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 - Dana Ludwig, "Research Data Browser"


Published on

  • Be the first to comment

UCSF Informatics Day 2014 - Dana Ludwig, "Research Data Browser"

  1. 1. Research Data Browser UCSF ITS - Academic Research Systems Research Data Browser
  2. 2. The goal is to accelerate the clinical scientific discovery process for UCSF investigators by enabling: • Accelerated hypothesis generation to a wider audience, prior to CHR approval, using de-identified data • Rapid Query – all data in memory, not disk • Visual Profiling of selected patients • Export to Excel or Flat File Overall Goal of Research Data Browser
  3. 3. When you access the Research Data Browser, the Select View is the first page displayed. There are two separate views of the same data: Patient View and Encounter View. Select View Page Data as of: – displays the production date. Most times this field is two days behind production.
  4. 4. Patient View – All Selections Tab The “All Selections” tab displays all variables that may be used for selecting patients
  5. 5. Patient View – All Selections Tab • Query for a Sex = Female by clicking the drop-down box for “Sex” and selecting “Female”
  6. 6. Patient View – All Selections Tab • The Total Patients matching your query conditions is always displayed • The “Current Selection” window summarizes what conditions you have already placed in your current query • You can use the Eraser icon to remove conditions • You can use “Clear Selections” to clear the query and start again • The variables that you have used in your selection will be displayed in Green
  7. 7. Patient View – All Selections Tab • For Diagnosis Name, type “Diabetes Mellitus” (quotes included) and it will select all diagnoses that contain “Diabetes” and “Mellitus” in the same name • You can accept the whole list, or click-and-drag the diagnosis names you want
  8. 8. Patient View – All Selections Tab • For selecting on lab values, you first select the lab test(s) such as “Creatinine” • Then select a value range such as “>3”. Using the “>” or “<“ symbol limits resulting values to only numeric, and only within the requested range
  9. 9. Patient View – All Selections Tab • Use the “?” Icon to show a definition of each variable in the selection box
  10. 10. Add Bookmark Make a Bookmark of the Current Selection. In the future, Bookmark applies query to the new dataset
  11. 11. Patient View Home Tab The first tab of the Patient View profiles your currently selected patients by age, diagnosis, discharge disposition and ordered medications
  12. 12. Using a Graph to Change Search Criteria 1. Click on the Skilled Nursing Homes (the green pie piece) on the Top 5 Discharge Disposition graph. 2. The Top 5 Discharge Disposition graph now only displays Skilled Nursing Homes information. NOTE: As you select your criteria on one tab, it carries over to all the other tabs.
  13. 13. Demographics Tab Profile Selected Patients by Demographics Select the Demographics Tab Note: All Tabs reflect the same Current Selection Use the Toggle button to switch between Age and Age Group
  14. 14. Encounters Tab
  15. 15. Diagnoses Tab
  16. 16. Medications Tab
  17. 17. Labs Tab
  18. 18. Procedures Tab
  19. 19. Exploration Tab – Custom Visualization and Export
  20. 20. Exploration Tab – Bar Chart of Creatinines > 3
  21. 21. Exploration Tab – Tabulate all Creatinines > 3
  22. 22. Exploration Tab – Export all Creatinines > 3
  23. 23. HIPAA De-Identified Data • Data has been de-identified per HIPPA Safe-Harbor definition • Patient ID, Encounter ID and other identifiers have been replaced by randomly generated numbers • Dates have a random offset of up to a year in the past and each patient has the same offset for all dates • Ages over 90 are shown as 90 • De-identified data is sensitive information. User’s agree to: • Not attempt to re-identify the data • Not share data with unauthorized individuals
  24. 24. Limitations of the Research Data Browser • Only data from encounters after June 1, 2012 are included (APeX go-live date) • Data is typically one day old – see the Data as of: field • Procedures and Med Orders are limited to orders manually entered in APeX by a provider (ORDER_METRICS):  No child orders  No orders created by an interface  No pending orders  No historical orders  No cabinet overrides
  25. 25. Why We Are Excited About RDB • Unleash the creativity of world-class clinical and informatics scientists • Quickly ask questions any time of day, without CHR approval or programmers or funding • Visually Discover things you weren’t looking for • We discover errors in the data • Analyze your (De-identified) data in your own way, with your own favorite tools, with individual data element-level export • SAS • SPSS • Stata
  26. 26. In order to access the Research Data Browser, you need to submit a request through the ARF – Account Request Form system and complete an Account Request Form (ARF). The link to the ARF system is Accessing the Research Data Browser
  27. 27. Giving Feedback to ARS Your feedback is important to us. Our goal is to: • Make the Research Data Browser as useful as possible. Your feedback as to how we can make the tool and underlying data more useful is appreciated. • Let us know how we can improve the tool, data training and other related materials The Academic Research Systems (ARS) staff is here to assist you in using the tool, validating your results and to help document any issues. Please use the Report an Issue link located in the upper right-hand corner of each page to report any issues or questions. You may also contact ARS at its-