The History of the History: How to Read Prior Records

Tuesday, 8:10 a.m., morning report. The resident presents the night's admission: a 79-year-old patient, syncope while climbing stairs, a contusion mark on the left shoulder, otherwise unremarkable findings on admission. "Known severe aortic valve stenosis." The plan reads accordingly: monitoring, cautious volume therapy, echocardiography, consultation with cardiology to plan an intervention — syncope with severe stenosis, the case seems clear. The chief physician asks a single question: "How do we know that about the stenosis?"

The answer leads into the paper trail. The referral note names the diagnosis. The last discharge letter, from 2024, names it too — carried over from a 2022 letter, which in turn cites an external echocardiography report. The original report is not in the chart. The diagnosis list contains I35.0, aortic valve stenosis — a code that knows no severity grade. The "severe" lives exclusively in the free text, and it has lived there for years.

The hospital's own echocardiography the next day measures a moderate stenosis: valve area 1.2 cm², mean gradient 24 mmHg. The requested original report from 2022 confirms: the stenosis was moderate then, too, with a recommendation for follow-up. Somewhere between report and letter, "moderate, needs follow-up" became "advanced," later "severe." No single error you could hold against anyone — more like friction loss: each transcription paraphrased slightly, and each paraphrase nudged the label. The consequence nearly arrived: with the "severe stenosis," the syncope was apparently explained, and the search for other causes could have ended before it began. The Holter ECG shows intermittent third-degree AV block. The patient receives a pacemaker, not a new heart valve.

The patient does not exist in this form; the mechanics are familiar to every hospital.

How diagnoses harden

Physician letters (Arztbriefe) are the main source of the collateral history, and they cite one another more often than they enclose original reports. There is system to this: §630f BGB — the German statute on treatment-record documentation — expressly names physician letters as a mandatory part of the treatment record [1]. What once appears in a letter thus persists, with an authority the original finding does not always merit. A systematic review in JAMA described as early as 2007 how patchy the transfer of information between hospital and continuing care is: letters arrive late, reports are missing, decisive details are lost at the interface [2]. In the 2019 Düsseldorf survey of general practitioners, vague wording and missing reasons are among the most frequently named deficiencies of physician letters [3] — precisely the soil on which a cautious assessment grows into a hard label.

Add to this that inherited diagnoses carry no expiry date. The anamnestic statement ("patient reports a heart valve condition") and the confirmed diagnosis (own report on file) often look identical in the record. How many diagnoses copied forward in German records would fail a source check, nobody knows — there is no robust number. There is only what every experienced clinician knows from her own practice: the allergy that was never tested, the "epilepsy" without a documented EEG, the renal insufficiency whose stage nobody has recalculated in years.

The history of the history

You can treat prior records as administrative goods that make the admission complete. You can also treat them as what they clinically are: a source with its own error profile, to be taken like an auscultation finding. Four questions do the essential work:

1. What does the diagnosis rest on — original report, quotation from a letter, or the patient's account?
2. When and where was it first made, and by what method?
3. Has anyone seen the original report since?
4. Which decisions hang on it today?

The fourth question sizes the effort: a diagnosis copied forward, on which anticoagulation, fitness for surgery or — as in the case above — the interpretation of a presenting symptom depends, justifies requesting the original report. An inconsequential secondary diagnosis may not. Clinical judgment, nothing new; it is just rarely applied to the record itself.

Practically, this means three things for a department. Inherited diagnoses are labeled as such at admission — confirmed with the report on file, or adopted from history; that is a convention, not a software project. Requesting original reports is residents' work with high teaching value: whoever has once watched a label crumble under the original report reads letters differently. And the ward round that asks "How do we know that?" trains people — it makes the admission what it is at its best: the moment a department rethinks the case instead of copying it forward.

The same care works in the other direction. The letter your own department writes today is somebody's prior record tomorrow — and will be copied forward with the same authority. Whoever lifts an inherited diagnosis into their own letter can pass its provenance along: "aortic valve stenosis known from history; external echocardiography report from 2022 on file, moderate." Half a sentence that spares the next colleague the search that took two days in the case above. This is how the chain you would like to find yourself gets built.

How the completeness of prior records affects admission quality and billability, we have described from a process perspective in prior-record completeness: admission quality begins before admission (in German); how merging two correct findings can produce a third, false one, in how two correct findings become a fabricated diagnosis (in German).

Four questions for the next admission

1. Which of this patient's inherited diagnoses is covered by an original report on file?
2. Which of them is currently steering a therapy or diagnostic decision?
3. For which of them is requesting the original report therefore worthwhile — today, before the decision?
4. Is the provenance of the diagnosis (confirmed / from history) documented so that the next reader can recognize it?

If you would like to read regularly about such questions of documentation quality: our weekly briefing Visite (German; English edition Grand Rounds is in preparation) covers documentation and AI in German healthcare — collegial in tone, no product brochure.

Sources

1. § 630f BGB — Dokumentation der Behandlung. https://www.gesetze-im-internet.de/bgb/__630f.html
2. Kripalani S, LeFevre F, Phillips CO, Williams MV, Basaviah P, Baker DW. Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care. JAMA. 2007;297(8):831–841. doi:10.1001/jama.297.8.831.
3. Bechmann S, Riedel J. Arztbriefe: Epikrise in der Krise? Studienbericht. Düsseldorf: Heinrich-Heine-Universität; 2019. https://docserv.uni-duesseldorf.de/servlets/DerivateServlet/Derivate-53751/Studie_DE-EN_Bechmann_pdf:a.pdf