Ten referrals become one decision-ready admission — in minutes.
The software analyses incoming referrals, reconciles them against the documentation requirements of the applicable OPS and DRG logic, and makes the result available to the interdisciplinary admission team on a shared board. Your hospital makes the admission decision — on better data, faster.
Before a single admission decision is made, the admission team has already invested hours sifting referrals whose completeness can only be assessed page by page. In the end, hospitals are penalised for gaps that were not created inside the hospital but in the referring practice.
≈ 10Referrals per actual admissionObservation across aiomics customer hospitals
8 – 22Pages per referral on averageObservation across aiomics customer hospitals
> 80 %of referrals contain incomplete or erroneous informationObservation across aiomics customer hospitals
The economic consequences extend beyond the effort itself. The German Medical Service audits around 1.4 million hospital invoices each year; roughly every second audited bill is challenged. In Bavaria alone more than 193,800 hospital invoices were audited in 2024; for challenged bills the average reduction was €1,469 per case. A significant share of these challenges relates to primary misallocation — the question of whether inpatient admission was warranted at all. Hospitals that substantiate that question at the moment of admission stand on firmer ground in any later audit.
As soon as a referral reaches the hospital — as PDF, fax, scanned document, or structured dataset — the system reads its content and reconciles it against the documentation requirements of the applicable OPS and DRG logic. Within minutes a structured view is available: which secondary diagnoses are substantiated, which would need to be added, which OPS prerequisites are met, which are missing, whether prior reports are complete.
For Hybrid-DRG-relevant cases — the 2026 catalogue covers 69 case rates with 904 OPS codes — those situations are flagged where the boundary between outpatient, short-stay, and full inpatient care must already be decided at admission. (KBV: Hybrid-DRG 2026)
02
Input channel: optionally integrated
On request, today's incoming faxes can be routed as PDFs directly into the system — avoiding media breaks and double entry. Deep integration with the hospital information system is not required; the application runs in the browser. Where an interface is desired, HL7v2, FHIR R4, and a document-based handoff are available.
03
Decision: interdisciplinary kanban board
Each analysed referral appears as a card on a board visible to the entire admission leadership — admitting physicians, bed management, patient coordination, nursing leadership, and social services where applicable. The card shows the key clinical facts, the documentation completeness status, and an automatically generated suggestion of which gaps to request from the referrer before admission. The team decides — admit, decline, query. The decision is documented on the card and sent back to the referrer immediately.
In the hospitals we work with, the admission rate rises by roughly two to three percentage points as soon as decision-ready referrals are answered within hours. Patients are placed at competing hospitals less often, simply because someone replied faster.
04
Admission: from fragments to a working document
When the patient arrives, the preparation is already done. Documents the patient brought along, the contents of the admission questionnaire, and — if desired — a transcribed initial history are merged into a structured admission document. Every statement is linked to its underlying source; missing information is marked as a gap rather than silently filled in. Junior physicians get an overview that saves them hours of sorting. Nursing, therapy, and social services work from the same data baseline from day one.
The result aligns with the documentation requirements of the relevant reimbursement paths — both for acute care (§ 301 paras. 1–3 SGB V) and for rehabilitation (§ 301 para. 4 SGB V) in data exchange with the statutory health insurers and the German Pension Insurance.
What this means economically
Three orders of magnitude can be reported from our work with customer hospitals. These are observations from live operations, not promises — the concrete impact depends on case mix, existing process quality, and discipline in referral follow-up.
≈ 0.5 FTE
Patient management relieved
Roughly half a full-time equivalent in patient management is freed up by the software's preparatory work. The remaining work shifts from sorting and searching to deciding and communicating.
+ 2–3 pp
Higher admission rate
Because decision-ready referrals are answered within hours instead of days. For a rehabilitation hospital with 1,000 admissions per year and an average reimbursement in the five-figure range, the magnitude is economically meaningful.
Fewer MD dialogues
Protection against primary misallocation
Because the documentation evidence relevant to reimbursement is examined and recorded at the moment of admission. This relieves the interface between medical controlling and payer — and protects revenue that would otherwise become disputed after the fact.
What the software explicitly does not do
The system makes no clinical decisions. It does not diagnose, recommend therapy, or suggest treatment. Admission decisions, treatment decisions, and responsibility for patient safety remain with the attending physicians.
The system does not invent content. Every statement in the admission document is linked to its original source; information that cannot be substantiated is marked as a gap rather than silently added. This separation is the basis on which the software is deliberately positioned outside the scope of medical device law.
Beforehand: a documentation audit
Hospitals that want to know where they stand before introducing the software can commission a documentation audit. Working from an anonymised sample of real cases, the system reviews admission and progress documentation for the gaps where a subsequent MD audit typically begins. The report identifies which case constellations carry the highest risk of primary misallocation, where the chain of evidence is thin, and where invoice reductions are likely. A dress rehearsal for the actual audit.
Legal framework and data processing
Processing entirely within the European Union (Frankfurt data centre).
Templates for the data processing agreement and DPIA are delivered with the product.
A works council information package per § 87 BetrVG is available for co-determination.
Supported data standards: HL7v2, FHIR R4, LDT 3.0, § 301 SGB V (paras. 1–3 and para. 4), ICD-10-GM, OPS.
Preparation for the EU AI Act (applicable from August 2026); AI literacy training for clinical staff is included.
Frequently asked questions
What does the aiomics Referrals & Admissions module do?
The software evaluates incoming referrals — referral letters, prior findings, post-acute rehab applications, medication lists — within minutes, checks them for completeness against the documentation requirements of the applicable OPS and DRG logic, and presents the result on an interdisciplinary kanban board. The admitting team decides admission or refusal on a shared data foundation. When admitted, the available sources are merged into a structured admission document for physicians, nursing, therapy, and social services; what is not substantiated is flagged as a gap.
How many referrals does a hospital review per actual admission?
In the hospitals aiomics works with, the ratio is typically around ten incoming referrals per admission. Documents range from 8 to 22 pages. By our observation, more than 80 percent contain incomplete or erroneous information — gaps that later become visible as primary misallocation in MD audits.
Is aiomics a medical device?
No. The Referrals & Admissions module is administrative data and documentation software. All hints are framed as suggestions for physician review. Admission, treatment, and reimbursement decisions are made exclusively by the attending physicians and the hospital's medical controlling team.
Does this work without HIS integration?
Yes. The system runs in the browser. Integration with the hospital information system is not required. Hospitals that want to route a direct input channel — for example fax-to-PDF — into aiomics can set this up optionally to avoid media breaks at admission.
What effects do you observe in customer hospitals?
Three effects: first, a marked reduction in processing time per referral, in the order of magnitude of half an FTE in patient management; second, an admission rate that rises by roughly 2 to 3 percentage points because decision-ready referrals are answered faster; third, fewer downstream case dialogues from primary misallocation because gaps are caught before admission. These figures are observations, not promises — the concrete impact depends on case mix and process discipline.
Which reimbursement paths are supported?
Acute care (§ 301 paras. 1–3 SGB V, GKV) and rehabilitation (§ 301 para. 4 SGB V, GKV and German Pension Insurance). Documentation requirements are applied per reimbursement path.
How quickly is a hospital productive after contract signing?
Because the system runs in the browser and does not require deep HIS integration, productive use in the admission department can start within a few weeks. Onboarding, training, and the AI literacy course are included in the license.
Where is the data processed?
Processing is entirely within the European Union, in a Frankfurt data centre. Data processing agreement, GDPR-compliant data protection management, and works council information are delivered with the product.
Can the audit logic be adapted to our hospital?
Yes. Hospitals can create as many of their own presets as they wish — for example for admissions to specific departments, for particular OPS constellations, or for the requirements of individual payers. On request, we run the initial calibration jointly with your medical controlling team so the presets fit your treatment focus from day one.
Can individual OPS codes be configured with their own rules?
Yes. The audit logic can be tied to individual OPS codes with their respective documentation and structural requirements — for early rehabilitation complex treatments as well as for OPS codes with structural minimum-requirement criteria. Presets can be updated after the annual OPS revision; the hospital decides from which admission date a new version applies.
Can the application be tailored to the requirements of the German Pension Insurance?
Yes. For rehabilitation in the DRV space, the documentation requirements per § 301 para. 4 SGB V in XML data exchange are honoured, including the SWE checklist and the three audit stages of the substantive dataset audit performed by the German Pension Insurance. Rehab-specific presets for post-acute rehabilitation (AHB) and phase-transition constellations can be created.
Can we check where our current gaps are before introduction?
Yes. We offer an upstream documentation audit on an anonymised sample. The audit shows in which case constellations the chains of evidence are weak today and where a subsequent MD audit is likely to start. Hospitals that later introduce the software receive the audit as setup foundation; hospitals that only commission the audit receive the report standalone.
Does the system invent content when the chart is incomplete?
No. Whatever appears in the admission document is substantiated by the available sources and linked to its original record. Anything that cannot be substantiated is flagged as a gap, not filled in. This separation is the basis of the software's regulatory positioning.
In 30 minutes we walk you through the journey of one incoming referral — from a PDF on the fax server to the team's admission decision. You will then see where the biggest gaps in your current process lie.
