The seven-category PT billing company scorecard
Print this, take it into every call, and score each vendor while they are still in the room. The weights are a starting point. Move them to match what actually hurts in your practice, then keep them fixed across every vendor so the comparison means something.
| Category | Weight | Evidence to request | What a 5 looks like | Risk signal, score it 1 | Score 0 to 5 |
|---|---|---|---|---|---|
| PT specialty controls | 20 | Live walkthrough of timed units, POC, GP, KX, CQ and authorization logic. | The team works a synthetic 53-minute visit out loud, names the facts it needs, applies the rule and shows where the exception surfaces and who owns it. | Only generic medical billing slides. | |
| Denial prevention | 18 | Root-cause taxonomy and examples of upstream rule changes. | They show a denial reason, the front-end control they changed because of it, and the recurrence rate afterwards. | Focus only on appeal volume. | |
| Daily visibility | 16 | De-identified queues with owner, age, reason and next action. | You get a login. The queue refreshes from the source system, and every line has an owner, an age, a precise reason and a next action. | Monthly totals without work-level detail. | |
| Transition | 14 | Named milestones, validation, parallel checks and acceptance criteria. | A written plan with a parallel run, a data validation step, an open-AR split and acceptance criteria you could hold them to. | A date without a control plan. | |
| Systems fit | 12 | Access and integration map for your current tools. | They map your EHR, practice-management system, clearinghouse, portals and reports before proposing any change, and separate required integration from optional replacement. | Migration assumed before discovery. | |
| Security | 12 | Access model, secure transfer, incident process and BAA pathway. | A named secure transfer path, least-privilege roles, a documented incident process and a BAA offered before any PHI moves. | Requests PHI through public channels. | |
| Governance | 8 | Operating cadence, escalation path, decisions and improvement log. | A named senior owner who stays after the sale, a written cadence, and a visible improvement backlog rather than recurring status slides. | No senior owner after sales. | |
| Weighted total | 100 | Multiply each weight by its score, add them up, divide by 5. That gives a number out of 100 you can compare across vendors. Change the weights before you score anyone. They are a starting point, not a standard. | |||
Two rules make the scorecard work. Score only what you were shown, never what you were told. And write the evidence next to the score, because in three weeks you will remember the impression and not the reason.
In-house, outsourced or hybrid: where each one honestly wins
Before you evaluate a vendor, decide whether you should be buying one at all. This is the section a billing company has no incentive to write, which is exactly why it belongs here.
When outsourcing is the wrong answer
There are four situations where hiring a billing company will not help, and may make things worse.
- The workflow itself is undecided. If your own team cannot say who verifies benefits, who owns an expiring authorization or when a charge is entered, no vendor can inherit a process that does not exist. They will invent one, and it will not be the one you wanted.
- Nobody internally will own the relationship. Outsourcing moves the work, not the accountability. If no named person on your side reviews the queue, chases decisions and answers the partner's questions within a day, the engagement decays quietly.
- The real problem is documentation. If notes do not support the minutes, or plans of care are certified late, a billing company can only get better at appealing the same denial. Fix the note first.
- You are a single site with a stable, capable biller. One person who knows your five payers and sits twenty feet from the clinicians is genuinely hard to beat. The case for change is thin until volume, payer count or geography grows past what one desk can hold.
The honest version of the hybrid model deserves one more sentence. It works when the line of demarcation is written down and testable. It fails when both sides believe the other owns denials.
1. Prove physical therapy billing depth
Ask the potential partner to work through a synthetic PT visit. Include timed and untimed services, visit limits, a plan-of-care milestone, approaching KX threshold and PTA participation. The team should explain which facts it needs, which rules it applies, where an exception appears and who owns it.
This exercise quickly distinguishes specialty operations from vocabulary. The right answer is not a memorized modifier stack. It is a traceable workflow that adapts to payer, date of service, setting, provider role and documented minutes.
You can check their arithmetic against the therapy units calculator for the 8-minute rule while they talk, and the outpatient therapy revenue integrity whitepaper works the same allocation, modifier and threshold logic end to end if you want the long version first.
2. Inspect the work, not only the outcome
Summary KPIs are useful, but they cannot replace operational visibility. Request a view of held claims, expiring authorizations, unposted payments, denials, appeals and aging AR. Each item should have a precise reason, owner, age and next action.
Ask what you can see without waiting for a monthly meeting. Confirm whether the data comes from source systems or manual presentation files, how often it refreshes and how discrepancies are reconciled.
3. Test denial-prevention intelligence
A billing company can recover claims while the same defect repeats. Ask how payer responses change front-end work. A mature approach classifies denials by payer and root cause, links them to the original workflow and measures recurrence after a rule, edit or training change.
Use the PT denial and authorization playbook as an evaluation script. The partner should connect eligibility, referral, authorization scope, visit consumption, POC status, claim readiness and payer feedback.
4. Map systems before discussing migration
Your EHR, practice-management system, clearinghouse, payer portals, document repositories and reporting tools already form an operating environment. A potential partner should map access, data ownership, interfaces, exports, update frequency and control points before recommending change.
For authorization, POC, charge, payment, denial and payer rule data.
Named roles, least privilege, review cadence and termination process.
Visit-to-charge, claim-to-acknowledgment and remit-to-posting reconciliation.
Separate required integration from optional replacement.
5. Require a controlled transition plan
A strong transition defines in-scope work, open AR ownership, payer access, data validation, work queues, training, escalation, parallel checks and acceptance criteria. It also identifies dependencies such as credentialing, enrollment, bank or ERA changes and client-side decisions.
Ask for the first 30 days in enough detail that both organizations know what happens when access is late, a report does not reconcile or legacy AR conflicts with the new workflow. A transition should preserve revenue continuity and evidence, not simply move a start date.
6. Make security concrete
Confirm the secure communication and data-transfer pathway before sharing PHI. Review the proposed business associate agreement process, access controls, workforce roles, logging, incident escalation, data retention and termination steps. Public web forms should not be used to transmit patient information.
Security statements are only useful when they connect to the actual systems and team. Ask who receives access, how it is approved, how privileged actions are controlled and what evidence is available during governance.
7. Define governance before signing
Governance should state who makes operating decisions, how often teams meet, what metrics are reviewed, how escalations move and where decisions are recorded. The relationship needs a visible improvement backlog, not just recurring status slides.
- Daily: production exceptions, submission failures and urgent payer or access issues.
- Weekly: authorization risk, claim holds, denial root causes, AR priorities and actions.
- Monthly: trends, payer behavior, control changes, capacity, risks and decisions.
- Quarterly: strategy, automation opportunities, scope and improvement roadmap.
What each pricing model rewards, and what it quietly punishes
Pricing is not a negotiation detail. It is the shape of the incentive you are buying, and the three common shapes behave very differently as you grow. The chart below uses illustrative inputs so you can see the shapes. Replace them with your own numbers before you conclude anything.
What it rewards. Getting paid. The partner earns nothing on a claim that never pays, which aligns the incentive better than any clause in the contract.
What it punishes. You. Every fee-schedule increase and every high-dollar payer raises the bill without any change in work. Ask what happens to the rate as volume grows, and get the answer in writing.
Watch for. Whether the percentage applies to patient payments, to credits and refunds, and to legacy AR you brought with you.
What it rewards. Predictability. Your cost per claim is a number you can budget, and a rich payer mix does not cost you more.
What it punishes. Persistence. Nothing in the fee makes anyone want to work a claim a fourth time. If you buy per-claim, buy an AR service level next to it.
Watch for. Whether a resubmission counts as a new claim, and whether denials, appeals, patient statements and posting are inside or outside the fee.
What it rewards. Control and proximity. The biller is in the building and can walk to the clinician.
What it punishes. Small volume and single points of failure. Below the volume where one salary is fully used, this is usually the most expensive option per claim, and one resignation removes the whole function.
Watch for. The loaded cost, not the salary. Benefits, payroll tax, software, clearinghouse fees, training, supervision and coverage for leave all belong in the number.
The comparison that actually matters is cost per collected dollar at your volume, in three years, not this month. Build the arithmetic yourself with your own claim count, your own average collected amount and a loaded FTE cost that includes software and coverage. Any vendor unwilling to model their fee against your growth curve has told you something useful.
Red flags, and what to ask instead
None of these is proof of a bad partner. Each one is a place to stop and ask a better question, and the answer is usually more informative than the flag.
Nobody can guarantee a percentage without knowing your payer mix, fee schedule, documentation quality and denial history. A guarantee is a sales device, and the exclusions usually make it unenforceable.
Ask instead: what is your net collection rate on outpatient therapy books like ours, how is it calculated, and what would make it lower here?
If the first data request is casual about how patient information moves, the operating discipline behind it is not better.
Ask instead: what is your secure transfer path, and can we sign the BAA before we send you anything at all?
A partner that recommends replacing your EHR or practice-management system before mapping it is selling a product, not diagnosing a problem.
Ask instead: which parts of our current environment actually block your workflow, and what is the smallest change that unblocks it?
The people who answer the specialty questions in the sales meeting frequently are not the people who will work your account.
Ask instead: who will be named on our account, what else do they work on, and can we meet them before we sign?
If numbers arrive as a monthly presentation, they were assembled by hand. Hand-assembled numbers cannot be reconciled to the source system, and they always arrive after the moment you needed them.
Ask instead: what can we see ourselves, how often does it refresh, and does it come from the source system or from a spreadsheet?
Legacy AR is where transitions go wrong. If the split is vague at the proposal stage, it will be a dispute in week eight.
Ask instead: exactly which dates of service do you own, which stay with us, who works the denials on the boundary, and how is that reported separately?
Logos on a slide are not references. A named contact at a comparable therapy practice is.
Ask instead: can we speak to a therapy client of similar size, and to one that left you?
If the team hedges on how remainder minutes are allocated across codes, they will hedge on your claims too. This is the single fastest specialty test there is.
Ask instead: walk us through 18 minutes of one timed code and 20 minutes of another. Then check the answer against the therapy units calculator for the 8-minute rule.
What a 90-day transition actually looks like
The three markers above the timeline are the gates that decide whether a transition works. Every failed migration this pattern is drawn from failed at one of them, and every one of them is visible weeks before the damage lands.
Two things are worth insisting on. Capture a baseline in week one, before anything changes, or you will spend month four arguing about whether performance improved. And do not let the parallel run be shortened because the go-live date is fixed. The parallel run is the only place where a defect costs nothing to find.
Questions to ask on the call
Eight questions, with what a strong answer sounds like and what a weak one sounds like. Ask them in this order. The first four test whether they know physical therapy. The last four test whether they will still be useful in month twelve.
- Which PT claim issues will stop before submission, and which rely on payer denial feedback?
They name specific pre-submission edits: authorization scope, certification status, unit-to-minute consistency, modifier logic. They also admit which categories they can only catch on the remit.
Everything is described as caught up front, with no examples.
- How do you calculate and audit timed-unit allocation?
Total timed minutes to a unit count, then full 15-minute blocks per code, then leftover units to the largest remaining minutes, with a written tie convention and a sample audit.
Per-code rounding, or a description that changes halfway through.
- How do you track POC certification, authorization consumption and KX threshold status?
Three separate trackers with separate triggers, plus the correct pairing: PT and SLP share one threshold and OT has its own.
One combined status field, or a threshold model that pairs OT with SLP.
- How are GP and CQ decisions made at the claim-line level?
GP on every PT line as a discipline identifier, and CQ decided per unit against the assistant standard, not applied to the whole visit.
Modifiers applied at the claim level or from a habit list.
- What can our team see daily, and who owns each exception?
A login, a refresh interval, and a queue where each line carries an owner, an age, a precise reason and a next action.
A monthly meeting and a promise to send something over.
- How will open AR and legacy denials be divided during transition?
A date-of-service boundary, a named owner on each side, a rule for claims that straddle the line, and separate reporting for legacy versus new.
We will figure that out during onboarding.
- What requires client approval, and what can your team operate independently?
A written decision-rights list: write-offs above a threshold, refunds, payer contract questions and anything that changes clinical documentation come to you. The rest they run.
We handle everything, so you do not have to worry about it.
- What happens in month four, when the transition team has moved on?
The named senior owner stays, the cadence continues, and the improvement backlog is reviewed on a fixed schedule with dates against items.
A support inbox and a reassurance.
How ASP-RCM approaches the evaluation
ASP-RCM starts with a current-state workflow map rather than assuming a software replacement. Our physical therapy billing service is designed to connect benefit checks, authorization, POC status, timed-unit logic, modifiers, documentation, claims, payments, denials and reporting in one governed operating model.
Bring a de-identified visit or claim path to the first conversation. We will map the control points, identify gaps and outline the scope required to close them. Do not send PHI through the public website form.
If the scorecard above tells you to keep billing in-house, that is a legitimate result and we would rather you reach it honestly than switch and regret it. The physical therapy billing knowledge center is free either way.