All articles

Turnaround time is set by the slowest handover.

Sequencing takes days. Cases take months. The difference is queues between people, and the rapid programmes show what happens when the queues are removed.

A laboratory measures its turnaround from sample receipt to report at fourteen weeks for a standard exome. The sequencer takes under a week of that. The pipeline takes a day. When the laboratory maps the remaining twelve weeks, it finds them in eleven places: a sample waiting to be batched, a VCF waiting for a scientist to be free, a candidate list waiting for a phenotype clarification from the clinic, a draft waiting for a second reviewer, a report waiting for a multidisciplinary meeting that happens fortnightly, a signed report waiting in an outbox.

Nothing in the twelve weeks is work. All of it is waiting for the next person.

What the field knows

The rapid genome programmes in intensive care are the clearest demonstration that most turnaround is organisational. Farnaes and colleagues (2018) and Clark and colleagues (2019) reported diagnoses from whole-genome sequencing in days for acutely ill infants, with Clark's group describing a pipeline in which phenotype extraction from the clinical record, variant interpretation and the hand-off to the clinician were each engineered to remove a wait. Kingsmore and colleagues (2019), in a randomised comparison of rapid genome and exome sequencing in ill infants, achieved median times to result measured in days for both, with the differences between approaches smaller than the difference between either and routine practice.

Stark and colleagues (2018), describing the implementation of rapid testing in Australian paediatric care, were explicit that the gain came from redesigning the handovers: a dedicated intake, parallel rather than sequential steps, a clinician available to answer the phenotype question the same day, and a reporting path that did not wait for a scheduled meeting. The sequencing chemistry was the same as the routine service used. Stark and Ellard (2022) argued that the rapid model should become the standard for critically ill children, and noted that the obstacles were capacity and workflow rather than technology.

The lesson transfers to the routine service with one qualification. A rapid programme spends people to remove waits, and a routine service cannot staff every case that way. What it can do is see where the waits are, remove the ones that cost nothing to remove, and make the remaining ones visible so that a case does not sit unseen.

Where practice breaks

The case has no single place. It is a sample in a tracking system, a VCF on a share, a spreadsheet of candidates, a draft in a document folder, an email thread with the clinic and a line in a meeting agenda. No one view shows where it is, so no one can see that it has been waiting eleven days for a phenotype clarification that would take the clinician two minutes to give.

Handovers are made by message. A scientist finishes a review and emails the second reviewer, who is on leave; the email waits. A draft is sent to the clinic for comment and the clinic replies to the wrong address. Each handover is a point at which the case can stop without anyone noticing, and the number of handovers in a standard case is in double figures.

Waits are measured as a total, if at all. A laboratory knows its median turnaround and rarely knows which step contributes most of it, because the steps are not timed. Improvement effort goes where it is visible, which is usually the pipeline, and the pipeline is already the fastest part.

Batching adds a wait that looks like efficiency. Samples wait to fill a run, cases wait to fill a review session, reports wait to fill a meeting agenda. Each batch is cheaper per case and slower per case, and the trade is made by default, with nobody deciding it and nobody adding up the days.

The clinic is outside the system. The phenotype clarification, the consent query, the question about whether a relative can be sampled, each crosses an organisational boundary by letter or phone, and the laboratory has no way to see whether the question has been received, let alone when it will be answered. For the unresolved case, which needs more of these exchanges than a straightforward one, the boundary is where the months go.

How Tulparus handles it

The case is one record with a state, and every handover is a transition on it. Intake, analysis, review, clinical query, draft, second review, sign-off: each has an owner, a time of entry and the people waiting on it. A dashboard for the service shows every open case by state and by age in that state, so that the case that has waited eleven days for a clarification is at the top of someone's list rather than at the bottom of an inbox.

Clinical queries live on the case itself, where a message about it would otherwise sit in an inbox. A scientist who needs a phenotype clarified raises it on the case; the clinician sees it in their own view of the same case, answers it there, and the case moves on. The question, the answer and the time taken are recorded, and a query that has waited beyond a threshold set by the service is escalated to the clinic's designated contact.

Review and sign-off do not wait for a meeting. A draft report is available to the second reviewer and to the clinician as soon as it exists, with the structured case behind it, and each records their review as an action on the case. The service can still hold a multidisciplinary discussion for the cases that need one; the cases that do not are not held hostage to its calendar.

The times are measured per step, by design, because the states exist. A service can see that its second review takes a median of nine days and its clinical queries a median of sixteen, and decide where to put effort. Nothing about the sequencing changes. The twelve weeks of waiting are made visible, and visible waits get shorter. For the laboratory in the opening case, the first month of measurement would show which three of its eleven waits account for most of the twelve weeks, and two of the three are usually fixable without a single extra person.

What you can do tomorrow

The first step is measurement, and a spreadsheet will do for a month.

  • Take twenty recent cases and write down the date each handover happened. The step with the longest median wait is the one to fix first, and it is rarely the pipeline.
  • Give the clinic a single, named route for laboratory queries, with an expected response time, and track how long each query waits.
  • Separate the cases that need a multidisciplinary discussion from those that need a second signature, and let the second group report without waiting for the meeting.
  • Publish the per-step times to the team monthly. Waits that are seen shrink without anyone being told to shrink them.

Sources

  1. Farnaes L, Hildreth A, Sweeney NM, et al. Rapid whole-genome sequencing decreases infant morbidity and cost of hospitalization. npj Genomic Medicine, 2018.
  2. Clark MM, Hildreth A, Batalov S, et al. Diagnosis of genetic diseases in seriously ill children by rapid whole-genome sequencing and automated phenotyping and interpretation. Science Translational Medicine, 2019.
  3. Kingsmore SF, Cakici JA, Clark MM, et al. A randomized, controlled trial of the analytic and diagnostic performance of singleton and trio, rapid genome and exome sequencing in ill infants. American Journal of Human Genetics, 2019.
  4. Stark Z, Lunke S, Brett GR, et al. Meeting the challenges of implementing rapid genomic testing in acute pediatric care. Genetics in Medicine, 2018.
  5. Stark Z, Ellard S. Rapid genomic testing for critically ill children: time to become standard of care? European Journal of Human Genetics, 2022.

Fourteen weeks, one week of work.See where the case is waiting.

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