Maywell · Practice Analytics

Snapshot data

Speed vs. value — where the fast money is

Each bubble = one diagnosis (bubble size = patient count). Up and left = converts to procedures fast and pays well. From 1,004 patients with a coded primary diagnosis in the masterfile; procedure = interventional CPT (ESI/TFESI, MBB/RFA, joint, SCS, nerve blocks).

The Top 20 — ranked by expected procedure revenue per patient, weighted for speed

"Conv" = % of patients with this dx who ever reach a procedure. "Days" = median first-visit → first-procedure. "EV/pt" = conversion × procedure $ per converter — what one more referred patient with this dx is worth in procedure revenue alone. "LTV/pt" = total collections per patient. Recommendation is driven by each dx's actual payer mix in your data.

How to hunt them

Referred but NOT booked — by organization

True booking status from the patient master (Master Analysis v4): referred patients with no first appointment and no billing ever recorded. These are people a referrer sent who never made it onto the schedule.

Referral sources, month by month — per organization

Monthly new booked patients per referring organization (top 12 by volume + all others; from the patient master). Top 6 shown by default — click legend entries to toggle any organization on or off and compare pattern changes over time.

Referral channels, month by month

Same view at channel level, full 32-month history.

The entire referral pool — all referring organizations

Monthly collections vs. collections per business day

Raw monthly totals mislead: months have 19–23 business days, a ±10% swing that has nothing to do with performance. The line normalizes it. 2026 months also adjust for provider vacation (from card data: Florida Feb 14–18 ≈ 3 business days; Salt Lake City Mar 4–7 ≈ 3 business days — edit VACATION_DAYS in this file as needed).

Normalization table — last 18 months

How to read this

Standing rules: (1) Billed charges appear nowhere on this dashboard — every dollar shown is either cash collected or model-expected collections (observed Medicare rates per CPT; commercial at observed commercial rates; NF at 1.57× / WC at 0.79× Medicare). (2) The NF/WC arbitration lane = 50% of the NF/WC-attributable open balance, tracked as its own line and never blended with collected figures.

Monthly collections — cash actually received

Net payments by month of service. Recent months understate final collections because claims lag ~1–2 months.

Expected collectible vs collected to date — the realistic revenue line

Blue = expected collections per Dr. M's model (services valued at observed Medicare rates; NF/WC at observed collection ratios; commercial at Medicare). Green = cash actually collected so far. The blue-over-green gap in recent months is open A/R, not lost revenue. Charges are excluded everywhere — billed amounts are not money.

Monthly: expected collectible vs collected to date

Blue bars = model estimate for that month's services. Green bars = cash collected so far. Grey line = Medicare value of work performed. Recent months' gaps are receivables working through claims + arbitration.

Collection gap by service month — 2026

Expected minus collected. This is the A/R worklist, oldest first.

Model parameters

Per Dr. M's spec, July 2026. Full CPT rate table lives in the master workbook.

Per-diem volumes vs plan benchmarks

E&M sessions and procedure encounters per business day (weekday-normalized, per the CIM methodology). Dashed lines = business-plan capacity: 30 E&M/day physician, 25/day APP hiring trigger.

Production per business day — expected mix vs actual collected

Stacked bars = model-valued production per weekday by band; navy line = actual dollars collected per weekday from the full payment file.

Monthly operating analysis — trailing 12 months

CIM-format volume table: E&M encounters, procedures deduped per claim, per-diem band. Full history in the master workbook.

Monthly production — 18 months history, 12 months projected (Jul 2026 – Jun 2027)

Solid = model-expected production to date. Dashed = three scenarios: flat run-rate, organic trend (current team: Dr. M 5 clinical days + Rebecca 4 clinical / 1 admin), and with the two new PAs ramping.

Scenario totals — next 12 months

Expected production Jul 2026 – Jun 2027; collections follow with the usual 1–2 month lag plus the arbitration tail.

Assumptions

Every input, stated. Change any of these and the projection moves.

New patients per month

By registration date (Athena). Bars = monthly intakes; line = 3-month rolling average.

New patients by acquisition channel — monthly

Channel captured at intake (most reliable for trends per data rules). Stacked by month.

Channel mix — last 6 months vs prior 6

Share of new patients by channel, H1 2026 vs H2 2025.

Channel mix — H1 2026

Where the last six months of new patients came from.

Retention funnel — patients reaching each visit depth

Full billing book — 2,479 billed patients; visits = distinct service dates.

Visit-depth distribution

How many total visits each patient reaches before dropping off.

60-day return rate by monthly cohort

Of patients whose first service was in a given month, the % who returned for a 2nd visit within 60 days. Cohorts after Apr 2026 excluded (immature).

Retention by payer class

2nd-visit rate and average visit depth by the patient's dominant payer.

Retention by referral channel

Channels with ≥15 billed patients.

Retention by referral source

Sources sending ≥15 billed patients, full billing book — who sends patients that stay.

Monthly collections by payer class

Cash received by payer class, by month of service. Payer detail covers the patient-linked claim set ($902K); the Overview and Expected-vs-Collected tabs carry the full $2.48M.

Revenue share by payer class

All-time collected, Dec 2023 – Jun 2026.

Payer class economics

Collected dollars, patients, and revenue per patient.

NF/WC arbitration lane — kept separate from collections

Per month of service: NF/WC-attributable open balance at real-world rates × 50% recovery assumption. This is the Rizzo arbitration track; none of it appears in the collections charts above.

Top referral sources — collected vs NF/WC expected (separate bars)

Full-book collections joined to each patient's referring organization from the master. Amber = source's NF/WC arbitration lane at real-world rates × 50%. Never summed.

Referral source economics

Sorted by collected. Rev/patient = collected ÷ unique billed patients from that source.

Collected revenue by channel type

Organization-level channel classification.

Channel economics

Collected, pipeline and revenue per patient by channel.

Collected vs expected by division

Service lines grouped per the coding compendium. Both bars now cover the full billing footprint — collected is actual payments; the gap is genuine open A/R.

Actual collections by division — monthly

Real payments by division, 2025–2026.

Behavioral health — psychotherapy vs screeners vs collaborative care

Per the compendium v2 split. CoCM codes (99492–94) belong to Lin Health going forward; historical Maywell-billed lines are flagged below.

Projected total collections by provider — collected + open expected + arbitration

Teal = collected to date. Blue = remaining expected on non-NF/WC claims (observed payer, model rates). Amber = NF/WC arbitration lane at 50% — separate, never blended. Full billing book.

Collected revenue by provider

Full-book collections attributed via each patient's rendering provider (panel-level) — absolute dollars, full $2.48M base.

Provider economics

Panel size, collections, and revenue per patient.

Panel stickiness

2nd-visit retention and average visit depth by provider panel.

Who actually pays: payer × code for the codes in question

From the masterfile claim lines. "Paid rate" = share of lines with any payment; "Avg when paid" = average payment on paid lines.

Code outliers — realization vs expected rate

Codes ≥40 lines where collections diverge from the observed-rate expectation. Realization = collected ÷ (expected rate × lines); paid share = % of lines with any payment.

Outlier realization multiples

Log-scale bars — right of 1.0× collects above expectation (NF/WC premium payers), left collects below.

Where every active patient sits in the funnel (orders + billing, Oct-25 → May-26)

Stage inferred from orders placed and procedures billed. Stage 4 is the money row: the physician has already ordered a procedure and it has not happened.

Ordered → performed: completion rate by procedure family

Encounter-note fields that would power the next layer of cost analytics

The masterfile has orders and billing but no note content. Since MedWriter generates the notes and the Xaia rules structure them, have every note emit a small JSON block alongside the narrative: conservative_care (documented y/n + weeks), imaging_reviewed (modality + date), pain_score, work_status, plan_procedure (family, level, laterality), utox_reviewed, opioid_agreement_current. Seven fields, near-zero scribe effort. What they unlock: time-and-cost per funnel stage (visits burned before the procedure decision), documentation-completeness scoring that directly defends 72275 and shrinks the J0665 denial pile, automated opioid-UTOX compliance (currently ~52%) without chart review, and a leading indicator — procedure intent stated in the note before the order is even placed. That closes the loop this tab starts.

Order utilization — Q1 2026

Orders and unique patients by category, from the quarterly orders report.

Opioid ↔ UTOX compliance by prescriber

Opioid patients in Q1 vs those with a urine drug screen anywhere in the orders window (Oct 2025 – May 2026).

Monthly order volumes

Opioid scripts, other medications, UTOX, DME and imaging orders by month.

In-house UTOX — annual revenue scenarios

Rates: presumptive 80307 ≈ $62, definitive add-on (G0480-class, 60% of tests) blended to ≈ $131/test — approximate Medicare; NF/commercial pay above. Volumes: current pace vs guideline quarterly testing on the opioid panel.

In-house DME — annual revenue scenarios

Blended ≈ $350/item across spinal orthoses, knee braces, TENS (conservative vs fee schedules). Capture scenario assumes in-office dispensing doubles fulfilled orders.

Requirements before go-live

Compliance path — route through Weiss Zarett before any buildout.

Market view — Long Island vs NYC (all venues combined) — monthly actual collections

NYC has occupied many addresses (3rd fl MIMA → 1st fl Quantum → 12th fl, the Madisons, FlatIron, FiDi, Brooklyn, Astoria) — this rolls every NYC venue and NYC ASC into one market line so the market, not the lease history, is what gets compared to Long Island.

Actual collections by location — monthly (2025–2026)

Real payments by service department, stacked — per-site cash production.

Collected vs expected by location — all-time

Actual payments now in, Oct 2023 – Jul 2026.

Location economics

Model-valued production and line volume by site.

Two-way partnerships — patients received vs sent back

Top 15 sources by patients received (all-time), against outbound referrals we sent to them.

Top outbound destinations

Where Maywell sends patients (213 unique patients, 29 destinations all-time).

Reciprocity gaps

Received a lot, sent little — the relationships carrying an imbalance.

Collected revenue by service family

Actual payments, full billing footprint ($2.48M), grouped into clinical families.

Monthly volume — procedures vs E/M visits

Line-item counts per month, full billing footprint. Procedures = injections, RFA, SCS, facet/MBB, SI joint, trigger point, nerve blocks.

Lifetime value by visit depth

Full billing book — average collected per patient by visit depth; the economic case for retention.

LTV curve

Average collected per patient at each visit-depth bucket.

Practice Performance Dashboard