Storing medical reports so they are readable later

People plan storage against the risk of losing a document. The more common outcome is different and less obvious: the document is exactly where it was put, and it cannot be read.

There are four versions of this and each has its own remedy.

The paper has faded, which is overwhelmingly the commonest and applies to a specific and predictable set of documents.

The document is present and cannot be identified — an undated report, or one whose patient name is illegible, which in a household file is effectively an anonymous document.

The digital copy exists and cannot be found, because it is one of several thousand photographs on a phone with a filename assigned by the camera.

Or the record is stored on media nothing can read any more: a disc for a drive no current computer has, or a download link that expired.

None of these is a storage failure in the sense of losing something. All four are capture and labelling failures, decided at the moment the document was received rather than years later when it was needed. That is the useful reframing: how a record is handled on the day it arrives determines whether it will be readable, and almost nothing that happens afterwards can improve it.

Thermal paper is the specific problem, and knowing which documents use it is most of the defence.

Thermal printing works by heating a coated paper rather than by depositing ink, and the coating continues to react to heat, light, friction and contact with certain plastics long after printing. The result is progressive loss of contrast, ending in a blank sheet. In Indian conditions — heat, humidity, a document kept in a bag or a plastic sleeve — this can happen well within a couple of years.

The documents most often printed this way are the small ones produced at a counter: pharmacy bills, payment receipts, some laboratory slips, and reports from certain machines. That list overlaps almost exactly with the documents an insurance claim depends on, which is why claim files assembled from old receipts so often contain blank paper.

The defence is not better storage. A thermal print kept perfectly still in a cool dark drawer still degrades, just more slowly. The defence is copying it to a stable medium promptly — a scan, a photograph, or a photocopy — while there is still contrast to copy.

Two secondary points. A thermal document should be kept out of plastic sleeves that can accelerate the reaction, and away from heat. And where a thermal receipt matters for a claim, a photocopy taken alongside the digital copy is worth having, because a paper copy on ordinary paper does not fade.

A digital copy is only as good as its legibility, and the difference between a usable capture and a decorative one comes down to a handful of things.

Capture the whole document, edge to edge, including the header with the facility name and the footer with any reference number or signature. A crop that removes the letterhead removes the evidence of where the report came from, which is precisely what someone will later want to know.

Keep the page flat and the camera parallel to it. A photograph taken at an angle distorts a table of results in a way that makes columns hard to attribute.

Avoid your own shadow and direct glare. On glossy paper the reflection commonly lands exactly on the values.

Check it at full size before putting the paper away. This is the step that gets skipped, and it is the only one that catches a capture that is unreadable, since a thumbnail of an unreadable page looks identical to a thumbnail of a readable one.

Multi-page documents belong in one file rather than as loose images — a discharge summary split across four photographs will be reassembled in the wrong order by somebody eventually. A scanning application that produces a single document per record solves this and also usually corrects perspective.

And capture both sides where anything is printed on the back, which on lab reports frequently includes the reference information.

The central constraint of a digital medical file is easy to state and easy to forget: the contents of a photographed document are not searchable. Nothing will find a report by looking inside the image. Retrieval depends entirely on the filename and the folder it sits in.

A convention that works has four parts, in this order: the person, the date written year first then month then day, what the document is, and the facility. Written that way, files sort themselves chronologically without any further effort, and a partial name typed into a search box finds them.

The date-first ordering matters more than it appears. Written with the year first, dates sort correctly as text; written any other way they do not, and a folder of a hundred reports ends up ordered by the day of the month.

Folders should mirror the paper file: one per person, then one per document type. Depth beyond that costs more than it returns.

The date in the filename should be the date of the document, not the date it was scanned. This is a small discipline with a large effect, because a set of files named by scanning date records the history of your filing rather than the history of the treatment.

And for a long-running condition, a plain spreadsheet listing each document's date, type, facility and filename gives the searchability that images cannot, without needing any particular software.

Imaging records fail differently from paper, and the failure is usually the storage medium rather than the image.

An imaging study has two parts that are frequently confused. The report is the written document, and it is the part that is read in most consultations. The image itself — the film, the disc, or a download — is what a specialist may want to examine directly. They serve different purposes and they separate immediately if stored apart, so they belong together with the report on top and the location of the image noted on it.

Three specific problems recur. Discs are supplied for drives that current computers no longer have, so a study handed over on a disc becomes unreadable through hardware rather than damage — copying the contents onto ordinary storage while a drive is still available is the only fix, and it has to be done before the drive disappears. Some discs carry proprietary viewer software that stops running on newer systems, while the underlying image files, usually in the standard DICOM format, remain readable by other viewers, so copying the whole disc contents rather than just launching the viewer is what preserves the record. And download links issued by hospitals and imaging centres commonly expire, sometimes within weeks, so a link received is a link to act on rather than to file.

A photograph of a film held against a light is not equivalent to the study, though it is far better than nothing and worth taking.

A single copy in one place is not storage, whatever the medium.

The minimum arrangement is the paper original in a known place and a complete digital copy somewhere that is not the phone that took the photographs. A phone is the single most likely thing in the arrangement to be lost, broken or replaced, and a file that exists only on it is a file with one point of failure.

Where the digital copy is kept in cloud storage, access control is the part that needs a decision rather than a default. Medical records are personal data of a kind treated as sensitive in most frameworks, and India's Digital Personal Data Protection Act, 2023 governs the handling of personal data, with the position for health data developing alongside the digital health infrastructure being built under the Ayushman Bharat Digital Mission. Practically, that means a shared folder with a link anyone can open is the wrong shape, and access granted deliberately to named people is the right one.

More than one person should be able to find the file. A perfectly maintained record that only its author can locate fails at exactly the moment it is needed, because that moment often involves its author being the patient.

And the purpose throughout is producing a complete, dated, legible record for a clinician to read — not equipping the household to read it. Storing a report well makes it available. It does not make it interpretable by whoever stored it.