Concierge Dentistry: Can the math work? pt 3

A 1,000-patient model of dental payment, prevention and financial incentives

In the first article in this series, we asked whether dentist burnout and patient dissatisfaction might partly be symptoms of the same underlying problem. In the second, we looked at the economics of the dental chair and at a payment system in which procedures are easy to measure, easy to reimburse and therefore economically important, while prevention, communication and time are much more difficult to monetize.

That left us with a hypothesis.

If some portion of dental revenue were attached to maintaining an ongoing relationship with a patient rather than performing an individual procedure, could the financial incentives of the practice become better aligned with the patient's health?

There is an obvious temptation at this point to construct a hypothetical membership practice, choose assumptions that make it profitable, and declare victory. That would not tell us much.

Instead, we are going to construct a deliberately simple model, state the assumptions, compare several payment structures and then change the assumptions to see what happens. If the model produces an inconvenient result, we keep the inconvenient result.

This is not an actuarial study and it is not evidence that one payment system causes better clinical outcomes. It is a deterministic cohort model, essentially a structured thought experiment. Its purpose is narrower: to examine the direction and magnitude of the economic incentives created by different ways of paying for dental care.

Start by defining the question

The unit of analysis in our model is not a filling, crown or appointment. It is the patient-year.

We will model 1,000 established adult patients for one year and ask four questions:

1. How much revenue does the practice receive?
2. What happens to practice economics if patients require less restorative treatment?
3. Who bears the financial risk when treatment needs vary?
4. Does the payment system reward or penalize the practice when prevention succeeds?

That last question needs an important qualification. Fewer procedures are not automatically evidence of better care. A practice could reduce procedures through excellent prevention, or it could reduce procedures by failing to diagnose or treat disease. Any real evaluation would therefore have to measure clinical outcomes alongside utilization, including untreated disease, periodontal progression, emergency visits, tooth retention and other measures of health.

Our model assumes that any reduction in treatment represents disease or treatment need that genuinely did not occur, not necessary care that was withheld.

That distinction becomes very important later.

What we know, and what we have to assume

There is enough public data to constrain the model, but not enough to pretend that we can reconstruct a representative dental practice from national averages.

The ADA Health Policy Institute reports that inflation-adjusted revenue per general dentist in urban practices averaged a median of about $695,000 during 2020-2024. HPI also reports average net income of $215,320 for general dentists in private practice in 2025. Those numbers come from different statistical measures and should not simply be subtracted from one another, but they give us a useful order of magnitude for calibrating a hypothetical practice. Ada Association

We also know that dental utilization is heterogeneous. In the 2023 Medical Expenditure Panel Survey, an estimated 341 million dental visits occurred nationally. About 190 million of those visits included a cleaning, prophylaxis, polishing or periodontal recall. Approximately 51.6 million included fillings, inlays, crowns or caps; 14.8 million included extraction or other oral surgery; 7.0 million included root canal treatment; and 4.9 million included periodontal scaling, root planing or gum surgery. Those categories are not mutually exclusive and they describe visits rather than patients, but they demonstrate that a dental population contains very different levels and types of treatment utilization. Medical Expenditure Panel Survey

Disease is heterogeneous as well. CDC surveillance estimates that 20.5% of adults ages 20-64 have at least one untreated permanent-tooth cavity. Among adults age 30 and older, NIDCR reports that 42.2% have some form of periodontitis, including 7.8% with severe periodontitis. These conditions overlap, so those percentages cannot be added together or converted directly into patient risk categories. They simply reinforce an important modeling requirement: patients cannot realistically be treated as economically identical units. CDC

With those constraints in mind, everything that follows will be labeled as a model assumption, not an industry benchmark.

Building the 1,000-patient practice

We will begin with 1,000 established adult patients and divide them into three economic risk groups. These are not diagnostic categories. They represent expected use of non-preventive dental care during the year.

Across the full panel, routine preventive and diagnostic care therefore represents $300,000 per year.

Additional treatment represents another $400,000:

400 × $100 + 400 × $400 + 200 × $1,000 = $400,000

Total annual collected revenue is therefore:

$300,000 + $400,000 = $700,000

That gives us an average of $700 per patient-year and places our hypothetical practice in roughly the same revenue scale as the HPI figures described above.

Again, none of the individual dollar amounts in this table should be interpreted as recommended dental fees. The ADA itself no longer publishes a fee schedule and specifically notes that it cannot set or recommend dental fees. We are constructing an internally consistent model, not a price list. Ada Association

For expenses, we will make another simplifying assumption. We will assign $430,000 annually to costs that do not immediately disappear when a restorative procedure is prevented: staffing, hygiene capacity, occupancy, technology, insurance, administration and other fixed or semi-fixed expenses. We will then assume that materials, laboratory expenses and other directly variable treatment costs equal 15% of the $400,000 in non-preventive treatment revenue, or $60,000.

Our model therefore begins here:

The $210,000 residual is deliberately close in scale to current HPI general-dentist income data. That is a calibration choice, not a claim that this is the expense structure of an average dental practice.

Now we can start changing the payment system.

Model 1: traditional fee-for-service

The first model is straightforward.

The practice receives $300,000 for preventive and diagnostic services and $400,000 for additional treatment.

Revenue is therefore directly related to the amount of care performed:

Annual revenue = preventive revenue + procedure revenue

If treatment need increases, revenue generally increases. If treatment need falls, revenue generally falls.

That does not mean a dentist wants patients to become sick. It means the economic model does not necessarily reward the practice when disease is prevented.

We can quantify that.

Suppose better prevention, earlier intervention, better home care or more patient engagement reduces the amount of non-preventive treatment genuinely required by the population by 20%.

Treatment revenue falls from $400,000 to $320,000.

Total practice revenue falls from $700,000 to $620,000.

Some expense disappears too. Using our assumed 15% variable treatment cost, $12,000 of expense is avoided.

The result is:

Revenue lost: $80,000

Variable expense avoided: $12,000

Reduction in modeled practice residual: $68,000

The patients are healthier in our hypothetical scenario, but the practice is financially worse off.

That is the incentive problem we have been discussing throughout this series, reduced to arithmetic.

Model 2: the familiar dental membership plan

Now consider something much closer to the in-office membership plans that already exist.

The ADA describes a typical structure in which the patient pays a fixed monthly or annual amount, receives preventive services without an additional charge, and then receives a discount on other procedures. In its illustrative example, the ADA shows annual fees of $150, $300 or $400 for two periodic exams, two prophylaxes and annual bitewings, combined with discounts of 10%, 15% or 20% on other procedures. The ADA emphasizes that these figures are examples rather than recommended pricing. Ada Association

For our model, we will use a 15% discount on treatment and solve for the membership fee rather than simply guessing one.

We want our hypothetical practice to begin with the same $700,000 revenue it had under fee-for-service.

Treatment that previously generated $400,000 will now generate:

$400,000 × 85% = $340,000

That means the recurring membership portion must produce the remaining $360,000.

For 1,000 patients:

$360,000 ÷ 1,000 = $360 per patient per year

or:

$30 per month

So Model 2 becomes a $360 annual membership that includes the modeled preventive and diagnostic care, with additional treatment billed at a 15% discount.

At baseline, the economics are intentionally identical:

Membership revenue: $360,000

Discounted treatment revenue: $340,000

Total revenue: $700,000

This is useful because it allows us to compare the incentives without giving one model a financial head start.

What happens when prevention succeeds?

Now apply the same reductions in treatment need to all three payment structures.

For the moment, our third model will be a fully prepaid patient-year model, which we will discuss more carefully in a moment. It receives $700 per patient per year regardless of the number of defined services performed.

Using a 15% variable treatment cost assumption, the results look like this:

This is the first result in the model that I find particularly interesting.

A conventional dental membership plan softens the fee-for-service incentive, but it does not eliminate it.

At a 20% reduction in treatment need, the traditional practice loses $68,000 of modeled residual. The membership practice loses $56,000.

That is better, but the direction has not changed.

The reason is simple. Most of the revenue associated with restorative treatment is still earned only when restorative treatment occurs. We changed the payment mechanics, but we did not fundamentally change the revenue function.

That suggests something important about the membership plans that already exist.

They may be valuable for patient retention, predictable cash flow, preventive utilization and reducing dependence on third-party payers. But a membership plan that mainly prepays predictable preventive services and discounts everything else is not the same thing as changing the underlying incentive structure.

It is still largely fee-for-service.

Model 3: paying for the patient-year

Now take the thought experiment one step farther.

Suppose the practice receives $700 per patient per year for a defined package of care, independent of how many procedures the patient ultimately requires.

For our 1,000-patient population:

1,000 × $700 = $700,000

If treatment need falls 20%, revenue remains $700,000 while our modeled variable treatment expense falls by $12,000.

The residual therefore increases from $210,000 to $222,000.

For the first time, the economics and the assumed clinical outcome point in the same direction. If patients remain healthy and require less treatment, the practice does better financially.

This concept is not imaginary. Dentistry already contains capitated models. Under a DHMO or capitation arrangement, the ADA explains that dentists may receive a predetermined monthly payment for each assigned patient and then provide specified services at no charge or reduced charge. The ADA also explicitly notes the other side of the bargain: the dentist assumes financial risk when the cost of caring for the assigned population exceeds the payment received. Ada Association

And that second sentence is where our apparently elegant solution begins to get complicated.

We have reversed the incentive, not eliminated incentives

Under fee-for-service, performing more treatment can generate more revenue.

Under full prepayment, performing more treatment generates more cost while revenue remains fixed.

If we are worried that one system can create an incentive toward overtreatment, we have to be equally willing to recognize that the other can create an incentive toward undertreatment.

That is why treatment volume cannot be our definition of success.

A prepaid practice that reduces fillings because fewer cavities develop is succeeding. A prepaid practice that reduces fillings because it fails to diagnose or treat cavities is not.

The accounting statement looks identical.

A serious evaluation would therefore need both economic and clinical endpoints. We would want to measure untreated caries, periodontal progression, emergency treatment, tooth retention and perhaps referral patterns alongside revenue and procedure counts. We would also want patient-reported outcomes, access measures and independent quality review.

The purpose of changing payment is not to create fewer procedures.

It is to make better health financially sustainable.

Those are different objectives.

The problem of risk

There is another complication hiding inside our 1,000-patient average.

Our low-utilization patient generates an expected $400 worth of care in the original model. Our high-utilization patient generates $1,300.

Yet under a flat $700 prepaid model, both would pay or generate the same $700 payment.

Economically, that is risk pooling.

The $700 price works only because we assumed a population containing 40% low-utilization patients, 40% moderate-utilization patients and 20% high-utilization patients.

Change the population and the arithmetic changes.

Suppose the plan becomes particularly attractive to people who already know they need substantial dental care. Instead of a 40/40/20 distribution, enrollment becomes 30% low utilization, 40% moderate utilization and 30% high utilization.

Under our original fee-for-service valuation, expected annual care for that population becomes:

300 × $400 + 400 × $700 + 300 × $1,300 = $790,000

The actuarially neutral patient-year amount is no longer $700.

It is $790.

A flat $700 payment is now $90,000 below the modeled value of expected care.

Nothing about the dentist's clinical philosophy caused the problem. The risk pool changed.

This is adverse selection, and it is one reason that transferring substantial treatment risk to an individual practice is not simply a matter of choosing a monthly subscription price. Once payment becomes disconnected from the actual amount of care delivered, population composition matters enormously.

The ADA makes this point directly in discussing capitation plans: profitability can depend on the oral health of the population, and the dentist bears financial risk when patient-care costs exceed the capitated payment. Ada Association

This also helps explain why many ordinary dental membership programs stop short of full prepayment. Predictable services such as examinations and cleanings are relatively easy to bundle. A root canal, implant, series of crowns or extensive periodontal treatment introduces a very different amount of risk.

By leaving those services outside the membership and merely discounting them, the practice avoids becoming the party responsible for absorbing most of that variation.

The tradeoff is that the original procedure incentive remains.

Capacity may matter more than the cash calculation

There is another dimension that a revenue model can easily hide: clinician time.

Suppose, purely for purposes of modeling, that our lower-utilization patients require an average of 0.25 dentist treatment hours per year, moderate-utilization patients require 0.75 hours and higher-utilization patients require 2.0 hours.

Under the original 40/40/20 population:

400 × 0.25 + 400 × 0.75 + 200 × 2.0 = 800 treatment hours

Now apply the 30/40/30 higher-risk population:

300 × 0.25 + 400 × 0.75 + 300 × 2.0 = 975 treatment hours

That is an additional 175 dentist-hours.

At eight clinical hours per day, it represents almost 22 additional clinical days.

The financial model may still show an apparently viable practice while the dentist experiences an additional month of clinical workload.

That matters because our original question was not simply whether a different payment system can remain solvent. We are trying to understand whether it could alter the conditions contributing to dentist burnout and patient dissatisfaction.

A model that improves cash-flow predictability while requiring the dentist to work substantially more hours has not solved the problem we set out to examine.

How much prevention would have to work?

We can use the same time assumptions to ask another useful question.

Our baseline population requires 800 dentist treatment hours. A 20% reduction in genuine treatment need would therefore eliminate approximately:

800 × 20% = 160 treatment hours

Suppose a prevention-oriented model allowed the dentist to spend an additional ten minutes per patient per year on risk assessment, education, consultation, care planning or follow-up.

For 1,000 patients, that requires:

1,000 × 10 minutes = 10,000 minutes

or approximately:

167 hours

Under our assumptions, the break-even point is remarkably close. A little more than a 20% reduction in future treatment hours would recover roughly the clinical time required to give every patient an additional ten minutes of dentist attention during the year.

That does not mean ten extra minutes of conversation will reduce treatment by 20%. We have no evidence here that it would.

The model tells us something different and more useful: it tells us the magnitude of clinical effect that would be required for the time investment to become capacity-neutral.

That is the sort of question a real study could test.

Testing whether our result depends on one convenient assumption

We assumed that 15% of restorative treatment revenue represents expenses that disappear when the procedure is prevented. That figure is hypothetical.

So we should change it.

Keeping the baseline practice residual at $210,000 and assuming a 20% reduction in treatment need gives us:

The exact dollar amounts move, but the conclusion does not.

Under fee-for-service, preventing a dollar of treatment eliminates a dollar of revenue while saving only the portion of that dollar that would have become a variable expense.

Under our membership model, preventing the same dollar eliminates 85 cents of revenue while saving the same variable expense.

Under the prepaid model, revenue does not change and the avoided variable expense becomes a financial benefit.

We can write that relationship more generally.

If:

d = the share of treatment revenue still dependent on performing the procedure

and:

c = the variable cost associated with that treatment

then the short-run change in practice economics from preventing a dollar of treatment is proportional to:

c - d

For traditional fee-for-service, d = 1.00.

For our 15%-discount membership plan, d = 0.85.

For full prepayment, d = 0.

As long as procedure-dependent revenue exceeds the variable cost avoided, preventing treatment reduces short-run practice income.

That is the incentive in its simplest form.

Could we simply charge a higher membership fee?

Yes, mathematically.

Our $360 membership was deliberately chosen so that the practice generated the same $700,000 at the original treatment level.

But suppose we expect the prevention-oriented model to reduce treatment need by 20%, and we want the practice to maintain the original $210,000 residual.

With treatment reduced from $400,000 to $320,000, the 15%-discounted procedure revenue would be:

$320,000 × 85% = $272,000

Modeled expenses would fall to $478,000.

To preserve a $210,000 residual, the practice would therefore require total revenue of:

$478,000 + $210,000 = $688,000

Subtract the $272,000 of treatment revenue:

$688,000 - $272,000 = $416,000

For 1,000 patients, the annual membership becomes:

$416 per year, or about $34.67 per month.

The corresponding break-even membership amounts in our model are:

This may be the most useful result of the exercise.

It demonstrates that changing incentives does not require pretending that prevention is free.

Someone has to pay for the clinician's time, the preventive infrastructure, the availability and the ongoing relationship.

The economic question is whether patients, employers, insurers or some combination of them would be willing to move some of the money currently attached to procedures into a recurring payment for maintaining health.

At a 20% reduction in treatment need, a pure fee-for-service practice in our model would collect $620 per patient-year. A hybrid membership priced to preserve the original practice economics would collect $688. A fully prepaid model would collect $700.

The healthier patient is not necessarily cheaper in the short term.

What changes is what the money is buying.

Instead of all economic value being attached to treatment delivered after disease appears, some of it is attached to access, prevention, continuity, communication and assuming responsibility for the patient's oral health over time.

That is a very different proposition from a discount plan.

This is where the analysis becomes a real research problem

If we were evaluating an actual practice rather than constructing a public thought experiment, this is where the work would become considerably more sophisticated.

We would want longitudinal patient-level data. We would segment patients by age, periodontal status, caries risk, smoking status, medical comorbidity, prior treatment history and other predictors of future utilization. We would examine the distribution rather than merely the average, because high-cost tails matter enormously in any prepaid system.

We would establish a baseline period, ideally several years, and distinguish predictable preventive utilization from low-frequency high-cost events. We would test for selection effects, patient attrition and changes in the composition of the enrolled population.

We would then compare payment models using both financial and clinical outcomes.

On the financial side, we would measure revenue per patient-year, contribution margin, revenue variance, clinical hours, utilization, write-offs and administrative cost.

On the clinical side, we would measure new caries, periodontal progression, emergency care, untreated disease, tooth loss and other outcomes appropriate to the population.

On the patient side, we would measure access, waiting time, continuity, out-of-pocket cost, retention, satisfaction and trust.

On the clinician side, we would measure hours worked, patients seen per day, schedule pressure, administrative burden and burnout.

Most importantly, we would establish those outcome measures before evaluating whether one payment model performed better than another.

Otherwise it becomes far too easy to choose a definition of success that happens to favor the model we already wanted to succeed.

A more sophisticated analysis could then use risk adjustment, matched cohorts, interrupted time-series analysis or difference-in-differences if practices changed payment systems at different times. With enough longitudinal data, survival analysis could examine time to major restorative events, while generalized models could compare utilization and cost distributions. A stochastic or Monte Carlo model could estimate how often a prepaid practice experiences unusually expensive patient populations rather than relying only on expected averages.

Those methods would give us better estimates.

They would not change the fundamental question.

So what did the model actually tell us?

It did not prove that concierge dentistry, membership dentistry, capitation or any other particular model is better.

In fact, it produced a result that cuts in several directions.

Traditional fee-for-service creates a clear financial penalty when successful prevention reduces future billable treatment. A familiar dental membership plan reduces that penalty, but does not remove it if most restorative care remains tied to procedures.

A sufficiently large patient-based payment can reverse the incentive. Once revenue remains stable while avoidable treatment cost declines, better health can become financially beneficial to the practice.

But the moment we do that, the practice begins assuming population risk. Patient mix matters. Adverse selection matters. Clinical capacity matters. Undertreatment becomes a legitimate concern. Quality measurement becomes essential.

The problem does not disappear.

It changes shape.

That may explain why the most interesting answer is probably not an ideological choice between fee-for-service and a fully prepaid practice. The design space between them is much larger. Dentistry already includes PPO fee schedules, ordinary fee-for-service, membership plans, capitation and emerging value-based approaches. ADA materials on value-based payment themselves describe a continuum from traditional fee-for-service through arrangements that connect payment to quality and ultimately to population-based payment. Ada Association

The question we started with was whether the economics of dentistry could be better aligned with patient care.

The answer from this model is more interesting than yes or no.

The alignment can be changed. But changing it transfers risk, changes what must be measured, and forces someone to explicitly pay for the parts of good dental care that the procedure-based system often treats as economically secondary.

That gives us a much better question for the final installment.

Not whether dentists should abandon insurance.

Not whether every practice should become a membership practice.

Not whether fewer procedures are inherently better.

The question is what a practical payment system would look like if we deliberately designed it around three objectives at the same time: healthier patients, sustainable practices and dentists who have enough time to provide the care they were trained to provide.