Epilepsy & Behavior 55 (2016) 113–119

Contents lists available at ScienceDirect

Epilepsy & Behavior

j ourna l homepage: www.e lsev ie r .com/ locate /yebeh

Use of antiepileptic drugs during pregnancy and lactation: Type ofinformation provided by searching Google

Tal Lavi-Blau a, Dana Ekstein b,⁎, Miri Y. Neufeld c, Sara Eyal a,⁎⁎a Division of Clinical Pharmacy, School of Pharmacy, The Hebrew University of Jerusalem, Jerusalem, Israelb Department of Neurology, The Agnes Ginges Center of Neurogenetics, Hadassah-Hebrew University Medical Center, Jerusalem, Israelc EEG and Epilepsy Unit, Department of Neurology, Tel Aviv Medical Center, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel

⁎ Correspondence to: D. Ekstein, Department of NeuroloNeurogenetics, Hadassah-Hebrew University Medical91120, Israel. Tel.: +972 2 677 7741; fax: +972 2 677 76⁎⁎ Correspondence to: S. Eyal, Institute for Drug RePharmacy, The Hebrew University, Ein Kerem, Jerusalem8667; fax: +972 2 675 7246.

E-mail addresses: tallavi04@gmail.com (T. Lavi-Blau),(D. Ekstein), mirin@tlvmc.gov.il (M.Y. Neufeld), sarae@ek

http://dx.doi.org/10.1016/j.yebeh.2015.12.0021525-5050/© 2015 Elsevier Inc. All rights reserved. This is

a b s t r a c t

a r t i c l e i n f o

Article history:Received 26 June 2015Revised 7 November 2015Accepted 2 December 2015Available online 13 January 2016

Surveys amongwomenwith epilepsy (WWE) show that they receive their essential pregnancy-related informa-tion frommany sources, including the internet. Our aimwas to assess the types ofwebsites provided by searchingGoogle for the use of four antiepileptic drugs (AEDs) during pregnancy and lactation. The search was performedon 40 computers used by health-care professionals, on 40 computers used by nonhealth-care professionals, andon 5 computers used by WWE in Israel and on 8 computers used by nonhealth-care professionals in the U.S.On each computer, a Google search was conducted for term combinations that included one AED name(“carbamazepine”,“valproic acid”, “lamotrigine”, “levetiracetam”, or “Keppra”) and “Pregnancy”, “Lactation”, or“Breastfeeding”. The top three and top ten websites retrieved in every search were mapped (a total of 45 and150 websites, respectively, from each computer). Across all searches in English, on both U.S. and Israeli com-puters, themajority ofwebsites listed among thefirst three andfirst ten resultswere those of independent healthportals. The representation of the Epilepsy Foundation website was 10% or less, and only a few results were ob-tained from theNIH's general public-orientedMedlinePlus. In Hebrew, results included almost exclusively Israelior Hebrew-translated websites. As in English, results from public-oriented, professionally-written websites inHebrew accounted for less than 50% of entries. Overall, the availability of readable and high-quality informationonAEDs used by pregnant and breastfeedingwomen is limited. Guiding patients towards accurateweb resourcescan help them navigate among the huge amount of available online information.

© 2015 Elsevier Inc. All rights reserved. This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords:InternetAntiepileptic drugsPregnancyBreastfeedingLactation

1. Introduction

Women of childbearing age account for about 30% of all people withepilepsy [1]. Although more than 90% of the women with epilepsy(WWE) can expect uncomplicated pregnancies, these women are atincreased risk of poor pregnancy outcomes. Seizure frequency may in-crease during pregnancy, and the seizures, as well as in utero exposureto antiepileptic drugs (AEDs), can increase the risk of adverse infant out-comes [2]. Therefore, the recommended treatment strategy toWWE in-cludes preconceptual and prelactation counseling [2]. Despite this clearneed, care remains suboptimal, and uninformed decisions may exposeWWE and their babies to unnecessary risks [3,4].

gy and Agnes Ginges Center forCenter, POB 12000, Jerusalem24.search, Room 613, School of91120, Israel. Tel.: +972 2 675

dekstein@hadassah.org.ilmd.huji.ac.il (S. Eyal).

an open access article under the CC

Surveys among WWE show that they receive their essentialpregnancy-related information frommany sources, including the inter-net [3]. This is consonantwith the fact that during the past two decades,the environment in which patients consume medical and health infor-mation has changed dramatically [5] and that today, patients and physi-cians alike collect this information in the World Wide Web. The PewInternet Project estimates that 59% of U.S. adults have looked onlinefor health information in the year proceeding the survey and thateight in ten online health inquiries started at a search engine [6]. AGoogle survey suggested that 86% of physicians have used the Internetto gather health, medical, or prescription drug information [7]. A morerecent survey, from 2013, found that 39% of physicians use generalbrowsers, such as Google, frequently and that another 41% use them oc-casionally [8]. Overall, Google Chrome has been the most used browseron the World Wide Web from April 2014 to April 2015 [9].

The aim of this study was to examine the type of top websitesthat are retrieved by Google for health professional (HP) and nonhealthprofessional (NHP) users, about the use of four AEDs during pregnancyand lactation, as they may reflect the sources of information availableon a quick browser search. We also included in this study searchesconducted by a small group of WWE of childbearing age. Since 2009,

BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

114 T. Lavi-Blau et al. / Epilepsy & Behavior 55 (2016) 113–119

Google personalized search has been extended to signed-out usersworldwide, such that it might rank specific websites higher on the re-sults page based on the user's search history [10]. Because of this“Google Personalized Search” [10], we assumed that different informa-tion will be provided to each group of users and conducted each searchon multiple computers.

2. Methods

The protocol for Google searches by WWE was approved by theHadassah Medical Center Institutional Review Board (protocol #0424-15). Women with epilepsy of childbearing age attending the epilepsyclinic of the Hadassah Medical Center were contacted by one of the au-thors (DE) and asked to participate in the study, if they had a personalcomputer and if they had used it for searching medical information onepilepsy. All women signed informed consent forms prior to conductingthe searches. The protocol for searches by HPs and NHPs was exemptfrom ethical approval. We conducted searches on 85 computers inIsrael (40 computers used by HP users (neurologists, pharmacists, andhealth-care students), 40 computers used byNHPusers, and 5 computersused by WWE of childbearing age) and on 8 computers used by NHPusers in theU.S. The computers, either desktops or laptops,were personalor used by amain user. On each computer, a Google searchwas conduct-ed for English search term combinations that included an antiepilepticdrug (AED) name and one of the words: “Pregnancy”, “Lactation” or“Breastfeeding”. With regard to these terms, we were interested in com-paring the outcomes of the search for “Breastfeeding” and “Lactation”, theformer term being much more commonly searched for, based onGooglefight [11], a website that returns a Google search frequencycount for each of two words/phrases [12,13]. The drugs chosen weretwo older AEDs (carbamazepine and valproic acid) and two newerAEDs (lamotrigine and levetiracetam). The generic name levetiracetamand the trade name Keppra were chosen in order to compare the resultsof searching for the trade versus the generic name of an AED. Searches byHPs and NHPs were conducted between October 2014 and May 2015.Searches by WWE were performed on September–October 2015.

We assumed that most of the NHP users, including theWWE, wouldprefer to conduct searches in their native language; therefore, we addedHebrew searches on 27 of the computers owned by NHP users in Israeland on the 5 computers owned byWWE. TheHebrew search term com-binations included the drug name spelled in Hebrew and the Hebrewphrases for pregnancy and lactation (“herayon” and “hanakah”, respec-tively). More details on the search strategy are provided as Supplemen-tary material.

The top 10websites of each searchweremapped, with a total of 150websites collected from the searches in English on each computer. Atotal of 100 results were added from the Hebrew searches on the com-puters on which these searches were performed. Separate analyseswere conducted for HP, NHP, and WWE computers. We calculated thefrequencies of each of the top ten websites among all search results,mapped the most frequent websites among the top three websites,and calculated the average ranking of selected websites.

Table 1Characteristics of HP and NHP computer users.

HP NHP

Israel

Men/women 14/26 19/21Age (years; median, range) 30, 27–64 (based on 26 users) 35, 17–94 (baHealth discipline Four neurologists, one ophthalmologist,

15 pharmacists, 20 studentsstudying for advanced degrees in healthsciences

Not applicable

Education Five — withouBSc or equivalone — with Ph

Statistical analysis was performed using the Wilcoxon matched-pairs signed-rank test (GraphPad InStat 3, La Jolla, CA). The significancelevel was set at p b 0.05.

3. Results

The characteristics of HPs and NHPs are provided in Table 1. Themedian age of theWWEwas 34 years (range: 31–37years), and theme-dian time since last delivery was 2.5 years (range: 1–5 years). Onewoman did not have children. One woman was pregnant at the timeof the search, and one was attempting pregnancy. The occupations ofall five women were unrelated to healthcare. Four of the women weretreated with lamotrigine, and the fifth was treated with levetiracetam.

Across all searches, results obtained from computers used by HPswere similar to those obtained from computers used by NHPs (Fig. 1).The most prominent website among the top three results was Drugs.com (Figs. 1A,B; Supplementary Fig. 1). The Epilepsy Foundationwebsite was among the first three listings in 10% or less of the pregnan-cy searches (Figs. 1A,B; Supplementary Fig. 1), and its fraction was neg-ligible in the breastfeeding or lactation searches (Supplementary Fig. 1).MedlinePlus was seldom retrieved among the first ten results. Searcheson computers with U.S. internet protocols (IPs) provided similar resultsfor the top three websites (Supplementary Fig. 1D–F).

Searches in Hebrew yielded different results compared with thosein English. For pregnancy, the first ten results in English includedAmerican and UK governmental websites (NCBI, FDA, and NHS), aswell as American and British organizations, health institutes, and healthportals (Fig. 2A). The results of searches in Hebrew included the IsraeliMinistry of Health as themost frequent website, followed by the largestIsraeli HMOwebsite (Clalit; mostly drugmonographs) and by forums ofthe Israeli Epilepsy Association (Eyal; forums; Fig. 2A). All the other topten websites were Israeli as well. The fraction of Wikipedia results waslow, albeit greater for searches in Hebrew than in English (Fig. 2A).When searches were conducted on computers owned by NHP in theU.S., results (in English only) were overall similar to those obtainedfrom searches in English on computers in Israel, with few exceptions(Fig. 2B). The British National Health Service website ranked amongthe top ten in Israel but not in the U.S. The results for breastfeedingand lactation were similar but not overlapping (Fig. 2C).

The first ten results varied among both searches and computers(Fig. 3). Different websites and their ranking were obtained for individ-ual AEDs (Fig. 3). For example, MedlinePlus was listed among the first10 sites almost exclusively for valproic acid (Fig. 3A), and rankings forKeppra and for levetiracetam could differ from each other (Fig. 3D). Amore extended analysis demonstrated that the first 10 results for leve-tiracetam and for Keppra were overall similar when searches wereconducted in English, whether the computers had U.S. or Israeli IP(Supplementary Figs. 2A,B), but significantly varied for searches inHebrew (Supplementary Fig. 2C). Fifteen percent of the results ofsearches in Hebrew for [levetiracetam] + [pregnancy] were irrelevant(mostly directing to information about behavioral disorders and a vari-ety of antipsychotic medications).

US

2/6sed on 35 users) 40, 30–60 (based on 7 users)

Not applicable

t academic degree, 22 — withent, 12 — with MSc or equivalent,D

One — with BSc or equivalent, one — with MSc.Education was not reported for the others.

Fig. 1. Top results of searches for [Pregnancy]+ [(AED)] on computers located in Israel and owned by health professionals (HPs) and nonhealth professionals (NHPs). A. Top three resultsfrom 40 computers owned by HPs. B. Top three results from 40 computers owned by NHPs. Results are shown as percent of all top three results across all computers and AED–pregnancycombinations. C. Top ten results from computers owned by HPs versus NHPs. Shown is the percent of all results retrieved on the first Google page across all computers and [Pregnancy]+[(AED)] combinations, excluding sponsored links. CFP, Canadian Family Physician; EMNE, epilepsy.med.nyu.edu (NYU Langone Medical Center Comprehensive Epilepsy Center); EpRes,Epilepsy Research UK; FDA, Food and Drug Administration; GSK, GlaxoSmithKline; MTB, MotherToBaby (a service of the nonprofit Organization of Teratology Information Specialists);NCBI, HIH's National Center for Biotechnology Information; NHS, UK's National Health Service; PatientUK, patient.co.uk (the link now leads to patient.info); UKTIS, UK teratology informa-tion services; WomensMentalHealth, Massachusetts General Hospital's Center for Women's Mental Health.

115T. Lavi-Blau et al. / Epilepsy & Behavior 55 (2016) 113–119

Forty percent of the top three results obtained byWWE searching inEnglish for information on AEDs and pregnancy represented indepen-dent websites. Less than a quarter represented governmental websitesor epilepsy-oriented organizations (Fig. 4A). Yet, the results obtainedby WWE had higher proportions of government-owned andorganization-owned websites than those acquired by other NHPs(Fig. 1B).When searches were conducted in Hebrew, themajority of re-sults consisted of independent websites or Wikipedia (in Hebrew;Fig. 4B). The ranking of the top ten results of searches conducted bytheWWE on their personal computers was almost identical to that ob-tained by other NHPs (Fig. 4C).When the content of the top ten Hebrewwebsites was evaluated, the percentage of websites containing irrele-vant information was much higher, for both WWE and NHPs (Fig. 4D).For example, some links to HMO websites led to information on pricesand not on special precautions. Similar trends were observed forbreastfeeding-related searches (Supplementary Fig. 3).

4. Discussion

Internet-based resources are increasingly used by patients and HP toaccess medical data, including drug information [6,7,14]. Among menwith epilepsy, the Internet was reported as the third most important

source of health information on epilepsy [15]. In another survey, 57%of patients from epilepsy clinics used the internet to find health infor-mation, and 30% searched for medication information [16]. In general,patients who are unable to obtain health-care services quickly useonline resources to obtain health information more than those notreporting difficulties obtaining services [17]. Web users are more likelyto view websites ranked among the first results on the engine, as theyonly visit the top 10 websites listed in search results [18]. In addition,users do not generally check the “about us” sections of websites orread disclaimers or disclosure statements and usually do not later re-member from which websites they retrieved information or whostood behind the sites [18]. Physicians in particular typically spendthree or less minutes to find answers [7]. Ninety-two percent of physi-cians click at the top of the page, and only eight percent click at spon-sored links [7]. It is important to know how search results are rankedin order to understand which information will reach patients andtheir health-care providers. Our results indicate that, if only thewebsites on the top of the results page are viewed, the physician is un-likely to obtain information directly from medical journals or the mostupdated information from regulatory authorities.

The results retrieved in our study represent a combination of generalhealth websites, epilepsy-specific websites, women health websites,

Fig. 2. The impact of language and search term on search results. A. Top ten results of searches for [Pregnancy] + [(AED)] in English (closed bars) and in Hebrew (patterned bars) on 27computers localized in Israel. **p b 0.01, significantly different from searches in English. B. Top ten results of searches for [Pregnancy] + [(AED)] in English on eight computers localized intheU.S. C. Top ten results of searches for [Breastfeeding]+ [(AED)] (closed bars) or [Lactation]+ [(AED)] (patterned bars) in English on eight computers localized in theU.S. Shown are theresults (means±SD) across all computers andAEDs for each search. *Significantly different from searches for [Breastfeeding]+ [(AED)], p b0.05, **Significantly different from searches for[Breastfeeding] + [(AED)], p b 0.01. FDA, Food and Drug Administration; NCBI, HIH's National Center for Biotechnology Information; NHS, UK's National Health Service; PatientUK,patient.co.uk (the link now leads to patient.info); UKTIS, UK teratology information services; Health.gov.il, the Israeli Ministry of Health's website; Clalit.co.il andMaccabi4u.co.il, websitesof two of the Israeli HMOs; Epilepsy.org.il, forums of the Israeli Chapter of ILAE; Doctors.co.il, forums of a Israeli Health Portal;Malam.org.il, forumof an Israeli organization of patientswithmental disorders. The Wikipedia pages available through the searches in Hebrew were in Hebrew.

116 T. Lavi-Blau et al. / Epilepsy & Behavior 55 (2016) 113–119

and broader scope websites (e.g., Wikipedia). One way in which Googleranks the search results is “link popularity” associated with the numberof hyperlinks leading to it from other websites [19]. According to AlexaInternet, an Amazon.com company which provides commercial webtraffic data [20], the top websites in the Health category are Nih.gov,Webmd.com, and Ncbi.nlm.nih.gov/pubmed. Also listed among thetop ten websites in this category are Mayoclinic.org (fourth rank; rank-ing changed along the study period) and Drugs.com (seventh rank)[21]. Epilepsy.com, which is linked from approximately 1600 sites[21], ranked third in most searches in English in our study, although re-sults were limited to forums rather than AED monographs. The MayoClinic monographs were seldom among the top ten results. In Hebrew,the most frequent top three results included governmental, HMO, andorganizational websites. The other websites on the first webpagecould be related to organizations of NHP (e.g., Malam) or blogsmanagedby either NHP or physicians, in which results did not always provide ac-curate information. For example, in response to a question posted onDoctorsOnly.co.il, Keppra was described as “not recommended for useduring pregnancy”.

In 2009, Google personalized search was extended to signed-outusers worldwide, and based on search history, it might rank specificwebsites higher on the results page the next time the user looks for aquery. Thus, the search results not only are based on the relevance ofeach web page to the search but also are customized based upon thelast 180 days of search activity, separately from Google Accounts and

signed-in Web History (which are only available to signed-in users)[10]. Based on this information, we originally estimated that the vari-ability among search results obtained from individual computerswould be high. Indeed, some variability was observed regarding theranking of individualwebsites (Figs. 2,3; Supplementary Fig. 2). Howev-er, the impact of the “personalized” searches appeared to be overall lesssignificant than expected and manifested mostly in the top three re-sults. Less variability was observed for the entire first page of results,as demonstrated by the small differences in results obtained from HPsvs. NHPs (Fig. 1C) and NHPs vs. WWE (Figs. 4C,D).

The effect of using the more common term “Breastfeeding” vs.“Lactation” for searches in English was modest only (Fig. 2C). Likewise,the results of searches that involved generic versus trade name werequite similarwhen searcheswere conducted in English (SupplementaryFigs. 2A,B), likely because information provided by the top rankedwebsites included both the generic and the trade name of the drug.For searches inHebrew, inmany cases (e.g., theMOH,HMOs and severalindependent websites), the generic name was provided in English, orseveral Hebrew spelling versions existed, whereas the trade name wasin Hebrew. Overall, for the searches in Hebrew, the results were obtain-ed from more reliable websites when they included a trade drug name(Supplementary Fig. 2C). This variability of results obtained byusingdif-ferent search options highlights the importance of guiding patientsabout patient-friendly, high-quality websites, in particular in countriesin which English is not an official language.

Fig. 3. Ranking of top ten results for selected websites. Shown are the results (means ± SD of ranks) of searches on eight computers used by nonhealth professionals in the U.S. for[Pregnancy]+ [(AED)]. A. Results for [Pregnancy]+ [Valproic acid]. B. Results for [Pregnancy]+ [Carbamazepine]. C. Results for [Pregnancy]+ [Lamotrigine]. D. Results for [Pregnancy]+[Levetiracetam] (circles) and [Pregnancy]+ [Keppra] (triangles). The circle or triangle size represents the number of computers onwhich the selectedwebsitewas represented among thetop ten results. *Significant differences between ranking of results for [Pregnancy] + [Levetiracetam] and [Pregnancy] + [Keppra], p b 0.05.

117T. Lavi-Blau et al. / Epilepsy & Behavior 55 (2016) 113–119

It was not within the scope of this study to evaluate the quality of re-sults retrieved by web users beyond assessing relevance. It has beenestablished that the most popular sites do not necessarily have charac-teristics that are the markers of quality preferred by HPs and that con-tent on these sites is not always readable [18,19]. Particularly, thereadability of websites providing information on epilepsy is limited[22–25]. In a recent study that evaluated the quality of online health in-formation about oral contraceptives from Hebrew-language websites,gaps were identified between evidence-based information and the in-formation available online. The HMOwebsites scored highest on credi-bility and usability compared with promotional websites and women'shealth websites [26]. Furthermore, one-third of Israeli primary carephysicians expressed concerns about the quality of the information re-trieved by their patients [27].

This study has some limitations, mostly related to the computerusers and location of the computers used for most searches as well asthe time effect. Most users were not persons with epilepsy, and fewwere women of childbearing age. Yet, the results obtained by WWEwere similar to those obtained by NHPs, in particular regarding thetop ten websites. In addition, NHPs other than WWE may representfamily members and friends who are concerned about the use of AEDsduring pregnancy or lactation. This is in line with the 2013 PEW surveydemonstrating that half of health information searches are on behalf ofsomeone else [6]. Another limitation is that available datamay be affect-ed by nonrepresentative sampling bias because the majority of userscompleted at least high school and were related to one or more of theresearchers. However, the overall variability between computers didnot seem to greatly affect the ranking. In addition, the time interval be-tween the first and the last searches was large (over a year), potentiallyaffecting our findings. At least for the top ten results, the time effect ap-peared to be minor.

Although the majority of searches were conducted on computersphysically located in a country inwhich English is not a native language,the results obtained for searches in English resembled those obtainedfrom computers located in the U.S. In addition, these results may be in-dicative of the type of search results retrieved by people in countries inwhich English is not an official language (currently encompassing themajor portion of the world population).

5. Conclusions

This study highlights the type of data expected to be observedby HPs and the potentially limited availability of patient-oriented,readable, and high-quality information on AED use by pregnant andbreastfeeding women. Health professionals should be aware thatwhen they use the links at the top of the search page, they are likelyto encounter data provided by health portals and not scientific/medicalorganizations. This point is even more important when searches areconducted by NHPs. Even in the U.S., short “basic” summaries targetedto patients and nonmedical users, such as those of MedlinePlus andthe Epilepsy Foundation, are seldom listed among the first ten results.Linking these sites to further search engines and websites may increasetheir accessibility. Medical and scientific societies, such as the AmericanEpilepsy Society and the American Academy of Neurology, can be in-volved in the task of providing easily readable and accurate informationfor patients, for example, by organizing educational campaigns, creatingfree apps for smartphones, interacting with Wikipedia, and providingfunding for translation of adequate sources of information to a wide va-riety of languages. Meanwhile, guiding patients towards accurate andreadable web resources can help them navigate among the hugeamount of available online information.

Fig. 4. Top results of searches for the combinations [Pregnancy]+ [(AED)] inHebrewand in English on five computers located in Israel and owned bywomenwith epilepsy (WWE). A. Topthree results from searches in English on computers owned byWWE. B. Top three results from searches in Hebrew on computers owned byWWE. Results are shown as percent of all topthree results across all computers and AED–pregnancy combinations. C. Top ten results from searches for the combination [Pregnancy] + [(AED)] in Hebrew on computers owned byWWE. D. Top ten results of the same searches, categorized by the website content. Websites were categorized as “irrelevant” if their content were not relevant to the search, regardlessof the website's category. Search results from computers owned byWWE are presented in (C) and (D) as closed bars. Also shown for comparison are the results obtained from computersowned by nonhealth professionals (patterned bars). Shown is the percent of all results retrieved on the first Google page across all computers and [Pregnancy] + [(AED)] combinations,excluding sponsored links.

118 T. Lavi-Blau et al. / Epilepsy & Behavior 55 (2016) 113–119

Author contributions

Tal Lavi-Blau contributed to the design, data collection, interpreta-tion, and manuscript writing. All other authors contributed to thestudy design, interpretation of data, and manuscript preparation. SaraEyal also supervised the collection of data.

Disclosure

None of the authors has any conflict of interest to disclose. TalLavi-Blau was supported by theHebrewUniversity School of Pharmacy.The funding source was not involved in study design; data collection,analysis, and interpretation; report writing; and decision-making withregard to submission of the article for publication.

Sara Eyal is affiliated with the David R. Bloom Centre for Pharmacyand Dr. Adolf and Klara Brettler Centre for Research in MolecularPharmacology and Therapeutics at the Hebrew University of Jerusalem,Israel.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.yebeh.2015.12.002.

References

[1] Adab N, Chadwick DW. Management of women with epilepsy during pregnancy.Obstet Gynaecol 2006;8:20–5.

[2] Tomson T, Battino D. Teratogenic effects of antiepileptic drugs. Lancet Neurol 2012;11:803–13.

[3] Bell GS, et al. Information recalled by women taking anti-epileptic drugs forepilepsy: a questionnaire study. Epilepsy Res 2002;52:139–46.

[4] McGrath A, Sharpe L, Lah S, Parratt K. Pregnancy-related knowledge and informationneeds of women with epilepsy: a systematic review. Epilepsy Behav 2014;31:246–55.

[5] Hesse BW, et al. Trust and sources of health information: the impact of the Internetand its implications for health care providers: findings from the first Health Informa-tion National Trends Survey. Arch Intern Med 2005;26:2618–24.

[6] Fox S, Duggan M. Health Online 2013. Pew Research Center; 2013 [http://www.pewinternet.org/2013/01/15/health-online-2013/. Retrieved May 30, 2015].

[7] Google/Hall & Partners. Connecting with physicians online searching for answers.U.S. http://www.gstatic.com/ads/research/en/2009_ConnectingwithPhysiciansOnline.pdf; 2009. [Retrieved May 30, 2015].

[8] Wolters Kluwer Health 2013 Physician Outlook Survey. Physician InformationSources. Wolters Kluwer Heath, Amsterdam, the Netherlands. http://wolterskluwer.com/binaries/content/assets/wk-health/pdf/company/newsroom/white-papers/wolters-kluwer-health-physician-study-executive-summary.pdf. Last accessed May30, 2015.

[9] StatCounter. Global stats. Statcounter. http://gs.statcounter.com/. [RetrievedMay 30,2015].

[10] Horling B, Kulick M. Personalized search for everyone. Posted: Friday, December 04,2009. http://googleblog.blogspot.co.il/2009/12/personalized-search-for-everyone.html. [Last accessed May 30, 2015].

[11] Googlefight. http://en.wikipedia.org/wiki/Googlefight. [Accessed May 16 2015].[12] Brigo F, Ausserer H. A Google fight between seizure and syncope. Seizure 2014;23:

86.[13] Brigo F, et al. Why do people Google epilepsy? An infodemiological study of online

behavior for epilepsy-related search terms. Epilepsy Behav 2014;31:67–70.[14] Law MR, Mintzes B, Morgan SG. The sources and popularity of online drug informa-

tion: an analysis of top search engine results andweb page views. Ann Pharmacother2011;45:350–6.

[15] Sare G, Rawnsley M, Stoneman A, Duncan S. Men with epilepsy— the lost tribe? Re-sults of a nationwide survey of men with epilepsy living in the UK. Seizure 2007;16:384–96.

119T. Lavi-Blau et al. / Epilepsy & Behavior 55 (2016) 113–119

[16] Escoffery C, et al. Use of computers and the Internet for health information by pa-tients with epilepsy. Epilepsy Behav 2008;12:109–14.

[17] Amante DJ, Hogan TP, Pagoto SL, English TM, Lapane KL. Access to care and use of theInternet to search for health information: results from the US National Health Inter-view Survey. J Med Internet Res 2015;17:e106.

[18] Eysenbach G, Köhler C. How do consumers search for and appraise health informa-tion on the world wide web? Qualitative study using focus groups, usability tests,and in-depth interviews. BMJ 2002;324:573–7.

[19] Fahy E, Hardikar R, Fox A, Mackay S. Quality of patient health information on the In-ternet: reviewing a complex and evolving landscape. Australas Med J 2014;7:24–8.

[20] Alexa. About. http://www.alexa.com/about. [Retrieved May 30, 2015].[21] Alexa. http://www.alexa.com/. [Updated November 4, 2015. Retrieved November 6,

2015].[22] Elliott JO, Shneker BF. A health literacy assessment of the epilepsy.com website. Sei-

zure 2009;18:434–9.

[23] Elliott JO, Charyton C, Long L. A health literacy assessment of the National EpilepsyFoundation Web site. Epilepsy Behav 2007;11:525–32.

[24] Brigo F, Otte WM, Igwe SC, Tezzon F, Nardone R. Clearly written, easilycomprehended? The readability of websites providing information on epilepsy. Ep-ilepsy Behav 2015;44:35–9.

[25] Brigo F, Trinka E. Google search behavior for status epilepticus. Epilepsy Behav 2015;49:146–9.

[26] Neumark Y, Flum L, Lopez-Quintero C, Shtarkshall R. Quality of online health infor-mation about oral contraceptives from Hebrew-language websites. Isr J Health Pol-icy Res 2012;1:38.

[27] Giveon S, Yaphe J, Hekselman I, Mahamid S, Hermoni D. The e-patient: a survey ofIsraeli primary care physicians' responses to patients' use of online information dur-ing the consultation. Isr Med Assoc J 2009;11:537–41.