GUIDE CMS & research

GUIDE CMS initiative: research and design behind GUIDE Companion

Explore CMS GUIDE requirements, research on caregiver support, and the questions to work through when planning a GUIDE Companion experience.

Concept illustration of AI displays and a digital human figure in a clinical room.

CMS GUIDE model at a glance

From CMS GUIDE goals to a practical Companion experience

CMS GUIDE is a dementia care delivery and payment model. It includes care navigation, caregiver education and support, and respite for eligible people. The participating program is responsible for meeting the model’s requirements.

GUIDE Companion can bring reviewed resources, communication, and agreed workflows around that work. We scope the tools, supported connections, access, and responsibilities with the participating organization. Software use does not establish CMS participation, eligibility, or payment.

Where a participating adult day center uses Day Center Companion, staff can connect intake, attendance, a reviewed family update, and closeout. The GUIDE program remains responsible for its respite arrangements and reconciliation.

  • Make the route to the care navigator and community resources clear.
  • Choose caregiver resources with the people who will use them.
  • Explain respite options using the participating program’s current eligibility and service rules.
  • Agree which records the program needs and who reviews them.

Supporting research

Adults and children smiling together outside a home.

Aligning steps in GUIDE projects

Plan one useful workflow with the people involved

Begin with the question a person living with dementia, a caregiver, or a team member needs answered. Map who can help, which resources are available, and what happens if the next step cannot be completed.

A standard branded preview follows within 24 hours of agreeing on the initial scope. Use sample information to explore the proposed workflow. Agree the schedule for a live pilot separately, including content review, access, supported connections, and training.

  • Discovery: hear from people receiving care, caregivers, clinicians, and community partners.
  • Preview: try the proposed resources and handoffs with sample information.
  • Pilot: review understanding, follow-through, access barriers, and the complete work required.
  • Next decision: continue, revise, pause, or expand based on what the team and participants learn.

Supporting research

A model ambulance crosses a miniature bridge between buildings on a desk.

Evidence under the surface

What the research helps us ask

A service day, a family update, and a navigation tool do different jobs. Evidence about one cannot establish the effect of another. The sources below help us choose questions for local evaluation.

For example, a reviewed update may help a family prepare a question. Ask whether it does, whether the question reaches the right person, and how much work the exchange creates for everyone involved.

  • Hear the person receiving care and the caregiver separately.
  • Keep findings attached to the people, setting, and intervention studied.
  • Treat app use and message delivery as activity records, not proof of understanding or reduced strain.
  • Use participant feedback and staff experience to decide what to try next.

Supporting research

Bar chart of health-system priorities for generative AI, led by charge capture and reconciliation, patient-data analysis, and workflow automation.

Learning alongside GUIDE teams

Return the learning to the people who contributed

Agree what feedback to invite, who will review it, and how participants will hear about any resulting change. A concern about confusing wording needs a response as much as a completed survey needs a count.

Partners may choose to contribute reviewed methods or findings to shared learning. Agree permission, attribution, and information boundaries first; participation does not grant access to private care records.

The Day Center research page brings together service-day evidence, a proposed evaluation approach, and dated funding sources. Read its findings in context: the evidence cited does not establish that software reduces costs or delays residential placement.

  • Choose a small set of questions that can inform a real decision.
  • Include understanding, access, follow-through, and staff and family effort.
  • Record what changed and what remains uncertain.
  • Check current CMS guidance when a program requirement changes.
Chart comparing generative-AI risks organizations consider relevant with risks they plan to mitigate; inaccuracy, cybersecurity, and intellectual-property infringement lead the list.

Ready when you are

See your connected solution, in your brand.

Book a 30-minute call to agree on a starting scope across your public care wall, care app, and partner portal. Try a working preview under your name, with sample people and notes, and explore how the configured workflows fit your team and community.