The payment structure is published. The operating costs are not. Here is how to build a realistic dementia program profit and loss before you are a year in and surprised by the numbers.
Key Takeaways
- GUIDE base rates differ significantly by tier and by whether the patient has a caregiver: a high-complexity patient with a caregiver generates $360 per month in the first six months and $220 afterward – nearly three times the low-complexity established rate of $65.
- The geographic adjustment factor, the health equity adjustment, and the performance-based adjustment can swing per-patient monthly revenue by more than 20 percent in either direction – all three belong in your financial model from the start.
- Respite reimbursement of up to $2,625 per patient annually (PY2026) is real revenue, but it flows only to moderate- and high-complexity patients with unpaid caregivers – your tier mix determines how much of it you can access.
- No published CMS staffing ratio exists for care navigators, and there is no CMS break-even analysis – you must build your own model against your own cost structure and payer geography.
CMS publishes the GUIDE payment rates in enough detail that building a per-beneficiary revenue model is straightforward. What is not published – and what the academic literature does not yet provide – is any reliable cost-side benchmark. CMS specifies no care navigator staffing ratio, no minimum technology budget, and no break-even patient census. That means every GUIDE participant operating today is running a program whose financial sustainability was estimated at enrollment and has not been tested since. For most programs, the honest answer to “is this profitable?” is still unknown.
Building that answer requires holding three things together: the payment rate structure, the adjustments layered on top, and the full cost of delivery – staff, technology, partner contracts, and the overhead your organization allocates to the program. If those three live in separate spreadsheets nobody reconciles monthly, you are not running a program. You are running a hope.
For real-world grounding: Northwell Health published its year-one GUIDE experience in Innovation in Aging in 2025, reporting 158 beneficiaries enrolled by end of year one, with a tier mix of 19 percent low, 73 percent moderate, and 8 percent high complexity. Over a third of their eligible patients used respite – additional revenue available only to moderate and high tier patients with unpaid caregivers. Your tier mix will differ, but the Northwell distribution is the best published reference point available while the formal Mathematica evaluation is still in progress.
The Payment Structure, in Numbers You Can Use
CMS pays through the Dementia Care Management Payment, a monthly per-beneficiary amount billed with one of twelve G-codes (G0519 through G0528 for community-dwelling patients, G0574 and G0575 for residential care community patients). One claim per patient per calendar month, standalone – no other HCPCS codes on the same claim. Patient coinsurance and deductible are waived; CMS pays 100 percent of the allowed amount.
Base rates before geographic adjustment: for patients with an unpaid caregiver, low complexity pays $150 in the first six months and $65 afterward; moderate pays $275 initially and $120 afterward; high pays $360 initially and $220 afterward. For patients without an unpaid caregiver, low pays $230 initially and $120 afterward; moderate and high combined pay $390 initially and $215 afterward. CMS has not published base rates for the residential care community tier.
Two things follow immediately. First, revenue per patient is highest in the first six months and drops sharply – a moderate-complexity caregiver patient goes from $275 to $120, a 56 percent reduction. If your census is growing, the decline is masked by new-patient revenue. If census stabilizes, it hits the P and L hard. Second, caregiver status is a material revenue variable. Build both scenarios into your model and track the actual caregiver rate in your population.
Three Adjustments That Belong in Your Model From Day One
The Geographic Adjustment Factor applies the Medicare PFS geographic adjustment based on service facility ZIP code. The maximum positive adjustment is 75 percent above the base rate – a high-complexity caregiver patient in the first six months can reach $630 instead of $360. If your program serves multiple ZIP codes, the weighted average matters for your P and L.
The Health Equity Adjustment begins in the second performance year. Programs at or above the 80th percentile of the CMS equity score receive an additional $15 per patient per month. Programs in the 51st through 79th percentile receive nothing. Programs at or below the 50th percentile are docked $6 per patient per month. The equity score draws on the National Area Deprivation Index, State ADI, Part D Low-Income Subsidy status, and dual eligibility rates. The $6 downward adjustment belongs in your base-case model, not in a downside scenario.
The Performance-Based Adjustment also begins in the second performance year. High-performing programs receive up to a 10 percent increase in total payments; poor-performing programs face up to a 3.5 percent reduction. A program with 150 aligned patients generating $180 per month average faces a swing of roughly $5,100 in monthly revenue between the best and worst PBA outcomes. That is a material operating lever, not a rounding error.
The revenue step-down at six months of alignment is not an edge case – it is the designed structure of the model. If your financial projection does not model it explicitly, the projection is wrong.
Respite Revenue: Real but Conditional
Respite reimbursement for PY2026 caps at $2,625 per patient annually. Three G-codes apply: G0529 for in-home respite per four-hour unit, G0530 for adult day center per day, and G0531 for facility-based care per 24-hour unit. Three conditions gate eligibility: moderate or high complexity, an unpaid primary caregiver, and community-dwelling status. Low-complexity patients are ineligible. If your population is heavily low-complexity, the respite line on your P and L is close to zero. The Northwell 35-percent utilization rate is a useful planning reference – but run the eligibility filter against your actual census before you model the revenue.
Building the Cost Side Without a Published Benchmark
CMS provides no staffing ratio, no technology cost guidance, and no reference for what a fully loaded dementia program costs to operate. Build from your own cost categories: care navigator compensation and benefits; clinician time for assessments, care plan development, and the 24/7 access function; third-party partner contract if applicable; technology and data submission infrastructure; and overhead allocation from your parent organization.
Many GUIDE participants – particularly physician practices without their own nursing or care management staff – work with specialized third-party companies whose nurses handle patient outreach and day-to-day care management on the practice’s behalf. This is legitimate and common. If you are in that category, the partner contract is likely your single largest expense line, and the economics of the program turn on whether the contracted rate leaves sufficient margin at your actual tier mix and census. Evaluating and structuring those contracts is a business and operations function, not a clinical one – and Guidestone helps participants think through that analysis without providing the care management services ourselves.
From the Field
A four-physician geriatric practice in the Mid-Atlantic reached 90 aligned patients by mid-year one. Revenue ran about 18 percent below projection. The causes: a tier mix weighted more heavily toward low complexity than assumed, the six-month step-down hitting a larger share of the census than modeled, and the geographic adjustment applying at a lower factor than estimated based on county-level rather than ZIP-level data. Guidestone rebuilt the revenue model patient-by-patient using the monthly beneficiary alignment file, mapped tier and alignment-length distributions against actual rates, and built a monthly P and L reconciling DCMP revenue to the care management partner invoice and internal staff cost. The model identified that adding 40 moderate-complexity patients would move the program from marginal to solidly positive – information the practice could act on through alignment outreach rather than discovering at year-end.
The Monthly Review That Makes the Model Useful
A P and L built once at program launch is a historical document within six months. Four numbers, reviewed monthly, tell you whether the program is tracking: total aligned census by tier, the share of census past six months of alignment, realized DCMP revenue against projected, and care delivery cost per patient per month against budget. If those four are not on a single page that someone owns before the month closes, the P and L is a reporting artifact rather than a management tool.
Advice Versus Execution in a Financial Model
A consultant can build the revenue model and hand you a documented spreadsheet. If your program has a finance-capable administrator who can maintain it and escalate when numbers diverge, that may be all you need – and it is often the more economical choice. We will say so when it fits.
What we see more often is a program where the financial model was built at enrollment and has not been touched since, because the person who would maintain it is also managing care navigator schedules, responding to CMS data requests, and handling the operational fires that GUIDE participants consistently report. That is the situation where a fractional executive adds value by staying in the work – sitting with the billing team to reconcile DCMP remittances, working the beneficiary alignment file into the P and L each month, and flagging the revenue step-down before it shows up as a quarter-end surprise. The label “fractional executive” is unregulated and some firms use it to describe work that is delivered as a report and handed over. Ask directly: who does the monthly reconciliation, and are they working in your systems?
Sources
- CMS GUIDE Participant Model Incentives Fact Sheet (payment tables) — https://www.cms.gov/files/document/guide-participant-model-incentives-factsheet.pdf
- CMS GUIDE Payment Methodology Paper — https://www.cms.gov/priorities/innovation/files/guide-payment-methodology-paper.pdf
- CMS MLN Fact Sheet MLN7172818, Guiding an Improved Dementia Experience Model (July 2026) — https://www.cms.gov/files/document/mln7172818-guiding-improved-dementia-experience-model.pdf
- Mongelli et al., Innovation in Aging (2025) – Northwell Health GUIDE year-one experience — https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12761895/
- CMS GUIDE Model main page (accessed August 2026) — https://www.cms.gov/priorities/innovation/innovation-models/guide
More in the Dementia Care Series
- What Nobody Warned You About Running a GUIDE Program — the operating problems every participant runs into, and which ones are actually fixable.
- Build It or Partner? Choosing Care Management for Your GUIDE Program — how to decide between building navigation in-house and partnering, and how to structure the deal.
- GUIDE Quality Metrics, the Performance-Based Adjustment, and the Reporting Burden — how GUIDE quality metrics are scored and what the performance-based payment adjustment actually costs or earns.
- GUIDE Alignment, Tier Assignment, and the Billing Mechanics That Cause Denials — the alignment and billing mechanics that determine whether your GUIDE revenue actually arrives.