Quick answer: PA software handles a multi-specialty, multi-payer MSO footprint by combining a centralized payer rule library that knows the payer-specific requirements for each specialty's most common CPT codes with a routing layer that pulls the right specialty-specific clinical context out of the source EHR for each request. The platform reads each PA at intake, identifies the specialty and payer rule set that applies, extracts the specialty-appropriate clinical attachments (op notes for ortho, imaging for cardiology, path reports for derm), and routes the work to the right specialty-trained coordinator at the MSO hub — without forcing every acquired practice onto the same EHR or workflow first.
Why multi-specialty, multi-payer MSO PA is qualitatively harder than single-specialty PA
PA at a single-specialty practice runs against one payer-rule library, one set of procedure codes, and one stable clinical documentation pattern. A 12-provider cardiology group works through cardiac imaging PAs, EP procedure PAs, and structural heart PAs against the same handful of commercial payers and Medicare Advantage plans the practice has worked for years. The PA coordinator team specializes, the documentation patterns repeat, and a clean workflow takes 90 days to build.
A multi-specialty MSO running cardiology, dermatology, GI, orthopedics, and an infusion clinic across acquired practices doesn't get any of that simplicity. Each specialty's PA workflow needs different clinical evidence, different payer-rule logic, and different sense of urgency. Cardiology runs imaging-heavy PAs with stress-test and rhythm-device documentation. Ortho runs surgical-procedure PAs with conservative-care documentation that's specific to each payer's medical-necessity criteria. Dermatology has biologics PAs at $30,000+ per request with payer-specific step-therapy requirements. GI runs colonoscopy and infusion PAs with specialty drug formulary considerations. The infusion suite runs high-dollar drug PAs where one missed PA means a patient sits in the chair without an approved therapy.
The same payer covers all of these specialties under different medical policies. The same procedure code can have different documentation requirements across payers. The same patient may need stacked PAs across specialties for one episode of care. The PA coordinator who's strong on cardiology cases is rarely strong on derm biologics. And the MSO is managing all of this across 5–15 acquired practices that each run a different EHR.
The 2024 CAQH Index pegs manual PA at $10.97 per request versus $5.79 fully electronic — but those numbers assume a workflow that's roughly the same across requests. At a multi-specialty MSO, the cross-specialty variance is what kills the math. PA software built for this environment has to absorb the variance at the platform layer rather than letting it land on the central coordinator team.
Why a single payer's PA criteria differ by specialty
The least-discussed but most important piece of multi-specialty MSO PA work is that the same payer maintains different medical policies for different specialties. The PA platform has to absorb this complexity rather than asking the central PA coordinator to know it.
A representative pattern across major commercial payers:
- Orthopedic surgery PAs typically require evidence of conservative care (typically 4–6 weeks of physical therapy and/or anti-inflammatory medication), imaging confirming structural pathology, and provider documentation supporting medical necessity. Payer-specific criteria vary on which procedures require PA and what conservative-care duration is required.
- Cardiology imaging PAs typically require evidence of clinical indication (chest pain, suspected coronary artery disease, structural concern), prior diagnostic workup, and provider documentation of the imaging study's role in management decisions. Each payer maintains specific medical policies on advanced cardiac imaging (cardiac MRI, cardiac CT, nuclear stress) versus standard echocardiography.
- Dermatology biologic PAs typically require evidence of disease severity (BSA involvement, PASI score, quality-of-life impact), failed prior therapies (topical, phototherapy, conventional systemics), and provider attestation to ongoing assessment. Each payer maintains step-therapy requirements that vary by biologic.
- GI infusion PAs typically require disease activity documentation (endoscopic findings, lab markers, symptom scoring), failed conventional therapy, and provider rationale for the specific biologic. Payer-specific criteria vary on which biologics are first-line versus reserved for treatment-resistant disease.
- Oncology infusion PAs typically require staging documentation, NCCN-aligned treatment rationale, prior-line documentation when relevant, and patient-specific clinical context. Most major payers cover oncology PAs at higher rates than other specialties but require deeper clinical documentation.
A multi-specialty PA platform absorbs this complexity in the payer-rule library. The central coordinator handling a derm biologic PA doesn't need to know cardiology imaging policy; the platform applies the right rule set automatically based on the PA's content.
How AI extracts specialty-appropriate clinical attachments
Once the platform knows which specialty's rule set applies, the next step is pulling the right clinical context out of the source EHR. This is where the AI extraction layer earns its keep at multi-specialty scale.
The pattern that works reads the EHR for the patient's recent clinical history, identifies the specialty-relevant attachments, and assembles them into a payer-specific request package. Different specialties need different attachments:
- Orthopedic cases need op notes, imaging reports, prior PT documentation, and conservative-care history.
- Cardiology cases need imaging reports (echo, stress, cardiac MR/CT), EKG tracings, lab work (lipids, BNP), and prior cardiac history.
- Dermatology cases need path reports (when applicable), photographs of affected areas (when applicable), prior topical and systemic therapy history, and disease severity documentation.
- GI cases need endoscopy reports, biopsy results, prior medication history, and symptom scoring.
- Oncology cases need staging documentation, tumor pathology, prior treatment lines, and current performance status.
The AI does this without manual chart navigation by the PA coordinator. What used to take 20–40 minutes of chart-pulling per PA at single-practice scale collapses to seconds of AI extraction with review for accuracy. The central coordinator's role shifts from chart navigation to exception handling and clinical-context review.
Honey Health's Prior Authorization agent is built around this specialty-aware extraction pattern. The agent reads the source EHR for each PA, identifies the specialty's required attachments, and assembles the payer-specific request package — across athenahealth, Epic, eClinicalWorks, NextGen, ModMed, and the long tail of specialty EHRs through desktop automation.
Routing PAs to the right coordinator at the MSO hub
The central PA pod at most multi-specialty MSOs isn't one homogeneous team. It's a small set of specialty-aware coordinators — typically one per major specialty area — who handle the cases that need their specialty's clinical context for review and escalation.
The platform's routing layer reads each PA and assigns it to the right coordinator. The routing decision considers:
- Specialty identification based on the diagnosis code, requested procedure, and clinical context — not just the originating provider's department
- Cross-specialty handling when the case spans multiple specialties (a cardiology workup that leads to an ortho consult that needs a surgical PA)
- Coordinator capacity across the network so PA work flows to available coordinators rather than queuing at the originating practice
- Escalation patterns for cases that need peer-to-peer or complex appeal work, routed to coordinators with specific specialty depth
The reporting layer surfaces per-coordinator performance, per-specialty first-pass approval rate, and per-payer rework patterns. MSO operations uses this data to identify which specialty/payer combinations drive the most denials, which coordinators are over-loaded versus under-loaded, and where additional payer-rule tuning would pay back.
Reporting roll-ups that drive operational decisions
At MSO scale across multi-specialty workflows, the reporting layer is where the platform's value proves itself or doesn't. Three reports drive most of the actionable operations decisions.
Per-specialty first-pass approval rate by payer and procedure. This is the headline metric for evaluating PA workflow quality across the MSO. When dermatology runs at 79% first-pass while ortho runs at 89%, the gap usually points to a rule-library coverage issue on a specific payer-procedure combination rather than a coordinator-skill issue. The reporting lets MSO operations target the rule library where it's leaking.
Per-site TAT distribution for routine versus complex PAs. Multi-site MSOs usually find that the same specialty performs differently at different sites — sometimes because of provider documentation patterns, sometimes because of payer-mix differences across the site's patient population. The reporting makes the variation visible and lets operations decide whether to target documentation training at the underperforming site or accept the variance.
Aged-PA exposure rolled up across the portfolio. PAs still pending past the payer's stated response window represent revenue at risk and operational drag. The MSO needs to see the aged exposure across every practice and every specialty in one number so the central team can intervene on the largest aged bucket first. This is the report that catches operational drift before it becomes a financial problem.
These reports don't exist inside any single EHR because the EHRs were built for single-site, single-specialty operations. The PA platform is the layer where the cross-specialty, cross-site, cross-payer view actually rolls up.
Practical limits worth being honest about
No PA platform — multi-specialty or otherwise — handles every case automatically. Three categories of cases consistently need human judgment at MSO scale.
Novel biologics and gene therapies without stable payer policies. When a new specialty drug launches, payers spend months building their utilization-management protocols. During that window, the platform flags cases as low-confidence and routes them to the PA coordinator for manual handling. Strong platforms surface this clearly rather than silently submitting against a stale rule library.
Experimental procedures and off-label use cases. Novel surgical techniques, investigational drugs in expanded-access programs, and off-label oncology uses don't fit standard payer rule libraries. The platform handles the workflow support, but the clinical case-by-case judgment stays with the practice's provider and the central PA coordinator.
State-specific Medicaid quirks. State Medicaid plans vary widely and update on their own schedules. Even mature payer-rule libraries occasionally lag the latest state-specific Medicaid update. The platform flags this when the rule library and the actual policy diverge, rather than asking the coordinator to discover the gap post-denial.
These limits aren't reasons to skip PA automation at multi-specialty MSO scale. They're reasons to require the vendor to be specific about exception handling, surface flag rates honestly, and design the workflow so exceptions get to the right coordinator quickly rather than getting lost.
Frequently asked questions
Can one PA platform really handle multiple specialties across an MSO?
Yes, if the platform was built for it. Multi-specialty PA software runs specialty-specific payer-rule libraries in parallel, with content-based routing that reads each PA request and applies the right specialty's logic. Single-practice tools retrofitted to multi-specialty MSO use rarely handle the cross-specialty routing well, which is why architecture matters more than feature checklist comparisons.
How does the platform handle cross-specialty referrals where multiple PAs stack up?
A working platform treats the patient — not the individual PA — as the unit of work. The platform maintains a longitudinal view of every PA in flight for that patient and triggers downstream PA work as upstream decisions land. The cardiology PA approves; the ortho consult PA fires automatically with the cardiology clinical context attached. Cross-specialty handoffs that used to require manual coordination between coordinator teams happen automatically.
What if an acquired practice's specialty isn't well-covered in the rule library yet?
Strong vendors tune the payer-rule library for the MSO's specific specialty mix during implementation. If a newly-acquired practice runs a specialty the library doesn't cover deeply yet, the vendor's implementation team should commit to building out that specialty's coverage as part of onboarding. Ask for this commitment in writing before signing.
How long does multi-specialty PA implementation take at a 10-practice MSO?
Plan for 3–6 months end to end. The first practice reaches go-live in 4–8 weeks depending on EHR pattern. Subsequent same-EHR practices layer on at a 2–4 week cadence each. New EHRs add 4–14 weeks each. The specialty-tuning phase runs in parallel during implementation, with each newly-onboarded specialty adding 2–4 weeks of rule-library calibration before reaching steady state.
How should MSO leadership measure whether the multi-specialty PA platform is working?
Track three KPIs monthly: per-specialty first-pass approval rate by payer and procedure (target steady-state above 85% across specialties within 6 months), median TAT distribution per site and specialty (target 50–70% reduction from manual baseline by month 4), and aged-PA exposure rolled up across the portfolio (target zero PAs aged past the payer's stated response window without an active workflow). If any of the three isn't moving in the right direction at a given specialty or site, the central team has specific data to escalate to the vendor.
