Dementia Care Needs a Capacity Contract: Finance the Household, Not Just the Referral

01 / THE ARGUMENT

Contract for usable capacity

A dementia diagnosis does not reveal how many hours of dependable help exist in the home. A referral does not create a worker, and a trained worker cannot supply hours that nobody funds. My thesis is that a viable community dementia business must contract for usable care capacity: who can respond, what responsibilities they hold, how family caregivers obtain relief, and what happens when the arrangement fails. The relevant unit is the person and their support network, including people who live without an available family caregiver.

This reframes the buyer's decision. A health plan, provider network, or community commissioner should ask what operating capacity a payment purchases before accepting a utilization forecast. Navigation has value, but a list of inaccessible services cannot substitute for service availability. Home care should also reflect the person's preferences; maximizing days at home is not a sufficient goal when the home arrangement becomes unsafe or unwanted.

02 / THE ARGUMENT

Better experience is a real but bounded result

The September 2025 prespecified D-CARE analysis compared health-system care, community-based care, and usual care in 2,176 person–caregiver pairs over 18 months. Cognition, function, patient quality of life, and overall caregiver burden did not differ significantly. Caregiver satisfaction improved: least-squares differences were 2.6 points for health-system care and 3.3 for community care on an 11–55 scale. The clinical importance of those satisfaction differences was not established; pandemic disruption also affected delivery. This analysis did not establish utilization savings. [1]

The operating inference is narrower than either enthusiastic adoption or dismissal. A buyer may reasonably purchase accessible advice and a better care experience. It should describe that purchase honestly and price it accordingly. Satisfaction is not interchangeable with caregiver relief, functional preservation, or lower hospital spending. A product promising all four needs evidence for all four. Because usual care differs across markets, even a reproducible improvement in one setting may shrink where existing coordination is strong.

03 / THE ARGUMENT

Economic perspective changes the apparent result

The October 2025 German InDePendent cluster trial tested expanded nursing responsibilities. Its efficacy analysis included 391 patients; the economic analysis included 332. At six months, the modeled unmet-needs ratio was 0.26 (95% CI, 0.17–0.40). Caregiver burden did not significantly improve. The payer analysis reported €1,425 higher costs and a 0.01 QALY gain whose confidence interval crossed zero; estimated cost-effectiveness probability was 47% at €160,000 per QALY. A caregiver subsample suggested societal savings driven largely by reduced informal care. Short follow-up and German delivery conditions constrain extrapolation. [2]

That difference between perspectives is the strategic issue. Freeing family time can create substantial value without releasing cash to the health plan. Conversely, a plan can lower its own spending by transferring work to families. Neither effect should be hidden inside a single return-on-investment number. Commissioners need separate payer, provider, and household ledgers, with an explicit decision about whose benefit justifies whose contribution. A favorable cost-effectiveness interpretation does not automatically identify a profitable service contract.

04 / THE ARGUMENT

Training is an input to capacity

A June 2025 quasi-experimental evaluation of online training for California home-care workers found improved dementia knowledge and self-efficacy, without reductions in worker distress or depression, or recipients' emergency visits and hospitalizations. The nonrandomized before-and-after design limits causal attribution. [3] A scoping review first published in December 2025 and assigned to the March 2026 Gerontologist issue analyzed 12 distinct training programs. Its literature search ended in June 2024; the recent publication date does not make its underlying evidence current through 2026. It maps an uneven training literature rather than establishing pooled savings. [4]

My proposed mechanism is practical: training helps someone recognize a problem; protected time allows a report; a reachable clinician interprets it; an available service addresses it. The chain fails at its narrowest point. Workforce budgets should therefore include paid learning time, supervision, predictable coverage, and a defined route for clinical escalation. These are testable design choices, not effects established by the training studies. A certificate count tells a board little about whether a household receives help on a difficult evening.

05 / THE ARGUMENT

Design the pilot around a constrained household

Consider a regional provider partnering with a home-care agency for a one-year dementia support pilot. The community-care director owns the service; a clinical lead owns escalation decisions; the agency operations lead owns staffing reliability. Enrollment records patient goals, existing paid hours, caregiver availability, and unmet needs. An assigned coordinator closes service requests, while a qualified clinician handles medical decisions. The contract reserves a defined amount of respite and rapid-response capacity instead of assuming referrals will be fulfilled.

The budget boundary includes assessment, coordination, clinical backup, paid worker training, replacement coverage, and contracted relief hours. It excludes speculative savings from future institutionalization. If local demand exceeds the pilot's capacity, report waiting time and unmet demand instead of silently excluding difficult households. Compare with a contemporaneous usual-care cohort using baseline need and support availability, or randomize phased access when appropriate. Preserve the right to move into a different care setting when it better serves the person.

06 / THE ARGUMENT

Use two economic tests and a falsification rule

The provider test is cash-based: contracted payments + attributable encounter margin − staff, partner, infrastructure, and coverage costs. The buyer test adds only its own verified avoided expenditures. The household account separately reports unpaid care hours, out-of-pocket costs, and caregiver outcomes; any monetary value assigned to time is a transparent analytic assumption. Do not count a family hour saved as both a cash saving and a quality gain without explaining the valuation.

The capacity thesis would fail if filled support hours and prompt clinical responses improved but patient goals, unmet needs, and caregiver outcomes did not improve against a credible counterfactual. It would also fail financially if adequate staffing required a payment the buyer could not sustain. Review cancellations, turnover, unresolved requests, and burden alongside claims. A fall in hospitalization accompanied by rising caregiver distress is a signal to investigate transferred work, not an automatic success.

07 / THE ARGUMENT

The executive decision

Purchase a bounded capacity contract and evaluate both household experience and organizational economics. Set aside funding for the labor that makes advice actionable, and keep savings claims conditional on observed results. Expansion should follow demonstrated reliability and patient value, with a payer arrangement that can survive the loss of optimistic assumptions. This creates a defensible home-care business: one that knows whose work it supports, whose budget it changes, and which outcomes it has actually earned.