AI Medical Intelligence
Practice EBITDA impact of medical assistant turnover using a $48,000 cost denominator
Each medical assistant departure removes roughly $14,860 from practice EBITDA, about 31% of the $48,000 fully loaded MA cost denominator, once vacancy c...
Practice EBITDA Impact of Medical Assistant Turnover: The $48,000 Denominator Playbook
Merry AI · Thoughtfully curated clinical briefs.
Clinically reviewed by Dr. Marcus Vance, MD (Medical Informatics Advisor) · Coding review by Kaif R. (Clinical Informatics Specialist) · Sources: CY 2026 CMS PFS, NCCI Policy Manual.
Each medical assistant departure removes roughly $14,860 from practice EBITDA, about 31% of the $48,000 fully loaded MA cost denominator, once vacancy coverage, recruiting, onboarding, ramp productivity, and lost provider throughput are counted. For a 12-MA practice at 35% annual turnover, that is about $62,400 in recurring EBITDA drag every year.
This playbook gives Practice Administrators and CMOs a reproducible cost model, an enterprise-value translation, the documentation mechanism that links MA turnover to coding leakage, and a 90-day containment protocol. All figures are modeled assumptions. Replace them with your own payroll, vacancy, and collections data before presenting them to a board or acquirer.
Each MA departure costs about $14,860, or 31% of $48,000
The $48,000 denominator represents one medical assistant's fully loaded annual cost: base wages plus payroll taxes, benefits, and paid time off. Dividing by 2,080 paid hours yields a loaded hourly rate of about $23.08. Every component below is priced against that single labor unit so the figures stay comparable.
| Cost component | Modeled assumption | Cost per departure | Share of $48,000 |
|---|---|---|---|
| Vacancy coverage premium | 6 weeks (240 hours) covered by overtime or agency at 1.5x, net of saved wages | $2,770 | 5.8% |
| Recruiting and screening | Job board spend, interviews, background checks, competency testing | $1,500 | 3.1% |
| Onboarding and training | 80 hours of new-hire time plus 80 hours of trainer time at $23.08 | $3,690 | 7.7% |
| Ramp productivity loss | 90 days at 70% productivity (30% of one quarter's loaded cost) | $3,600 | 7.5% |
| Lost provider throughput | 1 fewer visit per day for 30 clinic days at $110 net collections | $3,300 | 6.9% |
| Total | Excludes coding leakage and patient attrition | $14,860 | 31.0% |
Lost provider throughput is the component most administrators omit, and it is often the most volatile. When rooming slows, providers run behind, add-on visits disappear, and same-day access contracts. Because a provider's fixed costs do not fall during a vacancy, nearly every lost visit flows directly to the EBITDA line.
Callout: Turnover cost excludes coding leakage, which is modeled separately below.
A 12-MA practice at 35% turnover loses about $62,400 annually
Multiply departures by unit cost to size the recurring drag. A practice employing 12 MAs at 35% annual turnover loses 4.2 MAs per year. At $14,860 each, the annual EBITDA impact is approximately $62,400, which equals about 1.3 full MA salaries spent on churn rather than care delivery.
Against a $1.5 million EBITDA base, $62,400 represents roughly 4 percentage points of EBITDA lost to a single, controllable operational variable. Practices with higher net collections per visit, longer vacancy windows, or heavier reliance on agency staff will see proportionally larger losses.
At an 8x multiple, turnover erases roughly $500,000 of value
Acquirers and lenders value practices on normalized EBITDA. Recurring turnover cost is rarely added back because it is a structural operating expense, not a one-time item. At an illustrative 8x multiple, $62,400 of annual drag reduces enterprise value by approximately $499,000.
The reverse is also true: cutting turnover from 35% to 20% removes 1.8 departures per year, restores about $26,700 of EBITDA, and adds roughly $214,000 of enterprise value at the same multiple. Retention is one of the few levers that improves both margin and valuation without new payer contracts.
Callout: Every $1 of recurring turnover cost compounds through your valuation multiple.
Turnover degrades intake documentation, which suppresses E/M level and G2211 capture
New MAs routinely miss the intake details that support medical decision making (MDM): complete medication lists, chronic condition status, abnormal vitals, and care-coordination history. When those elements are absent from the record, providers code conservatively, and the encounter settles one E/M level lower than the work performed.
Medication reconciliation is a common failure point. Documenting long-term drug therapy with Z79.899 (ICD-10-CM) supports drug-therapy-requiring-intensive-monitoring logic and ongoing risk management. An elevated reading captured as R03.0 (ICD-10-CM) establishes a new problem requiring workup, rather than leaving an unaddressed vital sign in the flowsheet.
Z79.899 and R03.0 anchor defensible MDM and G2211 claims
The table below maps common intake indicators to 2021+ AMA office E/M MDM thresholds and the audit defense each requires. Under current guidelines, 99213 requires low MDM, 99214 requires moderate MDM, and 99215 requires high MDM, with two of three elements (problems, data, risk) satisfying the level.
| Clinical indicator | ICD-10-CM anchor | MDM threshold supported | G2211 relevance | Audit defense required |
|---|---|---|---|---|
| Long-term medication requiring monitoring (e.g., anticoagulant, DMARD, antiepileptic) | Z79.899 with primary condition | Moderate risk (99214); high risk if intensive toxicity monitoring (99215) | Strong: ongoing longitudinal management | Named drug, monitoring parameter, lab ordered or reviewed, plan adjustment |
| Elevated BP without hypertension diagnosis | R03.0 | New problem with uncertain prognosis supports low to moderate problem complexity | Moderate: supports continuing relationship when follow-up is planned | Repeat reading, technique noted, home monitoring or recheck plan |
| Two or more stable chronic illnesses | Condition-specific codes (e.g., E11.9, I10) | Moderate problem complexity (99214) | Strong: provider is continuing focal point | Status of each condition addressed, not merely listed |
| Chronic illness with exacerbation or side effects | Condition code plus manifestation | Moderate (99214); severe exacerbation supports high (99215) | Strong | Objective change documented, treatment modification recorded |
| Prescription drug management | Linked to treated condition | Moderate risk element (99214) | Supportive | Drug, dose decision, and rationale stated (continue, start, stop, adjust) |
| External records or independent historian | Not code-dependent | Data element toward moderate or high MDM | Supportive of longitudinal context | Source identified, specific findings summarized |
NCCI and MAC auditors look for problems that were actually addressed, not problems copied into a list. Each row above therefore pairs the code with an action: a lab reviewed, a dose changed, a recheck scheduled. That action is what converts an indicator into defensible MDM.
Merry AI Pro costs $648 per year, 1.3% of one MA
Merry AI Pro is priced at $54 per month billed annually, or $648 per year per seat. Against the $48,000 fully loaded MA denominator, that equals roughly 1.3% of one assistant's annual labor cost, and about 4.4% of a single $14,860 departure.
An AI scribe does not replace rooming, vitals, injections, or patient coordination. Its EBITDA contribution comes from three mechanisms: reducing provider and staff documentation load (a recognized burnout driver), shortening the competency ramp for new MAs, and keeping MDM elements captured consistently during vacancies when staffing is thin.
The break-even math is short. If deploying ambient documentation prevents one MA departure per year across a 10-provider group, the avoided $14,860 alone covers roughly 2.3 years of a 10-seat Pro deployment ($6,480 per year), before any coding recovery is counted.
Callout: One avoided departure funds a 10-seat deployment for over two years.
G2211 capture recovers $15,600+ per provider annually under modeled volume
CPT G2211 is the Medicare add-on code for the visit complexity inherent in serving as a patient's continuing focal point for care or ongoing care of a single serious or complex condition. Its national non-facility payment is about $16 per eligible office or outpatient E/M visit, varying by locality.
A primary care provider with roughly 975 eligible Medicare E/M visits per year recovers about $15,600 or more annually through consistent G2211 capture. The architecture for identifying eligible encounters and documenting the longitudinal relationship is detailed in CPT G2211 Complexity Capture Architecture.
Payment restrictions still apply in 2026. G2211 is not payable when the base E/M carries modifier 25, except when the E/M is reported on the same day as an annual wellness visit, vaccine administration, or a Medicare Part B preventive service. Documentation must show the ongoing relationship, not merely a return visit.
Browser overlay deployment avoids integration delays that extend new-hire ramp
Deployment speed matters because turnover is continuous. A tool that requires a multi-month EHR interface build delivers no value to the MA hired next week. A browser overlay works inside the existing EHR session, so new staff learn one workflow, not two. See EHR Browser Overlay Architecture for the technical comparison.
FHIR-based interoperability remains the long-term standard for structured data exchange, but overlay deployment lets practices capture documentation gains now while API integrations mature. For a practice losing $62,400 per year to churn, a six-month integration delay costs roughly $31,200 in unmitigated drag.
State AI scribe statutes require patient disclosure and clinician attestation
Ambient documentation must comply with state recording consent laws, including all-party consent states, and with emerging AI disclosure statutes. California AB 3030 requires disclosure when generative AI produces patient clinical communications, and California SB 1120 requires that utilization review medical necessity determinations be made by licensed clinicians, not algorithms.
Every AI-drafted note must be reviewed, edited as needed, and attested by the rendering clinician before claim submission. The billing provider remains responsible for accuracy under Medicare signature and documentation requirements. Current state-by-state requirements are maintained in the State AI Scribe Rules & Statutory Compliance Hub.
Callout: No attestation, no claim. Clinician review is the non-negotiable audit control.
A 90-day protocol cuts turnover cost and protects coding capture
Use this sequence to convert the model into operating decisions. Each step produces a number that belongs on the monthly EBITDA bridge.
- Days 1 to 15: Pull 24 months of MA separations, vacancy days, agency invoices, and overtime hours. Calculate your actual cost per departure against the $48,000 denominator.
- Days 16 to 30: Measure lost visits per provider during vacancy windows and multiply by net collections per visit. Add this line to the turnover cost model.
- Days 31 to 45: Audit 30 encounters per provider for missing Z79.899, R03.0, and chronic condition status documentation. Quantify E/M downcoding and missed G2211.
- Days 46 to 60: Deploy ambient documentation via browser overlay for a pilot cohort. Confirm consent scripts and attestation workflow against state requirements.
- Days 61 to 90: Compare pilot versus control on G2211 capture rate, E/M level distribution, MA onboarding hours, and staff-reported documentation burden. Present results as EBITDA dollars.
Report results in denominator terms. Expressing turnover cost, software spend, and coding recovery as percentages of $48,000 gives a board one consistent unit for comparing retention investments against hiring, agency, and technology alternatives.
Frequently asked questions about MA turnover EBITDA impact
How much does one medical assistant departure cost a practice?
Under the Merry AI model, one departure costs about $14,860, or 31% of a $48,000 fully loaded salary. Components include agency or overtime coverage, recruiting, onboarding hours, 90-day ramp inefficiency, and lost provider visits. Practices with longer vacancies or higher net collections per visit will exceed this figure.
Why use $48,000 as the cost denominator?
$48,000 approximates the fully loaded annual cost of one medical assistant, combining base wages with payroll taxes, benefits, and paid time off. Using one consistent denominator lets administrators express turnover losses, software spend, and staffing decisions as comparable percentages of a single labor unit on the EBITDA bridge.
Does an AI scribe replace medical assistants?
No. An AI scribe reduces documentation tasks, not rooming, vitals, injections, or patient coordination. Its EBITDA value comes from shortening onboarding, reducing burnout drivers that trigger resignations, and protecting coding capture during vacancies. Merry AI Pro costs $648 per year, roughly 1.3% of one MA's $48,000 loaded cost.
How does MA turnover affect G2211 revenue?
New MAs often miss medication lists, chronic condition updates, and longitudinal care context during intake. Without that documentation, providers hesitate to append G2211, which pays about $16 per eligible Medicare visit. Across roughly 975 eligible visits, consistent capture recovers $15,600 or more per provider each year.
What compliance rules apply when deploying an AI scribe in 2026?
Practices must obtain patient consent consistent with state recording and AI disclosure laws, require clinician review and attestation of every note, and keep audit trails. California requires disclosure for generative AI patient communications under AB 3030. Merry AI tracks state-by-state requirements in its statutory compliance hub.