THANK YOU FOR JOINING ISMPP U TODAY! The program will begin promptly at 11:00 am EST December 10, 2014 ISMPP WOULD LIKE TO THANK. . . . . . the following Corporate Platinum Sponsors for their ongoing support of the society 2 ISMPP ANNOUNCEMENTS • Don't forget to register for the 2015 European Meeting of ISMPP—Early Bird discount ends December 16! Check out the latest version of the brochure @ ismpp.org • Watch for information on the 11th Annual Meeting of ISMPP—registration opening soon! • • Interested in taking the March CMPP exam? Don’t miss the February 1 deadline. • Do you have social skills? Follow us on Twitter (@ISMPP) or join the conversation at ISMPP's LinkedIn group page. • This program qualifies for 1 credit towards recertification (professional responsibilities) Did you know your company can sponsor an ISMPP U webinar? If you're interested or would like more information, contact ismpp@ismpp.org. 3 FOR THE BEST LISTENING EXPERIENCE . . . To optimize your ISMPP U webinar experience today, please: • If using VOIP (Voice Over Internet Protocol) vs. your telephone, turn up the volume of your computer speakers • • • Use a hardwire connection if available Use the fastest internet connection available to you If you experience audio problems, consider switching from one access method to another (VOIP phone or phone VOIP) 4 SOCIAL MEDIA IN PUBLICATION PLANNING: CASE SERIES Speakers: Sarah Feeny, ISMPP CMPPTM Complete Medical Communications Ira Mills, PAREXEL Tom Rees, PhD, PAREXEL Moderator: Lisa Baker, PhD, ISMPP CMPPTM INTRODUCTIONS • • Faculty: Sarah L Feeny, ISMPP CMPPTM is Global Head of Scientific Direction at Complete Medical Communications, head-quartered in the UK with offices in NJ, USA. Sarah moved to Complete Medical Communications in 1998 and has extensive experience in the field of ethical publication planning and delivery. She is a charter member of ISMPP, past Chair of the European Committee of ISMPP (2008-2011) and has served on the ISMPP Board of Trustees (2011-2014). Ira Mills, PhD is a Senior Scientific Specialist at PAREXEL where he has worked for the past 3 years. Ira has broad experience in publications with 15 years at medical communication agencies and 20 years as a Research Scientist on the faculty at Brown and Yale University. Ira has published more than 40 peer-reviewed articles of original research in the area of cell signaling. Ira is an ISMPP member and has been an active presenter at recent ISMPP Annual Meeting poster sessions with 2 recent presentations focused on assessing social media in relation to publication activity. Ira is currently serving on the ISMPP Annual Meeting Committee. 6 INTRODUCTIONS, Cont'd. • Tom Rees, PhD is a Scientific Strategy Advisor in the Strategy & Innovations group at PAREXEL International. Tom has 18 years’ experience in medical communications and publications planning, and is a regular contributor to ISMPP meetings through poster and oral presentations and through his work on the ISMPP Workshop Committee. • Lisa Baker, PhD, ISMPP CMPPTM is Principal Medical Writer and Team Lead at The Envision Pharma Group, a publication planning agency. She received her PhD in research psychology from McGill University and has been a medical writer since 2006. Lisa’s work has included publication development and strategic publication planning for varied clients and therapeutic areas. She is an ISMPP Certified Medical Publication Professional™ (CMPP), a member of the ISMPP-U committee and a former member of the ISMPP Standards and Communications committees. 7 AN ANALYSIS OF TWITTER ACTIVITY OF PHARMACEUTICAL COMPANIES AND TWEETS RELATING TO PUBLICATION ACTIVITY Sarah Feeny, Global Head Scientific Direction, CMC Presented at: ISMPP Annual Meeting (April, 2014) Medicine 2.0 (October, 2014) WHAT WE WANTED TO KNOW • Are pharmaceutical companies using Twitter to provide information about congress abstracts/presentations and journal publications? • Twitter (created 2006) is an online social networking service that enables users to post and read short 140character messages called ‘tweets’ • Tweets can contain images and links to other online content • As of July 2014, Twitter had more 271 million active users (Twitter Q2 Annual Report) 9 APPROACH • • In Jan 2014, we identified the ‘top’ 20 pharmaceutical companies Identified Twitter accounts for each: @AbbotNews @Amgen @AstellasUS @AstraZeneca @Baxter_Intl @BayerHealthCare @BMSnews @Boehringer @GSK_conferences • • @JNJnews @Lillypad @Merck @Millennium_US @Novartis @NovoNordiskUS @Pfizer @Roche @SanofiUS One company (Teva) included didn’t have a Twitter account One account @DaiichiSankyo did not have any tweets 10 OVERALL ACTIVITY IN THOSE 3 MONTHS 11 7% OF ALL TWEETS WERE RELATED TO CONGRESS PRESENTATIONS OR PUBLICATIONS 12 COMPANIES WHO ARE NOT TWEETING ABOUT PUBLICATION ACTIVITY 13 COMPANIES WHO ARE NOT TWEETING ABOUT PUBLICATION ACTIVITY, Cont'd. 14 COMPANIES WHO ARE NOT TWEETING ABOUT PUBLICATION ACTIVITY, Cont'd. 15 IS ANYONE CONSUMING THIS INFORMATION? 16 USE OF LINKS AND DRUG NAMES Over half of the publication tweets contained links and the majority of links directed followers to a company-related website Only approximately 15% of tweets mentioned product names; however, 45.0% of tweets had links that mentioned products 17 CONCLUSIONS For the Twitter accounts analysed, over the time periods identified: 7.2% of tweets were considered related to journal publications or congress abstracts and presentation Over half of these tweets contained links and the majority of links directed followers to a company-related website Only approximately 15% of tweets mentioned product names; however, 45.0% of tweets had links that mentioned products Twitter activity varies greatly across top pharmaceutical companies 18 AUDIENCE QUESTION Do you have a Twitter account? A. Yes B. No AUDIENCE QUESTION 19 SOCIAL MEDIA ACTIVITY FOR OPTIMIZING PUBLICATION TIMING IN ACCORDANCE WITH DRUG APPROVAL Ira Mills, PhD Senior Scientific Specialist, PAREXEL AUDIENCE QUESTION Do you believe that tracking Twitter activity longitudinally across the evolution of a clinical trial could be useful in publication planning? A. Yes AUDIENCE QUESTION B. No C. I don’t understand the question 21 INTRODUCTION • • • • Social media Twitter conversations on oncology are ubiquitous • Treatment-related prostate cancer tweets are aligned with recent advances in prostate cancer drug development Contributions by oncologists are relatively sparse, but growing Oncologist-related Twitter activity varies by cancer type The proportion of prostate cancer tweets that are treatmentrelated is among the highest across different cancer types Reid BR, et al. The Social Oncology Project. Mdigital Life. http://www.wcgworld.com/mdigitallife. Accessed November 8, 2014. 22 OBJECTIVE To determine the feasibility of assessing registration trial publication timing relative to regulatory approval to generate the most “buzz” as measured by social media Twitter activity Registration Trial Publication (eg, NEJM) Regulatory Approval (eg, FDA) 23 METHODS • Twitter activity examined for the publication record timeline of an illustrative prostate cancer drug • Key events associated with this trial examined – IDMC recommendation of trial stoppage – NEJM publication – FDA approval notification – EMA approval notification 24 ILLUSTRATIVE TWITTER ACTIVITY WITH A RECENTLY APPROVED PROSTATE CANCER (PC) DRUG 25 METHODS (Cont’d) • As most Twitter activity was associated with the publication of the registration trial and US regulatory approval, we examined – The correlation of tweets associated with these 2 events for other prostate cancer drugs – The nature of Twitter users – The generalizability of our prostate cancer findings to a broader sample of oncology drugs 26 TWITTER ACTIVITY ASSOCIATED WITH NEJM PUBLICATION AND FDA APPROVAL OF PC DRUGS Prostate cancer drug NEJM publication date FDA approval date Enzalutamide (AFFIRM) August 15, 2012 Abiraterone (COU-AA-302) December 12, 2012 Radium-223 (ALSYMPCA) July 18, 2013 August 31, 2012 December 12, 2012 Sipuleucel-T (IMPACT) May 15, 2013 July 29, 2010 April 29, 2010 NEJM date to FDA date, Δ (days) NEJM tweets FDA tweets NEJM tweets/FDA tweets* -16 254 279 0.91 0 85 62 1.37 +64 119 116 1.03 +91 23 28 0.82 *FDA tweets (reference) = 1.0 27 CORRELATION BETWEEN NEJM TWEETS/FDA TWEETS AND AND NEJM DATE TO FDA DATE OF PC DRUGS 28 TWITTER ACTIVITY ASSOCIATED WITH JOURNAL PUBLICATION AND FDA APPROVAL OF ONCOLOGY DRUGS Oncologic Drug Indication Journal Pub date FDA date Δ (days) Journal tweets FDA tweets Journal tweets/FDA tweets Ibrutinib MCL NEJM 6/19/13 11/13/13 -148 406 265 1.53 Regorafenib GIST Lancet 11/22/12 2/25/13 -95 87 109 0.80 Pomalidomide MMyl Blood 1/13/13 2/8/13 -26 26 206 0.13 NSCLC JCO 7/1/13 7/12/13 -11 31 236 0.13 Lenalidomide MCL TLO 6/6/13 6/5/13 +1 0 118 0 Trametinib MMel NEJM 6/4/13 5/29/13 +6 15 126 0.12 Adotratuzumab MBC JCO 3/20/13 2/22/13 +26 0 81 0 Dabrafenib MMel Lancet 6/25/13 5/29/13 +27 12 126 0.10 Denosumab GCT TLO 7/16/13 6/13/13 +33 33 131 0.25 Obinutuzumab CLL NEJM 1/8/14 11/1/13 +69 33 160 0.21 Afatinib 29 CORRELATION BETWEEN JOURNAL TWEETS/FDA TWEETS AND JOURNAL DATE TO FDA DATE OF ONCOLOGY DRUGS 30 CHARACTERIZATION OF TWITTER PROFILE TYPE RELATED TO NEJM PUBLICATION AND FDA APPROVAL – ILLUSTRATIVE PC DRUG 31 LIMITATIONS • • Twitter search conducted by generic name • Data sampling conducted in general Twitter feeds with no specialization per medical professionals and journal/congress-related blogs Data represent oncology social media and may not be generalizable across therapeutic areas 32 CONCLUSIONS • • • • We observed episodic Twitter usage related to registration trial publication and regulatory approvals Focusing on these 2 events failed to show an association between timing of these events and respective Twitter activity Thus, managing timing of registration trial publication to FDA approval to potentially optimize publication impact may not be warranted The presence and visibility of medical professionals on Twitter, particularly of specialists, is limited and may reflect privacy and legal concerns 33 FUTURE TRENDS 34 2014 ASCO ANNUAL MEETING TWEETERS ASCO Daily News. May 30, 2014. http://am.asco.org/2014-asco-annual-meeting-tweeters. Accessed November 30, 2014. 35 #ASCO2014 CANCER TYPE/TOPIC HASHTAGS ASCO Daily News. May 30, 2014. http://am.asco.org/2014-asco-annual-meeting-tweeters. Accessed November 30, 2014. 36 ACKNOWLEDGMENTS • This study, presented at the 10th Annual Meeting of ISMPP (April 7-9, 2014, Arlington, VA), was authored by Ira Mills, Hajira B. Koeller, Nathaniel Hoover, Ken Youngren, and Sheelah Smith • The authors would like to thank Mary-Ellen Taplin, MD, Dana Farber Cancer Institute and Harvard Medical School, and Mary C. Burke, MD, UMass Medical School, for their insightful comments in the development of this study and for their critical review of the content 37 BAD NEWS TRAVELS FURTHEST: THE SOCIAL MEDIA IMPACT OF PUBLICATIONS AROUND TRIAL DISCLOSURE AND MEDICAL WRITING Tom Rees, PhD Scientific Strategy Advisor, PAREXEL BEN GOLDACRE LAUNCHED THE ALLTRIALS CAMPAIGN IN 2013 AS A CALL TO ACTION The AllTrials campaign: Calls for ALL clinical trials to be registered and ALL methods and results to be reported Launched in January 2013 Signed by 532 organizations, including: • • • • • Patient organizations GSK NHS BMA UK Clinical Research Collaboration 39 NEGATIVE OPINIONS OF THE PUBLICATIONS PROFESSIONALS ABOUND • Despite many good publications clarifying the issues, misconceptions remain widespread – Medical writers are “Ghost writers” – Negative trials aren’t published • Why isn’t the message getting through? – Perhaps the positive messages aren’t popular? 40 WHAT WE DID • Identified 7 representative peer-reviewed commentaries on publications practice published 2012–2014 – – – – – – – • Jones CW, et al. Non-publication of large randomized clinical trials: cross sectional analysis BMJ. 2013;347:f6104 Doshi P, et al. Restoring invisible and abandoned trials: a call for people to publish the findings BMJ. 2013;346:f2865 Bosch X, et al. Challenging Medical Ghostwriting in US Courts. PLoS Med. 2012;9(1):e1001163 Rawal B and Deane BR. Clinical trial transparency: an assessment of the disclosure of results of companysponsored trials associated with new medicines approved recently in Europe. Curr Med Res Opin. 2013;1-11 Chan A-W, et al. Increasing value and reducing waste: addressing inaccessible research. Lancet. 2014;383:257266 Mansi BA, et al. Ten recommendations for closing the credibility gap in reporting industry-sponsored clinical research: a joint journal and pharmaceutical industry perspective. Mayo Clin Proc. 2012;87(5):424-429 Woolley KL, et al. Time to Finger Point or Fix? An Invitation to Join Ongoing Efforts to Promote Ethical Authorship and Other Good Publication Practices. Ann Pharmacother. 2013;47:1084-1087 Used Altmetric.com to analyse social media impact 41 THE ALTMETRIC SCORE WAS HIGHEST FOR ARTICLES WITH A NEGATIVE PERSPECTIVE 42 NEGATIVE ARTICLES WERE MORE LIKELY TO BE COVERED IN NEWS STORIES AND BLOGS 43 MOST TWEETS WERE FROM MEMBERS OF THE PUBLIC Insufficient tweets to analyse Mansi, et al. 44 WHY THE DIFFERENCE IN ALTMETRICS? • The negative articles were published in high impact factor journals with wide readership (BMJ, PLoS Medicine) – Although Chan, et al was published in the Lancet • • The negative articles carried a clear message The negative articles appeal to “confirmation bias” 45 AUDIENCE QUESTION What kind of titles do you click on? AUDIENCE QUESTION A. Ones with numbers (5 things that will make you click!) B. They have between 5 and 9 words C. I never click on a title unless it has a negative spin D. Subheadings: they're a real incentive to click E. A good title is very clear 46 NEGATIVE HEADLINES ARE MORE CLICKABLE! http://www.outbrain.com/blog/2013/07/brainpower-how-negative-headlines-far-outperform-positive-ones.html Reproduced with permission. 47 DECONSTRUCTING A TITLE Negative 6-7 Words Subheading Non-publication of large randomized clinical trials: cross sectional analysis Jones CW, et al. BMJ. 2013;347:f6104 Very clear Clinical trial transparency: an assessment of the disclosure of results of company-sponsored trials associated with new medicines approved recently in Europe. Rawal B and Deane BR. Curr Med Res Opin. 2013;1-11 48 DID IT INFLUENCE THE DEBATE? TWEETS: #ALLTRIALS 2500 Daily tweets 2000 Doshi, et al Jones, et al Chan, et al 1500 Rawal, et al Woolley, et al 1000 500 0 Jan Mar May Jul Sep Nov Jan 49 DID IT INFLUENCE THE DEBATE? SENTIMENT: #ALLTRIALS 2500 Daily tweets 2000 Doshi, et al Jones, et al Chan, et al 1500 Rawal, et al Woolley, et al 1000 500 0 Aggregate sentiment score 2000 Jan Mar May Jul Sep Nov Jan 1500 1000 500 0 -500 -1000 50 DID IT INFLUENCE THE DEBATE? OTHER EVENTS WERE MORE IMPORTANT 2500 Daily tweets 2000 Doshi, et al Jones, et al Chan, et al 1500 Rawal, et al Woolley, et al 1000 500 0 Aggregate sentiment score 2000 GSK Jan announce Mar support 1500 Launch 1000 500 May Goldacre on BBC Jul BMA votes that nonreporting is misconduct Sept Nov Jan European agreement European agreement on Clinical Trials on Clinical Trials Regulation Regulation US launch 0 -500 -1000 51 CHALLENGE: AUTOMATING SENTIMENT 200 sample Automated sentiment: positive, neutral, negative Same 200 sample Manual coding: pro, anti, neutral climate change Only a 55% accuracy rate when trying to use machine learning to automate manual coding Billy Duffy, 1:AM conference, London, September 2014. Reproduced with permission. 52 SO WHAT ARE PEOPLE TWEETING ABOUT? “RESTORING INVISIBLE AND ABANDONED TRIALS: A CALL FOR PEOPLE TO PUBLISH THE FINDINGS” – DOSHI, ET AL 307 tweets in 5 days … then nothing! 53 "WE'VE FOUND EFFECTIVELY NO CORRELATION BETWEEN SOCIAL SHARES AND PEOPLE ACTUALLY READING." Haile T. What You Think You Know About the Web Is Wrong. TIME Business. March 19, 2014. Reproduced with permission. 54 "WE'VE FOUND EFFECTIVELY NO CORRELATION BETWEEN SOCIAL SHARES AND PEOPLE ACTUALLY READING." DO WE SHARE THE ARTICLES WE READ? Nope Haile T. What You Think You Know About the Web Is Wrong. TIME Business. March 19, 2014. Reproduced with permission. 55 WHAT DOES IT ALL MEAN? • For publications to have social impact, they need to have a clear message and to be in high-profile journals … • … but the titles need to confirm biases, otherwise they won’t be shared • Sentiment analysis is currently primitive and is easily skewed by retweets of titles 56 ACKNOWLEDGMENTS • This study, presented at the 10th Annual Meeting of ISMPP (April 7-9, 2014, Arlington, VA), was authored by Tom Rees (PAREXEL, Worthing UK), Euan Adie (Altmetric LLP, London) and Sheelah Smith (PAREXEL, Worthing UK). 57 FOR FOLLOW-UP CORRESPONDENCE WITH THE PRESENTERS: • Sarah Feeny • Ira Mills • Tom Rees – sarah.feeny@complete-mc.com or @iamsarahfeeny – ira.mills@parexel.com or @IraMills – tom.rees@parexel.com or @tomrees_meded See the complete posters (Feeny #1, Mills #5, Rees, #24) @ http://www.ismpp.org/10th-am---poster-presentations 58 QUESTIONS...... To ask a question, please type your query into the ‘Q&A’ chat box at the bottom right of your screen. Every attempt will be made to answer all questions. UPCOMING ISMPP U WEBINARS • January 2015 – Topic: Guideline Updates & ISMPP Partners – Date: TBA • February 2015 – – Topic: MPIP Mission & recent article Date: TBA 60 THANK YOU FOR ATTENDING! We hope you enjoyed today's presentation. Please take a moment to click on the link that will be provided and complete the survey. We depend on your valuable feedback as we develop future educational offerings. 61