key: cord-0994254-o7bg6nd5
authors: Munawar, Kamran; Sugi, Mark D.; Prabhu, Vinay
title: Radiology in the News: A Content Analysis of Radiology-Related Information Retrieved From Google Alerts
date: 2020-10-09
journal: Curr Probl Diagn Radiol
DOI: 10.1067/j.cpradiol.2020.09.010
sha: de1779fb149bc53e8cdfc6bd5cc190e931cf5ffd
doc_id: 994254
cord_uid: o7bg6nd5

INTRODUCTION: Radiology topics receive substantial online media attention, with prior studies focusing on social media platform coverage. We used Google Alerts, a content change detection and notification service, to prospectively analyze new radiology-related content appearing on the internet. MATERIALS AND METHODS: An automated notification was created on Google Alerts for the search term “radiology,” sending the user emails with up to 3 new links daily. All links from November 2019 through April 2020 were assessed by 2 of 3 independent raters using a coding system to classify the content source and primary topic of discussion. The top 5 primary topics were retrospectively evaluated to identify prevalent subcategories. Content viewing restrictions were documented. RESULTS: 526 links were accessed. The majority (68%) of links were created by non-radiology lay press, followed by radiology-related lay press (28%), university-based publications (2%), and professional society websites (2%). The primary topic of these links most frequently related to market trends (28%), promotional material (20%), COVID-19 (13%), artificial intelligence (8%), and new technology or equipment (5%). 15% of links discussed a topic sourced from another article, such as a peer-reviewed journal, though only 2 linked directly to the journal itself. 8% of links had content viewing restrictions. CONCLUSION: New radiology content was largely disseminated via non-radiology news sources; radiologists should therefore ensure their research and viewpoints are presented in these outlets. Google Alerts may be a useful tool to stay abreast of the most current public radiology subject matters, especially during these times of social isolation and rapidly evolving clinical practice.

Medicine has consistently been the subject of substantial media attention. [1] [2] [3] [4] [5] [6] [7] This also applies to the field of radiology, for example in the areas of radiation safety, 8 teleradiology, 9 mammography, 10 and artificial intelligence. 11 Modern analytic tools have recently been employed to analyze content trends in radiology-related online media searches 12 and the effects of media coverage on promoting research findings. 7 Furthermore, a prior study has demonstrated an association between media coverage and scientific citations of peerreviewed journal articles, irrespective of article quality. 3 Recent months have seen rapidly evolving practice and clinical guidelines due to the novel coronavirus SARS-CoV-2 (COVID-19) pandemic 13, 14 and there has been a strong emphasis on staying up-todate. Google Alerts is one way to do so, especially during times of social isolation; it is a content change and notification service launched in 2003 which notifies users of new results (e.g. web pages, news media, and scientific research) matching a particular user-specified search query. 15, 16 This service leverages Google Search, the most popular search engine 17 and website in the world 18 with over 1 trillion searches per year. 19 Google Alerts has previously been utilized to study new media content in the fields of primary care, 20 infectious disease, 21, 22 and midwifery. 23 Google Alerts has not previously been utilized in the field of radiology nor, to our knowledge, has there been any detailed and comprehensive content analysis of new radiology-related media on the internet. In order to fill this knowledge gap, we used Google Alerts to prospectively analyze new radiology-related media content on the internet.

This study was deemed exempt by our local Institutional Review Board.

On Google Alerts (Google, Mountain View, CA), a user can detail a search query utilizing standard search engine language, with or without Boolean operators (e.g. AND, OR) and may further specify the frequency ("as-it-happens," "at most once a day," or "at most once a week"), media source (automatic or specific [news, blog, web, video, books, discussions, and/or finance]), language, geographic region (country), and number of results ("all results" or "only the best results") ( Fig 1A) . When enabled, this query generates an automated email to the user matching the specified search parameters (Fig 1B) .

For this prospective study, an email alert was created on the Primary Investigator's (VP) Google Alerts account prior to November 1, 2019 for the search term "radiology," with daily frequency, from "automatic" sources, in English language, from any geographic region, and including "only the best results." This generated daily emails directly to the user containing up to three results per day and was run consecutively for the 6-month period of November 2019 through April 2020.

All links for the 6-month period were retrieved and accessed. The links were categorized by two of three raters (VP, KM, and MDS, with eight, one, and one year(s) of online media research experience, respectively) using an iteratively developed coding system to classify the content source and primary topic of discussion. Additionally, content viewing restrictions, including "paywalls" (i.e. requiring payment to view content), requirement for account creation, and requests to deactivate ad blockers before viewing content, were documented. For the purpose of ad blocker deactivation determination, the free, publicly available ad blocker AdBlock (www.getad block.com) was added to the web browser on which the link was loaded. When articles were unavailable (nonworking link or subscription-requiring service) an attempt was made to categorize the article based on the title or any available viewable content. The top five primary topics of discussion were then retrospectively evaluated to identify prevalent themes or subcategories. Note was also made of whether links were pertaining to the 2019 Radiological Society of North America (RSNA) Annual Meeting, which occurred on December 1-6, 2019. All discrepancies were re-reviewed until consensus was achieved.

The data were evaluated using standard summary statistics, computed in Google Sheets (Google, Mountain View, CA).

A total of 526 unique links were accessed. Four links were nonfunctioning during the coding process and unable to be categorized based on the title or available content. Of the remaining 522 links, a majority of the content was created by non-radiology lay press (n = 355, 68%), followed 

by radiology-related lay press (n = 145, 28%), university-based publications (n = 10, 2%) and professional society websites (n = 9, 2%) (Fig 2) .

The details of the coding system used to evaluate the content of each link is provided in Table 1 . The primary topic of the linked content was most frequently related to market trends (28%), promotional (20%), COVID-19 (13%), artificial intelligence (8%), new technology or equipment (5%), celebratory (5%), and legal matters (4%) (Fig 3) . COVID-19 first appeared in links on February 20, 2020, after which it represented the primary topic in 43% of posts, exceeding that of any other topic. COVID-19 represented 40% (64/163) of links in the months of March and April.

Results of our subcategorization of the top 5 primary discussion topics are displayed in Table 2 . Links relating to market trends most frequently discussed growth trends in information technology (48%), interventional radiology (21%) and radiology equipment (11%), and a minority discussed reimbursements (7%). Promotional links most frequently discussed new business partnerships (38%), career advancements (38%), and advertisements (24%). Links relating to COVID-19 most frequently discussed operational issues and guidelines (65%), COVID-19 related deaths (15%), imaging findings and guidelines (7%) and experimental research (6%). Links relating to artificial intelligence most frequently discussed perception (27%), implementation (20%), new software (18%), and new research (11%); only 9% (n = 4) discussed limitations of artificial intelligence.

A total of 15% (78/526) of all links discussed a topic sourced from another article, such as a peer-reviewed medical journal, however only two links directly sent the user to a primary peer-reviewed journal article. Seventeen links covered topics directly related to the 2019 RSNA Annual Meeting, all of which were posted on or after December 2, 2019.

A total of 43 links (8%) had one or more content viewing restrictions, most commonly requiring deactivation of ad-blockers (n = 19) or account creation (n = 15) and less frequently paywalls (n = 9).

Radiology continues to be a rapidly advancing field as clinical guidelines evolve and subspecialties continue to produce new guidelines. This has been further accelerated in recent months due to the COVID-19 pandemic and staying current with daily updates is as important as ever. This is the first study to prospectively assess new online searchable media in radiology, utilizing a tool that leverages the most popular search engine in the world. We identified a diverse set of topics present in online media and found that most links were directly to non-radiology lay press. The most common topics related to market trends, promotional content, COVID-19, and artificial intelligence. 

We identified a diverse set of sources providing new searchable radiology-related content. Links most frequently came from nonradiology lay press (68%), followed by radiology-related lay press (28%), with a much smaller fraction coming from university publications (2%), and professional medical societies (2%). In addition to the lay press predominance, only two links (<1%) over the 6-month period sent the user directly to a primary peer-reviewed medical journal article. These findings highlight the importance of staying well-informed of medical coverage by traditional non-radiology lay press outlets and/or making active attempts to publicize research or viewpoints in these forums. These mediums can be the primary guiding source for public perception of medical topics, rather than primary peer-reviewed literature that is most frequently viewed by physicians. This is especially important as prior work has shown that media articles can be sensationalized 24, 25 and may omit basic facts, which have the potential to mislead the public. 2, 26 Some links (15%) were secondary descriptions of primary peerreviewed journal articles. Of these, 30% were from non-radiology lay press. In addition, of the 17 articles covering topics directly related to the 2019 RSNA Annual Meeting, the majority (65%) were from nonradiology lay press. This implies that non-medical lay press outlets are paying significant attention to radiology literature and large national meetings as topics worthy of coverage.

Our study also highlighted that radiology-related online content encompasses a wide variety of topics. The most common of these were market trends (28%). Market trends have consistently been a topic of discussion in radiology literature, including topics such as market shares, the job market, and growth trends. [27] [28] [29] [30] For example, the 2019 Medicare Physician Fee Schedule proposed by the Centers for Medicare and Medicaid Services has been a topic of much discussion within the radiology community in recent months. 31, 32 To our surprise, however, we found that only 7% (n = 10) of all links discussed the topic of radiology reimbursements, which is popular in peer-reviewed literature. 31, 33 Rather, links related to market trends most frequently discussed recent and future market growth trends in the areas of information technology, interventional radiology, and radiology equipment.

The first link related to the COVID-19 pandemic was sent by Google Alerts on February 20, 2020, a full nine days before the first reported COVID-19 related death in the United States. 34 After this, 43% of links were related to coverage of the COVID-19 pandemic. Recent months have seen rapidly evolving practice and clinical guidelines, 13, 14 and our findings highlight the potential role that Google Alerts can play in assisting radiologists stay up-to-date during times of rapid change. Google Alerts provided links that covered a broad range of COVID-19 related topics. Most frequently, links offered insights into operational issues, such as ways to prepare imaging suites, protect staff, maintain practice volumes, and impending financial burdens. Several links brought awareness to COVID-19 related deaths of radiologists and radiographers. Surprisingly, only one link discussed the effects of the pandemic on radiology trainees, a topic that has received much interest in recent literature. [35] [36] [37] We also found a large proportion (20%) of promotional material via Google Alerts. These links often discussed new business partnerships, career accomplishments (such as a new hiring, promotion, or award), and advertised businesses and society meetings. Prior research has suggested that radiologists and radiology departments could utilize social media to promote their departments. [38] [39] [40] The results of our study demonstrate the additional role that non-social internet media plays in the promotion of departments and radiologists in order to increase visibility.

The fourth most common topic covered was artificial intelligence (8%). This is no surprise, as artificial intelligence is one of the hottest topics in the radiology peer-reviewed literature, with over 8000 articles published worldwide in the past two decades, a number which is rapidly rising. 41 The links we found via Google Alerts covered a broad range of topics, including perceptions, issues with implementation, new software, and new research. However, only a small percentage (9%) of links discussed the potential limitations of artificial intelligence. This is concerning, as the prospect of artificial intelligence is often cited as a point of concern for medical students considering radiology as a career choice and radiologists concerned about their job prospects, 42 despite well-documented limitations. 43, 44 This suggests that there should be more participation by radiologists to disseminate well-rounded information on artificial intelligence into the lay press.

Lastly, a small proportion (8%) of links had one or more content viewing restrictions, including requirement for account creation, deactivation of ad-blockers, and paywalls. This is helpful information to know, as widespread public access to health-related information is advantageous for the purposes of the expansion of scientific knowledge of the community, although there are obvious financial downsides for those providing this information for free. Our results parallel not only the rapid global expansion of access to the internet, which provides an enormous amount of information at no cost, but also the recent open access movement in the scientific community, calling for expanded access of journal articles, 45 with many diagnostic and interventional radiology journals following suit. 46 There are several limitations of our study. We evaluated links over a 6-month period, so these results may not be generalizable to the entire calendar year. Second, we used a blanket search term of 'radiology' and users may desire to choose their own search term relevant to subspecialty or imaging modality. Lastly, there is uncertainty regarding the details of Google's (and thereby Google Alerts') automated search algorithm, which may prioritize results based on its own preferences, a user's search history, and/or its advertisers, as it does in its standard publicly-available search engine. 47 It is also unclear if the links provided by Google Alerts correlate with changes in search result prioritization on Google itself.

New radiology-related media obtained from a popular search engine were largely provided by traditional non-medical news sources and infrequently described peer-reviewed literature. Google Alerts may be a useful tool that radiologists can use to design unique searches relevant to their practice area and stay abreast of the most current public radiology or subspecialty subject matters, especially during these times of social isolation and rapidly evolving clinical practice.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Media coverage of scientific meetings: too much, too soon?

Media reporting on research presented at scientific meetings: more caution needed

Importance of the lay press in the transmission of medical knowledge to the scientific community

Attitudes toward public dissemination of research

An uneasy relationship: the tensions between medicine and the media

Sources and coverage of medical news on front pages of US newspapers

Interaction between academic radiology and the news media: a potentially powerful and unpredictable processÀfive stories

Newspaper coverage of the risks and benefits of medical radiation imaging

ACR white paper on teleradiology practice: a report from the Task Force on Teleradiology Practice

The U.S. online news coverage of mammography based on a Google News search

International evaluation of an AI system for breast cancer screening

Public interest in imaging-based cancer screening examinations in the United States: analysis using a web-based search tool

COVID-19-New insights on a rapidly changing epidemic

COVID-19) outbreak: what the Department of Radiology should know

Alert Notification as a Service

Create an alert -Google Search Help

Search engine statistics 2017: Smart insights digital marketing advice

List of most popular websites

Google now handles at least 2 trillion searches per year

Supporting social prescribing in primary care by linking people to local assets: a realist review

The financial burden of public health responses to hepatitis a cases among food handlers

Tracking the global spread of vaccine sentiments: the global response to Japan's suspension of its HPV vaccine recommendation

Media representations of breech birth: a prospective analysis of web-based news reports

Sensationalism and the news media

Sensationalism in the media: when scientists and journalists may be complicit collaborators

Fake news and clickbait -natural enemies of evidence-based medicine

White paper: corporatization in radiology

The job market in diagnostic radiology 1999: updated findings from a help wanted index of job advertisements

Trends in PET scanner ownership and leasing by nonradiologist physicians

Peripheral arterial interventions: trends in market share and outcomes by specialty

How Radiology Maintains Relative Value Units But Could Lose Big in Reimbursement: The Power of the Conversion Factor

Real and Perceived Opportunities for Radiologists to Respond to Value-Based Payment Reform

The Alternative Payment Model Pathway to Radiologists' Success in the Merit-Based Incentive Payment System

Washington State Report First COVID-19 Death

Radiology residency preparedness and response to the COVID-19 pandemic

The impact of COVID-19 on radiology trainees

Coronavirus disease 2019 (COVID-19) and radiology education-strategies for survival

The power of promotion: using social media to promote a Radiology Department

Using social media to share your radiology research: how effective is a blog post?

Enriched audience engagement through twitter: should more academic radiology departments seize the opportunity?

Global Trend in artificial intelligence-based publications in radiology from

The impact of artificial intelligence in medicine on the future role of the physician

Artificial Intelligence (AI) for the early detection of breast cancer: a scoping review to assess AI's potential in breast screening practice

How far have we come? Artificial intelligence for chest radiograph interpretation

Open access: the changing face of scientific publishing

Open access journal policies: a systematic analysis of radiology journals

Getting to the top of Google: search engine optimization