Difference between RIF, LDS, and PUF data files

Difference between RIF, LDS, and PUF data files

Current Version Date: 
08/10/2016

Public Use File (PUFs), also called Non-Identifiable Data Files, have been edited and stripped of all information that could be used to identify individuals. In general, the PUFs contain aggregate-level information.   

 

Limited Data Set (LDS) files also contain beneficiary level protected health information similar to the RIF files.  In fact, many of the RIFs have an LDS equivalent. LDS files are considered identifiable because of the potential to re-identify a beneficiary.  The difference, however, between RIF and LDS is that selected variables within the LDS files are blanked or ranged. LDS requests require a DUA, but do not go through a Privacy Board review.  LDS files are available as a 100% or 5% random sample file. The DUA-Limited Data Sets (LDS) page on the CMS website describes the ways in which the LDS files may be used.

 

Research identifiable files (RIFs) contain beneficiary level protected health information (PHI). Requests for RIF data require a Data Use Agreement (DUA) and are reviewed by CMS’s Privacy Board to ensure that the beneficiary’s privacy is protected and only the minimum data necessary are requested and justified.  The Researchers page of the CMS website provides the criteria for the release of these data. 


 

Table 1. Overview of file difference by privacy level

 

Public Use File

Limited Data Sets

Research Identifiable

Requires Privacy Board Review?

No

No

Yes

Requires a Data Use Agreement?

No

Yes

Yes

Files include beneficiary-level data?

No

Yes

Yes

Researchers may request customized cohorts (e.g. Diabetics residing in MN)?

No

No

Yes

Data can be linked at beneficiary level to non-CMS data using a beneficiary identifier?

No

No

Yes[1]

Claim run off period[2]

NA

NA

Annual file: 6-month run off

Quarterly file: 3-month run-off

Annual file: 12-month run off

Quarterly file: 3-month run off

 

 

 

The RIF and LDS files both contain beneficiary-level data, however, some variables included in the RIF data may be presented differently (ranged or absent) in the LDS counterpart. See Table 2 for the key differences.

Table 2. Variable differences between RIF and LDS files

Variable

File

Limited Data Set

Research Identifiable File

Unique Beneficiary Identifier

Claims & Enrollment files

MedPAR

Encrypted identifier

NO identifier

Encrypted identifier

Encrypted identifier

Health Insurance Claim (HIC) or

Social Security Number (SSN)

Claims & Enrollment files

Not included in file

Included as an add-on with special permission only

Dates (MM/DD/YYYY)

Claims files

MedPAR

Included as of CY2010[3]

Quarter and Year only

Included

Included

Claim from date

Claims files

Not included

Included

Claim through date

Claims files

Included

Included

Beneficiary Zip Code[4]

Claims & Enrollment file

MedPAR

County & State

State Only

Included

Included

Beneficiary Date of Birth

Claims, MedPAR & Enrollment file

Not included. Age year or age range[5]

Included

Date of Death

Enrollment file

Included, for validated dates of death only[6][7]

Included

NPI/UPIN for person level provider

Claims files

 

MedPAR

As of 2013, the real NPI is included[8]

NOT included

Included

 

NOT included

Facility provider number[9]

Claims files & MedPAR

Included

Included

NPI of the facility

Claims files & MedPAR

Included

 Included

 


 

Appendix

NPI/UPINs are encrypted in the LDS files from 1999-2012. A free crosswalk is available for request.  The crosswalk includes encrypted to unencrypted UPIN and NPI numbers back to 1999.

Please see the DUA - Limited Data Sets (LDS) page under "Requesting LDS SAF NPI/UPIN Crosswalk" for ordering instruction.

 



[1] The inclusion of patient identifiers linkable to outside data requires CMS approval. Without this approval, the RIF patient identifiers are not linkable to outside data.

[2] More detailed information about the run off periods and availability are found in RIF Medicare Quarterly Data article and Medicare Limited Data Set (LDS) Quarterly Claims and Enrollment Data

[3] LDS files include dates as of 2010. For 2009 files, CMS provides the dates as a separate file.  Prior to 2009, the files present dates as a quarter and year.

[4] The Medicare Current Beneficiary Survey (MCBS) LDS and Health Outcomes Survey (HOS) LDS files contain zip code, date of birth, and date of death.

[5] See footnote 2 above.

[6] See footnote 2 above.

[7] Based on a ResDAC analysis of the 2012 RIF Master Beneficiary Summary file, 4% of the death dates are not validated.

[8] NPI/UPINs are encrypted in LDS data files from 1999-2012. A free crosswalk file is available for request to identify individual providers back to 1999.   See the Appendix for the record layout and ordering information.

[9] The facility provider number is also called the CMS Certification Number (CCN) or the Medicare provider number and identifies the institutional facility.

Article Number: 
148
Author(s): 
This work was performed under CMS Contract Number HHSM-500-2005-00027I.