Private Pay or Insurance: Run the Numbers Before You Pick a Side

Ask whether a therapy practice should be private pay or insurance-based and you will mostly hear identity arguments: independence and simplicity on one side, access and steady referrals on the other. Both sides argue from the two numbers that matter least — the cash fee you could charge and the contracted rate you would accept. Neither number is what you take home. What you take home is what each model actually collects, divided by every hour the model consumes: the sessions, the documentation, and the very different quantities of administrative work each one demands to turn a visit into money. This article works that comparison the way an accountant would, ending in a single calculation — break-even fill — that converts the whole debate into arithmetic you can run with your own numbers this week.

A private-pay fee of $140 and a contracted allowed amount of $95 look like a $45 argument for private pay, and that is where most comparisons stop. But the sticker numbers describe different products. The insurance rate arrives with a referral stream attached: families find you through the payer directory, a physician sends patients because you take their coverage, and the schedule fills with little marketing effort. It also arrives with a collections apparatus you must staff — verification before the first visit, claims after every visit, posting, denials, and patient balances after the plan pays its share. The private-pay fee arrives with none of that overhead and none of that demand. You collect at the time of service, and you also recruit every single family yourself, at full price, in a market where most of your competitors accept insurance.

So the honest comparison is not $140 versus $95. It is two complete systems, each with its own collection rate, its own admin load, and its own referral physics. Practices in the same discipline land on opposite answers because those three variables differ by market: a pediatric practice in a suburb where families expect to use their benefits faces different physics than an adult-rehab practice near employers with thin therapy coverage. The method below exists precisely because the answer is local.

Before comparing models, measure what your practice — or your model of the practice you are planning — actually produces. All four numbers come from records you already have: remittances, your calendar, and whatever you use to track referrals. If you are pre-launch and have no records, build each number as a range from the payer fee schedules you have been offered and the fees comparable local practices post publicly, and treat the pessimistic end as the planning case.

1. 01

   Collected dollars per visit, by payer

   Not the billed charge and not the contracted rate — what lands in the bank per visit after denials, underpayments, and patient-balance leakage. Pull three recent months of remittances per payer, divide total collected by total visits. For the private-pay side, the same number is your fee times the share of sessions you genuinely collect for, which card-on-file policies push close to full.

2. 02

   Admin minutes per visit, by payer

   Count every touch that exists only to get paid: eligibility verification, authorization requests, claim preparation, rejection rework, payment posting, statements, and phone calls about balances. Attribute the time per visit, and count it wherever it happens — owner evenings are still hours. Private pay has its own line here too: invoicing, receipts, and superbill preparation for families who file out-of-network claims themselves.

3. 03

   Days from visit to cash, by payer

   The lag between delivering a session and being able to spend the money it earned. Time-to-cash does not change what a visit is worth, but it decides how much operating reserve the model forces you to hold, and a payer that pays slowly is quietly charging you the cost of floating your own payroll.

4. 04

   Referral volume each model can sustain

   The hardest number and the decisive one. For insurance: how many of your current referrals arrive because you are in-network — directory lookups, physician offices that filter by coverage, families who would not or could not pay cash. For private pay: how many families per week your non-payer channels — word of mouth, schools, your website, self-referral — actually deliver at full fee. Count referral sources for a month before trusting your intuition; owners are reliably wrong about this one in both directions.

With those four numbers in hand, the comparison stops being ideological. The table below is the full ledger — what each model earns, what it costs to run, and what it demands of you that the other does not. Read it with your own numbers in the margins, not as a scorecard where one column is supposed to win. The last two rows are the ones sticker-price comparisons always omit, and they are usually where the decision actually gets made.

Notice what the ledger does not contain: a universal answer. Every row is a variable, and three of them — collected per visit, admin minutes, and sustainable referral volume — combine into one decisive figure. That figure is the subject of the worked example below.

Run Maya’s three steps with your own records and the debate usually collapses into something much less dramatic than the identity argument: a per-panel decision, made annually, with a number attached to each panel. The models are not rivals; they are line items.

Collection rates and admin minutes are measurable and fixable. Referral volume is neither quick to build nor honest to estimate, which is why it deserves the most skepticism in your model. In-network status is, functionally, a marketing channel someone else operates: the payer’s directory works around the clock, and every physician front desk that filters referrals by coverage is routing patients toward the practices on the list. Leave the network and that channel does not shrink — it disconnects. What remains is what you have built independently: referrers who send because of outcomes rather than coverage, schools and preschools that know your work, families who tell other families, and whatever your website earns. Those channels are excellent and slow. They compound over years, not quarters.

This is also why the transition, when practices make it, is rarely a leap. The common path is sequenced: measure per-panel economics, drop the single worst panel, hold fill steady with strengthened direct channels, and only then consider the next panel. Each step tests the referral question with real stakes but survivable downside. A practice that terminates every contract at once is betting the entire schedule on channels it has never had to rely on exclusively — and discovering the answer with no way back except re-credentialing, which takes months.

One payer sits outside the whole framework. Physicians and a defined list of other practitioner types may formally opt out of Medicare and privately contract with beneficiaries — but physical therapists, occupational therapists, and speech-language pathologists are not on that list. For covered outpatient therapy services delivered to a Medicare Part B beneficiary in private practice, federal law requires the therapist to enroll in Medicare and submit the claim; in most circumstances you may collect only the applicable deductible and coinsurance from the patient. “We are private pay” is not a lawful answer to a Medicare beneficiary seeking a covered service from a therapist who has simply not enrolled.

Dropping payers removes claims, not obligations. Under the No Surprises Act, providers must give uninsured and self-pay patients a good faith estimate of expected charges when a service is scheduled or when they ask for one — and if the eventual bill runs at least $400 over the estimate, the patient can take it to a federal patient–provider dispute resolution process. For a therapy practice this is close to free to do well: episodes are predictable, so an estimate covering the evaluation and the expected visit cadence, refreshed when the plan of care changes, satisfies the requirement and doubles as the financial conversation that prevents disputes in the first place. Families filing their own out-of-network claims will also ask for superbills, so a clean, itemized receipt process is part of the private-pay admin line, not an afterthought.

The binary question — private pay or insurance — dissolves once the numbers exist, because the numbers are per payer. A panel that collects well, pays promptly, and feeds the schedule is worth its admin cost. A panel that underpays, authorizes grudgingly, and consumes appeal hours is a bad client you have mistaken for infrastructure. Review the ledger once a year, per panel, and act on the worst line only. Before any contract decision, assemble the file:

“

“Payers are not an identity. They are line items — and line items get reviewed annually, kept when they earn it, and dropped one at a time.”