Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

UCSF Informatics Day 2014 - Ida Sim, "Informatics Technologies: From a Data-Centric View to a Hypothesis-Centric View"

1,071 views

Published on

  • Be the first to comment

UCSF Informatics Day 2014 - Ida Sim, "Informatics Technologies: From a Data-Centric View to a Hypothesis-Centric View"

  1. 1. Clinical and Translational Science Institute / CTSI at the University of California, San Francisco Informatics Technologies: From a Data- Centric View to a Hypothesis-Centric View Ida Sim, MD, PhD Prof Medicine, Co-Director of Biomedical Informatics, CTSI
  2. 2. • Where is the wisdom we have lost in knowledge? • Where is the knowledge we have lost in information? • Where is the information we have lost in Big Data? The Rock, T.S. Eliot. 1934
  3. 3. What is Data, Information, Knowledge? MRN: 000-00-00-0 9.8
  4. 4. Data-Information-Knowledge MRN: 000-00-00-0 HgbA1c 9.8%
  5. 5. Data-Information-Knowledge MRN: 000-00-00-0 HgbA1c 9.8% (normal 4.3% to 5.6%)
  6. 6. Data + Context = Information • Data -- discrete, atomic, raw observations with no inherent structure • Information – data related to other data in an interpretive context that is meaningful for a particular task, e.g., - normal range is meaningful for clinical care - charge is meaningful for payment capture
  7. 7. Knowledge • Data -- discrete, atomistic, raw observations with no inherent structure • Information – data related to other data in an interpretive context that is meaningful for a particular task • Knowledge – generalizable statements that allow you to make predictions about individuals - e.g., people with diabetes are at higher risk of cardiovascular disease
  8. 8. Knowledge Through Hypothesis Testing • Data supports or refutes hypotheses • Hypotheses in translational medicine – T1: about mechanisms of disease – T2/T3 clinical research: treatment, diagnosis, quality of care • within traditional health care settings • beyond hospital and clinic
  9. 9. Hypotheses-Centric View of EHR Data • Observational studies – more prone to confounding and other biases – informatics needs include:1 • clean data and metadata • accurate cohort selection and comparison (see Research Browser) • Expect a Big Data flood of observational findings – many non-health data scientists flocking to health data – need to learn epi/biostats and domain experts to deal with confounding, etc. • Interventional trials will be at a premium to verify observational findings
  10. 10. Observational vs. Interventional • In 2013, 84% of UCSF studies were non- interventional • What's the right balance? – are PIs doing observational research for the right reasons? UCSF 2013 Clinical Research Studies Total = 1,380
  11. 11. Informatics for Interventional Research • Traditional trials, parallel to clinical care – OnCore interoperation with APEX • Virtual, community-based trials – e.g., recruited & screened via web, lab testing in their community, run-in phase with e-diaries, informed consent via web, study medication shipped directly to participants1 • Point-of-care (POC) research – e.g., POC randomization2 , alerts, patient support and care coordination, capturing self-reports via MyChart/sensors, etc – care, QI, and research are on a spectrum3 1 Orri, et al. Contemp Clin Trials 38 (2014) 190–197; 2 Fiore, L. et al. Clin Trials, 2011 8:183-95; 3 Pletcher, et al. JAMA Int Med 2014 174:668-670.
  12. 12. Functional View of POC E-Research Requests • APEX: – programmatically sending patient-specific messages to targeted clinicians via their Inbox – integration of existing risk calculators into APEX (with automated uploading of clinical/lab data) – launching external websites from within APEX – offering structured note templates within APEX (e.g., PHQ-9) – introducing reminders (pop-ups) in APEX – adding documents to Notes or Scanned Outside Documents – insertion of new patient education materials into the After Visit Summary • MyChart: – inserting patient-facing materials into MyChart – recruiting study subjects via MyChart and securing informed consent – mass messaging patients through MyChart – deploying surveys through MyChart
  13. 13. DGIM Survey Choices (n=16/28 researchers) APEX Requests Votes for "Top 5" Votes for "Top 1" Integrate existing risk calculators into APEX 13 4 New structured note templates within APEX 8 2 MyChart Requests Votes for "Top 5" Votes for "Top 1" Deploying surveys through MyChart 9 2 Recruit study subjects via MyChart 8 3 Insert patient-facing materials into MyChart 8 1 Other Top 1 choice with 2 votes: launching external website from within APEX
  14. 14. Next Steps • Build on data management foundation to increase hypothesis-centric use of informatics at UCSF – make point-of-care and other interventional research easier, faster, cheaper • Next CTSI proposal will emphasize e- research infrastructure – interventional, observational, community-based • Need input from you!

×