CMS Ensured Nearly All Part D Drug Records Contained Valid Prescriber Identifiers in 2016

WHY WE DID THIS STUDY

Prescriber identifiers are a valuable program integrity safeguard. They enable CMS and Part D plan sponsors to determine if legitimate practitioners have prescribed drugs for enrollees. Plan sponsors are required to include prescriber identifiers on the Part D prescription drug event (PDE) records they submit to CMS. The Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) requires that, beginning in 2016, pharmacy claims for covered Part D drugs must contain valid prescriber National Provider Identifiers (NPIs). Additionally, the law requires the Secretary of the Department of Health and Human Services to establish procedures for determining the validity of these prescriber NPIs. The law also requires OIG to submit to Congress a report on the effectiveness of these procedures no later than January 1, 2018. This evaluation report fulfills OIG's MACRA mandate.

HOW WE DID THIS STUDY

We compared the prescriber NPIs on calendar year 2016 PDE records submitted to CMS for covered drugs to the NPIs in CMS's National Plan and Provider Enumeration System (NPPES) file. NPPES is the system of record for health care providers' NPIs. We considered prescriber NPIs to be invalid if (1) they did not appear in the NPPES file, or (2) they appeared in the NPPES file but had been deactivated before January 1, 2015, and remained deactivated through 2016. We reviewed CMS documentation and its responses to a questionnaire regarding its procedures to determine the validity of Part D prescriber NPIs.

WHAT WE FOUND

Of the 1.5 billion PDE records that plan sponsors submitted to CMS for covered drugs in 2016, we found only 147 records that contained invalid prescriber identifiers. These records represented $19,122 in Part D payments. The 147 PDE records were associated with 70 invalid prescriber identifiers, which accounted for a small percentage (0.005) of the 1.4 million unique prescriber identifiers on the PDE records in our review. Specifically, 1 invalid NPI was not listed in the NPPES file, and the remaining 69 invalid NPIs had been deactivated more than 1 year prior to the dates of service on associated PDE records.

CMS has system edits in place to review PDE records to determine whether prescriber identifiers are valid NPIs. These edits reject PDE records (1) that do not contain a prescriber identifier that is an NPI in CMS's current NPPES file and (2) where the date of service is more than 1 year after a prescriber NPI has been deactivated. Regarding the one NPI that was not listed in the NPPES file noted above, CMS stated that this invalid identifier had bypassed its PDE system editing and that CMS is in the process of determining if its edit logic can be modified. For the PDE records associated with the 69 deactivated prescriber NPIs noted above, the dates of service occurred during a time when the date-of-service edit was operating as an informational edit rather than a reject edit. This suggests that CMS accepted these PDE records because the edit flagged-rather than rejected-these records. This edit was changed to a reject edit in May 2016. Additionally, CMS has provided plan sponsors with the procedures they should follow to ensure the validity of prescriber NPIs.

WHAT WE CONCLUDE

The system edits that CMS currently has in place to check PDE records are effective in ensuring the validity of the vast majority of Part D prescriber NPIs.