Attribution Modeler

My Folder

Product Overview

Attribution Modeler is aimed at clinical directors, operational directors, clinical-program guidance teams, and operational guidance teams through implementation in Health Catalyst applications. For patient populations of interest Attribution Modeler provides attribution based on configurable methods, weighting, scoring and ranking.


  • Configurable patient populations, attribution methods, weighting, scoring and ranking.
  • Summary table listing all providers for each patient, with rank and confidence percent.
  • Override capability on attribution, to activate or inactivate attributed values as needed.


Benefits Include:

  • Ability to attribute patients to providers.
  • Ability to attribute patients to locations.
  • Ability to see all providers or locations associated with patient’s care.
  • Ability to override attribution.


Attribution Methods (Configurable)

  • Weighted Physician Score: Highest Weighted Score based on a 12 month look back period. PCP provider gets 20 points per visit, Specialist Provider gets 15 points per visit, the most recent visit get 10 points, and every Order placed by an Ordering Provider gets 1 point.
  • Dartmouth Method: Patients are assigned a provider based on the most visits in the highest priority specialty. Each provider is classified as (any visit to a higher priority specialty trumps the lower): 1-PCP providers 2-medical specialists 3-surgical specialists.
  • Medicare Method: Patients assigned a provider based on charges of 1) PCP service (internal medicine, general practice, family practice, and geriatric medicine 2) Specialist Provider and certain non-physicians (nurse practitioner, clinical nurse, and physician assistant).
  • Weighted Location Score: Highest Weighted Score based on a 12 month look back period. Each visit gets 1 point, with most recent visit getting extra 2 points.

Starting Populations of Patients (Configurable)

  • Multiple Encounters: Pulls max encounter timestamp per patient where provider primary specialty name in value sets specified. Outpatient only, patient list output.
  • Medicare Patients: Pulls max discharge timestamp per patient where financial class = ‘Medicare’ and total charge amount > 0. Inpatient only, patient list output.

See Sample Screenshots of Attribution Modeler

Data Sources

  • Sources that feed the Essentials Layer, such as the EMR(s), as Attribution Modeler feeds from the Essentials Layer.

Attribution Modeler: A Deeper View


In a healthcare system looking for potential areas of improvement, it often helps to start with the attribution of patients to providers for accountability of care and outcomes.

What types of problems does Attribution Modeler address?

Healthcare organizations are seeking to hold providers accountable for the care of the patients attributed to them, but most patients will see many providers in any given timeframe so identifying the provider with main responsibility for the patient’s care can be a difficult task.

Anticipated Improvements

Develop more standard attribution methods over time for all patient settings, and National healthcare improvement initiatives.

Success Measure Examples

  • Standardize the methodology in which a provider’s patient panel is created across practices and units.
  • A standards\ transparent model generates greater buy-in with physician groups that are frustrated with limited or inadequate attribution models.
  • Attribution may be applied across different types of caregivers. Nurses, physicians, and even behavioral health specialists can all share in the patient attribution this provides points of contact to coordinate care between providers