Key points
- DefinitionA DiGA (Digitale Gesundheitsanwendung) is CE-marked medical device software that a physician or psychotherapist prescribes and that German statutory health insurance (GKV) reimburses in full. It must first be listed by the federal institute BfArM through the Fast-Track procedure, which decides within 3 months of a complete application § 139e SGB V.
- MarketAbout 60 DiGA are listed. From September 2020 to the end of 2025, 1.9 million DiGA were prescribed or approved, and the GKV paid about €400 million.
- Main hurdleProof of a positive care effect through a comparative study run in Germany. Of 74 DiGA listed since 2020, 16 were removed.
- TimingLegal deadlines add up to 3 months of review, a trial of up to 24 months and 3 months for the final decision. CE marking, certificates and the study come first; their duration is [to be estimated].
- CostBfArM fees range from €150 to €9,900 per procedural act. One consultancy models about €2.67 million of spending before launch to be confirmed.
- PriceThe manufacturer sets the price within caps for the first 12 months; a negotiated price applies afterwards. In 2025, the average was €544 at entry and €227 after negotiation.
- DirectionAt least 20% of the price tied to outcomes since 2026, volume discounts from 2027, continuous outcome measurement, and pressure from the funds to end the free first-year price.
What a DiGA is
In law, a DiGA is a medical device of risk class I, IIa or IIb whose main function rests essentially on digital technology. It supports the detection, monitoring, treatment or alleviation of diseases, or the detection, treatment, alleviation or compensation of injuries or disabilities § 33a(1) SGB V.
The GKV pays only if the DiGA is in the BfArM directory and is used after a prescription, or with the health fund's approval on proof of the indication.
- CE-marked medical device under the EU Medical Device Regulation (MDR) G1. Under MDR Rule 11, software that informs diagnostic decisions is at least class IIa M1.
- Used by the patient, alone or together with a clinician. A tool used only by the physician is not a DiGA in BfArM practice to be confirmed G2.
- Tied to a disease. Wellness, fitness and primary prevention apps are not DiGA to be confirmed.
- Legal exclusions: products that control an active therapeutic product, applications tied to one specific aid or medicine, and general everyday items § 33a(1) SGB V.
- No advertising in the application (federal health portal gesund.bund.de).
Class IIb has been eligible since the Digital Act (Digital-Gesetz, in force 26 March 2024). A class IIb DiGA must prove medical benefit at application and cannot use the trial phase § 139e(2) and (4) SGB V. BfArM must evaluate this route and report by 31 March 2027.
The market in figures
Use is growing fast: 695,000 activation codes were redeemed in 2025, up 63% on 2024, and GKV spending passed €170 million in the year.
Codes redeemed rose more than tenfold between 2020-21 and 2025
Activation codes redeemed per year, statutory health insurance (GKV)
By the end of 2025, 1.9 million DiGA had been prescribed or approved and 1.6 million had been used, for about €400 million in total. The manufacturers' association puts DiGA at about 0.05% of GKV spending to be confirmed. The largest cumulative spends were Vivira (back pain, €26.9 million), zanadio (obesity, €26.2 million), Kalmeda (tinnitus, €21.5 million) and Endo-App (endometriosis, €17.2 million).
| State of the directory | Count |
|---|---|
| DiGA listed since 2020 | 74 |
| Removed | 16 (22%) |
| Listed at the end of 2025 | 58: 48 permanent, 10 on trial |
| Listed in 2026 | About 60; counts differ between sources to be confirmed in the directory |
| Moved from trial to permanent | 34 |
| Proof available at listing | Fewer than 1 in 5 overall; one third of new listings in 2025 |
Sources: GKV-Spitzenverband DiGA report 2025 and press release of 8 April 2026; AOK. Mental health is the largest category in the directory to be confirmed.
Conditions for listing
The application must prove three things: safety, functionality and quality of the product; data protection and data security; and positive care effects § 139e(2) SGB V. The first two are hard prerequisites. The third can be deferred through the trial phase, for classes I and IIa only.
| Requirement | What is required | When | ID |
|---|---|---|---|
| CE marking | MDR class I, IIa or IIb, quality management system, technical documentation | Before application | G1 M1 M3 |
| IT security | BSI certificate under Technical Guideline TR-03161 for app, web and backend, penetration tests included | Required since 1 January 2025 § 139e(10) | G5 T6 |
| Data protection | Certificate under Art. 42 GDPR based on BfArM test criteria; processing based on explicit consent | Required since 1 August 2024; BfArM can set a deadline while no certification scheme is available § 139e(11) | G4 D1 |
| Security management | Certified information security management system (ISMS) in addition to the BSI certificate to be confirmed | Before application | |
| Data location | Processing in Germany, the EU or EEA, or a country with an adequacy decision DiGAV § 4(3) | Before application | H1 H5 |
| Interoperability | Open standards; export to the electronic patient record (ePA) with consent | Before application | G6 |
| Quality and AI | DiGAV quality requirements, with specific rules for AI systems to be confirmed; declaration on whether the EU AI Act applies to be confirmed | Before application | G7 M2 |
| Positive care effects | Quantitative comparative study run in Germany | At application, or at the end of the trial | G3 |
Positive care effects
Either a medical benefit (health status, duration of illness, survival, quality of life), or a patient-relevant improvement of structures and processes of care (patientenrelevante Struktur- und Verfahrensverbesserungen) § 8 DiGAV. The second route covers nine areas: coordination of care, alignment with guidelines, adherence, access to care, patient safety, health literacy, patient autonomy, coping with daily difficulties caused by the illness, and reducing the burden of treatment for patients and their relatives.
Study requirements
The study must, under § 10 DiGAV:
- Show that using the DiGA is better than not using it. The comparator is no treatment, treatment without a DiGA, or another DiGA already permanently listed.
- Be quantitative. Retrospective comparative studies are accepted; prospective studies are an alternative.
- Be run in Germany, or prove transferability to the German care context.
- Be registered in a WHO-recognised study register and published in full within 12 months of completion.
- Record duration and frequency of use alongside.
In practice, all 35 DiGA permanently listed by mid-2024 relied on randomised controlled trials to be confirmed. The ordinance allows lighter designs, but no permanently listed DiGA had used them by then.
The Fast-Track procedure
BfArM only reviews a complete application. Everything before submission is the manufacturer's work.
Without proof at submission, the trial leaves 12 to 24 months to deliver it
Fast-Track procedure, § 139e SGB V
Without proof at submission, a class I or IIa DiGA enters the directory on trial, paid by the GKV, and has 12 to 24 months to deliver the evidence. After a removal, a new application is possible after 12 months with new evidence; a second trial listing is never allowed § 139e(4) SGB V.
Before and around submission
- Set the intended purpose, the indication (ICD-10 codes) and the MDR class.
- Optionally book a paid advisory meeting with BfArM, ideally before the study is designed.
- Obtain CE marking, the BSI TR-03161 certificate and the data protection certificate.
- Choose the route. Permanent listing: the comparative study is complete. Trial listing: a plausible rationale, an evaluation concept written by an institution independent of the manufacturer, and the results of a systematic data analysis.
- Submit online through the BfArM portal. Forms exist in German and English. An incomplete file leaves 3 months to complete it.
- Assign the DiGA to a price group and return the form to the joint body of funds and manufacturers within two weeks.
Timeline
The legal deadlines are fixed. The preparation before submission is not, and it decides the overall schedule.
After submission, the law fixes every step
From project start to permanent listing, trial route; phases not to scale
One provider reports queues of up to 24 months at notified bodies in Germany to be confirmed. One trial year is rarely enough: by the end of 2024, 43 DiGA had needed an extension, and most DiGA take longer than the planned year to reach permanent listing. Going straight to permanent listing avoids the trial, but moves the study before submission, and therefore before any revenue.
Costs
BfArM fees are small. The study, certification and team carry the cost.
| BfArM act | Fee | Basis |
|---|---|---|
| Advisory meeting | €250 to €5,000 | § 27 DiGAV |
| Decision on listing (permanent or trial) | €3,000 to €9,900 | § 25 DiGAV |
| Final decision after the trial | €1,500 to €6,600 | § 25 DiGAV |
| Trial extension | €1,500 to €4,900 | § 25 DiGAV |
| Significant change to the product | €1,500 to €4,900 | § 26 DiGAV |
| Change to directory information | €300 to €1,000 | § 26 DiGAV |
| Removal | €200 | § 26 DiGAV |
Overall budget. The consultancy research2guidance models about €2.67 million before launch for a standalone developer: strategy, software, clinical study, MDR compliance, cybersecurity, BfArM submission and launch preparation. The same model reaches about €13 million over the first five commercial years. It also states that only 7% of companies that consult BfArM end up with a listed product to be confirmed.
Items to quote (no reliable public price list found):
- Notified body for CE marking in class IIa or IIb.
- TR-03161 assessment by a BSI-recognised lab, then the BSI certificate.
- Data protection certification and ISMS certification.
- Comparative study: sites, independent evaluator, statistics, registration and publication.
- Person responsible for regulatory compliance under the MDR; product liability insurance.
Cash flow. The GKV pays from the first day of listing, trial included; this is what usually funds the study. By the end of 2025, the funds had pre-financed €119 million above negotiated prices: €56 million from first-year price gaps and €63 million from extended trials.
Pricing and reimbursement
For the first 12 months, the manufacturer's actual price applies within caps; after that, the price negotiated with the GKV-Spitzenverband applies to all funds § 134 SGB V. In 2025, the average entry price was €544 and the average negotiated price €227, about 58% lower. Entry prices range from €119 to €2,077.
First year: manufacturer price, capped
- Group caps apply from the 2,001st activation code; above 10,000 codes they drop by a further 25%.
- A DiGA on trial is capped at 80% of its group cap.
- Exempt from caps on request: DiGA for rare diseases, DiGA whose main function relies on artificial intelligence, and other justified cases (§ 3h of the framework agreement).
- Below a threshold of €0.84 per day (€75.60 for 90 days), the price applies permanently without negotiation.
| Group "mental disorders, medical benefit", from 1 October 2026 | Cap per day, VAT incl. | 90 days |
|---|---|---|
| Permanent listing | €6.66 | €599.40 |
| Permanent, above 10,000 codes | €5.00 | €450.00 |
| Trial | €5.33 | €479.70 |
| Trial, above 10,000 codes | €4.00 | €360.00 |
After the first year: negotiated price
- No agreement within 9 months of listing: the arbitration board sets the price within 3 months.
- The negotiated price applies after the first year, often retroactively: the manufacturer then refunds the difference to the funds.
- Since 1 January 2026, at least 20% of the price must depend on outcomes.
- From 1 January 2027, a volume discount applies: at least 2% above 3,000 units a year, at least 30% above 100,000.
- The manufacturer discloses its prices to self-payers and in other European countries.
Obligations after listing
A permanently listed DiGA must now measure real use continuously and report it to BfArM, which publishes it. Every significant change to the product must be reported.
Continuous outcome measurement (anwendungsbegleitende Erfolgsmessung, AbEM)
- Applies to permanently listed DiGA § 23a DiGAV.
- Data collected every quarter from the third quarter of 2026, with a standard questionnaire annexed to the DiGAV.
- Sent to BfArM every six months, anonymised and aggregated; first submission by 15 April 2027 to be confirmed.
- Content: duration and frequency of use, patient satisfaction, patient-reported health status § 139e(13) SGB V.
- BfArM publishes usage frequency once it covers at least 200 people in a quarter to be confirmed.
- A DiGA that proved only a structural and process improvement does not report health status § 23a DiGAV.
- Validated questionnaires become mandatory from 2028 to be confirmed.
Changes to the product
Significant changes must be reported to BfArM, which decides within 3 months whether to update the directory or remove the DiGA § 139e(6) SGB V. If a change was not reported, BfArM sets a deadline, usually no longer than 4 weeks, and can impose a penalty payment of up to €100,000. Manufacturers report that every renewal of the BSI certificate is currently treated as a significant change to be confirmed.
BfArM now also publishes the reasons when a DiGA is removed to be confirmed.
Prescription and patient access
The patient gets the DiGA free of charge with a prescription or the fund's approval, then activates it with a code. General practitioners prescribe most DiGA, and most users are women.
- Indication. A documented diagnosis with an ICD-10 code covered by the DiGA, and no listed contraindication.
- Prescription. A physician or psychotherapist prescribes, one DiGA per form, with no cap per patient. Alternatively, the patient applies to the fund directly with proof of the indication.
- Format. E-prescription has been possible since early 2026, but the obligation remains suspended; the paper form (Muster 16) stays valid.
- Activation. The fund issues a code; the patient downloads the app or opens the web version and activates it. No co-payment is charged today.
- Medical follow-up. For some DiGA, BfArM defines required medical services, which physicians bill separately.
Two safeguards apply. Physicians may not steer prescriptions to a manufacturer or make arrangements with one. Manufacturers may not make agreements with drug or aid makers that limit patient choice § 33a(5) and (5a) SGB V. Privately insured patients depend on their contract.
Reforms in 2026 and 2027
The framework is tightening on prices and on proof of real use. Ending the free first-year price, as the funds demand, has not been adopted so far.
| Date | Measure | Effect on DiGA | Status on 9 Oct 2026 |
|---|---|---|---|
| 14 and 15 Oct 2026 (planned) | Amendment by the CDU/CSU and SPD groups | Price reduction based on the share of users who stop within the first 28 days, from 1 January 2027 | Committee and Bundestag votes expected |
| July 2026 | Data and Digital Innovation Act (GeDIG), government bill | Opens DiGA to remote monitoring; data exchange with the ePA; the GKV-SV annual report replaced by data sent to the ministry | Adopted by the federal cabinet to be confirmed; in Parliament |
| 30 Jul 2026 | Contribution Rate Stabilisation Act (GKV-BStabG) | Mandatory volume discount from 2027 | In force |
| 27 Jul 2026 | EU Digital Omnibus on AI, Regulation (EU) 2026/1744 | AI Act high-risk obligations for AI in medical devices apply from 2 August 2028 to be confirmed | In force |
| 30 Mar 2026 | Report of the FinanzKommission Gesundheit | Recommends ending the trial phase and the free price, and introducing patient co-payments | Recommendation only |
| 1 Feb 2026 | Second amendment to the DiGAV | Continuous outcome measurement, AI Act declaration, wider study requirements | In force |
| 1 Jan 2026 | Digital Act, price provision | At least 20% of the price tied to outcomes | In force |
| 26 Mar 2024 | Digital Act (DigiG) | Class IIb eligible; legal basis for outcome measurement | In force |
On the funds' side, the AOK federal association asked in August 2026 for proof of efficacy at listing and a patient co-payment of at least 10%. This is the main regulatory risk for any plan that relies on the trial phase.
Case study: attexis (adult ADHD)
attexis shows the typical route for a mental health DiGA: a German randomised trial, direct permanent listing, then a pharmaceutical partner for sales.
| Manufacturer | GAIA, Hamburg; marketed under licence by MEDICE |
| Indication | Adult ADHD, 18 and over, ICD-10 F90.0 and F90.1 |
| Approach | Digital cognitive behavioural therapy, no mandatory human support |
| Evidence | Randomised trial with 337 adults: attexis plus usual care against usual care alone; primary endpoint ASRS score at 3 months |
| Status | Permanent listing in August 2025 |
| Price | €599.40 for 90 days, equal to the mental health group cap |
| Medical services | None required from the prescriber |
In the same field, ORIKO ADHS-Therapie (adult ADHD) and hiToco (parents of children with ADHD) are listed on trial. No DiGA for autism appeared in the sources consulted; to be checked directly in the BfArM directory.
What this means for ClinAutAI
DiGA is the reimbursement model in view for later, on the patient side. It is not part of the prototype or the V0. The question arises at V1 at the earliest, once physicians, researchers, psychotherapists and patients have tested the product.
What the law allows
The legal definition includes the detection of diseases, and the structural and process route explicitly covers access to care and coordination of care § 8 DiGAV. A tool that prepares adults for an autism assessment fits these terms in principle.
Points to settle before any DiGA work
- Indication. Prescription relies on an ICD-10 code, but a person preparing for assessment has no F84 diagnosis yet. Prescribers of attexis are told that a suspected diagnosis can support the prescription; whether this works for a pre-diagnostic tool must be clarified with BfArM.
- User. Only the patient-facing part (preparation and modules) can be a DiGA G2. The AI synthesis for physicians is a separate medical device software M1.
- Evidence. The natural route is access to care. Demand is documented: more than 12 months' average wait across 15 centres surveyed, with about 70 adult assessment services in Germany to be confirmed, and 12 to 16 months before a first appointment at the Heidelberg adult autism clinic. If the tool contains a screening instrument, BfArM also asks for a diagnostic accuracy study G3.
- Risk class. A tool that prepares and guides without making the diagnosis is more likely to stay in class IIa, which keeps the trial option. Class IIb requires proof of medical benefit at submission G1 M1. To be confirmed with a notified body.
- AI. A DiGA whose main function relies on AI can apply for an exemption from price caps. In return: AI Act declaration, traceability requirements, and high-risk obligations from 2 August 2028 M2.
- Data location. If anyone accesses patient data from Canada, the adequacy decision and the DiGAV rules on processing outside Germany must be checked early H1 H5.
- Legal manufacturer. The entity that holds CE marking, the quality system and the submission: Valirax, Serefin or a dedicated company. To be decided before any regulatory spending.
Choices in the prototype and V0 that keep the route open
These choices cost little now and avoid rework later:
- A clear boundary between coordination and the regulated AI module T1.
- Hosting in Germany or the EU, and AI models hosted in the EU without training on patient data H1 H3.
- Explicit consent before any processing, with separate consent for research D1.
- Data export designed for the ePA G6.
- A development process an auditor can follow T5.
- TR-03161 as the target of the external penetration test T6.
A post-diagnostic variant (psychoeducation and daily life support for diagnosed autistic adults) would follow the attexis model: a clear indication, and "coping with daily difficulties" as the evidence area.
For comparison: France (PECAN)
France has a similar early-access scheme, PECAN, but no digital therapeutic had been reimbursed through it as of March 2026 to be confirmed.
| Criterion | DiGA (Germany) | PECAN (France) |
|---|---|---|
| Products | Digital medical devices, classes I to IIb | Digital therapeutic devices and remote monitoring |
| Assessors | BfArM alone | HAS (presumed innovation) and ANS (compliance), then a ministerial decision within 30 days |
| Early reimbursement | Trial of 12 months, extendable by 12 | 1 year, not renewable |
| Exit condition | Proof of positive care effects at the end of the trial | Application for standard reimbursement within 6 months (therapeutic) or 9 months (remote monitoring) |
| Reimbursed products | About 60 listed | Remote monitoring solutions; no digital therapeutic as of March 2026 |
Sources
All sources checked on 9 October 2026. Statements marked to be confirmed rely on the secondary sources in the last group.
Official texts and public authorities
- § 33a SGB V: entitlement, definition
- § 134 SGB V: prices, version in force since 30 July 2026
- § 139e SGB V: directory, procedure, certificates
- § 8 DiGAV and § 10 DiGAV: positive care effects, studies
- § 23a DiGAV: outcome measurement
- § 25, § 26 and § 27 DiGAV: fees
- Federal Ministry of Health, DiGAV page: amendment of 1 February 2026
- Public service portal, DiGA application: documents, fees
- gesund.bund.de: patient information
- BfArM DiGA directory
Health insurance data
- GKV-Spitzenverband, DiGA report 2025 and press release of 8 April 2026
- GKV-Spitzenverband, caps and threshold from 1 October 2026
- GKV-Spitzenverband, product caps from 1 April 2026
- GKV-Spitzenverband, caps and thresholds and group assignment form
- AOK, DiGA data and analysis
- Barmer on the DiGA report 2025 and on the report 2024
- KV Bayern, prescribing DiGA (January 2026)
- KV Baden-Württemberg, e-prescription stays optional (July 2026)
Clinical and product information
- Ärzteblatt, attexis trial; Pharmazeutische Zeitung, attexis listing; GAIA news; psychenet profile; MEDICE prescriber flyer
- Heidelberg University Hospital, adult autism clinic (HEAS)
Secondary sources
- Covington, DiGAV amendment 2026; Rödl & Partner, outcome measurement; Rödl & Partner, DiGA explained
- Physitrack, DiGA reform 2026; MedtecLIVE, DiGA regulation 2.0; BIOPRO, DiGAV summary
- Pharma Deutschland, statement on BSI certification; fireup.pro, technical requirements and AI Act 2028; Pure Global, AI Act dates; PwC, TR-03161
- Forvis Mazars, BStabG and GeDIG; Bitkom, amendment on early stops; zm-online, vote planned for October; SVDGV, GeDIG; SVDGV, share of GKV spending
- Ärzteblatt, AOK position; Pharmazeutische Zeitung, FinanzKommission
- research2guidance, cost of a DiGA; JMIR 2024, DiGA market analysis; PMC 2024, DiGA in practice; nexvyra, directory count; leinetal24, codes per year
- Correctiv, adult autism waiting times
- mind Health, PECAN (March 2026); French ministry, PECAN launch; AP-HP, PECAN order of 10 September 2024