Contribution Margin per Visit: Finding the Caseload That Pays

Picture two clinicians in the same practice, both booked solid, both finishing their notes on time. One of them generates most of the money that pays the rent, the biller, and the owner. The other one — through no fault of her own — roughly breaks even, because her schedule happens to be dense with the payer that allows the least and the visit type that eats the most unbilled time. A full schedule is an attendance achievement, not a financial one. Whether the practice thrives is decided by which visits fill the schedule, and most owners have never computed the number that answers it: the contribution margin of each visit type, under each payer, per hour of clinician time. This article builds that matrix from records you already have, then walks through what to do with the answer — including the version of the answer nobody wants, where the payer sending you the most patients is the one paying you the least.

Every margin analysis dies at the first input if you feed it the wrong revenue number. The fee on your charge master is fiction; the allowed amount in the contract is closer but still optimistic. The number that belongs in the matrix is the expected collection per visit: what a visit of this type, billed to this payer, has actually deposited on average — payer portion plus the share of patient responsibility you genuinely collect — measured from your own remittance and posting history. That one number quietly absorbs your denial rate with that payer, your patient-collection curve, and every payment rule the contract applies before the check is cut.

The payment rules are worth understanding rather than just absorbing, because two of Medicare’s are large enough to reshape a per-visit average on their own — and many commercial payers run their own versions of both, which only your remittances will reveal. First, the multiple procedure payment reduction: when more than one “always therapy” procedure or unit is furnished to the same patient on the same day, Medicare pays the practice-expense portion of the highest-valued procedure in full and cuts that portion to 50 percent for the rest. A visit billed as three timed units does not collect three times the single-unit amount. Second, the assistant differential: Part B pays 85 percent of the otherwise applicable amount for PT and OT services furnished in whole or in part by a PTA or OTA — flagged with the CQ or CO modifier — so a schedule staffed heavily with assistants earns a different per-visit number than the same schedule run by therapists. Neither rule is a complaint; both are arithmetic your matrix has to carry.

Private pay looks refreshingly simple next to all of this — the rate is the collection, minus whatever you discount and fail to collect — which is exactly why the matrix matters: it puts insurance visits and private-pay visits in the same units for the first time, instead of letting the clean number and the murky one live in separate mental accounts. If you are still deciding what those private rates should be, that is its own analysis, covered in how to set private-pay rates; this article takes whatever rates you have and asks what each visit contributes.

Contribution margin asks a narrower question than the profit-and-loss statement: if this visit happens, what does it collect, and what does delivering it cost? Only the costs that scale with the visit belong on the cost side — everything else is fixed overhead the margins collectively have to cover. In a therapy practice the variable cost is dominated by one line: paid clinician time, fully loaded with payroll taxes and benefits, counted for everything the visit consumes. That is the session itself, plus the documentation it obligates, plus for evaluations the scoring and report writing that happen after the family leaves. A 45-minute treatment slot with a 15-minute note is an hour of clinician cost; if your notes routinely take longer than that, a documentation time audit will change your margin numbers before any payer negotiation does.

The remaining variable costs are smaller but real: consumable materials, the per-claim cost of billing — a clearinghouse fee, a percentage if you outsource, a slice of the biller’s time if you don’t — and card processing on the patient share. What does not belong here: rent, the front desk, software subscriptions, your own owner draw. Those are fixed, and loading them onto individual visits at this stage buries the comparison you are trying to make. The matrix first tells you which visits contribute most per hour; whether the total contribution covers the fixed costs is the practice-level question that follows, and it connects directly to the cash-flow model and the KPI dashboard if you already run those.

The whole analysis is one spreadsheet: a row for each combination of visit type and payer that occurs on your schedule, and six passes over it. A solo practice has maybe eight rows; a multi-discipline group might have thirty. Expect the first pass to take an afternoon and every later refresh to take an hour, because the structure survives — only the numbers move.

1. 01

   Lay out the matrix: one row per visit type, per payer

   List every visit type you deliver — evaluation, individual treatment, group, teletherapy if it pays differently — and cross it with every payer, private pay included. Resist the urge to average across payers “to keep it simple”: the entire finding of this analysis usually lives in the spread between payers for the same visit type, and averaging erases it before you can see it.

2. 02

   Fill in expected collection per visit, from remittances

   For each row, divide total actual collections — payer payments plus collected patient share — by the number of visits, using your twelve months of postings. Do not use the contract rate and do not use billed charges. This average already carries your denial rate, the multiple-procedure reduction, assistant-modifier reductions, and the fraction of patient responsibility that never arrives, which is exactly why it is the honest input. For rows too rare to average, price them from the contract and mark them as estimates.

3. 03

   Assign the variable cost per visit

   Cost each row as clinician hours consumed times fully loaded cost per hour, plus the small per-visit items: materials, per-claim billing cost, card fees on the patient share. Hours consumed means session plus documentation, and for evaluations the report; use measured times where you have them and honest estimates where you don’t. If different clinicians deliver the same row at different costs — a senior therapist versus an assistant-delivered visit — split the row, because the assistant row also collects differently under the 85 percent rule.

4. 04

   Compute margin per visit, then normalize to the clinical hour

   Contribution margin is expected collection minus variable cost. Then divide by the clinician hours the visit consumes, and rank the rows by margin per hour. This normalization is the step most practices skip and the one that changes conclusions: a $150 evaluation that consumes two and a half hours with its report contributes less per hour than a $90 treatment visit that consumes one, and no per-visit view will ever show you that.

5. 05

   Weight by volume to build the margin map

   Multiply each row’s margin per visit by its annual visit count. Now you have two rankings that rarely agree: where your hours go, and where your margin dollars come from. The gap between them is the finding. A payer can be 40 percent of your schedule and 15 percent of your contribution — which means the practice is, in effect, spending its scarcest resource subsidizing its weakest contract.

6. 06

   Decide the move for each quadrant, then refresh annually

   Sort rows into four quadrants by margin per hour and volume. High margin, high volume: protect it — these slots get priority when the schedule is tight. High margin, low volume: grow it deliberately through intake priorities and referral development. Low margin, high volume: this is the payer conversation, covered below. Low margin, low volume: simplify or stop, unless it serves a clinical mission you are choosing on purpose — a legitimate choice, but make it knowingly. Re-run the matrix when contracts renew, when you hire, or annually, whichever comes first.

This is the result the matrix produces most often, and it is uncomfortable precisely because the payer earned that volume honestly: referrals flow, families depend on the coverage, and the panel probably carried the practice through its first years. The wrong responses are the two reflexive ones. Dropping the panel overnight destabilizes half the caseload, breaches the continuity your patients were promised, and may violate the contract’s notice provisions. Doing nothing quietly commits your clinicians’ hours to subsidizing your weakest contract indefinitely. The workable path runs between them, in escalating order of commitment.

First, audit before you act: confirm the low margin is the payer’s doing and not yours. A high denial rate, slow-motion authorizations you absorb as unbilled admin time, or uncollected patient responsibility can make a mediocre contract look like a terrible one — and those you can fix without a negotiation. Second, renegotiate with the volume as the opening fact, not the grievance; a practice delivering thousands of visits a year is a network-adequacy asset, and the mechanics of that conversation are covered in negotiating payer contracts. Third, rebalance at the margin rather than the core: intake priority, referral development, and schedule design shift the mix a few percent per quarter with no harm to existing patients — pointed at whichever rows your own matrix ranked highest, not at a rate someone quoted in a forum. Only after those are exhausted does panel exit belong on the table, and if it reaches the table, the private-pay-versus-insurance analysis is the framework for deciding it — with the contract’s notice period, patient transition obligations, and your state’s continuity-of-care expectations mapped before anything is signed.

The analysis only earns its afternoon if it changes standing decisions — otherwise it is trivia with a spreadsheet. Three places it should land. Intake: when the waitlist is long, its order should reflect the margin map as well as clinical urgency and fit, because “first come, first scheduled” is an accidental policy of growing whichever payer refers fastest, and a deliberate approach to waitlist management can hold both goals at once. Scheduling: protect slots for the visit types the per-hour view ranked highest — in most practices that means defending evaluation capacity, since evals gate every downstream treatment plan anyway. Hiring: the matrix prices a hire honestly, because a new clinician fills whichever rows intake feeds them; model the offer against the mix they will actually inherit — including the 85 percent assistant differential if the role is a PTA or OTA — not against your best contract.

One caution against overcorrecting: contribution margin is one axis of a practice, not the practice. A row can earn its place clinically — the Medicaid panel that is the reason you opened, the complex evaluations nobody else in the county does — and choosing to keep low-margin work is a legitimate act of mission, provided the rest of the matrix knowingly funds it. The failure mode this article exists to prevent is not low-margin work; it is unexamined low-margin work, discovered years late, after it has silently set the practice’s salaries, prices, and stress level.

“

“A therapy practice does not run out of patients or purpose. It runs out of clinician hours — and the only question the margin matrix asks is whether each of those hours is funding the mission or quietly taxing it.”