All articles

The case lives in email and spreadsheets.

The laboratory system holds the sample and the report. Everything between them, candidates, queries and decisions, lives in files owned by individuals.

A senior scientist leaves a laboratory after eleven years. Her successor inherits a shared drive with four hundred folders named by case number, each holding a candidate spreadsheet in a format that changed three times over the decade, an email export for some cases and not others, and a document called 'notes' that is sometimes empty. Forty of the cases are still open.

For each of the forty, the successor has to work out what was done, what was considered, who was asked what and whether they answered. For several, the honest answer is that nobody now knows.

What the field knows

Laboratory information systems were built around the sample and the result. They track accessioning, batching, sequencing and the issued report, and they do it well, because those are the steps accreditation inspects. The interpretation that happens between the variant call and the report, the part that takes the most skilled time, is largely outside them. The laboratory standards (Rehm et al., 2013) require that the interpretation be documented and retained; they do not say where, and in most laboratories the answer is wherever the scientist put it.

The consequences have been described from several directions. The reanalysis literature (Wright et al., 2018; Deignan et al., 2019) notes that reanalysis is only cheaper than a first analysis if the first analysis left a usable record. Work on diagnostic error in medicine generally identifies handover and information loss between people as a leading cause, and the undiagnosed case, which passes through more hands over more years than any other, is the most exposed to it.

Other knowledge-intensive professions moved their working record out of personal files a generation ago. Engineering keeps its reasoning in issue trackers and design records that survive the engineer; law keeps the matter file as the firm's property rather than the associate's. The working record of a genomic investigation is still, in most places, a personal spreadsheet. The difference is not that genomics scientists are less careful. It is that nobody built them the matter file, so each of them built their own.

A spreadsheet is a good tool and a poor system of record. It has no owner but its author, no history but its file name, no link to the sample, the report or the clinician's reply, and no way for two people to work on it at once without one of them losing. Each of those is a property a case needs and a spreadsheet cannot have. The problem is not the scientist's diligence. The most careful spreadsheet in the laboratory is still a document that one person can read and that stops being a record the day that person is unavailable.

Where practice breaks

The record leaves with the person. Spreadsheets on a personal drive, decisions in a personal inbox, reasoning in a personal memory: when the scientist moves on, each goes with them or is left behind in a form only they could read. The opening case is the ordinary end of an ordinary career.

Nothing links the pieces. The candidate list cannot point at the clinician's answer to a phenotype query, because one is a file and the other is an email. The decision to dismiss a gene cannot point at the paper that justified it. A reader has to hold all of it in their head at once, which is possible for the person who built it and for nobody else.

Open cases become invisible. A case that is waiting on a reply, a sample or a reanalysis lives as a row somewhere, and there is no view across all the rows. The forty open cases in the opening story were open for years without anyone being able to list them, and some of them would have been solved by information that arrived and was filed under a different case number. A clinician's letter with a new finding, a sibling's sample, a reclassification notice: each reached the laboratory and none reached the case.

Audit becomes reconstruction. When a report is questioned, the laboratory assembles the story from files, emails and memory, and the assembly is itself a judgement about what happened. A laboratory that cannot produce the working record on request is relying on the goodwill of the person asking.

Reanalysis inherits all of it. The scientist reopening a case two years on finds a spreadsheet with a column of gene names and a colour code whose meaning left with its author. The dismissals have to be redone, the queries re-sent, and the reanalysis costs as much as the first analysis did, which is the main reason most services cannot afford to do it.

How Tulparus handles it

The case is a single record owned by the organisation, and everything that happens to it happens on it. Phenotype, pedigree, samples and files; candidates with their evidence and their standing; hypotheses; queries to the clinic and the answers; decisions with their authors and reasons; the report versions. There is no separate spreadsheet, because the structure the spreadsheet was standing in for is the case.

Every element is linked to what it rests on. A dismissed candidate points at the evidence that dismissed it; a hypothesis points at its missing items and the actions proposed for them; a query points at the finding it concerns and the answer it received. A reader opening the case for the first time can follow any decision to its ground without asking the person who made it.

Open work is visible across the service. Cases waiting on a reply, a sample, a review or a reanalysis trigger appear in one view by state and age, with an owner. A scientist leaving hands over a list of cases with their current state and open items, and the successor opens each case to find the full working record exactly where it was left.

The record is tenant-owned and exportable. An auditor, a regulator or a successor service can be given the working record of a case as a whole, with its history, which is what the retention requirement always meant and what a shared drive could never deliver.

What you can do tomorrow

Short of a new system, the laboratory can decide that the working record belongs to the laboratory.

  • Agree one shared case-note template and location, owned by the service, and make its use part of sign-off. The template matters less than the ownership.
  • Keep clinical queries and their answers in the case note, copied from the email, with the date. The email is the message; the note is the record.
  • Maintain one list of open cases with their state and what each is waiting for, and review it monthly.
  • Before anyone leaves, walk their open cases into the shared note with them. It is the cheapest reanalysis you will ever do.

Sources

  1. Rehm HL, Bale SJ, Bayrak-Toydemir P, et al. ACMG clinical laboratory standards for next-generation sequencing. Genetics in Medicine, 2013.
  2. Wright CF, McRae JF, Clayton S, et al. Making new genetic diagnoses with old data: iterative reanalysis and reporting from genome-wide data in 1,133 families with developmental disorders. Genetics in Medicine, 2018.
  3. Deignan JL, Chung WK, Kearney HM, Monaghan KG, Rehder CW, Chao EC. Points to consider in the reevaluation and reanalysis of genomic test results: a statement of the American College of Medical Genetics and Genomics (ACMG). Genetics in Medicine, 2019.
  4. Singh H, Giardina TD, Meyer AND, Forjuoh SN, Reis MD, Thomas EJ. Types and origins of diagnostic errors in primary care settings. JAMA Internal Medicine, 2013.

Four hundred folders, one record.See a case the next person can read.

A demo takes 20 minutes, screen shared, on synthetic cases. Bring the question from one of your own.