Quick answer: A multi-specialty group on athenaOne should segment prior authorization automation by specialty line rather than treating the network as one workflow, because the PA mix (drug versus procedure versus DME), payer concentration, and clinical documentation requirements vary dramatically across cardiology, dermatology, orthopedics, GI, endocrinology, and primary care. The pattern that works pairs athenaOne's native Authorization Management for the routine workflow with a third-party AI agent on the specialty-specific long tail — dermatology biologics, orthopedic imaging and surgery, GI procedures, endocrinology GLP-1s — where the native rule library runs thin.
Why a single PA workflow fails at multi-specialty scale
A single-specialty practice on athenaOne runs against one payer-rule library, one set of dominant procedure codes, and one stable clinical documentation pattern. The auth team specializes, the workflow repeats, and a clean operation takes 90 days to build.
A multi-specialty group running cardiology, dermatology, orthopedics, GI, endocrinology, and primary care across the same athenaOne instance doesn't get any of that simplicity. Each specialty has a different PA mix, a different dominant payer concentration, and different clinical documentation requirements:
- Dermatology runs heavy on biologics PAs at $30,000+ per request, with payer-specific step-therapy criteria across Humira, Stelara, Skyrizi, Cosentyx, and the rest of the biologic shelf.
- Orthopedics runs heavy on imaging PAs (cardiac MRI, advanced shoulder/knee imaging) and surgical procedure PAs with conservative-care-tried documentation requirements that vary by payer.
- GI runs procedure PAs for colonoscopies and EGDs plus high-volume biologic infusion PAs for IBD treatment.
- Endocrinology runs the GLP-1 PA flood that's exploded across the payer landscape since 2024, with payer-specific BMI thresholds, comorbidity requirements, and prior-treatment documentation.
- Primary care runs broader-volume but lower-complexity PA mix — routine imaging, common medications, durable medical equipment.
The auth coordinator who's strong on cardiology cases is rarely strong on derm biologics. Rotating staff across specialties creates first-pass denial rates that compound. The AMA's 2024 prior authorization survey reports physicians and their staff spend 13 hours per week on PA work, with the multi-specialty groups facing the highest variance in that burden by service line.
Step 1: Quantify PA volume and cycle time per specialty
Before any automation decision, the multi-specialty group needs a clean baseline per specialty. Three numbers matter, captured for at least 30 days from athenaOne.
PA volume by specialty. Pull data filtered by ordering department and procedure code groups. The shape that emerges usually surprises operators — one or two specialties drive most of the total PA volume, with the long tail of low-volume specialties handled by the same shared auth team.
Median turnaround time by specialty. Each specialty's TAT distribution tells you where the operational pain is concentrated. Dermatology biologics often run 7–14 days; cardiology imaging routinely lands at 2–4 days; primary care medication PAs typically close in 24–72 hours. The variance is the actionable signal.
First-pass approval rate by specialty. Capture this from athena's denial reporting. The specialties with the lowest first-pass approval rates are usually the ones where athena's native rule library doesn't model the payer-specific criteria deeply enough — and where layered automation pays back fastest.
For most multi-specialty groups, this baseline exercise lands at one or two specialties driving 50–70% of total PA volume, with the long tail of low-volume specialties handled by the same shared auth team. That distribution determines the rollout sequence.
Step 2: Decide which specialties get native, layered, or hybrid
athenaOne's native Authorization Management handles the routine 60–75% of PA volume across most specialties. The honest framing is that it's strong on the deterministic core and thinner on specialty-specific long tails where payer-specific clinical criteria need deeper modeling.
The pattern that works at most multi-specialty groups is a three-bucket categorization per specialty:
Native-only: Specialties where athena's Authorization Management covers most of the workload cleanly. Primary care medication PAs, routine imaging, common DME. The native workflow handles these without significant gaps; adding a third-party agent on top doesn't pay back.
Layered: Specialties where native handles the routine but a third-party AI agent closes the long-tail gap. Dermatology biologics, orthopedic surgical PAs, GI biologic infusions, endocrinology GLP-1s. The native rule library catches the obvious cases; the AI agent picks up the payer-specific clinical criteria that don't fit the deterministic rule set.
Specialty-AI primary: Specialties where the native rule library is thin enough that the layered approach inverts — the AI agent becomes the primary PA workflow, with native serving as the integration backbone. This bucket is unusual but shows up at multi-specialty groups with heavy oncology infusion volume or high-cost specialty drug concentration.
The right categorization depends on the practice's specific specialty mix and payer footprint. Most multi-specialty groups end up with primary care in the native-only bucket and at least one of derm, ortho, GI, or endocrinology in the layered bucket.
Step 3: Centralize the auth team or keep specialty-embedded?
The operating model question — centralized PA team versus specialty-embedded coordinators — is where multi-specialty PA automation rollouts succeed or fail.
The centralized model has one shared PA pod handling volume across every specialty. The team specializes by exception type (one coordinator on medical necessity, one on step-therapy, one on peer-to-peer prep) rather than by specialty. The platform's content-based routing handles the specialty-tuning of each PA's clinical evidence and payer-rule lookup. Cost: one shared team across the group; benefit: specialization by exception type with the platform handling specialty-tuning.
The specialty-embedded model keeps PA coordinators inside each specialty's local team. Cardiology has its own coordinator; dermatology has its own; orthopedics has its own. Each coordinator builds deep specialty expertise on payer behavior, documentation patterns, and provider preferences. Cost: linearly scaling headcount with specialty count; benefit: deep specialty context.
The hybrid that works at most multi-specialty groups: centralized AI processing of the routine 75–85% of PA volume, with specialty-embedded coordinators handling the judgment cases (peer-to-peers, complex denials, novel payer policies). The platform owns the queue mechanics; the specialty coordinators own the cases that need their clinical context.
Honey Health's Prior Authorization agent is built around this hybrid pattern for multi-specialty groups running on athenaOne. The agent applies specialty-specific clinical extraction and payer-rule modeling per PA, with content-based routing that directs cases to the right coordinator pod. The same architecture extends across the rest of the back office — fax triage, referral intake, eligibility verification, refill management, denial management, payment posting, and data fetching — so the multi-specialty group can layer in additional automation without changing vendors.
Step 4: Handle cross-specialty patient handoffs without losing the auth
The hardest PA pattern at a multi-specialty group is the cross-specialty patient who needs stacked authorizations. A new cardiology consult uncovers a finding that needs orthopedic evaluation, which triggers a surgical procedure that needs its own PA. Each step has its own payer-rule logic, documentation requirements, and timeline.
Under manual workflows, this is where cases get lost. Each specialty's coordinator works the PA in front of them; nobody owns the longitudinal view of which authorizations need to be in place before the patient can move to the next step of care. Patients drop out of the workflow because no one called them between the cardiology consult and the ortho consult.
The pattern that works 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, surfaces dependencies across specialties, and triggers downstream PA work as upstream PAs reach their decision points. The cardiology PA approves; the ortho consult PA fires automatically with the cardiology clinical context attached.
For multi-specialty groups running on athenaOne, this longitudinal view requires either deep integration with athenaOne's scheduling and clinical workflow data or a third-party PA agent that maintains its own patient-level state across specialties.
Step 5: Measure per-specialty performance, not network averages
The reporting layer is where multi-specialty PA automation rollouts prove value or hide it. Three KPIs matter monthly, broken out by specialty:
Per-specialty first-pass approval rate by payer and procedure. This metric surfaces where the rule library is leaking. When dermatology runs at 79% first-pass while orthopedics runs at 89%, the gap usually points to payer-specific criteria coverage on biologics rather than coordinator skill.
Per-specialty median TAT. The variance across specialties tells you which workflows are running cleanly and which are dragging. Specialty groups whose TAT consistently exceeds the rest of the network are where additional automation tuning or specialty coordinator capacity is needed.
Aged-PA exposure rolled up across specialties. PAs still pending past the payer's stated response window represent revenue at risk. The multi-specialty group needs to see the aged exposure across every specialty in one number so the central team can intervene on the largest aged bucket first.
CMS's 2026 Interoperability and Prior Authorization Final Rule mandates 7-day standard and 72-hour urgent response windows for most Medicare Advantage and Medicaid managed care plans starting in 2027. Multi-specialty groups need per-specialty visibility to verify payer compliance and identify cases where escalation is warranted.
Frequently asked questions
Can one PA platform handle every specialty in our group?
Yes, if the platform was built for multi-specialty operations. Look for content-based routing that reads each PA's diagnosis code and procedure to apply the right specialty's payer-rule logic, specialty-tuned clinical extraction that pulls op notes for ortho cases versus imaging reports for cardiology versus path reports for dermatology, and per-specialty reporting that surfaces variance by service line.
How long does multi-specialty PA automation take to roll out on athenaOne?
For a 5–8 specialty group on athenaOne, plan for 3–6 months end to end. Start with the anchor specialty (usually the highest-volume one), validate the workflow in 6–8 weeks, then add specialties at a 2–4 week cadence each. The AI's specialty-classification layer learns each new specialty's denial patterns during the per-specialty onboarding.
Should we centralize the auth team when we adopt automation?
The hybrid model usually wins at multi-specialty scale. Centralize AI processing and the routine queue across the network; keep specialty-embedded coordinators for peer-to-peers and complex denials. Most multi-specialty groups reduce centralized auth team headcount slightly while preserving specialty-specific expertise where it's most needed.
How does the platform handle specialty handoffs where one patient needs multiple PAs?
The pattern that works treats the patient as the unit of work rather than the individual PA. The platform maintains a longitudinal view of every PA in flight for that patient, surfaces dependencies across specialties, and triggers downstream PA work as upstream PAs reach their decision points. Cross-specialty cases that previously fell through the cracks get worked end to end.
What if our highest-volume specialty isn't the one with the worst PA performance?
Sequence the rollout by operational pain, not by volume. The specialty with the most aged PAs, the lowest first-pass approval rate, or the highest staff burnout typically deserves first attention regardless of total volume. Start where the operational improvement will be most visible; volume rollout follows.
