Yesterday, CMS addressed why many records have been deleted from Open Payments. As we wrote, many of the problems are tied to the system's reliance on the NPPES database.
"Records submitted to Open Payments are considered invalid if:
- 1. The NPI, state license information, and first and last names given for the physician in the record do not all exactly match against CMS data sources.*
- 2. The physician in the record has an NPI according to CMS data sources*, but the NPI is not included in the record, in accordance with 42 C.F.R. §403.904(c)(3)(ii).
- 3. The physician in the record validly has no NPI and the state license information and first name and last name given in the record do not all exactly match against CMS data matching sources.**
- 4. The physician in a general or research payment record submitted by an applicable GPO does not also have an ownership interest record submitted for that same GPO, in accordance with 42 C.F.R. §403.906(b)(6).
- 5. The legal business name and TIN of the teaching hospital in a general or research payment record do not exactly match to the corresponding information in the CMS list of approved teaching hospitals.
- 6. A research payment record in which any physician covered recipient principal investigators’ NPI, state license information, and first and last names given in the records do not exactly match against the CMS data sources*, in accordance with 42 C.F.R. §403.904(f)(1)(v).”
* National Plan and Provider Enumeration System (NPPES) and the Provider Enrollment, Chain and Ownership System (PECOS)
“CMS will contact affected applicable manufacturers and applicable GPOs, using the contact information associated with their Open Payments registration profile, to provide additional instructions on how to correct and re-submit their records after September 30, 2014.”
“Resubmitted payment records will be made available for review and dispute at a later date. Timelines for when physicians may view the resubmitted files and begin review and dispute actions is forthcoming, however this will not occur before the September 30, 2014 public posting deadline. As such, these resubmitted payment records will not be published in the September 30, 2014 data publication.”
As the “*” indicate, CMS states that many of the problems resulted from faulty NPPES matching. However, CMS notes that “Truven” also provided insight into some of the mismatched data. Truven Health Analytics provides healthcare analytics and data solutions. According to several sources, they do not necessarily keep master lists of physicians in order to properly match data as a basis to reject companies’ submission and must have relyed on an additional source.
In this first year of data submission, we notice some issues that need to be addressed going forward. For one, doctors’ addresses change throughout the year— it is difficult to match the address with the often outdated NPI database. Also, if CMS is relying on multiple data sources, it is important for the submitters to know the data sources and the date time stamp of which those sources were made. It is problematic to hold manufacturers responsible for what was initially a correct physician identifier.
Second, we spoke with companies who experienced scenarios where doctors get married and changed their last names. In some cases these doctors get a new NPI; in other situations, they bill Medicare under their old name.
Physicans can also have multiple NPI numbers. One physican described how she had looked up her NPI data and had six NPI numbers assocated with her. The first year she was in fellowship she applied for an NPI number, then at each location that she has practiced since. She was an academic physican so her path took her several places throughout the country.
The issues can be summed up in this announcement on the PECOS database:
Also in looking up Truven, we found CMS’ December 2013 webinar, which stated that “[a]fter the matching process is complete and if there are any records that could not be attributed with a sufficient degree of confidence, reporting entities will receive a follow-up email with information about unsuccessful matches.” We are seeing this process now. However, CMS also states on the same slide that:
“Reporting entities will have the opportunity to make corrections to the data or proceed with submission as entered. Inability to reliably match a submitted record to a physician or teaching hospital in the reference sources will not result in record rejection and will not prevent submission or attestation.”
This notice process evidently came after the deadline for submission. CMS makes no mention of penalizing companies in the slides for “mismatched” data. Our article yesterday showed that state license numbers often did not match up with the NPPES database, despite the fact that state license numbers are the more reliable physician indicator. We will wait to see CMS’ timeline for allowing manufacturers either to reconfigure their submissions to the NPPES’ version of the state license numbers, or if CMS will have an alternative solution.