Use Case 19: Out-of-network patient leakage impact

From Demand-Driven Open Data
Jump to: navigation, search

Want to be notified when this use case or any of its datasets change?
You’re signing up to be notified on changes to this use case and associated data sources
* indicates required

Requirements

Use case summary

  • Title: De-identified claims data to analyze referral patterns and out-of-network patient leakage impact
  • Work item: demand-driven-open-data/ddod-intake#19
  • Status: Inactive due to lack of specifications from requester.

Description

  • Need to obtain claims datasets at the provider-level in order to perform analysis of provider referral patterns and patient leakage out of given ACO networks to determine leakage’s downstream impacts on financial performance and care quality.


Value

  • Value to customer: Richard Ferrans, MD
  • Value to industry/public:
    • Repatriate referrals to a given ACO network to better control costs, improve financial performance, and improve care coordination
    • Promote referrals to physicians with superior clinical outcomes
    • Develop comprehensive view of referral patterns to better understand and promote patient access to care
    • Reduce inappropriate or duplicative care; promote preventative services and lower-cost care

Customer

  • Referral management and population health SaaS company
  • Richard Ferrans, MD
  • Fred Trotter

Current data and limitation

  • Data source: Detailed claims data is available internally at CMS (for Medicare/Medicaid) and payers.
  • Challenge: A complete dataset of provider-level claims are not easily available outside of payers for claims leakage analysis


Specifications

There are 3 datasets needed:

  • De-identified patient data, yet still in a format to allow for longitudinal patient tacking
  • Adjudicated claims to understand the financial impact of leakage
  • Roster of MSSP ACO and Pioneers ACO providers by NPI Number


De-identified patient data

  • Purpose: Longitudinal patient tacking
  • Description: Roster of MSSP ACO and Pioneers ACO providers by NPI Number
  • Fields: _______
    • Patient ID: Need an arbitrary pseudoidentifier to allow for longitudinal patient tacking
    • ______(needs to be filled out)____
  • Update frequency: _______
  • Joins between datasets: _______

Adjudicated claims

  • Purpose: To understand the financial impact of leakage
  • Description: Roster of MSSP ACO and Pioneers ACO providers by NPI Number
  • Fields: _______
    • Provider NPI Number
    • ______(needs to be filled out)____
  • Update frequency: _______
  • Joins between datasets: _______

MSSP ACO and Pioneers ACO providers

  • Purpose:
  • Description: Roster of MSSP ACO and Pioneers ACO providers by NPI Number
  • Fields: _______
    • Provider NPI Number
    • ______(needs to be filled out)____
  • Update frequency: _______
  • Joins between datasets: _______

Solution

Workaround

"...ACO members are required to release their providers on their websites.", per Fred Trotter

  • However the NPI field is missing from such data.
  • Ultimately, there would be a benefit from CMS releasing NPIs belonging to each ACO with each Monthly CCLF (Claims and Claims Line Feed) file release. (Source: @FTrotter)

Solution #1: CMS's referral data

Warning: Potential data problem as of 10/5/2015

DocGraph had a recent data retraction associated with the Teaming dataset mentioned here.
Details and possible fix is posted at http://www.docgraph.org/notice-teaming-data/.

CMS now has physician referral data available at https://questions.cms.gov/faq.php?faqId=7977 that in essence shows referrals from one physician to another.

  • It was provided as a response to FOIA requests and analytics efforts by Fred Trotter at DocGraph.
  • It works on the following definition of referral:
A referral occurs when an organization or provider participating in the delivery of health services to the same patient in up to one-month period, after another organization or provider participated in providing health services to the same patient.


Files for download

  • Physician Referral Patterns - 2009  zip
  • Physician Referral Patterns - 2010  zip
  • Physician Referral Patterns - 2011  zip
  • Physician Referral Patterns - 2012 - 2013 30 day interval  zip
  • Physician Referral Patterns - 2012 - 2013 60 day interval  zip
  • Physician Referral Patterns - 2012 - 2013 90 day interval  zip
  • Physician Referral Patterns - 2012 - 2013 180 day interval  zip
  • Physician Referral Patterns - 2012 - 2013 365 day interval  zip
  • Physician Referral Patterns - 2013 - 2014 30 day interval  zip
  • Physician Referral Patterns - 2013 - 2014 60 day interval  zip
  • Physician Referral Patterns - 2013 - 2014 90 day interval  zip
  • Physician Referral Patterns - 2013 - 2014 180 day interval  zip
  • Physician Referral Patterns - 2013 - 2014 365 day interval  zip

Solution #2: DocGraph

Various versions of referral data is also available from commercial and open source entities. One example is DocGraph / CareSet. For those who don't want to do perform manual manipulation of text files, these data sources provide data that is cleaned up, pre-processed, and made accessible in a way that's easier to work with.

Want to be notified when this use case or any of its datasets change?