Read this first
Nobody who works on this site is a clinician, and nothing we rate is clinical. The agencies on our ranking provide everyday, non-medical support at home. None of them diagnose or treat memory loss, and neither do we. That belongs with a doctor.
This page shows our working: what we scored, where each number came from, who pays us, and how to tell us we got something wrong. The ranking itself is on the front page.
How we rate in-home memory support, and where every number comes from
A family choosing help for someone whose memory is changing is not really shopping for a service. They are choosing who will be in the kitchen at four in the afternoon, and who will answer the phone when the plan stops fitting. So our criteria are about those things, and not about anything a caregiver would need a licence to do.
The five criteria
Each agency is read against the same five criteria, with equal weight. No agency on the ranking carries a numeric score from us.
| Criterion | What earns a high score | What we check |
|---|---|---|
| Local ownership and accountability | An owner who can be named, based in the St. Louis area, and a real person on the phone | The company's own About and contact pages |
| The first conversation | A first call that is a conversation about the household, with support that can start small | How the company describes getting started, on its own site |
| Plan flexibility | Hours, evenings and overnight support that change when needs change, without a new contract each time | Service descriptions, stated minimums, overnight and around-the-clock options |
| Non-medical clarity | A company that says plainly what it does in the house and what it does not, so a family never mistakes support for treatment | Whether the site separates non-medical support from anything clinical |
| Family communication | Contact on the family's terms: optional phone numbers, no forced sign-ups, information given away freely | The company's own forms, guides and contact options |
How the order comes together
The five ranked agencies carry their real Google rating and review count and a short qualitative label instead of a score, because we would rather print a number someone else measured than one we made up. Their order reflects our editorial reading of the same five criteria against each company's own website: how central memory support is to what the company describes, whether it is locally owned, how clearly it keeps non-medical support separate from anything clinical, and how established its local footprint is. That is judgment, and we are saying so.
Who could be ranked
- A row in our Google Maps file of St. Louis area home care listings, with 20 or more Google reviews.
- A public website that describes dementia, Alzheimer's or memory support as a service, in its own words, read on September 12, 2026.
- Non-medical support at home on offer: routine, companionship, personal care or respite, not only skilled services.
- Excluded by publisher policy: FreedomCare, which is a Medicaid consumer-directed program rather than an agency, and a small number of other listings. FreedomCare may appear in a pure count with that label.
The lists
The lists on the front page are counts, not verdicts. Each states its item set, its sort (Google review count, highest first, ties broken by rating) and its pull date in a methodology box on the page. A list page needs at least ten rows to be published, and a ranking needs at least five. Offices of the same company with separate Google listings get separate rows, labelled by location.
The survey figures
The state tables on this site come from the Centers for Disease Control and Prevention's Alzheimer's Disease and Healthy Aging Data, resource hfr9-rurv, collected through the Behavioral Risk Factor Surveillance System. The dataset is public domain and attributed to the CDC Division of Population Health. Our copy, engine/data/cdc_cognitive_decline.json, was pulled September 10, 2026.
The row selector is exact: Class = Cognitive Decline; YearStart = 2019 and YearEnd = 2022; age stratification 65PLUS; second stratification OVERALL; and a value the CDC did not suppress.
The dataset files two families of rows under YearStart 2019, one ending 2019 and one ending 2022, and for 22 jurisdictions they disagree on at least one question. We use the YearEnd 2022 family only, on every page, and never mix in the other to fill a gap. For Missouri the two families happen to agree on all four questions.
We drop the five regional roll-up rows, drop suppressed rows (Puerto Rico is suppressed on these questions), and keep a jurisdiction only if it has a value for all four questions, so every table ranks the same 46 places. Missouri and its eight neighbours uses a smaller set fixed before any value was read, ranked on Q30 only.
These are self-reports gathered by telephone. Nobody in the survey was examined or diagnosed, the CDC does not present these figures as a measure of dementia, and no state percentage says anything about the person you care for. Every table prints the CDC's confidence interval, and where intervals overlap heavily we say the difference is not meaningful.
Photographs
Photographs on this site are supplied by New Plan Care Solutions and credited to it. None shows a clinical setting, and none is presented as a picture of any other company's staff or clients.
Who supports this site
This site is supported by New Plan Care, a St. Louis in-home care company. New Plan Care is not on the ranking and is not in any table on this site.
Pages are written by the Memory Care at Home STL desk. There is no named medical author, because there is no clinician on this site and inventing one would be worse than admitting it.
Corrections
If a rating or review count here does not match the pull date, if a company's website no longer says what we quote, or if a CDC figure does not match the resource using the selector above, that is a defect and we fix it. We re-check the row against the file or the live page, correct the page, and change its date. A company that believes its label is unfair can point us to the page on its own website that shows otherwise, and we will read it.