For a programme director, GUIDE is less interesting as a Medicare announcement than as the first time the coordination work your team already does has a payment attached to it.
GUIDE — Guiding an Improved Dementia Experience — launched on 1 July 2024 as a voluntary CMS model running eight years. Roughly 390 organisations are participating, with an established cohort that began in July 2024 and a second cohort delivering from July 2025.
The model pays for a defined bundle: comprehensive assessment and care plan, ongoing caregiver support and training, care coordination, and respite of up to $2,500 per year per eligible beneficiary.
Read that list again from the perspective of a team already running caregiver support. Most of it describes work that is currently done because someone thinks it matters, funded out of general programme money or not funded at all.
A named navigator becomes a funded role rather than a stretched one. Most programmes have someone playing this part informally. GUIDE makes it a defined function with a payment behind it, which changes what you can staff and what you can promise.
Respite arrives without a state queue. If you operate anywhere with a waiting list — Florida's Alzheimer's Disease Initiative had roughly 17,081 people waiting as of mid-2025 — a second funding route with different eligibility is materially useful. Families can be in both.
The caregiver becomes a payable subject, not an adjacent one. That is the structural shift. Programmes have always known the caregiver determines whether the care plan works; GUIDE is the first Medicare model to fund acting on that at scale.
Not trivial, and worth being clear-eyed about:
The honest answer depends on your existing caseload and on geography.
If you already serve a meaningful number of community-dwelling people with dementia on Traditional Medicare, the model largely funds work you are doing. If your population skews Medicare Advantage or institutional, the addressable share may be smaller than the headline suggests — and that is worth counting before committing, not after.
Geography matters too. Participation is uneven, and in a region with several participants the differentiating question becomes referral flow rather than eligibility.
A model test is not a permanent programme. GUIDE runs eight years from 2024 and what follows has not been decided. Build the capability so it survives the model, not so it depends on it.
1. How many of your current families are Traditional Medicare, community-dwelling, with a documented dementia diagnosis? That number is the size of the opportunity, and most organisations have not counted it. 2. Who is your navigator, and what are they doing today instead? If the answer is nobody, the model is a hire, not a margin. 3. Who else near you already participates? A published participant list exists; the answer changes whether you are entering a market or joining one.
CMS publishes the GUIDE model details and the participant list. LeadingAge state affiliates and Area Agencies on Aging generally track local participation, which is faster than reading the national list if you only need to know your own region.