This study attempts to determine the representation of ophthalmologists and optometrists when a Google search for “eye doctor near me” is made from each county in the US.
Key Points
Question
Are ophthalmologists well represented in a Google search of the phrase “eye doctor near me,” compared with their proportion in each county in the US?
Findings
A Google application programming interface was used to search the phrase “eye doctor near me” from the centroid of every county in the US. The overall mean proportion of ophthalmologists represented by the Google search (28.91%) was less than their real proportion (37.58%).
Meaning
Google search of the phrase “eye doctor near me” may underrepresent ophthalmologists, potentially undervaluing the role ophthalmologists play as eye care professionals.
Abstract
Importance
In order to continue to clarify and maintain their role as eye physicians and surgeons, ophthalmologists may want to understand how they are viewed in the public eye and online.
Objective
To determine the representation of ophthalmologists (OMD) and optometrists (ODs) when a Google search for “eye doctor near me” is made from each county in the US.
Design, Setting, and Participants
This population-based cross-sectional study used publicly available data on OMDs and ODs and a Google search application programming interface (API) to search the phrase “eye doctor near me” from the geographic coordinates of each county centroid in the US (searched June 30, 2021). The top 10 sites and 3 Google map links, excluding physician ratings sites, were recorded. Data from the US Centers for Medicare and Medicaid Services were used to estimate the real number of OMDs and ODs per county.
Main Outcome and Measures
The primary outcome was the mean proportion of OMDs listed by Google search as compared with the real proportion of OMDs for the US overall and for each state and county.
Results
A total of 2955 counties from 52 states and territories were included. The overall mean proportion of OMDs (OMDs with ODs) from the Google search of all counties was 4726.97 of 16 345.93 (28.91%), which was also less than the real proportion of ODs (15 778 of 41 975 [37.58%], a difference of 8.67%; 95% CI, 37.13-38.05%; P < .001). OMDs were underrepresented by Google in 35 of 52 states and territories (67.3%).
Conclusions and Relevance
In most counties in the US, Google search of the phrase “eye doctor near me” may underrepresent ophthalmologists. Ophthalmologists may want to pursue search engine optimization to try to achieve balanced representation online.
Introduction
As the US population ages and visual impairment continues to grow, reliable access to eye care is evermore imperative.1 Access to eye care is influenced by both patient predisposing factors like age, race, socioeconomic status, and geography and by the availability of ophthalmologists and optometrists.2 In light of the growing demand for equitable access to eye care, US health care policy makers have had to decide how eye care is distributed, and more importantly, by whom. The language with which eye care is described in the US is confusing and often counterproductive. When the phrase “what is an eye doctor” is searched online, an enormous number of search results are generated. One of the first results from WebMD defines the qualifications and capabilities of both optometrists and ophthalmologists, but later suggests “for primary eye care, you may wish to start with an optometrist. From there, they may refer you to an ophthalmologist if needed.”3 In our opinion, the public’s perception of the term eye doctor shapes the demand for ophthalmologic services. It is critical that we first understand how ophthalmologists are represented across the country to define our value to the eye care system.
Our study aims to characterize how the term eye doctor represents ophthalmologists across the US. We used a Google search engine application programming interface (API) to search the phrase “eye doctor near me” from the geographic longitude and latitude of every county centroid in the US. We found the proportion of ophthalmologists (OMDs) as represented by the top 10 sites and top 3 Google Map links from the Google search. We then compared this proportion with the real proportion of OMDs per county, as estimated from publicly available data from US Centers for Medicare and Medicaid Services (CMS).
Methods
The current cross-sectional, retrospective study was designated a nonhuman participants research study by the Wills Eye Hospital institutional review board. The research adhered to the tenets of the Declaration of Helsinki. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline were followed.
Retrieving Google Search Results
A script was written in Python, version 3.9 (Python Software Foundation) that used SerpAPI4 to retrieve Google search results for the query “eye doctor near me,” for all 3222 county and county equivalents in US (searched on June 30, 2021, at 4:45 pm Eastern Standard Time). SerpAPI is an API that can scrape search result parameters including link, description, and if applicable, rating. SerpAPI was also used to spoof location by inputting the longitude and latitude of the centroid of each county. Search results were split into Google Maps links and website results. All webpages associated with the top 3 Google Maps links and the top 10 website results were collected. The top 3 Google Maps and top 10 websites were chosen to standardize a number across all counties and to roughly represent the typical layout of the first page of a Google search result. The total number of OMDs and optometrists (ODs) associated with each website or Google Maps link were manually determined. Websites which collate a large number of practices, such as Yelp, Whitepages, or Healthgrades, were excluded. Paid promotions (advertisements), usually at the top of Google’s search results, were also excluded. Google search results with invalid or dead end links were excluded.
Estimated Real Proportion of OMDs and Sociodemographic Information by County
Publicly available data from CMS’s Medicare Physician and Other Practitioners—by Provider and Service were used to estimate the real proportion of OMDs and ODs per county.5 County-level sociodemographic data were concatenated from the 2014 to 2018 US Census Bureau's American Community Survey. Variables included were percentage male, age (categorized into younger than 18 years, 18 to 64 years, and 65 years of age or older), race and ethnicity (White, Black/African American, Hispanic, American Indian/Alaskan Native, Asian, Native Hawaiian/Pacific Islander, or other [this category is listed by the US Census Bureau, which people select if they feel if they do not fit into a given race or ethnicity]), education level (less than high school, high school completion, some college, associates degree, or bachelor’s degree, or higher), and income at or less than the federal poverty level (FPL). We also documented Rural-Urban Continuum Codes (RUCC) for each county, as designated by the US Department of Agriculture.6 RUCC codes (on a scale of 1 to 9) are based on population density and proximity to a metropolitan area. RUCC code 1 indicates counties in metro areas of 1 million population or more. RUCC code 9 indicates completely rural or less than 2500 urban population and not adjacent to a metropolitan area. RUCC codes were binarized to 1 to 3 representing urban and 4 to 9 representing rural.
County-level visual impairment data were collected from the US Centers for Disease Control and Prevention Vision and Eye Health Surveillance System. According to the Vision and Eye Health Surveillance System, visual impairment was defined as best-corrected visual acuity loss of 20/40 or less and blindness was defined as best-corrected visual acuity loss of 20/200 or less.7
Primary Outcomes
The primary outcome was the Google proportion of OMDs. Since the final number of Google links (both Google Maps and search results links) for each county was not equal (after excluding collated links, ads, etc), the mean estimate of OMDs per county was found (number of OMDs found by Google search/number of total valid links for each county). Google proportion of OMDs was then defined as mean OMD with mean OD per county. We compared the Google proportion of OMD with the real proportion of OMD (total number of OMDs [OMDs with ODs] per county) in both relative (mean Google proportion of OMDs with real proportion of OMDs) and absolute (Google proportion of OMDs with real proportion of OMDs) terms.
Secondary Outcomes
Google also gives user ratings (of 5 stars) and reviews for each Google map link. In counties with at least one Google rating for both an OMD and OD, we compared the mean Google ratings and reviews for counties where mean number of Google OMDs was more than mean number of Google ODs vs when mean number of Google OMDs was less than mean number of Google ODs.
Statistical Analysis
Descriptive statistics were performed for categorical variables reporting proportion and for continuous variables reporting mean with standard deviation. The normal distribution of data was assessed using the Shapiro-Wilk test. For normally distributed data, continuous variables were analyzed with an independent 2-sample t test, and for nonparametric data, continuous variables were analyzed with Mann-Whitney U test. The χ2 test was used to compare proportions between the study groups.
Univariate and multivariate generalized linear models were performed to examine the association between variables of interest (age between 18 to 64 years, percentage male, white race and ethnicity, education level less than high school, RUCC code, percent visual impairment, percent below FPL) and underrepresentation of OMDs (mean Google proportion of OMDs with real proportion of OMDs more than 0). All covariates with univariable significance of P less than .10 were added to the multivariable model. Statistics and modeling were done using SPSS Statistics, version 25.0 (IBM). All P values were 2-sided but there were no adjustments for multiple analyses.
Results
A total of 2955 of 3222 counties (91.7%) from 50 states, Puerto Rico, and the District of Columbia were included in the final analysis. The national Google proportion of OMDs was 4726.97 of 16 345.93 (28.91%), which was less than the mean real proportion of OMDs (15 778 of 41 975 [37.58%]; 95% CI, 37.13%-38.05%; χ2; P < .001) (Table).
Table. OMD Relative and Absolute Representation in Google Search of the Phrase “Eye Doctor Near Me”a.
| Characteristic | No./total No. |
|---|---|
| National level | |
| Google OMD proportion, mean (%) | 4726.97/16 345.93 (28.91) |
| Real OMD proportion | 15 778/41 975 (37.58) |
| State-level relative OMD representationb | 33/52 (67.30) of states, OMDs are underrepresented |
| County-level absolute OMD representationc | 773/1124 (68.7) of counties, OMDs are underrepresented |
Abbreviation: OMD, ophthalmologists.
2955 of 3222 Counties used in analysis.
50 States, Puerto Rico, and District of Columbia.
Defined as real proportion of OMDs less than Google proportion of OMDs in counties with at least 1 real OMD. A total of 1124 of 3222 counties were included.
To demonstrate the degree of overrepresentation or underrepresentation of OMDs, the Google proportion of OMDs were divided by the real proportion of OMDs per county and then averaged over the state. The map shows the level of OMD representation in all counties. In summary, OMDs were disproportionately underrepresented in 33 of 52 of states and territories (67.30%) (Figure).
Figure. Ophthalmologist (OMD) Representation on Google Searches of the Phrase “Eye Doctor Near Me”.
Purple tones represent states where OMDs are relatively overrepresented on Google searches, white is proportional representation, and red tones represent states where OMDs are relatively underrepresented.
We performed a subanalysis to evaluate the sociodemographic variables associated with OMD absolute underrepresentation in counties with at least 1 real OMD. Absolute OMD underrepresentation was defined as real proportion of OMDs more than Google proportion of OMDs. A total of 1124 counties were included, of which 773 had OMD underrepresentation (68.7%). In univariate analysis, Latino ethnicity (odds ratio [OR], 1.19; 95% CI, 1.11-1.23; χ2; P = .001) and being below poverty (OR, 1.13; 95% CI, 1.07-1.18; t test; P = .01) were associated with of absolute OMD underrepresentation. However, in multivariate analysis there were no associations with OMD underrepresentation.
Google also gives user ratings (of 5 stars) and reviews for each Google map link. A total of 1105 counties had at least 1 Google rating for both an OMD and an OD. The overall mean (SD) rating for all counties was 4.43 (0.52), and the mean (SD) number of reviews was 72.88 (99.1). In 846 of 1105 counties (76.6%), there was a higher mean number of Google ODs than Google OMDs. Despite the overall underrepresentation of OMDs on Google in these counties, there was no difference in mean (SD) ratings in counties with more Google OMDs (4.39 [0.55]) vs counties with more Google ODs (4.45 [0.51]) (P = .11). Additionally, there was a trend toward fewer mean (SD) number of reviews in counties with more Google OMDs compared with ODs (67.08 [92.30] vs 74.65 [101.10]; mean difference, −7.57; 95% CI, −6.22 to 21.36; P = .09).
Discussion
These findings suggest, across the US, ophthalmologists are underrepresented as eye physicians by Google search. This underrepresentation appears to exist in most states and territories (67.3%). In an era when clear definitions of clinical and surgical responsibilities are key to protecting the field of ophthalmology and patients alike, one may want to understand how patient vernacular is associated with how eye care is sought. Understanding where we are least well represented, especially in the states and counties highlighted in the Figure, is crucial for ophthalmologists’ market viability and continued growth.
Being featured prominently in a Google search may connote greater authority or popularity, potentially resulting in a greater volume of traffic to a website.8 Google is the search engine used for roughly 89% of all internet searches.9 Of these searches, over 60% of all clicks go to the top 3 websites listed.10 A study from search engine optimization (SEO) experts revealed that the first nonadvertisement website listed on a Google search generates a click-through rate (the number of clicks that a website link receives, divided by the number of times a website link is shown) of 39.6%, which is twice as much as the second-ranked website.11 Google search ranking is clearly an important indicator of society’s value of a website or organization.
SEO, or the techniques used to improve rankings in search results, can be the starting point to change the dialogue surrounding who is an eye physician.12 Great content alone is no longer the determining factor for being featured in internet searches. The proper SEO keywords need to be embedded within the content. Google recognizes authorities in certain arenas based on not only content, but clicks and also keyword searches.13 The American Academy of Ophthalmology (AAO) recommends every ophthalmologist evaluates their website and practices SEO. AAO recommends choosing 3 popular and simple key phrases per page, making webpages simple and well organized, and cross-linking (getting other websites to link to yours and vice versa).12 Developing a network of ophthalmology websites which are optimized, linked to each other, and linked to AAO, is a relatively simple way to rebrand eye physicians.
In univariate analysis, Latino ethnicity and having a greater percentage of the population falling below FPL was associated with OMD underrepresentation. It may be that in counties with greater populations of non-English speakers, patients seek optometrists to a greater extent. Alternatively, optometry practices may market themselves better to non-English speakers. Moreover, in lower-income counties, where patients may be underinsured or uninsured, people may seek optometry practices as a first line due to perceived reduced cost. Interestingly, in multivariate analysis there was no association of sociodemographic variables with OMD underrepresentation. This may indicate that there are other factors like practice size, practice advertisement, patient health care literacy, and local scope of practice laws which have greater affect on public vernacular and search traffic.
Despite the underrepresentation of OMDs in most counties, we found that there was no difference in average Google Maps customer ratings or numbers of reviews in counties where there were OMD predominant search results compared with counties with OD predominant search results. This suggests that OMDs may elicit good patient reviews despite underrepresentation. While customer ratings and reviews may not be controllable, it is possible to continue to augment the number of ratings and reviews that ophthalmologists receive on Google Maps. By encouraging patients to post positive reviews on Google Maps, practices can optimize their Google Maps ranking (Google favors websites with a greater number of positive reviews). Ultimately, being highly ranked and reviewed on Google Maps is important because the top 3 Google Maps listings also display in regular Google Search, greatly increasing visibility of a practice.14 Given that Google Maps is the most common navigation application used, having a higher ranking facilitates patient access.15
Strengths and Limitations
Other studies have used Google searches to evaluate website content and quality for various patient-facing materials on ophthalmologic disease.16,17,18 However, ours is the first, to our knowledge, to use Google search to understand ophthalmology’s purview in American culture. Our study also has limitations. The CMS Provider database excludes practitioners who do not accept Medicare or who only see children or young adults. The CMS data may, therefore, underestimate the real proportion of both OMDs and ODs who see only children or younger patients, or those who do not take Medicare. The data do not take into account that there may be differences among practitioners who are associated with academic centers, some of which may focus less on advertising as safety net hospitals. As ophthalmologists may be more likely to be at an academic center, it is possible that ophthalmologists are thereby less represented on Google. Furthermore, the database relies on the geographic location of the primary office and does not include other offices; this may underestimate the number of OMDs and ODs in counties where practitioners have a larger number of offices. This may differentially affect some rural counties. To reduce the manual burden of classifying the number of ODs and OMDs per search link, we excluded collated results (results from websites like Yelp, etc). This may lead to some amount of misclassification bias if ODs and OMDs are differentially represented in collated links, compared with a Google search. Furthermore, while we excluded paid promotions to focus on representation through SEO, advertisements are becoming increasingly important. Thus, even when a practice maximizes SEO, the search result may still rank quite low as ads are filling a greater amount of space at the top of Google searches. Thus, it may behoove future studies to analyze the proportion of OMDs and ODs who are paying for Google Ads. We used county geographic centroid as a surrogate location for the whole county. The centroid may not granularly represent how eye physicians are represented at a more local level, especially since the Google algorithm bases search results on specific location. Additionally, searching the phrase “eye doctor near me” may include results of OMDs and ODs from other counties. We chose to include all of the search results, even those from different counties, to be most representative of a search result. Ideally, we would have calculated the real proportion of CMS ODs to OMDs for the top 13 nearest eye care locations to each county centroid. However, given the limitations of the CMS data, such that only the primary office location is given, this network analysis was not possible. Thus, comparing Google “eye doctors near me” searches with CMS county doctors was the best surrogate for comparing online presence to reality. We accepted the misclassification bias that may be inherent to this choice.
Conclusions
Our study suggests that ophthalmologists are underrepresented on Google searches across the country. To maintain and expand our share of the eye care market, ophthalmologists can consider SEO marketing techniques. We recommend optimizing searchability for each office location. We suggest creating programs to survey patient satisfaction and to monitor and respond to this online feedback. Hiring an in-house marketing consultant can be helpful. If this is not possible, SEO may be outsourced to a marketing firm. Ophthalmologists should continue to educate the public on their role as eye physicians and surgeons. Improved searchability appears to be critical to provide a clear distinction between the various categories of eye physicians.
References
- 1.Varma R, Vajaranant TS, Burkemper B, et al. Visual impairment and blindness in adults in the United States: demographic and geographic variations from 2015 to 2050. JAMA Ophthalmol. 2016;134(7):802-809. doi: 10.1001/jamaophthalmol.2016.1284 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Zhang X, Andersen R, Saaddine JB, Beckles GL, Duenas MR, Lee PP. Measuring access to eye care: a public health perspective. Ophthalmic Epidemiol. 2008;15(6):418-425. doi: 10.1080/09286580802399102 [DOI] [PubMed] [Google Scholar]
- 3.WebMD . Eye doctors: optometrists and ophthalmologists. Accessed September 19, 2022. https://www.webmd.com/eye-health/eye-doctors-optometrists-ophthalmologists
- 4.SerpApi . Google search API. Accessed on September 19, 2022. https://serpapi.com/
- 5.US Centers for Medicare and Medicaid Services . Medicare physician & other practitioners—by provider and service. Accessed September 19, 2022. https://data.cms.gov/provider-summary-by-type-of-service/medicare-physician-other-practitioners/medicare-physician-other-practitioners-by-provider-and-service/data/2019
- 6.US Department of Agriculture . Rural-urban continuum codes. Accessed September 19, 2022. https://www.ers.usda.gov/data-products/rural-urban-continuum-codes.aspx
- 7.US Centers for Disease Control . Prevalence estimates. Accessed September 19, 2022. https://www.cdc.gov/visionhealth/vehss/estimates/index.html
- 8.Search Engine Watch . Is it important for SEO to rank first in 2018? Accessed September 19, 2022. https://www.searchenginewatch.com/2018/08/17/is-it-important-for-seo-to-rank-first-in-2018/
- 9.Statista . Worldwide desktop market share of leading search engines from January 2010 to July 2022. Accessed September 19, 2022. https://www.statista.com/statistics/216573/worldwide-market-share-of-search-engines/
- 10.HubSpot . Ultimate list of digital marketing stats. Accessed September 19, 2022. https://www.hubspot.com/marketing-statistics
- 11.Smart Insights . 2022 comparison of Google organic clickthrough rates (SEO CTR) by ranking position. Accessed September 19, 2022.https://www.smartinsights.com/search-engine-optimisation-seo/seo-analytics/comparison-of-google-clickthrough-rates-by-position/
- 12.American Academy of Ophthalmology . Optimizing your ophthalmology practice’s website to show up in search results. Accessed September 19, 2022. https://www.aao.org/eye-health/resources/search-engine-optimization
- 13.Search Engine Journal . The three pillars Of SEO: authority, relevance, and experience. Accessed September 19, 2022. https://www.searchenginejournal.com/seo-guide/search-authority/
- 14.WordStream . How to rank higher on Google Maps in 10 steps. Accessed September 19, 2022. https://www.wordstream.com/blog/ws/2020/07/02/how-to-rank-higher-on-google-maps
- 15.The Manifest . The popularity of Google Maps: trends in navigation apps in 2018. Accessed September 19, 2022. https://themanifest.com/app-development/trends-navigation-apps
- 16.Kloosterboer A, Yannuzzi NA, Patel NA, Kuriyan AE, Sridhar J. Assessment of the quality, content, and readability of freely available online information for patients regarding diabetic retinopathy. JAMA Ophthalmol. 2019;137(11):1240-1245. doi: 10.1001/jamaophthalmol.2019.3116 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Redick DW, Hwang JC, Kloosterboer A, et al. Content, readability, and accountability of freely available online information for patients regarding epiretinal membranes. Semin Ophthalmol. 2022;37(1):67-70. doi: 10.1080/08820538.2021.1913192 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Patel AJ, Kloosterboer A, Yanuzzi NA, Venkateswaran N, Sridhar J. Evaluation of the content, quality, and readability of patient accessible online resources regarding cataracts. Semin Ophthamol. 2021;36(5-6):384-391. doi: 10.1080/08820538.2021.1893758 [DOI] [PMC free article] [PubMed] [Google Scholar]