
You may have heard that Germany's new rules for health apps mean you now need Diagnostics as a Service. New rules did take effect this year, and they do require measurement and twice-yearly reports. None of the entries they ask for is measured from a body, though. So a product built on that rumour pays for a blood test the reports never ask for.
The apps in question are called DiGA, short for "digitale Gesundheitsanwendung", which is a digital health application a doctor can prescribe and the statutory system reimburses. The rumour treats two separate subjects as one, so this article takes them one at a time. The first subject is what the new reporting asks for. The second is where a measured blood value belongs in the DiGA rules.
The German reporting rules for prescribed health apps collect minutes of use, sessions, drop-off and questionnaire answers. They collect no blood value and no measurement of the body. Source: the Second DiGAV amendment ordinance of 27 January 2026, new paragraphs 23a to 23e.
An ordinance called the Second DiGAV amendment took effect on 1 February 2026. It added ongoing reporting for apps that are already permanently listed.1
"Permanently listed" is a term of art that decides who the reporting applies to. A DiGA is in the directory either provisionally, while its evidence is still being produced, or permanently, once that evidence has been accepted.2 The reporting duty is on the permanent group only.
For that group the cycle is fixed: makers collect every quarter and report twice a year. The first report covers the last two quarters of 2026, so on that cycle collection started on 1 July 2026.2
Yes, though the rule doing that is older than the new reporting and works differently. Paragraph 134 of the social code, book five, says that from 1 January 2026 a price agreement must set at least a fifth of the payment as success-dependent.3
The ordinance does not join that price rule to the new reporting. The health ministry's draft of the ordinance says that how the new measurement is taken into account when the success-dependent share is set is not part of the ordinance.4
That question is left to the price negotiators instead. Paragraph 134 does have makers hand the measurement results to the payer side for those negotiations, so the two will meet at the negotiating table.3
The timing keeps them apart for now. The price duty started in January 2026, while the first reports are not due until April 2027. As we read those two dates, whatever success means in a price agreement signed today, it cannot yet mean the new data.
In the evidence a maker needs to get listed in the first place. The new ordinance changed almost none of that evidence.
The definition of a positive care effect is unchanged. It names two routes: a medical benefit, or one of nine named improvements in how care is organised.5
A medical benefit has to show in illness, death or quality of life, and a measured value such as HbA1c is an accepted route to it. The care-process route asks for one of nine named improvements and no clinical measurement at all. Sources: paragraph 8 of the DiGAV and the BfArM DiGA guide version 3.6.
The first route is where a blood value belongs. A medical benefit has to be visible in endpoints that matter to the person, which the rules name as health status, the length of an illness, survival and quality of life.5
A measured clinical value can be a route to one of those endpoints. The regulator's own guide uses lowering HbA1c in a diabetes app as its worked example.6
One change did reach study design. A new paragraph 10(8) of the DiGAV requires a study to record how long and how often people use the product alongside whatever else it measures.1
That is a usage record added to the study, not a measurement removed from it.
So the case for a measured value comes from the endpoint you choose for listing. A clinical value is an endpoint the regulator's own guide accepts.
The first reports are due on different dates, depending on which data they cover. Usage figures are first reported on 15 April 2027. One question about the person's health and five satisfaction questions are first reported on 15 April 2028. Questionnaires for the specific condition start on 1 January 2028, and the first report on them is due on 15 April 2029.2
The ordinance took effect on 1 February 2026, collecting usage data started on 1 July 2026, the first report is due on 15 April 2027, and questionnaires for the specific condition start on 1 January 2028. Sources: the ordinance itself and BfArM.
Not the app, and not the company that makes it. A team that decides on a measured endpoint therefore has three things to arrange: where the person goes to be drawn, who is allowed to order the test, and how the product receives the number.
A study needs the same three arrangements as a shipped feature. So a team that builds them for a trial has already built them for the product.
Those three arrangements give you the questions to ask, and each has an answer you can verify rather than accept.
The last question is the easiest to skip, yet it decides whether your evidence holds. Two values measured by different methods are not a before and an after. A supplier who does not raise that unprompted has not thought about your problem.
Aniva does the measuring in that arrangement. Because both measurements run through one service rather than two, the second value is measured the same way as the first. The collection network is Aniva's own rather than a subcontracted one. The people using your product see your name on all of it.
Which endpoint to choose, and which of the two routes to listing your evidence should take, is a decision for your own regulatory adviser rather than for a supplier.
Whether your endpoint is a measured value is the deciding question. If it is, book a demo.
This article describes German rules for prescribed health apps as they stand for companies buying blood testing. It is general information and not legal or regulatory advice, and your own listing is a question for your own adviser.

Aniva handles the lab, the logistics and the report. You keep the patient relationship and your own branding.
Twenty minutes, and you will know whether it fits your setup.