Search de-identified data from clinic visits at several Harvard-affiliated hospitals. SHRINE

SHRINE Frequently Asked Questions

  1. Who is eligible to use SHRINE?
  2. I am not employed by a participating hospital. However, I do have an academic appointment within, and am employed by, Harvard University. Can I access SHRINE data?
  3. What are the steps for obtaining data for individual patients at each of the hospitals?
  4. What is the quality/completeness of the data at each hospital?
  5. What kinds of data are available through SHRINE?
  6. What timeframe is covered by available data?
  7. SHRINE currently returns aggregate results only. Are there plans to provide more specific data sets in the future?
  8. I ran the same query several times. Why are the results different each time?
  9. I get a result of “No Results Available” from my query. What does that mean?
  10. Is IRB approval required to request data?
  11. How is the SHRINE IRB structured across independent institutions?
  12. What is the Data Steward?
  13. How do I obtain approval for my query topic?
  14. What is the difference between a query topic and a query?
  15. How can I learn about formulating a data query before actually entering the system?
  16. Is my query activity monitored?
  17. Do I need a separate approval from the Data Steward for each query topic?
  18. What is i2b2?
  19. Can I select for patients taking particular medications?
  20. Can I get patient samples?
  21. Can I search directly through the text of medical records?
  22. How is patient privacy protected?
  23. What are the rules governing access to data?

  1. Who is eligible to use SHRINE?

    SHRINE is a service available to Harvard Medical School faculty (Instructor or above) and Fellows employed by one of the five participating hospitals. Fellows must be sponsored by an approved faculty member. See Getting Started for details.

  2. I am not employed by a participating hospital. However, I do have an academic appointment within, and employed by, Harvard University. Can I access SHRINE data?

    If you are not employed by a participating hospital, you may access SHRINE data indirectly under the following conditions:

    • Employed by Harvard University (e.g., HMS, Harvard T.H. Chan School of Public Health, HBS, FAS, SEAS)
    • Appointed as Instructor or above, or as Fellow
    • Collaborate with a hospital-based SHRINE user in the medical specialty of your interest, who must initiate and receive approval for a Query Topic and perform queries on your behalf.

  3. What are the steps for obtaining data for individual patients at each of the hospitals?

    SHRINE does not provide direct access to information about individual patients at any of the hospitals. To follow up with a potential cohort of interest, you will need to identify a collaborator at the institution where the patients were seen.

    If you are not currently collaborating with an investigator at that institution and are not aware of others who might be doing similar research, you can use other resources on the Harvard Catalyst website to make a connection. For instance, Harvard Catalyst Profiles is a research social networking system that allows you to conduct keyword searches for collaborators at Harvard-affiliated institutions. You can search Harvard Catalyst Profiles using the Search box at the top right corner of every page on this website.

  4. What is the quality/completeness of the data at each hospital?

    The system currently has a complete set of data for demographics and diagnosis. There are partial data sets for medications and laboratory values. Data begin in 2001 and the quality/completeness reflect what was available based in each institution's electronic medical record system.

  5. What kinds of data are available through SHRINE?
    • Total numbers of patients at each participating SHRINE hospital meeting your inclusion or exclusion criteria for demographics (age, gender, race, etc), diagnoses (ICD9/Billing Codes), labs, and most frequently prescribed medications.

    • Not all data points will be available on all patients.

    • Individual patient records are not available through this process.

    Return to top

  6. What timeframe is covered by available data?

    The timeframe at the participating SHRINE hospitals for Demographics, Diagnoses, Medications, and Lab Tests covers the period 1/1/2001 - 6/30/2014.

  7. SHRINE currently returns aggregate results only. Are there plans to provide more specific data sets in the future?

    Future major releases of SHRINE may include access to more detailed data. There are currently no plans to provide full datasets.

  8. I ran the same query several times. Why are the results different each time?

    As an additional way to protect patient privacy, we blur the aggregate patient counts so they are an estimate of the actual number. If you run the same query more than once, you will receive slightly different results each time. The system is designed to lock out users who run the exact same query too many times. If you would like to find out more about the obfuscation process please see "A Security Architecture for Query Tools used to Access Large Biomedical Databases" (Murphy, SN and Chueh, HC, Proc AMIA Symp. 2002:552-6).

  9. I get a result of “No Results Available” from my query. What does that mean?

    If the user experiences issues with queries where one or more institutions never returns a result, or returns “no results available,” it may be a result of a query that is linked to an extremely high number of codes at local institutions. Such issues are most likely to appear when several commonly used medications are queried together. While queries of individual terms should never lead to this issue, querying more than 5 common medications may cause this result.

  10. Is IRB approval required to request data?

    No, the IRB at each participating institution has declared by an expedited review that aggregate data counts are not "human research" and as such are exempt from the IRB review process.

  11. How is the SHRINE IRB structured across independent institutions?

    SHRINE has been independently approved by the IRBs at each of the participating institutions and at Harvard Medical School.

    Return to top

  12. What is the Data Steward?

    The Data Steward refers to both an application, the Data Steward application, and a person, the Data Steward administrator. The Data Steward application collects information about the types of queries an investigator would like to perform in SHRINE.  The Data Steward administrator reviews each request to perform queries and, if appropriate, approves them.  Before approving a query, the Data Steward administrator may contact investigators to get more information about how they hope to use SHRINE.  Query requests MUST be submitted and approved before a registered investigator can begin using the system.

  13. How do I obtain approval for my query topic?

    This process is an electronic workflow that consists of the following steps:

    1. Initiate the process by accessing your institution's SHRINE portal and providing the information requested, including your eCommons name and password.
    2. Your academic appointment and employment status will then be automatically verified.
    3. Once your status has been authenticated, you will be asked to review and agree to a Terms of Access Agreement prior to submitting your data request form to the Data Steward.
    4. The Data Steward administrator will review your request for scientific appropriateness.
    5. Check your SHRINE log for approval to proceed to the query tool. The review process should be complete within two working days.
  14. What is the difference between a query topic and a query?

    A query topic is the general subject that you as a researcher are interested in studying by seeking access to defined subjects from across the hospitals. For instance, "irritable bowel syndrome", "pancreatic cancer", "diabetes type 1". (More detailed examples are found on Getting Started: Query Topic Approval.)

    Once your general query topic is approved by the Data Steward administrator, you may conduct as many individual queries as you like to learn more about your general topic. For instance, you may wish to look at gender differences, breakdown by race, in combination with other diagnoses, by age range, etc. As long as these individual queries are related to your general query topic, you are free to explore your topic in whatever depth is currently available.

  15. How can I learn about formulating a data query before actually entering the system?

    See the SHRINE User Guide, particularly the section on How to Construct a Query.

  16. Is my query activity monitored?

    Each of your individual queries is logged and will be reviewed by the Data Steward for compliance with your approved query topic. Violation of the Terms of Access Agreement will result in a misconduct review by your institution and, if appropriate, by Harvard Medical School.

  17. Do I need a separate approval from the Data Steward for each query topic?

    Yes.

    Return to top

  18. What is i2b2?

    i2b2 (Informatics for Integrating Biology and the Bedside) is an NIH-funded National Center for Biomedical Computing based at Partners Healthcare. i2b2 has developed a software suite (typically shorthanded to "i2b2") that can be installed "on top of" an existing research patient data repository to enable easy and logical manipulation of the data contained within that repository. Each of the Harvard SHRINE hospitals has installed a version of that software suite and thus can be easily linked by a web-based query (SHRINE).

  19. Can I select for patients taking particular medications?

    The system shows a subset of medications for those practices that record this information. A complete set will be included in the next release of the software.

  20. Can I get patient samples?

    No, not directly through SHRINE.

  21. Can I search directly through the text of medical records?

    SHRINE does not allow direct access to patient records.

  22. How is patient privacy protected?

    The system architecture is configured with the following protections in place:

    • Each user of the system needs to be authenticated at their individual institution to verify employment and faculty status.
    • All communications are encrypted using standards approved by the W3C Consortium.
    • Queries return only aggregate counts.
    • Aggregate numbers are blurred (or obfuscated), so that the counts returned are an estimate of the number of patients meeting the queried upon criteria at each institution. The obfuscation algorithm is based on what has been published by researchers at Partners HealthCare in "A Security Architecture for Query Tools used to Access Large Biomedical Databases" (Murphy, SN and Chueh, HC, Proc AMIA Symp. 2002:552-6).
    • No personally identifiable patient information ever leaves an individual institution.
    • Institution-specific user log-in credentials never leave an individual institution.
    • Users must register the topics they would like to query with the Data Steward application. The Data Steward administrator manually reviews all query requests to make sure they are in compliance. Actual query histories are logged and audited on a regular basis to ensure that there have been no violations of the Terms and Conditions.
  23. What are the rules governing access to data?

    Read the Terms of Data Access agreement to understand the rules governing access to data.


    Return to top

SHRINE National Pilot

A pilot implementation of SHRINE as a demonstration of the feasibility of a national research network.