IT Innovation Contest

A team-based contest for creative IT solutions

Disseminating UCSF research & connecting with disease communities via social media automation

Proposal Status: 

Synopsis: Develop a suite of datafeed-driven disease-specific UCSF Twitter feeds (e.g. @UCSFDiabetes) to promote and disseminate UCSF disease-specific research more widely and engage thousands of members of disease communities (patients, family members, students, funders, health care providers, policymakers). Apply social media outreach best practices to extend reach. Project builds on a successful demo, and is designed to be potentially scalable to dozens, perhaps hundreds, of disease areas, to support existing campus communications work in disease-specific communities.


 

Relevance

61% of U.S. adults search for health information online, and 59% of adult Internet users have looked online for information about specific diseases or treatments (Pew Research, 2009, 2011). On the other hand, peer-review journal articles and professional presentations are still the two major methods used by researchers to disseminate their work (Chen et al., 2010; CTSA Consortium 2008). As a result, “scientists are failing at communicating science to the (wide) public” (The Welcome Trust, 2001; Wilcox, 2012).

 

Description

We propose a novel automated mechanism that leverages a data feed strategy via Twitter to achieve three goals: 

  1. disseminate UCSF disease-specific research more widely
  2. promote science advances in the biomedical field by employing the channels in which potential target audiences are currently engaged (e.g. patients, family members, students, funders, health care providers, policymakers)
  3. provide a time-efficient mechanism that leverages state-of-the-art technology to support the ongoing science dissemination efforts of UCSF research institutes and programs (e.g., UCSF AIDS Research Institute, Helen Diller Family Comprehensive Cancer Center)

 

Demand

In our experience, disease-specific communities crave access to research findings, but often don't know where to look. A recent online experiment indicates that more people look at research articles if these are promoted on social media. While prior work in the field has often focused on using social media to disseminate research among researchers, there are a wide variety of important additional audiences (e.g. emerging donors) who are also online, yet not sufficiently aware of relevant UCSF research and the UCSF brand. Twitter is a particularly good medium for this outreach because of the significant presence of disease communities, as evidenced, for example, by the widespread use of disease-specific hashtags, making it easy for users to categorize and search for disease-related content.

 

Are people really using disease-specific hashtags? According to Hashtracking, in the past 24 hours, there have been:

  • 119 tweets referencing #rheum (rheumatoid arthiritis), reaching an audience of 46k Twitter accounts
  • 125 tweets referencing #psoriasis, reaching an audience of 146k Twitter accounts
  • 210 tweets referencing #alzheimers, reaching an audience of 210k Twitter accounts
  • 960 tweets referencing #AIDS, reaching an audience of 1.2 million Twitter accounts

 

Our Prior Work Proving the Concept:

Anirvan Chatterjee, our team’s technology expert, created a test Twitter account @EpilepsyScience which demonstrates the concept: @EpilepsyScience is intended to distribute links to PubMed articles relevant to epilepsy. Relevant hashtags are automatically added, and URLs are shortened. After a year online, @EpilepsyScience now has over 350 followers, despite the fact that the account was essentially a one-time test, with tweets posted on only 2 of the last 365 days.

 

Criteria

We will select disease areas based on several criteria:

  • UCSF’s areas of excellence (e.g., cancer, diabetes, heart and vascular, neurology and neurosurgery, immunology and infectious diseases, stem cells, transplant services)
  • under-represented, neglected diseases (as identified by national research priorities) that would benefit from wider exposure and dissemination of ongoing research
  • existing UCSF outreach activities on Twitter (e.g. @UCSFCAPS, @UCSFDC) that could be supported or supplemented with the proposed project

 

Automated Twitter feeds are by no means a replacement for existing hand-curated UCSF Twitter accounts, such as @UCSFScience (1,100 followers) or @UCSFCAPS (53 followers). Instead, the project will provide a new approach to disseminating UCSF research widely which we will develop, implement, and evaluate. If the project is successful, we hope to be able to share best practices and/or supplemental automated content to existing UCSF Twitter feeds.

 

Deliverables 

Over the project period of seven weeks, we will:

  • Launch six to ten pilot Twitter accounts that provide disease-specific live feeds of new UCSF publications.
    • Each pilot Twitter account will be associated with a given disease subject area (e.g. epilepsy, psoriasis, HIV/AIDS), making sure to use naming and descriptor hashtags that leverage our understanding of the actual keywords that Twitter users are using.
  • Develop a hosted solution to automatically feed relevant content on new UCSF research to these Twitter accounts on a daily basis
    • New publications will be tracked either directly from PubMed, or via the UCSF Profiles publications database, by making use of keywords and MeSH associations.
    • News stories may be automatically sourced by searching UCSF RSS feeds for relevant keywords.
    • Relevant hashtags will be added to tweets to enhance discoverability. Hashtags will be added in at least two ways: by marking existing text with hashtags (e.g. "Types of sleep problems in adults living with #HIV/#AIDS"), and adding relevant trailing hashtags where none exist in the text ("CD4 cell count and viral load monitoring in patients undergoing antiretroviral therapy #HIV")
  • Implement and refine promotional strategy
    • Follow relevant Twitter accounts to help publicize the account's existence and engage target audiences
    • Do direct outreach to influential Twitter users in specific disease communities. We will use free online listening / analytics tools, such as SocialMentions and Klout, to identify these influential users.
    • We track the relevant literature on Twitter outreach strategies, and will evaluate and implement other strategies, as appropriate.
  • Evaluate the success of the project.
    • We will focus primarily on the number of followers, and secondarily on the number of links clicked and retweets per posted tweet.

 

Impact on UCSF's Mission and Community

  • Reach over 2,000 people (e.g., patients, family members, students, funders, health care providers, policymakers) with current updates about relevant UCSF research
  • Pilot a novel method to reach target audiences in a targeted and measurable manner through an automated technology-driven approach combined with a social media strategy that can be substantially scaled up to potentially benefit groups across the UCSF campus and at other research institutions.
  • If the pilot is successful, follow-up steps might include:
    • rolling out additional disease-targeted Twitter feeds
    • handing over ownership of the new Twitter accounts to relevant groups on campus
    • integrating automatic publication updates into existing Twitter accounts used by groups on campus (e.g. automatically publishing links to UCSF HIV/AIDS-related papers to @UCSFCAPS)
    • sharing and publishing on the development, implementation, and evaluation of this novel approach

 

Risk Assessment

  • Technology: Because this project builds on a successful proof-of-concept project (the @EpilepsyScience Twitter experiment), we anticipate no technological risks.
  • Social Media: While the @EpilepsyScience experiment was successful in drawing attention from a wide variety of patients, advocates, and medical professionals with relatively low effort, we can't know for sure whether other audiences will be more responsive, or less. We will be distributing the risk by choosing a variety of disease areas and implementing a more comprehensive promotional strategy, incorporating outreach to influential Twitter users.

 

Team Members and Roles

  • Anirvan Chatterjee (20% effort during pilot period)
    • technologist, member of epilepsy online disease community, Twitter user since 2007
    • @anirvan on Twitter
  • Katja Reuter (20% effort during pilot period)
    • scientist, science writer & editor, social media strategist overseeing 12 UCSF Twitter accounts
    • @CTSIatUCSF on Twitter
  • Bradley Voytek (advisory/strategy role)
    • neuroscientist with a strong interest in science outreach, data-driven methods, and social media for research
    • @BradleyVoytek on Twitter

 

Chatterjee and Reuter's time during working hours has been pre-approved. We also anticipate using time outside of working hours.

Comments

A very imaginative use of Twitter to promote UCSF science and brand! I like the concept and hope that it really takes off. Use of automation is key as hand curating would take significant effort.

Hi Dr. Terrazas - Thanks for the comment. We look forward to more feedback, also regarding how we can improve this idea.

I like this idea, especially for rare diseases. What will an automated system do with relevant articles that are not available free online?

Hi Sarah, great to hear that you think this idea could be helpful. In most cases, a summary of an article will be available online even for those articles that are not fully accessible publicly. The feed will provide a link to that summary. However, we hope that as UCSF adopts the open-access policy, more opportunities will emerge to develop this approach further. See also: 1) http://www.ucsf.edu/news/2012/05/12056/ucsf-implements-policy-make-resea... 2) http://www.library.ucsf.edu/help/scholpub/oapolicy

I can speak to this from personal experience. As someone with very limited exposure to biomedical research prior to joining UCSF, it was a revelation to me that research results are at all accessible. I can think of several occasions where as a patient or family member, I would have benefited from being able to access PubMed or Cochrane Summaries. Even being able to read just an abstract is a substantial improvement over feeling like you have zero access to research data — but full access is so much better. This project benefits substantially from our new open access policy. Open access *removes* barriers to access; projects like this are the next step, *encouraging* access, allowing us to showcase UCSF's work and people.

I also think this is a good idea, and agree with the earlier comment about it supporting UCSF's position as a leading research institution. It may also be helpful to create an additional Twitter feed that combines all types of research conducted at UCSF to highlight the University's overall research output.

I like the idea. Maybe @UCSFRawScience, to distinguish it from the human-curated @UCSFScience (https://twitter.com/UCSFScience)?

Would be really interested in how to integrate/promote/etc with current @UCSFCancer hand-curated account, which often posts latest research article links. Keep me posted!

Hi Karen - Thanks for the feedback. We'd be very interested in working with you on a pilot to demonstrate how this approach can advance existing Twitter accounts, fine-tune the strategy based on your organizational communication goals, and help you save time.

Terrific, just let me know when you are ready to discuss. cheers!