CMS built five metrics into GUIDE that directly affect payment – including two developed specifically for this model. Understanding the measurement methodology is not an academic exercise; it determines whether you receive a bonus or a penalty.
Key Takeaways
- Five metrics drive the performance-based adjustment: high-risk medication use, quality of life outcome, caregiver burden via the Zarit Burden Interview, total per capita cost, and long-term nursing home rate – two of the five were developed specifically for GUIDE.
- The performance-based adjustment can swing total GUIDE payments by up to 13.5 percentage points – a 10 percent bonus for top performers against a 3.5 percent penalty for the bottom – but it does not begin until the second performance year, giving programs time to build their measurement infrastructure.
- The Zarit Burden Interview and quality-of-life outcome data must be collected prospectively from caregivers and patients – they cannot be reconstructed from clinical notes or claims data, which means your workflow must capture them at the point of care.
- The Northwell year-one report identified changing program requirements and data submission system constraints as two of their primary operational challenges; quality reporting infrastructure is a program management problem before it is a clinical one.
The GUIDE performance-based adjustment is not a small number. The best-performing programs receive up to a 10 percent increase in total DCMP payments. The worst face a 3.5 percent reduction. At a census of 150 patients generating an average of $180 per month, the difference between the top and bottom of the distribution is more than $5,000 in monthly revenue – a material budget variance in either direction.
The adjustment does not begin until the second performance year for each track, which means Established track programs (year one began July 1, 2024) are already in the window where it applies. New Program track participants, whose first delivery year began July 1, 2025, have more runway, but the data collection infrastructure that feeds the adjustment must be built and running long before the measurement period closes. The metric that required no workflow in year one requires a workflow in year one to produce a score in year two.
The formal federal evaluation is being conducted by Mathematica under a 2024 through 2034 contract; no interim report has been published as of August 2026. The quality scoring methodology described here comes directly from CMS payment and program documentation.
The Five Metrics in Plain Terms
Use of High-Risk Medications in Older Adults. A claims-based measure tracking whether patients receive medications on the geriatric high-risk list – benzodiazepines, first-generation antihistamines, anticholinergics, and related classes. Externally calculated from Medicare claims, so it does not add documentation burden. The operational implication is still real: your interdisciplinary team should have a defined medication reconciliation protocol, because this metric will reflect whatever your prescribers do regardless of whether you are watching it.
Quality of Life Outcome for People with Neurological Conditions. A patient-reported outcome measure administered directly to patients. It is not extractable from clinical notes. Someone on your team must administer the instrument at the required intervals and record results in a submittable format. Collecting patient-reported outcomes in a cognitively impaired population requires explicit protocol decisions about who administers it, when, and what happens when the patient cannot complete it reliably.
Zarit Burden Interview. A 22-item validated instrument measuring caregiver burden – the degree to which caregiving affects the caregiver’s health, well-being, social life, and finances. CMS developed this as a new metric for GUIDE. It is administered to the caregiver, not the patient, and cannot be reconstructed after the fact. For programs using a third-party care management partner for caregiver contact, the partner’s workflow must include Zarit administration at the right intervals, with data flowing back in a submittable format. Verify this is in the contract before the measurement window opens.
The Zarit and quality-of-life instruments cannot be reconstructed from clinical notes. If your workflow does not collect them prospectively at the right intervals, your performance-based adjustment score will be built on incomplete data – and there is no fix after the measurement window closes.
Total Per Capita Cost. A claims-based measure of total Medicare expenditure for aligned GUIDE patients versus a risk-adjusted expected level. Programs that reduce unnecessary utilization and prevent hospitalizations score well. A small retrospective study (n=85 matched pairs, Innovation in Aging 2025) reported a 22 percent decrease in total cost of care in the GUIDE group versus a 2 percent increase in the comparison group. Treat that as preliminary – the formal Mathematica evaluation will tell us how it holds at scale. But it suggests the mechanism works.
Long-Term Nursing Home Rate. The rate at which aligned patients transition to long-term nursing home placement, also compared against a risk-adjusted expected level. Developed specifically for GUIDE. The same preliminary study reported a 28 percent decrease in hospital days in the GUIDE group; nursing home placement was not separately reported at that scale, and this metric will be closely watched in the formal evaluation.
What the Reporting Burden Actually Looks Like
The Northwell year-one paper is candid: their challenges included constant workflow adjustment, database and submission system constraints, changing program requirements, and needing external partners to supplement home visits and respite. Weekly data review and structured recurring meetings emerged as essential – not things they had in place from the start.
The claims-based metrics – high-risk medications and total per capita cost – are externally calculated and do not add direct workflow burden. The three outcome measures require structured data collection woven into the care navigator’s and clinician’s workflow at defined points. If those collection moments are not built into the workflow explicitly, the data submitted to CMS will be incomplete and the PBA score will reflect it.
Building the Infrastructure Before the Measurement Window
- Map when each instrument is required. Confirm administration intervals from CMS documentation. Build those intervals into the patient record and care navigator workflow as calendar triggers, not as optional clinical tasks.
- Assign ownership. Every instrument needs a named role, a backup, and an escalation path when the patient or caregiver cannot complete it. Ambiguity about ownership is the primary reason outcome data is missing at measurement time.
- Audit your partner contract. If a third-party company handles caregiver contact, the contract should specify which instruments they administer, at what intervals, in what format, and with what turnaround. This is a contract term, not an implementation assumption.
- Build a monthly quality dashboard. Track instrument completion rates alongside DCMP billing metrics. A care navigator at 70 percent Zarit completion needs coaching before the measurement window closes, not after the score arrives.
- Model the health equity adjustment separately. A program that improves its PBA score through strong clinical outcomes but falls below the 50th percentile equity threshold still faces the $6 per patient per month equity penalty. The two adjustments are calculated independently and are additive.
One financial interaction to flag: GUIDE DCMP and respite payments count toward MSSP ACO expenditures for patients aligned to both a GUIDE participant and an ACO. If your organization is in both programs, financial modeling for quality performance should account for both simultaneously, and legal or compliance counsel should review the interaction before you set program scale targets.
From the Field
A geriatric-focused practice in the Southeast with 120 aligned GUIDE patients reached the end of its first performance year with incomplete Zarit data on roughly 40 percent of its caregiver population. Care navigators had been having caregiver support conversations – but nobody had built the instrument into the workflow as a required data collection event. The conversations happened; the structured data did not. Guidestone built a quality reporting infrastructure alongside the practice: a monthly instrument completion dashboard, a care navigator checklist with Zarit and quality-of-life administration as explicit steps at defined intervals, and a weekly data review meeting where the program director could intervene before the measurement window closed. The practice also added Zarit administration to its third-party partner scope of work, with submission format requirements written into a contract amendment. By the midpoint of year two, completion rates were above 85 percent – enough to support a meaningful PBA score rather than an incomplete rating.
Consulting Versus Execution in Quality Infrastructure
A consultant can design the quality reporting workflow, write the care navigator checklist, and spec the partner contract amendment language. If your program has a director with the bandwidth to implement those designs and maintain the monthly review cadence, consulting may be the right and more economical engagement – we will say so if it fits.
What the Northwell experience suggests is that the reporting burden is not a one-time setup problem. It is a continuous operational discipline: weekly data reviews, monthly completion audits, workflow adjustments as the program matures and CMS requirements evolve. A fractional operations leader works inside the program’s data systems, sits in on the weekly review, and adjusts the workflow when completion rates slip rather than flagging it in a report. Ask any firm you engage to be specific about which mode they are actually operating in. “Fractional executive” is an unregulated label; the test is whether they are working in your systems or delivering to them.
Sources
- CMS GUIDE Payment Methodology Paper — https://www.cms.gov/priorities/innovation/files/guide-payment-methodology-paper.pdf
- CMS GUIDE Participant Model Incentives Fact Sheet (payment tables) — https://www.cms.gov/files/document/guide-participant-model-incentives-factsheet.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
- CMS GUIDE Model FAQs (accessed August 2026) — https://www.cms.gov/priorities/innovation/guide/faqs
- Mongelli et al., Innovation in Aging (2025) – Northwell Health GUIDE year-one experience — https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12761895/
- Mathematica, Evaluation of the Guiding an Improved Dementia Experience Model (2024-2034) — https://www.mathematica.org/projects/evaluation-of-the-guiding-an-improved-dementia-experience-model
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 Program Economics: Modeling Per-Beneficiary Revenue and Building Your P and L — the financial model that determines whether your GUIDE program is an asset or a cost center.
- GUIDE Alignment, Tier Assignment, and the Billing Mechanics That Cause Denials — the alignment and billing mechanics that determine whether your GUIDE revenue actually arrives.