The First Words on the Ambulance Ramp.

A paramedic brings a patient into the emergency department and speaks six words. The nursing triage nurse types a primary ICD code based on those words within two minutes. The emergency physician, reviewing that code thirty minutes later, accepts it or modifies it based on her assessment. The admitting consultant six hours later writes an admission note that typically mirrors the triage ICD rather than rebuilds it from scratch. Six days later, the coder closes the case against the primary diagnosis and assigns the DRG. Reimbursement follows. And the reimbursement was effectively determined, in a meaningful fraction of cases, by those six words at the ambulance ramp.

This is the most under-audited interface in the German hospital revenue chain. The paramedic is an employee of a different organisation, governed by a different set of protocols, trained in a different curriculum, and speaks directly to a triage function operating under its own time pressures. Nobody in the hospital formally owns this interface. No quality committee audits the handover phrasing. No coding department traces which handover phrases systematically route cases into lower-reimbursement DRGs. The interface is invisible to every function that would have the standing to question it.

The CAPTURE layer of CuraOS includes this interface because the revenue impact, at the scale of a regional hospital’s annual emergency admission volume, is measured in the low seven figures. Not because hospitals want to inflate coding beyond what the clinical complexity justifies — but because the clinical complexity that was genuinely present gets systematically lost in the handover language.

This is the observation: the first words on the ambulance ramp are clinical shorthand, not documentation. Treating them as documentation, without a structural bridge between paramedic language and coding-appropriate clinical description, costs the hospital revenue it genuinely earned. The fix is operational, not technical, and takes about ninety seconds per handover.

Why the handover phrase matters so much downstream.

In the default German emergency department workflow, the triage ICD code is not a preliminary estimate that gets rebuilt as information emerges. It is an anchor. Subsequent clinical assessments, admission notes, nursing documentation, and physiotherapy records all tend to orient to the primary diagnosis that was entered at triage. Changes do happen — imaging reveals unsuspected pathology, lab results shift the picture, specialist review re-frames the presentation — but the orientation persists. A case admitted under R29.6 (“tendency to fall”) stays oriented toward falls workup even when the underlying pathology is a femoral neck fracture with anticoagulation-related complications and geriatric multimorbidity. A case admitted under S72.0 (femoral neck fracture) immediately orients toward fracture management with attention to the co-occurring complexities that subsequent documentation captures.

The cascade is well-documented in the German coding literature [1,2]. Primary diagnosis determines DRG grouping [3]; DRG grouping determines reimbursement; documentation of complexity depends on the clinical orientation the team took from admission onward. A case that was never oriented toward its complexity from the start produces documentation that is structurally unable to support the higher-complexity DRG, even when the complexity was genuinely present.

This is why the difference between “fell at home” and “fall with suspected hip fracture on anticoagulation” is not a pedantic distinction. It is the structural difference between a case oriented toward I47C (femur, uncomplicated, BWR 1.389, €6,104) and one oriented toward I34Z (hip fracture with geriatric co-management, BWR 2.629, €11,552) [3]. The difference of roughly €5,450 per case does not reflect over-coding in the second pathway. It reflects accurate coding of the complexity that was genuinely present, captured because the clinical orientation from admission forward supported the documentation chain it required.

The five handover phrases that systematically under-code.

Across the audit work I have been involved in, five specific handover patterns appear repeatedly and each routes the case toward a lower-reimbursement DRG than the underlying clinical picture would support. The patterns are not paramedic errors — they are the shorthand paramedics are trained to use for clinical prioritisation at the scene. They become problems only when the receiving hospital has no structural bridge between the shorthand and the documentation that coding will later require.

The five patterns are not exhaustive. They are the most operationally frequent. A rigorous audit of a hospital’s emergency admissions over a rolling quarter typically surfaces the five plus another three or four specialty-specific patterns. The pattern is almost always that the paramedic handover is clinically appropriate for the paramedic’s role — prioritisation, rapid disposition, immediate safety concerns — but is lexically inadequate for the coding chain that will follow. The gap needs a bridge, not a re-training of paramedics.

Why no function currently owns this interface.

Three structural reasons keep the ambulance-ramp interface invisible.

It is organisationally exterior. The paramedic service is not part of the hospital’s org chart. The hospital’s quality and coding functions have no authority over paramedic training, and no one in the hospital has ever been formally tasked with managing this interface. The emergency department leadership typically focuses on the clinical receiving function rather than the upstream communication structure. The interface exists in a jurisdictional gap.

The revenue impact is invisible at the individual case level. A single case under-coded by €5,000 is not visible in any operational report. The revenue leak only becomes legible at aggregate scale across a quarter or a year, and the aggregate is not attributed back to specific handover patterns in any standard coding report. Without the attribution, the leak is not identified; without the identification, no remediation can be designed.

The fix requires a new 90-second interaction in the ED workflow. The structural bridge between paramedic shorthand and coding-appropriate clinical description is an enriched handover interaction at the ramp — a ninety-second structured exchange in which the receiving clinician or an appropriately trained ED coder elicits the specific clinical features that paramedic shorthand tends to omit. This is not difficult, but it adds a defined step to a workflow that is optimised for speed. Installing it requires leadership commitment that most EDs have never been asked to make.

What the Schlüchtern CAPTURE work showed at this interface.

At Main-Kinzig-Kliniken Schlüchtern, the geriatric department established an enriched ambulance-ramp handover protocol in 2020 as part of the broader CAPTURE restructuring under the research programme with Prof. Dr. Rainer Sibbel at Frankfurt School [4,5]. The protocol is a ninety-second structured exchange covering eight clinical features that paramedic shorthand tends to omit: medication lists (particularly anticoagulation), baseline cognitive status, living situation, recent acute events, acute clinical findings at scene, vital signs trajectory in transit, any witnessed fall mechanism, and suspected primary diagnosis. The exchange is documented on a single-page form that becomes part of the admission record.

The protocol does not replace paramedic handover. It enriches it. The paramedic still speaks the clinical shorthand that prioritises immediate care. The receiving clinician ensures that the eight features are captured before the paramedic crew leaves. The form moves with the patient into the admission workflow. Coding downstream has the structural basis to support complex-treatment DRGs when the clinical picture warrants them.

The pattern generalises across specialties. A cardiology service can install an enriched handover for chest pain presentations. A respiratory service can do the same for dyspnoea. A stroke unit’s handover enrichment has different specific features but the same structural logic. Every service that receives ambulance admissions has its version of the interface; every service that has not audited it is losing revenue at it.

What to do on Monday.

Commission a retrospective audit of fifty consecutive emergency admissions from the most recent completed quarter. For each, trace the cascade from paramedic handover narrative (typically available in the ambulance service’s electronic record) through ED triage code, admission primary diagnosis, and final coded DRG. For each case, the audit question is: did the paramedic handover contain enough information to route the case toward its actual clinical complexity, and was the information captured into the clinical orientation from admission onward?

The audit typically produces a specific distribution. A subset of cases will show the full cascade intact — handover rich, triage appropriate, admission oriented correctly, DRG matching complexity. Another subset will show a handover that did not contain enough information for appropriate routing, with coding that accurately reflects the documentation but under-captures the clinical complexity. A third subset will show a handover that did contain the information but the hospital’s capture broke down at the triage-to-admission interface. The three sub-patterns have different remediations.

For handover-insufficient cases, the remediation is the enriched ramp protocol described above. For triage-to-admission capture failures, the remediation is a structured admission-review discipline that does not simply mirror the triage ICD but re-assesses the primary diagnosis with the full clinical picture. Both remediations are operational workflow changes, not technical systems changes.

Install the enriched handover protocol in a single-department pilot. Geriatric emergency admissions or internal-medicine admissions typically produce the clearest signal because the patient population carries more clinical complexity on average. After ninety days, re-audit. Cases coded under the enriched protocol should show measurably higher complexity capture than the pre-protocol baseline. The revenue signal will not appear in the monthly coding report until the new protocol has been running for a full DRG closure cycle — typically six to eight weeks — so allow a full quarter before judging the pilot.

Do not frame this as paramedic re-training. Frame it accurately: the hospital is installing a structural bridge between the clinically-appropriate shorthand paramedics speak and the documentation-appropriate description coding requires. The bridge is the hospital’s responsibility to build, because the revenue impact is the hospital’s to manage. Paramedics, asked to cooperate with a ninety-second structured handover rather than to change how they speak, almost always cooperate readily. The friction in most pilots is internal, not external.

Six words on the ambulance ramp. A ninety-second enriched handover. The difference between I47C and I34Z, between F41B and F41A, between under-capture and accurate capture. The first words determine the DRG that closes six days later. The hospital that audits the interface captures what it clinically earned. The hospital that does not, loses the revenue quietly — case by case — in increments too small to see in any single report, too systematic not to sum to seven figures annually across meaningful volume.