at SciVerse ScienceDirect

The Breast 22 (2013) 24e27

Contents lists available

The Breast

journal homepage: www.elsevier .com/brst

Original article

Who’s talking about breast cancer? Analysis of daily breast cancer postson the internet

Edel M. Quinn a,*, Mark A. Corrigan a, Seamus M. McHugh b, David Murphy c, John O’Mullane c,Arnold D. Hill b, Henry Paul Redmond a

aDepartment of Academic Surgery, Cork University Hospital, Wilton, Cork, IrelandbDepartment of Surgery, Beaumont Hospital, Dublin, IrelandcDepartment of Computer Science, University College Cork, Cork, Ireland

a r t i c l e i n f o

Article history:Received 21 December 2011Accepted 5 May 2012

Keywords:InternetBreast cancerDiscussion forumsOnlinePatient information

* Corresponding author. Tel.: þ353 214922371; faxE-mail address: edelquinn@rcsi.ie (E.M. Quinn).

0960-9776/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.breast.2012.05.001

a b s t r a c t

Objective: Breast cancer is the cancer most commonly searched for on the internet. Our aim was to assessdaily new breast cancer related posting on the internet.Methods: We analyzed numbers of new daily posts for common cancers for one month and subsequentlyanalyzed content of 1426 breast cancer related posts. We also assessed use of online discussion forumsfor breast cancer related dialogue.Results: Breast related topics had significantly more posts per day compared to others (mean 66.7,p < 0.01). Most posts were on media sites (65.8%). Accuracy levels were high (87.5%) but significantlylower where posted on blogs and discussion forums (p < 0.001). Anonymous posts were common (55%)and less likely to be accurate (p < 0.001). Use of discussion forums has exponentially increased over thelast five years (p < 0.001).Conclusions: The internethasbecomeaprimary forumwithinwhichhealth information, particularly relatingto breast cancer, is both sought and shared. Increasingly information is provided by patients themselves.

� 2012 Elsevier Ltd. All rights reserved.

Introduction

Breast cancer is the commonest cancer occurring in women,affecting approximately 47,400 women per year in Ireland and theUnited Kingdom alone.1,2 Many recent high visibility mediacampaigns have included the internet as a source to raise revenuefor breast cancer charities and increase public awareness.3,4 Arecent study by McHugh et al.5 has shown that breast cancerinformation is the most commonly searched for cancer informationon the internet.

Internet searches using the term “breast cancer” return over 52million results, which can vary in type and content. Similar highnumbers of results are returned when other common cancers suchas colorectal, lung and prostate cancer are searched for.5 Studieshave shown that the majority of health-related information on theinternet is unregulated, with registration of websites with qualityassurance bodies such as ‘Health on the Net’ (HON)6 entirelyvoluntary and therefore infrequent.7

: þ353 214546400.

All rights reserved.

The internet is frequently accessed by patients8,9 as a forum forcommentary on all issues in health, with commentators rangingfrom health professionals to interested individuals with no healthbackground frequently posting content. The onset of “blogging” hasallowed individuals from all backgrounds to post health-related dataon the internet that is widely available to the public and totallyunregulated. Many internet sites allow public responses to health-related articles, news stories and blogs which are then returned asresults to patients searching for breast cancer information.

The aim of this study was to assess for the first time the number,content and accuracy of new internet postings relating to breastcancer posted online worldwide in the English language over a onemonth period. We further sought to compare the number of breastcancer related postings with postings relating to colon, rectal, lungand prostate cancer. Finally we aimed to assess the frequency of useof online discussion forums as a medium for dialogue regardingbreast cancer.

Methods

The free online program ‘Google Alerts’ (Google Inc, MountainView, CA, USA) was used to gather data. An “alert” set up for

Table 2Numbers of new internet postings in a thirty day period for breast, colon, rectal,prostate and lung for each of cancer, tumor and tumour. Mean, minimum andmaximum corresponds to the number of new posts per day over a thirty day period.

Breast Colon Rectal Prostate Lung

Total 2002 1772 800 1813 1766Cancer 1490 1468 674 1492 1496Tumor 196 68 25 91 185Tumour 396 255 101 296 127

Mean 66.73 59.07 26.67 60.43 58.87Std deviation 9.57 8.34 10.17 7.88 6.48Minimum 49 38 14 44 38Maximum 88 79 66 86 67

The italic font denotes the breakdown of total number of posts into those related tocancer, tumor and tumour respectively.

E.M. Quinn et al. / The Breast 22 (2013) 24e27 25

a specific term (i.e. breast cancer) causes the user to receivea hyperlink emailed to them every time the term is posted on anysite worldwide on the internet. Clicking on the hyperlink thenallows the user to view the new posting containing the term as itappears online. This covers approximately 99% of all internet posts.

A daily alert was set up over a four week period from 11thJanuary 2011 until 9th February 2011 for the following terms:“breast cancer”, “breast tumor”, “breast tumour”, “colon cancer”,“colon tumor”, “colon tumour”, “rectal cancer”, “rectal tumor”,“rectal tumour”, “prostate cancer”, “prostate tumor”, “prostatetumour”, “lung cancer”, “lung tumor” and “lung tumour”.

New posts for “breast cancer” for a 28 day period were subse-quently analyzed in further detail. Google Alerts automaticallyclassifies links into three categories e “news” “blog” and “web”. AllWebPages were further sub-classified into one of fifteen categories(Table 1) based on the “about us” section of the website or equiv-alent. Each webpage was then analyzed for topic of discussion,anonymous or named poster, presence of references to peerreviewed literature, presence of HON certification, accuracy ofcontent as compared to nationally accepted breast cancer guide-lines10,11 and facility for comments to be posted in response to thetopic.

Finally, again using the search engine ‘Google’, we assessed thenumber of new posts in online discussion forums containing theterm “breast cancer” each year for the years 2006e2010.

Data was entered into Microsoft Excel (Microsoft Corporation,Redmond, WA, USA) and analyzed using SPSS Version 18 (IBMCorporation, Somers, NY, USA) using student’s t-test for continuousvariables and Fisher’s Exact test for categorical variables withp < 0.05 considered statistically significant.

Results

Overall, there were 8153 new internet postings relating to thefour common cancers. Breast cancer related topics had the largestnumber of new internet postings in the four week period(n ¼ 2002; 24.6%), followed by prostate related topics (n ¼ 1813,22.2%). Of the remainder, colon cancer related internet postingsaccounted for 1772 (21.7%) with rectal cancer accounting for 800(9.8%) and lung cancer 1766 (21.7%) (Table 2).

On a daily basis this equated to a mean of 66.7 � 9.5 (range49e88) breast cancer related postings per day (Table 2). Thiscompared with a mean of 58.8 � 6.5(range 38e67) for lung cancer,59 � 8.3 (range 38e79) for colon cancer, 26.7 � 10.2 (range 14e66)for rectal cancer and 60.4 � 7.9 (range 44e86) for prostate cancer.

In a subgroup analysis of new postings related to “breast cancer”over a 28 day period 1426 WebPages were further assessed. The

Table 1Subcategories of breast cancer related websites analyzed.

Website categories

MediaPatient advocacy groupBlogCharityCommercialDiscussion forumEducationalGovernmentalHealth facility runIndividually runInterest group runPeer reviewed journalProfessional standards groupSocial networkOther

commonest website types of those analyzed were media sites(n¼ 939, 65.8%), followed by blogs (n¼ 216, 15.1%) and commercialsites (n ¼ 100, 7%) (Fig. 1). The most common geographic origin forwebsites was the USA (64.5%, n¼ 920) followed by Europe (n¼ 131,9%) and Asia (n ¼ 83, 6%) (Fig. 2).

Of the 1426 breast cancer related postings on internet Web-Pages, 747 (52.4%) were based upon clinical treatments andmanagement issues, while 679 (47.6%) were not. Furthermore, 73(9.7%) of the clinically related postings were inaccessible or illegibleto the authors due to expired hyperlinks or non-English languagepostings leaving 649 clinically related postings for analysis. Thecommonest clinical topic discussed was adjuvant therapy (n ¼ 150,20%), followed by risk factors (n ¼ 170, 15.6%) and pathogenesis ofbreast cancer (n ¼ 55, 7.3%) (Table 3).

Of the 649 clinically related breast cancer postings, 292 (45%)were posted by named identifiable individuals and 357 (55%) wereanonymous. Five hundred and sixty-eight (87.5%) posted informa-tion which was accurate when compared to accepted guidelines.Media sites, peer reviewed journals and social networking siteswere more likely to have an identifiable individual as the author ofthe posting (p< 0.001, p < 0.001 and p¼ 0.001 respectively). Blogs,discussion forums and commercial sites were less likely to have anidentifiable author (p < 0.001 for all). WebPages with an identifi-able author were more likely to contain accurate information(p < 0.001) and were also more likely to reference peer reviewedliterature (p < 0.001).

Only 23 (3.5%) sites showed registration with the qualityassurance body Health on the Net (HON). Overall 429 (66.1%) didnot contain any references to peer reviewed supporting literature.Websites referencing peer reviewed literature were also more

Fig. 1. Categories of 1426 websites which had new postings related to breast cancer ina four week period e media (n ¼ 939), blogs (n ¼ 216), commercial (n ¼ 100), interestgroup run (n ¼ 60), discussion forum (n ¼ 30), charity (n ¼ 19), peer reviewed Journal(n ¼ 10), governmental/educational/health (n ¼ 20), other (n ¼ 14).

Fig. 2. Place of origin for 1426 websites analyzed (USA ¼ 970, Europe ¼ 131,Unknown ¼ 202, Asia ¼ 83, other America ¼ 35, Global organizations ¼ 18, Australia/New Zealand ¼ 19, Africa ¼ 8).

Fig. 3. Number of discussion form posts per year relating to the term “breast cancer”from 2006 to 2010.

E.M. Quinn et al. / The Breast 22 (2013) 24e2726

likely to contain accurate information (p< 0.001). Blogs (p< 0.001),discussion forums (p< 0.001) and individually run sites (p¼ 0.015)were less accurate with blogs and discussion forums also less likelyto reference peer reviewed articles (p < 0.001 and p ¼ 0.003respectively). Educational sites, governmental sites and socialnetwork sites were more likely to reference peer reviewed litera-ture (p ¼ 0.004, p ¼ 0.047 and p < 0.001 respectively). Commercialsites were less likely to be HON accredited compared to other sites(p ¼ 0.003).

Of non-clinical topics, the most common postings related tofundraising (n ¼ 200, 29.5%), personal stories (n ¼ 189, 27.8%) andraising awareness (n ¼ 114, 16.8%) (Table 3).

Use of discussion forums to discuss the topic “breast cancer” hasincreased on a yearly basis over the last five years, with significantlymore new posts occurring per month in 2010 than in 2006(p < 0.001) (Fig. 3).

Discussion

Our study demonstrates the high frequency of internet postingsin relation to breast cancer with over 2000 new postings in a four

Table 3Topics discussed 1426 new breast cancer postings over a 28 day period.

Topics discussed (n ¼ 1426)

Medical 52.4% (n ¼ 747) Non-medical 47.6% (n ¼ 679)

Adjuvant therapy 150 20.0% Fundraising 200 29.5%Risk factors 117 15.6% Personal stories 189 27.8%Pathogenesis 55 7.3% Raising awareness 114 16.8%Overview 47 6.3% Services 54 7.9%Treatment/surgery 38 5.0% Celebrity news 48 7.0%Lifestyle issues 36 4.8% Advertising 22 3.2%Symptoms 25 3.3% Research meetings 10 1.4%Diagnosis 24 3.2% Funding 7 1.1%Screening 21 2.8% Stocks/shares/

companies6 .08%

Survival 21 2.8% Page redirect 6 .08%Pathology 14 1.9% Other 23 3.3%Case report 13 1.7%Inflammatory breast

cancer11 1.5%

Male breast cancer 9 1.2%Prevention 8 1.0%Alternative therapy 8 1.0%Staging 6 .8%Other medical 25 3.3%Other 46 6.1%Illegible/inaccessible 73 9.7%

week period. Similarly, use of discussion forums for breast cancerdialogue is rapidly increasing on a yearly basis. Breast cancer topicsaremore frequently posted per day than any of the other commonlysearched cancers. Such high volume posting for all the commoncancers demonstrates how widespread use of the internet hasbecome for the worldwide population in sharing health andspecifically cancer related information.

The commonest sites which daily upload new information onbreast cancer and related topics are media sites. This is likely due tothe fact these sites are routinely updated daily and new informationin relation to cancer will be reported in themedia as it occurs. Everyday of our search breast cancer related stories appeared ona minimum of 26 different media sites. This shows the high volumereporting of breast cancer related information in the media and useof the internet to deliver news stories worldwide. However, it mustbe acknowledged that while media sites are updated mostfrequently they may not contain the most in-depth discussion ofcancer related topics. The next most common sites encounteredwere blogs. Again blogs are frequently updated daily which mayaccount for their prominence in our study. Other sites such asgovernmental sites, health and educational facility run sites are notupdated daily and therefore do not feature as prominently in ourresults, although may be a source of more accurate information.

Nearly half of all breast cancer topics discussed in the four weekperiod were not related to clinical management, with 22% of allbreast cancer posts related to fundraising and raising awareness.This illustrates the high profile that breast cancer awarenesspossesses, with the internet used as a powerful marketing arena.Personal stories and celebrity news also featured strongly in non-medical posts (27.8% and 7% respectively) which highlight thepublicity generated by personal stories of both famous and non-famous breast cancer patients. Publication of personal storiesfurther generates awareness of breast cancer among readers ofthese posts.

Over half of all medical posts were posted anonymously.Therefore the authors are non-contactable and their credentialsand background cannot be checked. This again illustrates theunregulated nature of health information available on the internet.In our study only 3.5% of all posts were from websites registeredwith the quality assurance body HON. Over half of medical postsallowed readers to comment on the topic; again these commentsare largely unregulated and individuals are free to post anycomments they wish which will be available to future readers. Ina review of online health information reliability, Adams et al. alsoidentified the potential concern raised by lack of disclosure ofauthorship and information anonymity online.12

An accuracy rate of 87.5% was detected for medical related posts.While this is encouraging, the accuracy rate was website

E.M. Quinn et al. / The Breast 22 (2013) 24e27 27

dependent. Blogs, discussion forums and individually run siteswere all significantly less likely to contain accurate informationthan other sites. Blogs and discussion forums were significantlymore likely to contain information from anonymous posters, withno reference to peer reviewed information which correlates withtheir reduced accuracy levels. These sites are unregulated sitesconsisting predominantly of personal unsubstantiated opinionsand should not be recommended to patients.

The need, or indeed the ability, to regulate health-relatedinformation on the internet has been a widely debated topic inrecent times. As our results show, there is an increasing volume ofinformation appearing online on a daily basis and the majority ofthis information is unregulated. However, effective “policing” ofthis data is likely unfeasible and perhaps it may be more appro-priate to direct individual patients seeking advice to known reliablewebsites. The phenomenon of “apomediation” has arisen since thedevelopment of “Web2.0” (the system of enhanced functionality ofwebpages, which allows more end-user generated content). Apo-mediation refers to the process whereby individuals use an “expertsource” to help guide their online experience; this source can bea person, group or technological tool that points users to infor-mation sources deemed to be relevant, credible and trustworthy.13

Examples of this would be where users provide ratings or recom-mendations judging quality and relevance of webpages to allow“good” information to be highlighted prominently to other users,14

or groups or networks which highlight links to relevant informa-tion. In this way online health-related information searching andanalysis becomes more patient-centred and patient-controlled.

Some limitations do apply to the results in this study. We useda single search engine to provideWebPages for analysis, which doesnot guarantee 100% coverage of all internet postings. This mayaccount for some selection bias. Similarly the number of new postsper day can vary with breaking news stories or topical issues thatare prevalent at the time of an internet search. However we did notencounter recurrent posts on one major topic of interest over thefour week period analyzed and therefore feel the results analyzedare representative of baseline internet coverage of daily cancerrelated postings. Our search was only conducted in the Englishlanguage, which again limits the extent to which our results wouldapply internationally. We compared the accuracy of WebPages toUK and Ireland guidelines, which again may not have full applica-bility worldwide. However, these guidelines are based on a broadinternational evidence-base and we used two guideline documentsto increase the accepted accuracy of WebPages, where the guide-lines differed.

Conclusions

Breast cancer is the most commonly discussed cancer on theinternet on a daily basis. The majority of posts are on media sites,

originate in the USA and almost half do not relate to medicalaspects of breast cancer. Raising awareness and fundraising forbreast cancer have a strong profile on the internet. A broad range ofclinical topics are also discussed. Online discussions in forumsregarding breast cancer are increasing on a yearly basis. Blogs,discussion forums and individually run sites have low informationaccuracy rates, and tend to be anonymously posted withoutreferences.

Adequate provision of accurate, referenced information fromidentifiable posters remains an issue to be addressed in the futureas internet usage is likely to continue to increase. Directing patientsto reliable information online, perhaps through the use of apome-diation, is to be encouraged, in order to guide their access to therelevant, accurate information which is available.

Conflict of interest statement

The authors declare no external funding sources or conflicts ofinterest.

References

1. National Cancer Registry of Ireland. Incidence, mortality, treatment and survival,http://www.ncri.ie/data.cgi/choose-methods.php; 1994 [accessed 15.11.10].

2. Cancer Research UK. Breast cancer statistics: key facts, http://info.cancerresearchuk.org/cancerstats/types/breast/#incidenceCancer; 2010[accessed 15.11.10].

3. Bender JL, Jimenez-Marroquin MC, Jadad AR. Seeking support on facebook:a content analysis of breast cancer groups. J Med Internet Res 2011;13:e16.

4. Jacobsen GD, Jacobsen KH. Health awareness campaigns and diagnosis rates:evidence from national breast cancer awareness month. J Health Econ2011;30:55e61.

5. McHugh SM, Corrigan M, Morney N, Sheikh A, Lehane E, Hill A. A quantitativeassessment of changing trends in internet usage for cancer information.World JSurg 2011;35:253e7.

6. Berland GK, Elliot MN, Morales LS, Algazy JI, Kravitz RL, Broder MS, et al. Healthinformation on the internet: accessibility, quality and readability in English andSpanish. J Amer Med Assoc 2001;285:2612e21.

7. Ni Riordain R, McCreary C. Head and neck cancer information on the internet:type, accuracy and content. Oral Oncol 2009;45:675e7.

8. Diaz JA, Griffith RA, Nq JJ, Reinert SE, Friedmann PD, Moulton AW. Patients’ useof the internet for medical information. J Gen Intern Med 2002;17:180e5.

9. Avantunde AA, Welch NT, Parsons SL. A survey of patient satisfaction and use ofthe internet for health information. Int J Clin Pract 2007;61:458e62.

10. Clinical Guidelines Committee Royal College of Surgeons in Ireland. Breastcancer management clinical guidelines. Royal College of Surgeons in Ireland,http://www.rcsi.ie/files/surgery/docs/20101221084028_Breast%20Cancer.pdf;2000 [accessed 06.07.11].

11. National Institute for Health and Clinical Excellence. Breast cancer (early &locally advanced): diagnosis and treatment, http://guidance.nice.org.uk/CG80;2009 [accessed 28.08.10].

12. Adams SA. Revisiting the online health information reliability debate in thewake of “web 2.0”: an interdisciplinary literature and website review. Int J MedInfrom 2010;79(6):391e400.

13. Esyenback G. Medicine 2.0: social networking, collaboration, participation,apomediation, and openness. J Med Internet Res 2008;10:e22.

14. Deshpande A, Jadad A. Trying to measure the quality of health information onthe internet: it is time to move on? J Rheumatol 2009;36:1e3.