Two Clinics, One Plan of Care, Very Different Totals. Where the Number Is Actually Set

A regional physical therapy group with eleven clinics ran a comparison after a patient complaint, and the finding was awkward enough that they kept it internal for a month. Two locations had produced good-faith estimates for what was, on paper, the same plan of care: a lumbar strain, twelve visits over six weeks, the same evaluation code, the same discharge criteria. One estimate came in roughly a third higher than the other. Nothing had been billed incorrectly, no clinician had padded anything, and both estimates would have survived an audit. The plans of care were identical. The totals were not.

What separated them was a set of decisions nobody had described as pricing decisions. At the higher clinic, the scheduling template ran forty-five-minute slots, which meant more timed units billed per visit than the sixty-minute template used at the other site, where two patients overlapped and manual therapy time was shorter. The front desk at one location collected a deductible estimate from the clearinghouse eligibility response; the other used a printed table last revised the previous plan year. Two schedulers, one template committee, one benefits check. The clinicians who wrote the plans of care had touched none of it.

The plan of care is not the estimate, and different people own each

Practitioners are trained to think in courses of treatment: an evaluation, a working diagnosis, a frequency, a reassessment point, a discharge standard. That is a clinical object, and it is the clinician's to hold. The estimate is a financial object built on top of it, and inside any organisation larger than a single-provider office it is assembled by people with different job titles, different software, and different incentives. Revenue cycle owns the payer terms. Operations owns the visit template. The front desk owns what the patient is told. When the total surprises someone, the surprise almost always originates in a handoff between two of those owners rather than in the clinical judgment underneath.

The first discipline, then, is to stop asking what a course of treatment costs and start asking which role changed the number last. In the physical therapy group, the answer was the template committee, a quarterly operations meeting that had adjusted slot lengths at four clinics to improve throughput. The change was defensible on access grounds and had been evaluated on access grounds alone. Nobody at that table had modeled what it did to units per visit, and therefore to the twelve-visit total a patient with a high-deductible plan would actually pay out of pocket before benefits engaged.

Four variables move the total, and only one of them is clinical

Visit count is the variable everyone watches, because it is the one the clinician writes down and the one the payer authorizes. It is also the least volatile in practice, since most plans of care land in a familiar range for a familiar diagnosis. The content of each visit moves the number harder. Timed codes, units, whether a modality is delivered by the licensed provider or an aide, whether two patients share a treatment hour: these are scheduling and staffing choices, and they change the per-visit charge without changing a word of the clinical documentation. A group that standardizes plans of care but not templates has standardized the wrong layer.

The third variable is the patient's own coverage position at the moment care starts, which is why the same plan produces wildly different patient responsibility across two people sitting in the same waiting room. Deductible remaining, coinsurance, visit limits, whether the clinic is in network under the specific product rather than the carrier's name on the card. The fourth is drift: cancellations, holidays, a reauthorization that arrives late and pushes three visits into a new plan year with a fresh deductible. Drift is the one that turns an accurate estimate into a wrong one, and it is nobody's job unless somebody is assigned it.

Build the judgement into a role, not into a person's memory

The group's fix was procedural rather than exhortatory. They wrote a one-page estimating protocol that named, for each of the four variables, the source of truth and the person accountable for it. Eligibility came from the live payer response and never from a printed table. Units per visit came from a template-to-units mapping maintained by operations and republished whenever a slot length changed anywhere in the group. The estimate itself was generated by the front desk from those two inputs, with a required reforecast at the reassessment visit, which is where a course of treatment usually gets longer or shorter in reality. One page. Four owners.

They also changed what went to the template committee. Any proposed change to slot length or staffing mix now arrives with a line showing its effect on units per typical visit, so the throughput argument and the patient-cost argument get made in the same room at the same time. That single addition is what converts scattered institutional knowledge into judgement an organization can actually exercise, because it puts the trade-off in front of the group that holds the decision instead of leaving it to be discovered by a patient six visits in. The Centers for Medicare and Medicaid Services oversees the fee schedules and price transparency requirements that shape how these estimates must be built and disclosed, and a protocol written to that standard tends to survive contact with commercial payers too.

Then decide who the estimate is actually for

Practitioners often assume the estimate exists to inform the patient, and it does, but inside a larger provider it does more work than that. It is the document a scheduler uses to decide whether to book six visits or twelve. It is what a case manager reads before approving continued care. It is what a referring physician's office hears when a patient calls back to say the course sounded expensive, and it is what determines whether that referral pattern continues. Written well, it protects the clinical plan from being quietly shortened by financial anxiety that nobody voiced.

Which points at the question worth putting to your own organization. When a patient decides to complete a course of treatment or stop at visit five, whose number are they responding to, and who inside the building last changed it? In most groups the honest answer names a scheduling template, a benefits table, or a quarterly operations decision, none of which sits with the clinician who owns the plan of care. Naming those owners out loud, and giving each one a source of truth, is the whole of the work.

The eleven-clinic group now reconciles estimate to final patient responsibility on a sample of completed courses each month, and reads the variance as a signal about handoffs rather than a billing error. Where the two diverge, someone can say which of the four variables moved and who moved it. That is a smaller ambition than pricing accuracy and a far more useful one, because it makes the number explainable to the person paying it.