By Christoph Eberhardt · CEO, DUX Healthcare · 19 April 2026
Most international digital therapeutics (DTx) teams arrive in Germany with DiGA (Digitale Gesundheitsanwendung — Germany’s certified digital health application) as their anchor concept. It is the most visible reimbursement path, it has a clean BfArM process, and it has a public product directory. It is also one of seven regular paths under German statutory reimbursement, and for a meaningful share of products it is not the right one. This hub maps the path choice: which German reimbursement route fits which product, what trade-offs each path forces, and how the decision actually gets made. DUX has its deepest operational experience on DiGA (mHealth Suite, five DiGAs under contract); for HMV, DiPA, selective contracts, NUB, ZPP and the Innovation Fund we work mandate-specifically and reference public sources where the secondary literature is reliable. Where DUX has direct experience, claims are marked DUX position.
Why path choice comes before product choice
Three common mistakes in early-stage product planning:
- Finish the product, then look for the reimbursement path. A team has a clinical concept, builds an MVP, and sits down with the DiGAV for an afternoon during regulatory scoping. Three weeks later it turns out the indication is not codable (no DiGA path), or the target group is care-dependent people (DiPA path), or the application extends a medical aid (HMV path). The MVP was built for the wrong market.
- “We’ll go DiGA because everyone goes DiGA.” DiGA is a specific path with narrow admission criteria (DiGA Guide v3.6 Sec. 2.1): medical device, low-to-moderate risk class, digital main function, codable indication, demonstrable positive care effects. A product that meets three of five criteria is not a DiGA – it is a candidate for one of the other paths.
- Force the path switch late. “We’ll apply as a DiGA, and if the BfArM rejects we’ll pivot to HMV.” Practically it costs a year, a full redesign of the evidence strategy, and usually a reposition of the intended use.
The seven paths are not interchangeable. They live in different German social codes (SGB V for statutory health insurance, SGB XI for long-term care), have different gatekeepers (BfArM, GKV-Spitzenverband, state hospital associations, individual sickness funds, G-BA Innovation Committee), different evidence thresholds, and different price mechanics. The manufacturer chooses the path before building the product – or builds for the wrong market.
DiGA (§ 33a SGB V) – the visible path, not always the right one
Measured by press coverage and application numbers, the DiGA under § 33a SGB V (Sozialgesetzbuch V – Social Code Book V) is the dominant path. It is also the path with the tightest procedural logic: BfArM application under § 139e SGB V and the DiGAV, decision within three months of receipt of complete application materials (§ 139e para. 3 sentence 1 SGB V), listing in the public DiGA directory, reimbursable for all statutorily insured persons by physician prescription or sickness-fund approval (§ 33a para. 1 sentence 2 no. 2 SGB V).
The scope is narrowly specified (§ 33a paras. 1 and 2 SGB V): medical devices of low risk class (Class I/IIa) and – since the 2024 DigiG law – also Class IIb, whose main function rests essentially on digital technologies and which support detection, monitoring, treatment or alleviation of diseases, or detection, treatment, alleviation or compensation of injuries or disabilities. Explicitly excluded under § 33a para. 1 sentence 5 SGB V: devices that control active therapeutic products, digital applications intended for use with a specific medical aid or medicinal product, and general consumer goods.
Three exclusion criteria teams regularly overlook:
- No primary prevention. DiGA Guide v3.6 Sec. 2.1.4 is unambiguous: “Primary-preventive digital applications cannot be listed in the directory.” Secondary and tertiary prevention qualify – but only where the risk factor is codable as a diagnosis. A nutrition app for healthy people is not a DiGA. A nutrition app for patients with hypertension or type-2 diabetes can be.
- Not a pure care application. If the target group is care-dependent people, informal caregivers or professional carers and the intended use is care support, the correct path lives in SGB XI, not SGB V. The BfArM rejects the filing not as “substantively wrong” but as procedurally outside the DiGA gate.
- Not a patient portal. Appointment booking, shop integration, community – without a medical intended use, not a DiGA. A DiGA may contain ancillary services but must segregate them architecturally (DiGA Guide v3.6 Sec. 2.1.3).
When DiGA is the right path: when the product is a CE-marked medical device that treats or alleviates a codable disease, when evidence (or a credible trial-phase evidence strategy) is at least sketched, and when the goal is broad reimbursement – all statutorily insured people, national reach, by prescription. DiGA is the path with the highest effort and the highest reach; that is not a contradiction, it is the deal.
The DUX DiGA knowledge area – application process, DiGAV, AbEM, data protection, pricing – covers DiGA from the inside. This hub covers when it is not the right choice.
DiPA (§ 40a SGB XI) – care focus, different rules
The DiPA (Digitale Pflegeanwendung – digital care application) under § 40a SGB XI was introduced in 2021 as the sister regime to DiGA – with one decisive difference: it lives in SGB XI (the long-term-care code), not SGB V (health insurance). That is the structural dividing line.
DiPA address care-dependent people, their relatives, and professional carers in home care – not, as DiGA do, all insured people. The entitlement covers applications “used by care-dependent people, or in the interaction of care-dependent people with relatives, other volunteer caregivers, or authorised outpatient care services, to reduce impairments of autonomy or capabilities of the care-dependent person, or to counteract a worsening of care needs” (§ 40a para. 1 SGB XI). Excluded: applications of general daily use, outpatient-care operations management, knowledge transfer, or benefit application/administration (§ 40a para. 1a sentence 2 SGB XI).
The gatekeeper is the BfArM (as for DiGA); the ordinance is the DiPAV (parallel to the DiGAV); the directory is the DiPA directory under § 78a para. 3 SGB XI. Reimbursement runs through the long-term-care fund – up to €40 per care-dependent person per month for the application, plus up to €30 per month for supplementary support services by outpatient care providers (§ 40b para. 1 SGB XI). First approval is limited to six months (§ 40a para. 2 sentences 3–5 SGB XI).
The evidence threshold differs substantively from DiGA: under § 11 para. 1 DiPAV, retrospective comparative studies are permitted to demonstrate care benefit – a different world from the prospective-comparator logic of § 10 DiGAV. Where the application qualifies as a medical device, it must additionally be of low risk class (§ 40a para. 1b SGB XI).
DUX reading: the DiPA path is regulatorily clean but the market is materially smaller and slower than the DiGA market. Three reasons: the care-fund benefit logic is a different economic frame; the target group has different adoption barriers; and the product-build experience in care software is historically thinner than in therapy software. Teams that explicitly build for care – relief for family carers, cognitive activation in dementia, outpatient-care documentation – find the right frame here. Teams that build “something for older people” should not default to DiPA just because the cohort is 65+.
Hilfsmittelverzeichnis / HMV (§ 33 SGB V) – when software counts as a medical aid
The Hilfsmittelverzeichnis (HMV – Germany’s Medical Aids Directory) under § 33 para. 1 SGB V – the entitlement to medical aids “to secure the success of medical treatment, prevent an impending disability, or compensate for a disability” – has historically been the path for physical aids: walking frames, blood-glucose meters, hearing aids, CPAP devices. The directory is maintained by the GKV-Spitzenverband under § 139 SGB V; prescribability is governed by the Hilfsmittel-Richtlinie (HilfsM-RL) – G-BA version of 20 February 2025, in force from 16 May 2025. Software long figured here only as an accessory. As sensor hardware and analytic software fuse, the HMV path has become structurally more relevant for digital solutions – in a configuration neither DiGA nor DiPA captures: when the software is an integral component or essential accessory of a medical aid. HilfsM-RL § 2 defines medical aids expressly as “physical means or technical products” and includes “accessory parts without which the base products cannot be operated”.
Products that work in the HMV: the readout app of a continuous-glucose monitor, the companion app of a hearing aid with fitting functions, the telemetry software of a pacemaker. The software is part of the medical-aid product; billing runs through the medical-aid positional number.
The delimitation from DiGA is conceptually clean but operationally fuzzy. DiGA Guide v3.6 Sec. 2.1.2 discusses the grey zone directly: if additional devices listed as medical aids under § 33 SGB V are required for the intended use of a DiGA, those devices fall under the medical-aid entitlement, not the DiGA entitlement.
DUX reading: HMV is the underrated path for sensor-plus-software combinations. A company building a wearable with clinical intended use, an implant companion, or a continuous-monitoring system with a hardware component should not automatically default to DiGA. HMV listing is effortful (own application procedure, own review, own contract mechanics with the sickness funds) but delivers a different economic mechanism: no one-sided first-year price-setting, but a broader, more stable reimbursement base that scales with the medical-aid ecosystem. For non-EU companies, HMV inherits the same EU-legal-manufacturer precondition as DiGA (MDR Art. 10/11); see International Markets.
ZPP – prevention without a disease diagnosis
The Zentrale Prüfstelle Prävention (ZPP – Central Prevention Certification Office) is the right address for certified prevention courses under § 20 SGB V – exactly the category systematically excluded from DiGA (DiGA Guide v3.6 Sec. 2.1.4). Prevention courses target healthy people and aim to prevent disease onset (§ 20 para. 1 SGB V), not treat an existing disease.
The mechanics differ structurally: no BfArM, no mandatory MDR certificate, no comparative study for positive care effects. The ZPP certifies on behalf of the sickness funds against uniform criteria set by the GKV-Spitzenverband under § 20 para. 2 SGB V. These criteria live in the Leitfaden Prävention des GKV-Spitzenverbands (2025 edition); for digital offerings, chapters 5 (individual behavioural prevention, in-person and ICT-based courses) and 7 (digital prevention and health-promotion offerings) apply. Certified courses are then eligible for subsidy (§ 20 para. 5 SGB V); sickness funds reimburse participants fully or partially. Since 1 January 2026, the nationwide standard participation certificate is mandatory (Leitfaden Prävention 2025 Sec. 5.3.9).
For digital solutions: an app-based exercise, nutrition or stress-prevention offering can be ZPP-certified, but it runs not as a medical device, not as a DiGA, and not via prescription – it runs as a course offering with subsidy.
DUX reading: ZPP is regularly underrated because it lacks “BfArM glamour”. For primary-prevention providers, it is often the only regular path – and it is not a bad place when the product genuinely sits in primary prevention. The classic mistake: a primary-prevention concept tries to position as DiGA (“the first three months are prevention, then it is secondary prevention”) and is either rejected substantively or stumbles on indication codability. The honest alternative is ZPP – not as a “DiGA substitute”, but as an independent path for a different product goal.
NUB (§ 6 para. 2 KHEntgG) – innovation reimbursement in hospitals
NUB (Neue Untersuchungs- und Behandlungsmethoden – New Examination and Treatment Methods) is an inpatient-care mechanism, governed by § 6 para. 2 KHEntgG (Krankenhausentgeltgesetz – Hospital Fees Act). Not to be confused with the G-BA’s NUB methods in outpatient care – different regulatory logic.
The hospital NUB logic: for new methods “that cannot yet be adequately remunerated by DRG case flat rates or supplementary charges … and that have not been excluded from financing under § 137c SGB V”, the contracting parties shall agree “time-limited, case-related charges or supplementary charges outside the revenue budget” (§ 6 para. 2 sentence 1 KHEntgG). Before the agreement, the hospital must – by 31 October each year – obtain an assessment from the InEK (Institut für das Entgeltsystem im Krankenhaus – Institute for the Hospital Payment System) on whether the method can already be billed via existing DRGs (§ 6 para. 2 sentence 3 KHEntgG). The charge is hospital-specific and time-limited; after expiry the method either enters the DRG catalogue or drops out.
For digital solutions, NUB is relevant when a product in the hospital setting supports a new diagnostic or therapeutic method not yet part of standard care. Typical examples: imaging-augmentation AI, decision-support for complex surgery, monitoring solutions for specialty indications.
DUX reading: NUB is not a permanent reimbursement path but a market-entry window. Teams that understand NUB as “an easier DiGA” misread the mechanism – NUB addresses the hospital budget, not the consumer market. It is a way to build inpatient traction, gather evidence, and prepare either a G-BA assessment or DRG integration. A stage, not a destination.
Selective contracts (§ 140a SGB V) – direct with insurers, bypassing BfArM
The selective contract under § 140a SGB V – the statute text calls it “special care” – is the quietest and most flexible path. Sickness funds may, under § 140a para. 1 SGB V, “conclude contracts on special care for insured people” with the providers in paragraph 3. Eligible contracting parties expressly include providers of digital services and applications under § 68a para. 3 sentence 2 no. 2 and 3 (§ 140a para. 3 sentence 1 no. 8 SGB V) and manufacturers of medical devices in the sense of the MDR (§ 140a para. 3 sentence 1 no. 6 SGB V). The contracts may deviate from provisions of Book IV SGB V, KHG and KHEntgG as long as they pursue quality, effectiveness or cost-efficiency improvements.
The structural difference to all other paths: no BfArM, no G-BA, no federal-level GKV-Spitzenverband. The sickness fund decides per contract, the product is shaped contract-specifically, reimbursement is tied to the contract – and to insured people of that fund, who must formally opt in (§ 140a para. 4 sentence 1 SGB V).
- Strength: fast market access (negotiate with one fund), possibility to tailor the product to a single indication, care research in live operation, no fast-track procedure costs, niche with a clear contracting partner.
- Limitation: no nationwide reimbursement, no transfer to other funds (each new fund = new contract), political-negotiation logic rather than regulatory-path logic, no public directory, correspondingly less marketing leverage.
DUX reading: selective contracts are underrated precisely because they lack regulatory glamour. For teams building in a niche indication, for products that do not slot into DiGA logic, and for market validation in year one, selective contracts are often the more realistic option than a six-figure DiGA preparation. A team using a selective contract as a wedge can build substantial market understanding in twelve months while a DiGA candidate is still in the evidence phase. This is particularly true as a prelude phase: selective contract today, DiGA filing in 18–24 months, once evidence from the contract has matured.
Innovation Fund (§ 92a SGB V) – funding, not reimbursement
The Innovation Fund sits at the G-BA, is fed from statutory-health-insurance funds, and funds in two distinct lines regulated separately in § 92a paras. 1 and 2 SGB V: new forms of care (para. 1 – cross-sectoral care models going “beyond the existing standard of care”) and care research (para. 2 – knowledge generation for guideline and G-BA decisions). The annual funding volume per § 92a para. 3 SGB V is €100 million in 2026 and €200 million per year from 2027 – 80 % for new forms of care, 20 % for care research.
The decisive point: the Innovation Fund does not remunerate routine care. It is a funding instrument – money flows for the duration of the project (up to four years, § 92a para. 3 sentences 6–7 SGB V), not per unit delivered. After project completion, the Innovation Committee issues recommendations on transfer to standard care (§ 92b para. 2 sentence 8 SGB V) – rarely immediately, often not at all. § 92a para. 1 sentence 7 SGB V says so directly: “There is no entitlement to funding.”
DUX reading: the Innovation Fund is a strategic bridge, not a destination. It can carry an evidence study that would otherwise be unfundable; it can validate a care model that then feeds a selective contract or a DiGA filing. Teams planning the Innovation Fund as their sole reliance build a project, not a market.
The decision logic in three questions
For products that do not produce a clear single candidate from the descriptions above, three questions in order:
- Is there a codable disease that the product addresses – and are we building a CE-marked medical device? If yes → DiGA is the first candidate. If no, go to question 2.
- Is the target group care-dependent people, informal caregivers, or professional carers – and is the intended use care support? If yes → DiPA is the first candidate. If no, go to question 3.
- Is the product part or accessory of a medical aid (sensor, implant, assistive technology)? If yes → examine HMV. If no → primary prevention (ZPP), inpatient special case (NUB), sickness-fund niche (selective contract), or funding need (Innovation Fund) – in that order.
Not exhaustive. Not legally binding. This is the filter to apply in the first twenty minutes of a reimbursement-path conversation to rule out the obviously wrong paths before discussing the right one.
When the first path fails
It happens: an application runs into a wall. BfArM rejects a DiGA, a medical-aid listing is denied, an NUB status request comes back negative, a sickness fund withdraws from selective-contract negotiations. The first instinct is usually to try the same path again. Sometimes right, often expensive.
Three structural points:
- DiGA rejection rarely means “wrong path” – usually it means “wrong application”. Most rejections are procedural (incomplete filing, data protection not implemented, evidence package misaligned with claimed indication). The fix is “same path, better prepared”. For incomplete filings, § 16 para. 2 DiGAV grants a supplementation deadline of up to three months. After a negative decision on trial admission, a fresh application is possible only after 12 months, and a repeat preliminary listing for trial is excluded (§ 139e para. 4 sentences 9 and 10 SGB V) – the statutory lock that ends many second filings.
- A path switch is rarely a copy. Moving from DiGA to HMV is not “the same application in a different legal frame”. Intended use changes, evidence requirements change, often target group and architecture change. Realistic time loss: six to twelve months plus repositioning. The switch plan belongs before the first filing.
- Several paths in parallel are possible but non-trivial. A selective contract with one fund alongside a DiGA filing works – it requires clean separation (the same service cannot be billed twice). But each path has its own maintenance obligations and operational load compounds.
Bottom line: the “plan B” should already exist in the original filing.
Frequently asked questions
Can one product pursue multiple reimbursement paths in parallel?
Yes, under three conditions: intended uses must not contradict each other, billing logics must be cleanly separated (no double billing of the same service), and the operational overhead must be accounted for.
In practice, parallelism most often takes the form of selective contract plus DiGA preparation: a provider starts with one or two funds in a niche indication, collects real-world evidence under the selective contract, and files the DiGA in parallel. Less common but realistic: HMV for the hardware-coupled product plus DiGA for a software-focused variant, managed as distinct directory entries. Not sensible: DiGA and DiPA for “the same product” – the intended uses structurally exclude each other because one addresses a disease and the other care support.
My DiGA was rejected – which path now?
First read the reasoning. Procedural rejections (incomplete application, missing data-protection building block, evidence package too narrow) are reparable – for incomplete filings, § 16 para. 2 DiGAV grants a supplementation deadline of up to three months. After a negative decision on trial admission, a fresh application is possible only 12 months later, and a repeat preliminary listing for trial is excluded (§ 139e para. 4 sentences 9 and 10 SGB V).
Substantive rejections (intended use outside scope, indication not codable, positive care effects structurally not demonstrable) are a signal to change path. Realistic alternatives in order: HMV (if a hardware share exists), selective contract (if niche or fund partner available), ZPP (if primary prevention), DiPA (if care context). The Innovation Fund is rarely the right answer after a DiGA rejection – it funds evidence work but does not solve the intended-use problem.
Does HMV make sense for a non-EU company?
Yes, under the same precondition as DiGA: HMV requires a CE-marked medical device, and the MDR requires the legal manufacturer to be established in the EU. Non-EU companies therefore face the same structural question on HMV as on DiGA – EU legal manufacturer (via subsidiary, partnership, or acquisition) is a precondition, not an optional step.
What HMV does not share with DiGA: no BfArM procedure, no DiGAV-specific data-protection and cybersecurity catalogue, no AbEM mechanism. The application goes to the GKV-Spitzenverband, the assessment focuses on function, safety and medical necessity, and the reimbursement logic is that of the medical-aid ecosystem. For sensor-plus-software products from outside the EU, HMV is often the straighter route – once the legal-manufacturer question is answered. See International Markets.
Why is the DiPA market so much smaller than the DiGA market?
DiPA was introduced in 2021 – later than DiGA (2019/2020) – and the market has grown materially more slowly. The GKV-Spitzenverband reports from the payer perspective that DiPA uptake has lagged expectations; the industry side points to structural barriers in target-group access and digital adoption among the care population. Both readings are partly right.
Three structural reasons add to that: the benefit logic of the long-term-care fund differs from that of the sickness fund – DiPAs are not prescribed by a physician but approved by the care fund with a six-month limit on first approval (§ 40a para. 2 sentences 3–5 SGB XI), and end-user awareness is lower; product-build experience in care software is historically thinner than in therapy software; and the reimbursement ceiling is capped at €40 per month (§ 40b para. 1 no. 1 SGB XI) – materially below typical DiGA negotiated prices. For individual products the DiPA path is nonetheless the only correct one – precisely when the intended use is genuinely care. Market size is not an argument against path fit.
Sources and attributed positions
This knowledge area consistently separates legal framework from position. The frameworks – §§ 20, 33, 33a, 92a, 139, 139e, 140a SGB V, §§ 40a, 40b SGB XI, DiGAV, DiPAV, Hilfsmittel-Richtlinie (version 20 February 2025), Leitfaden Prävention (2025 edition), § 6 para. 2 KHEntgG – are authoritative and are the basis. The positions are two, and neither is neutral:
- The GKV-Spitzenverband represents the payer perspective (GKV-Spitzenverband DiGA Report 2025). Its criticism of first-year free pricing, of the evidence base, and of marketing practices flows from its mandate on contribution-rate stability, not from neutral market observation.
- The SVDGV (Spitzenverband Digitale Gesundheitsversorgung) represents the industry perspective (SVDGV DiGA Report 2025). Its emphasis on trial-phase successes and market development flows from industry interest, not from independent evaluation.
Both reports belong with attribution and read as two sides of the same calculation, not as independent confirmation. A comparably broad, publicly documented report tradition with contrasting payer and industry perspectives does not exist for DiPA, HMV (for digital entries), ZPP, NUB, selective contracts, or the Innovation Fund. For those paths the work relies on the legal texts and on DUX’s project experience.
Where DUX fits – and where DUX does not make the path choice
DUX does not build reimbursement paths. DUX builds products. The role in the path question is tightly bounded: helping teams reach a defensible path decision in the first weeks of a project – based on intended use, MDR class, target group, evidence status, and market ambition – and building the product so it is viable on the chosen path. For DiGA, DUX has the deepest operational experience (mHealth Suite, five DiGAs platform-engineered); for HMV, DiPA, selective contracts, and NUB, the work runs mandate-specifically.
DUX does not recommend a path because it happens to build it often. If a team arrives with a primary-prevention concept, the conversation is about ZPP. If a team builds in care, the conversation is about DiPA. If the path question concludes that the product does not fit any of the seven regular routes, that is what DUX says.
For the strategic positioning of digital health products in the German market, see /scale/ – DUX’s work on market-access and reimbursement strategy.
Path positioning