DemographicsAge Children vs. adults
Older adults?Sex:
In the Medicaid database, for example, women may be over-represented.
Race:Does your study require an over-representation of
racial minorities? If so, does the source that you are considering over-
samples minorities? Does the coding allow you to have a more granular
classification of race?
Geographical representationDoes your study require for you to have
adequate representation of individuals residing in underserved areas Rural areasAppalachian areasSpecific states, or regions of the country?
Consider the eligibility criteria to be included in a given data sourceMedicare database
Age (older adults)DisabilityEnd-stage Renal Disease
For example, if your study involves the analysis of data for pregnant women and children, the Medicare database may not be suitable.
Medicaid databaseIncome threshold (low income)DisabilityDual Medicare-Medicaid enrollment
Consider the services covered (or not covered) in a given database
What is covered by a given program, and what does that database offer?If you are interested in looking at medication
use by older adults, then you can’t use Medicare data in the pre-Part D era.
Inclusion/exclusion criteriaWhose data are included in your database?
Consider the fact that claims data are not available for Medicare beneficiaries who are enrolled in managed care. data will be representative of fee-for-service
beneficiaries only.
Gaps in enrollment (Medicaid)How do you account for gaps in enrollment in
Medicaid on/off enrollment do you limit your study population to those
with Continuous enrollment? Enrollment for 80% / 60% of the time in a given
study period?
Considerations in longitudinal studiesIndividuals willing to participate in
longitudinal studies different from those unwilling to participate (?)
Consider mortal and non-mortal attritionIf you limit your study population only to those
who are alive an extended period of time, then you’re excluding the sicker ones (those more likely to die)
Are those dropping out of the study for reasons other than death different from those who stay in the study?
GeneralizabilityWill the inclusion/exclusion criteria affect
your studyBiases introduced?
Managed care population healthier than the fee-for-service population?
Medicaid beneficiaries enrolled in Medicaid full time Sicker than those enrolled part time? Better navigators of the system?
Greater awareness of the eligibility rules / enrollment or re-enrollment procedures?
Connected to a case manager?
Are the measures that you’re looking for available in the database?For example,
Functional / cognitive status is not available in Medicare or Medicaid data
Readmissions may be derived from HCUP data only with certain databases
How are the measures derived? How good are they?Consider comorbid conditions derived from
claims dataWhat does the presence or absence of certain
conditions mean?Consider other measures
Individual’s race (is it truly self-reported??) prenatal visits vs. birthweight Cause of death
Availability of identifiers in both datasetsWhich identifiers do we have in common?
Social security numberNameDate of birthSex
Do we have permission to use them to link the data?
Requesting data from the Center for Medicare and Medicaid Services
DO IT THROUGH THE RESEARCH DATA ASSISTANCE CENTER: RESDAC (resdac.umn.edu)
Contact ResDAC representatives as you are planning for your study/writing grant applicationRepresentative to offer guidance and data
expertise, and provide cost estimate for your data
Cost estimate to be included in grant application (shows that you’ve done your homework)
Cost estimate will make it possible for you to make an educated estimate of the cost associated with your data needs
1) Written request letterThe purpose for which the data are needed A brief description of the methodology in
which the data will be used Delineation of the data requirements Criteria for data selection or searches
2) Study plan or protocol (typically includes grant application and notice of award (NOA)
The study requires individually identifiable records
The study is of sufficient importance to warrant effect, or risk, on beneficiary privacy
There is reasonable probability that use of data will accomplish purpose, i.e., project is soundly designed and properly financed
3) Data users’ agreementEnsure the data will be used only for the
specific purpose stated in the agreement Develop and implement the appropriate
procedural, technical, and physical safeguards to prevent unauthorized use
Not release any files without prior CMS approval
Return or destroy file(s) by the date specified Not publish or release information that would
permit the identification of a beneficiary
4) IRB documentation IRB: date on which the alteration or waiver of
authorization was approved; HIPAA requires a potential research subject to sign a
HIPAA authorization in order to use their protected health information in a research study. In circumstances where obtaining individual authorization is not practicable, researchers may seek a full or partial waiver of the authorization requirement from an IRB.
A statement that the IRB has determined that the alteration or waiver of authorization, in whole or in part, satisfies the three criteria in the Rule;
A brief description of the protected health information for which use or access has been determined to be necessary by the IRB;
A statement that the alteration or waiver of authorization has been reviewed and approved under either normal or expedited review procedures; and
The signature of the chair or other member, as designated by the chair, of the IRB, as applicable.
The waiver criteria are: (A) The use or disclosure of protected health information
involves no more than a minimal risk to the privacy of individuals, based on, at least, the presence of the following elements; An adequate plan to protect the identifiers from
improper use and disclosure; An adequate plan to destroy then identifiers at the
earliest opportunity consistent with conduct of the research, unless there is a health or research justification for retaining the identifiers or such retention is otherwise required by law; and (3)
Adequate written assurances that the protected health information will not be reused or disclosed to any other person or entity, except as required by law, for authorized oversight of the research study, or for other research for which the use or disclosure of protected health information would be permitted by this subpart;
(B) The research could not practicably be conducted without the waiver or alteration; and
(C) The research could not practicably be conducted without access to and use of the protected health information.
5) Evidence of funding6) CMS data request form7) Privacy Board Review Summary Sheet8) Request letter of support from the project officer
What do you do when you receive the data?The data arrive on CDs/DVDs (they used to
come on cartridges or reel tapes)Go over the documentation, and prepare an
INVENTORY of the dataHave they sent you everything that you had
requested?Does their documentation of the data and the
data match?
What to do, con’tdStart reading the tapes/files
Record layout through ResDACAre the files in good shape? Any “corrupt”
materials?Does the number of records in each file
correspond to the number of records documented in the letter by CMS (or contractor)
VERY IMPORTANT!!!!!!!!!NOTIFY CMS IMMEDIATELY of any
inconsistencies between what has been requested and what has been sent
NOTIFY CMS IMMEDIATELY of any files that you are unable to read (even after assistance from ResDAC or others)
Requesting data from (Ohio) MedicaidData user’s agreement required to
conduct studies examining specific research questions
Study protocol to be shared with the Ohio Department of Job and Family Services (ODJFS), which administers the Medicaid program
ODJFS then to issue a $0 contract with the university
Requesting data from the Ohio Department of Health (ODH)Prepare study protocol justifying your need
to use patient identifiersProtocol to be reviewed by the ODH
Institutional Review Board (IRB)Attendance of the Investigator at the IRB
meeting desirable
Your Institution’s IRBObtain their approval prospectivelyData management protocolQualification of personnel handling the data
Obtaining permission to link data: Logistics and other issues to consider
Agencies’ perspective:TRUSTUsers’ knowledge of the data; proficiency;
hypotheses to testMeasures to protect patient confidentiality
Obtaining permission to link data: Logistics and other issues to consider, cont’d
Users’ perspective:Underestimation of what it takes agencies
to release dataUnderestimation of the agencies’ concerns
on how the data should be used and protected
Proposed approachDIALOGUE
Agencies to explain their concerns to potential users
Users to explain The benefits of the proposed studies Measures to protect patient confidentiality
Agencies’ perspectiveAssess the benefits of the studies
proposed Assess the ability of researchers
to carry out the proposed studiesto abide by rules of protecting patient
confidentialityAssess the willingness of researchers to
establish a collaborative relationship with the agency
Guide the researchers through the agency’s process of IRB reviewEngage the researchers in the
process
Users’ perspectiveExplain the proposed study:
Protocolbenefits relative to public health
Provide details on the measures you and your staff will be taking to protect patient confidentiality
Participate in the process of IRB review. Agency staff like to see documents prepared in a format that’s ready to be presented to the IRB. Some would like to see the researcher present at the IRB meeting.
Users’ perspective, cont’dDialogue between agency and user(s)
does not end upon gaining access to the data. It should be ongoing Regular meetingsDefinition of study cohorts and variablesFindings/ interpretation of resultsDissemination of results – share
abstract/manuscript draftsAuthorship/ acknowledgmentsInform the agency staff on study progressRespect and benefit for the staff’s
knowledge of the data -- They use the data on a daily basis; the user may be new to the data
How to protect patient confidentialityRestricted access to data
All individuals accessing the data to understand the seriousness of the measures adopted to protect the data
All individuals to sign data users’ agreementProtocol in place on:
Electronic access of the data (password protected)
How to protect data tapes and printed materialsLocked, fireproof cabinets
Use of online data only through secure internet connections (e.g., firewalls)
No contact to be initiated with patients or providers identified through the study (as it is in our study)
NO DOWNLOADING OF THE DATA TO YOUR COMPUTERS