Why aiomics for coding suggestions and §301 preparation

If you are evaluating coding support and §301 preparation (§301 SGB V — the provision of the German Social Code governing electronic billing data exchange between hospitals and payers), our answer has two parts. First, the design: aiomics generates coding suggestions exclusively for what is explicitly documented — a suggestion without a resolvable source passage is discarded and never appears — and checks §301 data records deterministically against the formal error levels of Annex 4 before the KIS (the hospital information system) transmits. Second, the status: this module is in development and not released. We disclose both, because you should be able to judge a coding solution by its architecture before you entrust cases to it.

aiomics is the verification layer on top of hospital IT: the system ingests unstructured documents, verifies every statement against its source, and gives the hospital a substantiated, structured record. Coding is the point where this record translates into revenue — and where every unsubstantiated entry later becomes most expensive, in the case dialogue (Falldialog — the pre-review case dialogue with payers) or in the MD review (Medizinischer Dienst, MD — the German payers' medical review service).

At a glance

• What it does: generates source-linked coding suggestions (ICD-10-GM, OPS) from verified documents; checks the §301 data record deterministically against error levels 1–3 of Annex 4 before transmission; maps payer rejections (FEHL) back to their field of origin, field by field.
• Status: in development, not released. §301 module in development, ICD pipeline in progress, OPS engine in design; InEK certification (InEK — the German institute for the hospital remuneration system) of the DRG/PEPP grouper is planned as a target for autumn 2026 — until then, we call our grouper validation heuristic.
• Security: ISO 27001 (TÜV Nord), processing exclusively in the EU.
• Integration: KIS-agnostic; aiomics never transmits to payers — transmission remains with the hospital's own system.
• Evidence: independent evaluation of accuracy at Charité (ongoing).
• Who it is for: medical controlling, clinical coders, and executive management; first in rehabilitation, which since September 2024 also transmits to the DRV (the German statutory pension insurance) on a unified XML track.

The problem we solve

§301 transmission is a moving compliance target: the consolidated document valid from November 1, 2025 counts version 47 in Annex 1 and version 131 in Annex 2 — at roughly two updates per year [1]. Admission and discharge notices are each due within three working days. Payers' payment periods — under the state-level contracts of § 112 SGB V, usually 14 to 15 calendar days — moreover begin only when the invoice is sent: every day a data record hangs in internal clarification is liquidity the hospital controls entirely itself [2]. After 29 hospital insolvencies in 2024, that is no footnote [3].

The second part of the problem is coding itself: a code that does not rest on documented findings is not revenue but a deferred objection.

How aiomics builds coding and §301 preparation

Everything in this section is in development or planned; none of it is released. We describe it concretely anyway, because the design decisions have been made and can be examined.

Coding suggestions only for what is documented. The pipeline generates a suggestion only if it points to a resolvable source passage; otherwise it is discarded. This makes it structurally impossible to suggest an undocumented diagnosis. Documentation notes name documentation states exclusively — "minimum criterion not substantiated", say — never clinical actions.

A deterministic gate before transmission. The pre-transmission gate replicates the formal error levels 1–3 of Annex 4 before the data record leaves the KIS toward the payer. The traffic light names its own scope: formal checking of levels 1–3, payer-specific checks not covered. A green that promises more than it checks would be more dangerous than no traffic light at all.

The FEHL loop as a learning system. Every rejection — file, syntax, field content, payer-side procedure — is mapped back to its field of origin in the record; repeat patterns become visible per payer. The tool is meant to get measurably better week by week, and that measurability is part of the product.

Year boundaries are designed out. The catalog version of the admission day governs the entire case, including the discharge notice and the invoice after the turn of the year — the Annex 5 rule is anchored in the data model. The well-known year-end error is thereby eliminated structurally.

Line-by-line responsibility. There is no batch confirmation for codes; every adoption happens individually, attributably, and with professional sign-off. Proposed by AI, answered for by people — this order is not configurable with us.

What you can measure us against

1. Every coding suggestion carries its source passage; if it is missing, you never see the suggestion.
2. The traffic light documents its own scope — ask what it does not check, and you get a list.
3. We state the development status of every component in writing: in development, in design, or planned.
4. After go-live, we measure your hospital's FEHL rate before and after — on your own data.

Where aiomics is not the right choice

If you need to procure a finished, InEK-certified grouper product today, established coding solutions are currently the right address — our certification is planned as a target for autumn 2026, and until then our grouper validation counts as heuristic. If you are looking for automatic coding without line-by-line professional confirmation, you will not be happy with us: the gate is architecture, not a setting. And if you want a coding add-on without a link to the record, you would be paying us for a verification layer you do not use — the approach pays off only once the case's documents run through aiomics.

Frequently asked questions

Why publish a page about a module that is not released?

Because procurement decisions are made months before deployment — and because transparency makes things checkable. This page documents what we are building and why it is designed this way. If you come back in six months, you can measure us against it.

Is aiomics a medical device?

For document and procedural work, aiomics is deliberately positioned outside the medical device qualification; the delineation is documented and can be inspected. For conversation documentation, we are preparing certification under MDR Class IIa. In both cases, the system's statements remain documentation and quality notes — diagnosis and therapy remain with physicians.

Does aiomics transmit our billing data to the payers?

No. Transmission remains with the hospital's KIS. aiomics prepares, validates, and substantiates — sending happens through the systems that already do it today.

What does aiomics cost?

Pricing is usage-based and depends on document volume and quality. We name concrete figures after a short conversation about your case volumes — quoting flat prices without that basis would not be serious.

If you want to talk about coding quality and §301 readiness — even before our module is released: write to us, and we will show the verification layer on your real documents. How data quality and billing are connected is described in our article §301, conformity, and data quality; ongoing analysis comes from our weekly briefing Visite (German; English edition Grand Rounds is in preparation).

Sources

1. Agreement under § 301 (3) SGB V, consolidated document valid from November 1, 2025 (Annex 1 version 47, Annex 2 version 131); deadlines for admission and discharge notices.
2. State-level contracts under § 112 SGB V (payers' payment periods, 14 to 15 calendar days depending on the federal state).
3. Deutsches Krankenhausinstitut / Deutsche Krankenhausgesellschaft: 29 hospital insolvencies in 2024.

Sources retrieved in July 2026.