysenbach et al [24] conclude that "the individual's risk of encountering an inadequate site on the Web is a function of both the proportion inadequate information on the Web (P) and the inability (I) of the individual (or his tools) to filter the inadequate sites." Because of the large public interest in LASIK it is the opinion of the authors that all LASIK Web sites should adhere to strict standards regarding what is published. The public is more likely to trust information from a government/educational site, but all the UK sites located in our study were commercial. It suggests that the UK government and/or ophthalmology institutions need to produce a good quality Web site on LASIK for the public to use.

Deciding on the most appropriate method used to control/guide health care information on the Web is difficult. Current methods include filtering tools, quality labels, codes of conduct, and user-guidance labels [25]. The National Electronic Library for Health uses "methodical, organized human reviewing, selection and filtering based on well-defined quality appraisal criteria" for their site [26]. There is a subjective bias in this system and one might argue that filtering systems violate human rights by acting as a censor of health information. The eHealth Code of Ethics of the Internet Health Coalition [27] is an example of a code of conduct. The coalition is an organization that has developed a set of quality criteria for those wishing to produce health-related Web sites. If the public knows that a Web site adheres to a certain standard when it was produced then one could avoid misleading Web sites that adopt misleading tactics such as picking up the client machine's current date and displaying it after the "last updated" remark, thus giving a false impression of the currency of the information [26]. Net Scoring [28] is an example of a user guidance system that enables users to check if a site and its contents comply with certain standards by accessing a series of questions from a displayed logo. This requires a considerable amount of effort on the part of the user. It is the opinion of the authors that the best solution is to have accredited third-party approval "stamps" that would be clearly visible on the Web sites. For example, the LASIK Web sites might have a comment after the description of the complications that says "This information has been reviewed and approved by the Royal College of Ophthalmologists, UK." Selection of a correct approach is controversial and some authors such as Delamothe [29] argue that "rating the quality of medical Web sites may be impossible " and that "one option is to rate the process by which the content was produced rather than the content itself - a medical journal's Web site containing peer-reviewed material would rate higher than a commercial site selling miracle cures for cancer."

The results of this study are somewhat limited in that part of the evaluation is subjective; however, this is balanced with some important objective assessments (such as the number of sites which mention complications). the evaluation is subjective; however, this is balanced with some important objective assessments (such as the number of sites which mention complications).

Many LASIK Web sites are giving minimal, dated, poor-quality, and inaccurate information on its complications and this can mislead the public. In many cases there is no authorship of the information and in those cases where authorship is clear the information may not come from a recognized authority. Authors should publish information that is easy to understand and locate. The explanations of the complications should be supplemented by the excellent multimedia capacity of Web sites with many images and links. One solution is for authorities such as the Royal College of Ophthalmologists to evaluate UK ophthalmic sites and offer to provide a stamp of approval for those that fit their good-quality criteria. We conclude that the quality of information on the Internet on the complications of LASIK is poor and needs to be addressed. As Wyatt [4] says, "unless we evaluate the quality of clinical sites and their effects on users, we risk drowning in a sea of poor quality information."