Attribution: Whose Patient Is It?

Every population-based arrangement rests on a list: the patients this organization is accountable for. Benchmarks, savings, and quality scores are all computed on the list. Yet most patients never formally choose a primary care provider. Attribution is the rule set that turns messy reality into a definitive list.

The main methods

Method	How it works	Weakness
Claims-based	Assign to whoever delivered the plurality of primary care	Thin evidence for scattered utilizers; no-visit patients drop out entirely
Voluntary alignment	Patient designates a clinician; designation trumps claims	Low response rates; works as an overlay, not a replacement
Enrollment-based	Every member assigned from day one (Medicaid managed care)	The assignment may be an administrative default, not a relationship

Prospective or retrospective

• Prospective: the list is fixed before the year starts, from prior-year claims. Providers know exactly who to manage from January, but the list goes stale as patients switch doctors.
• Retrospective: the list is built after the year ends from where care actually happened. Accurate, but providers spend the year accountable for a shifting cloud they cannot fully identify.

Hybrids (prospective list, retrospective reconciliation) are common. The trade-off is fundamental: certainty versus accuracy, and no rule gets both.

Why the plumbing moves millions

• The denominator drives the results. Which patients count changes measured spending, quality, and savings directly. Excluding patients without recent primary care visits flatters results with zero change in performance, while conveniently removing many of the hardest-to-reach people from everyone’s denominator.
• Attribution invites selection. If attribution follows visits, then who gets appointments edits the denominator. Easy access for engaged, healthier patients and friction for complex ones is quiet self-selection.
• Churn erodes investment. Care management is built around the attributed list; patients who churn off before longitudinal work pays off are one of the most underappreciated reasons interventions underperform their pilots.
• Primary care gets the slot. A patient seeing one PCP, three specialists, and two urgent care centers has one attribution slot. It goes to primary care, the closest thing the system has to a coordinating center.

Worth remembering: there is no correct attribution rule, only trade-offs. When a value-based result looks surprising in either direction, attribution is one of the first places to look.

Key takeaways

• Attribution builds the accountability list every other number depends on.
• Claims-based, voluntary, and enrollment methods trade accuracy, intent, and cleanliness.
• Check attribution rules first when results look too good or too bad.