Three electronic search engines (google.de, bing.de/yahoo.de, ask.com) were used. Searches were performed on April 28 and 29, 2019, using different search strategies, as google.de offers only limited options to combine multiple search terms with Boolean operators. Search terms represent different combinations of the German words for repair restoration(s), composite(s), and dentist(s) (Multimedia Appendix 1). A computer running macOS 10.14.4 (Apple Inc) connected to the internet in Germany was used. Cookies and browser history of Firefox Quantum 66.0.3 (Mozilla Foundation) were cleared and the default setting of each search engine was used.

In total, 2864 webpages were displayed as “most relevant” sites (google.de: 1299, bing.de/yahoo.de: 1295, ask.com: 270), and the search was not expanded beyond this number of displayed webpages assuming saturation. Also, advertisements (ie, websites from page owners paying a fee to have their website prominently displayed) were not additionally assessed.

For the purpose of this study, only websites from dentists were included. Patients are likely to be looking at these sources only while searching for a person able to deliver appropriate dental care. Therefore, websites from or associated with dental laboratories or supply and materials companies, forums and blogs operated by nondentists, dental regulatory bodies, dental schools and clinics, research agencies, or otherwise public bodies were excluded. Notably, however, patients may well find these informative, too.

The remaining 820 webpages were screened in full text. Webpages containing irrelevant information were excluded, leaving 74 webpages that were potentially eligible. Finally, after removal of duplicates, 50 websites fulfilled the inclusion criteria: (1) page freely accessible, (2) German language, (3) posted by a dental practice or practice cooperation, (4) mentions repairs. Websites containing multiple eligible webpages (ie, published under the same domain or published from the same practice) were jointly assessed as one website. The full search workflow is shown in Multimedia Appendix 2.

The following parameters were collected from the websites, if available: (1) practice name, (2) URL, (3) country, (4) practice location (rural, town [<100,000 inhabitants], or city [≥100,000 inhabitants]), (5) practice setting (single practitioner, multiple dentists, or practice cooperation), (6) dentist’s gender (female, male, mixed [in case of multiple dentists or practice cooperation]), (7) dental society memberships, (8) year of examination or approbation, and (9) information regarding repairs (Multimedia Appendix 3). Information regarding dental society memberships and year of examination or approbation were cross-referenced from dental societies’ member information pages (ie, German Society of Dentistry and Oral Medicine [DGZMK], Swiss Dental Association [SSO]) or curriculum vitae published elsewhere (ie, in dentists’ dissertations and public profiles at the social networking sites XING or LinkedIn), if information was not already listed on dentist websites. In case of multiple dentists or practice cooperations, the average years of examination or approbation was used.

Website quality regarding information on repairs was systematically assessed across three different domains: (1) technical and functional aspects (Table 1), (2) comprehensiveness of information (Table 2), and (3) generic quality and risk of bias (Table 3). Assessment was independently performed by two authors (PK, AFB). Discrepancies were resolved through discussion.

The established and validated LIDA instrument (version 1.2) [18] was used to assess items in domain 1 and DISCERN instrument [19] was used in domain 3. In dentistry, such tools have been successfully applied to evaluate the quality of information on websites regarding dental caries [20,21], periodontitis [22], root canal treatment versus implant placement [23], and orthodontics [24-29]. For this study, both LIDA and DISCERN have been slightly modified to uniformly score all domains on an ordinal scale as 0 (never or no), 1 (sometimes or partially), and 2 (mostly, always, or yes).

To assess items in domain 2, a structured checklist with 6 subdomains was developed by the authors focusing on the evidence (2.1); advantages (2.2) and disadvantages (2.3) of repair restorations; restorations (2.4) and defects (2.5) suitable for repairs; and technical implementation of repairs (2.6). The same 3-point ordinal scale was used. As the number of items within each domain differed, an overall quality score was calculated by averaging the quality scores (relative percentages) of all three domains, assuming them to be of equivalent importance. This score was used as primary outcome parameter. Quality scores (relative percentages) of all three domains were also assessed individually and used as secondary outcomes.

For each domain, a quality score (relative percentage: website score on all of the respective items divided by the maximum possible score sum) was calculated. Furthermore, an averaged overall quality score based on all three domains was calculated. As data were not normally distributed according to the Shapiro-Wilk test, descriptive statistical analysis contained median, quartiles, and ranges.

Differences in website scores between the three domains were analyzed using Wilcoxon signed-rank tests with a Bonferroni-Holm correction. Generalized linear modeling was used to assess the impact of practice-specific parameters on domain-related quality and the averaged overall quality score: (1) practice location (rural, town, or city); (2) practice setting (single practitioner, multiple dentists, or practice cooperation); (3) dental society membership (yes or no); and (4) year of examination or approbation. A multivariable analysis was performed and covariates entered simultaneously. Only main effects without interaction terms were tested. If no information regarding dental society membership was available, we scored this as no. If year of examination or approbation was not available, websites were treated as randomly missing and excluded from the regression analysis (n=4). Statistical analysis was performed using SPSS Statistics for Macintosh version 26.0.0.0 (IBM Corp). Statistical significance was set at P<.05.

In the D-A-CH countries (Germany, Austria, Switzerland), about 90% of the population has access to the internet [31]. Information on health-related aspects is increasingly assessed online, often using search engines [14,15]. Due to the broad access to the internet, operating a website has become the standard for most companies and businesses including dentists.

Regarding dental health, information on dental practice websites is of special interest as patients are likely to access those websites while searching for information and an appropriate dentist. Therefore, our study focused on dentist websites only. Websites were identified using different search engines with a combined market share of more than 99% in Germany [32]. The search was performed using consumer search engines only as patients are unlikely to use scientific databases (eg, Medline).

We found that only a small number of dentists included information about repairs on their websites. Dentist websites showed sufficient quality regarding technical and functional aspects but were not seen as fully trustworthy (generic quality was low, and there was a high risk of bias present). Comprehensiveness of repair-specific information was also rated low. This is in line with previous studies assessing the quality of websites regarding different dental health-related information [20,22,23,25,26,28,29]. Dental health-related information was not comprehensive and of lower quality than websites' technical and functional aspects.

A number of reasons for these findings are conceivable. First, dentists might not have enough time to create and maintain a content-comprehensive website. Dentists might also feel that informing patients online is not necessary or that it is not their task to supply patients with comprehensive health care information on their websites. The perceived lack of financial gain from providing such content online may add to this. Also, provided online content might need to be discussed with patients at the next appointment, which may be seen as a waste of time. Data from the United States demonstrated that physicians perceive appointments as less efficient and more difficult if patients have already gained information online [33]. In contrast, insufficient knowledge regarding repairs among dentists is unlikely to be a reason, as repairs are frequently taught at dental schools in Germany and all over the world [1,2]. However, dentists might regard implantology or orthodontic information to be of more importance than information on repairs, resulting in higher quality scores concerning technical aspects and generic quality and risk of bias of these websites (also measured using LIDA and DISCERN) [23,27,29].

We did not find any significant association between practice-specific parameters and website quality scores. We therefore must reject our hypothesis. This is a surprising outcome, as a range of parameters including those related to the individual practitioner and their practice seem to impact on repair behavior [1]. For example, low dentist density (ie, in a rural area), more experience or knowledge (ie, being a member of the dental associations), fewer years since dental school graduation, and working in larger group practices (ie, with multiple dentists) have been found to facilitate repairs. Also, we assumed website quality would be higher in younger dentists being more comfortable with technology and in larger group practices (with higher budgets for an online presence and marketing). Notably, a previous study on dentist websites and their quality also failed to demonstrate significant associations with most of such practice-specific parameters [22]. We mainly ascribe this to the fact that the overall quality was too poor throughout different websites, and dentists generally do not seem to prioritize providing information on repairs on their website regardless of their background or practice environment.

This study has a number of limitations. First, the relatively small number of included websites (n=50) must be noted. Notably, the sample size was not based on a formal sample size estimation but guided by a previous study [22] and the availability of websites. Our study might have been underpowered, and the lack of significant associations should hence be interpreted with caution. Second, we focused on German-speaking websites only. It is possible, albeit unlikely, that websites in other languages present a higher quality (eg, with regard to periodontitis, both German- and English-speaking websites showed a low quality of information) [22,34]. Last, we used established and validated criteria to assess technical aspects and generic quality and risk of bias but developed an assessment checklist for the repair-specific quality and comprehensiveness on our own. The validity of this checklist was not formally tested, and using another checklist may lead to different results.

To overcome the shortcomings of dentist websites, a number of interventions are conceivable. Regulatory and legislative bodies might enforce better information standards. Professional dental bodies might assist dentists by providing high-quality information suitable for adoption on dentists’ websites. Alternatively, dentists could provide links to other validated websites or organizations able to provide comprehensive information, such as dental research societies, thereby reducing the burden for the individual dentist to provide and maintain high-quality information. We did not check society websites as it can be assumed that information presented is both trustworthy and comprehensive.

In conclusion, only a minority of dentist websites informed patients about repair restorations. Despite sufficient technical and functional quality, the websites that did mention repairs were not comprehensive and prone to a high risk of bias. Dentists are encouraged to provide better and more trustworthy health information, including but not limited to repairs. Professional or regulatory bodies might assist dentists by providing high-quality information suitable for adoption on dentist websites. In the meantime, patients must be aware of the limitations and should seek information regarding repairs elsewhere.