Concept paper Puteaux, September 8, 2026
This concept note is initially presented in the context of the German home-care sector, but the underlying principles are designed to be applicable across home-care systems.
Germany is moving toward increasingly electronic documentation and billing of home-care services (elektronische Leistungserfassung). But digitising the existing process does not, by itself, make billing reliable.
Today, the operational chain is largely built around the provider: Provider schedule → Caregiver → eLNW time record → Billing
This chain can reliably establish that a caregiver worked. It does not necessarily establish that the beneficiary received the care they requested under their care plan.
This distinction becomes critical when the provider cannot organise its workforce according to the beneficiary's needs.
A beneficiary requests: 08:00–09:00 every day. The provider cannot staff this request consistently. It reorganises its workforce according to caregiver availability. Caregivers may work longer visits on fewer days. The provider's eLNW can therefore contain perfectly valid records of working time.
But the beneficiary's requested visits may remain: 08:00–09:00 — MISSED, day after day.
The problem is not that the eLNW is wrong. The problem is that the eLNW does not contain the beneficiary's original request as its reference point. The two systems can therefore produce two different, equally correct, representations of reality.
The current logic can answer: Did the caregiver work?
Reliable home-care billing needs to answer: Was the care requested by the beneficiary actually delivered?
A caregiver may work six hours. The beneficiary may nevertheless have missed every one of the seven one-hour visits they requested. Both facts can be true at once. The weakness is not inaccurate data — it is the absence of a single service-level reference connecting the beneficiary's request to the delivery and the billing.
Germany should establish a clear principle: for each home-care service, the beneficiary's requested service and its delivery status constitute the primary service record. The service record follows the care request through five states:
REQUESTED → ACCEPTED → DELIVERED
or
REQUESTED → ACCEPTED → MISSED → REPLACED → DELIVERED
REQUESTED — what did the beneficiary request? ACCEPTED — has a caregiver taken responsibility for the visit? DELIVERED — was it actually delivered? MISSED — if not, it is recorded as missed. REPLACED — was the care subsequently recovered?
This five-state record — not the working-time log — becomes the reference used throughout the rest of this note.
Operational flexibility: delivery windows and regional adaptationDefining a requested visit (e.g. 08:00–09:00) does not imply rigid, down-to-the-minute scheduling that would be unworkable in daily care operations.
To accommodate traffic, operational contingencies, and caregiver travel time, the service record architecture incorporates a dynamic validation window:
A visit transitions to MISSED only when this defined tolerance window expires without a caregiver check-in, preventing false alarms while maintaining an objective trigger for replacement care.
Germany should not attempt to reconcile two competing definitions of the same service indefinitely. It should define which record answers which question:
If the beneficiary requests 08:00–09:00 and that visit does not occur, the record must stay MISSED. It should not disappear because the provider subsequently reorganises its workforce, and it should not automatically become "2 hours delivered the next day."
For a dependent person, a missed requested visit is not a data discrepancy — it may mean an essential care has not been provided. A recorded missed visit should trigger the appropriate replacement mechanism, so the care is recovered while it still matters to the beneficiary, not merely flagged after the fact.
Under a provider-centred model, the provider determines what can be scheduled according to available caregivers. Under a beneficiary-centred model, the beneficiary's request remains visible even when the provider cannot fulfil it. This does not prevent the provider from managing its workforce — it prevents the provider's workforce constraints from becoming invisible to the beneficiary, the payer and the public authority.
Billing then follows the service actually delivered, including replacement care: when a visit is missed, a replacement is mobilised, accepted and delivered, and the caregiver who actually performed it — potentially an autonomous caregiver mobilised specifically to avoid a break in care — becomes the identifiable provider of that service for billing purposes.
Funding should be traceable to the care actually delivered, including replacement care, rather than exhausted by an undelivered provider commitment. This note does not attempt to determine whether a given replacement is reimbursable under existing SGB XI / Pflegekasse rules — that remains a matter for the applicable contractual and funding framework. Its purpose is to create the digital service record that makes the full chain auditable: requested → committed → delivered or missed → replacement → delivered replacement → corresponding billing.
One question will immediately follow any regulator reading this proposal, and is worth addressing directly rather than leaving implicit.
Who bears the cost when a replacement visit may be more expensive than the original commitment? German reimbursement does not work like a flat hourly rate with a margin that could simply be reallocated — home-care billing runs through negotiated service bundles (Leistungskomplexe under § 89 SGB XI) capped by a monthly Pflegesachleistung budget per Pflegegrad, so there is no single hourly margin to redirect between a provider's employee and a replacement caregiver.
Germany already has a more direct lever than a provider-margin arrangement: § 77 SGB XI allows a Pflegekasse to contract directly with an individual qualified caregiver (Einzelpflegekraft) for home care. Where this route is used for a replacement visit, the caregiver bills the Pflegekasse directly at their own agreed rate — the cost question is settled without needing to route funding through, or subsidise it from, the original provider's margin. Today § 77 is a discretionary, rarely used channel; this proposal's contribution is to make it the natural funding path specifically for replacement visits triggered by a recorded missed request, which would need to be worked out with the Pflegekassen and provider associations rather than assumed.
The data footprint required to operate this model is deliberately narrow. This is an argument for a GDPR-compatible design, not a reason to defer the question.
A clearly defined data controller, purpose, retention period and access right for the beneficiary or their legal representative should be specified from the outset.
The important point is that data minimisation is not a future safeguard to be promised after deployment. It is a structural property of the service record itself.
Germany should establish a regulatory requirement for an electronically traceable beneficiary-level service record, building on the existing move toward elektronische Leistungserfassung and complementing existing quality-assurance mechanisms (MD Qualitätsprüfungen). For every relevant home-care service, the system should be able to establish:
1. One source of truth for beneficiary care
The beneficiary-level service record is the authoritative record of requested and delivered care.
2. No double clocking
The caregiver should not have to record the same working time twice.
3. Missed care is an operational event, not an accounting discrepancy
A missed requested visit triggers the appropriate replacement process.
4. Billing follows the service record
Billing follows the service actually delivered — including replacement care. When an accepted visit is not delivered, the missed visit is recorded and the applicable replacement mechanism is triggered; if an autonomous caregiver successfully delivers the replacement, that caregiver becomes the identifiable provider of the replacement service for billing purposes.
5. Protecting providers while safeguarding care continuity
This beneficiary-centred model is not designed to penalise established providers (Pflegedienste) operating under severe staffing constraints. Rather, it creates a transparent safety net when a provider reaches operational capacity.
Home-care systems need a clear service-level reference. The beneficiary's recorded service request should remain the reference for the care that was expected: delivered when the visit takes place, missed when it does not — and, when care is missed, triggering the appropriate replacement mechanism.
Only then can electronic billing move beyond: Hours worked → Billing, toward: Care requested → Care accepted → Care delivered → Care replaced if missed → Billing.
This does not prevent providers from organising their workforce as they consider appropriate. It establishes a different principle: workforce scheduling must not silently redefine the care requested by the beneficiary.
The political choice is therefore simple: Do not ask beneficiaries to adapt their care needs to the provider's workforce availability.
Make service delivery accountable to the beneficiary's recorded request. When that commitment cannot be fulfilled, make the failure visible, trigger replacement care, and allow the caregiver who actually delivers that replacement to be identified and billed.
That is the foundation of reliable billing — and of reliable home care.
Chi Minh PHAM
Founder, YouTime — home care infrastructure
https://www.linkedin.com/in/chiminhpham/