A boy is referred to genetics at two with a developmental disorder. By nine he has had a microarray, a panel, an exome and a reanalysis, four admissions for investigations that a diagnosis would have made unnecessary, a muscle biopsy, two metabolic work-ups and visits to six specialties. His parents have had two further children without a recurrence risk they could rely on. His file is seven centimetres thick and his case is open.
The laboratory's accounts show four tests. Nothing in any budget shows the seven years, and nobody has been asked to account for them, because an open case is not a line item anywhere in the health system that holds it.
What the field knows
The diagnostic odyssey has been measured. Studies of rare disease patients report a typical interval of years from first symptom to diagnosis, with multiple misdiagnoses and many specialist consultations along the way, and the costs of that interval have been estimated from several directions. Analyses of early genomic testing in children (Stark et al., 2017; Schofield et al., 2017) compared the cost of sequencing first with the cost of the traditional sequence of investigations and found that the sequence, with its imaging, biopsies and admissions, cost more than the genome it delayed, before any value was placed on the earlier answer.
The value of the answer has been measured too. In critically ill infants, rapid genome sequencing changed management in a substantial share of diagnosed cases and reduced hospital costs enough to pay for the test (Farnaes et al., 2018). In outpatient paediatric neurology, a diagnosis changed management, including surveillance, treatment and referrals, in a large fraction of cases, and ended investigation in most of the rest. The value of ending investigation is easy to underrate: it is the biopsy that is not done and the admission that does not happen. For families, a diagnosis provides a recurrence risk, access to support and trials, and an end to the search, each of which has been given a value in willingness-to-pay studies.
What the economic literature has not often measured is the cost of the case after the test, when the test was negative. The case does not stop costing anything because the laboratory has reported. The investigations continue, the specialist visits continue, and the reanalysis that could end it is unfunded, because reanalysis is not a test and has no tariff. A service that is paid per test has, on paper, no reason to look at the case again, and the scientist who does so is doing it in time that belongs to something else. Ewans and colleagues (2018) argued that reanalysis is cost-effective when it is done; the barrier is that nobody is paid to do it.
Where practice breaks
The cost is distributed, so it is invisible. The family pays in time, travel, lost work and uncertainty. The clinic pays in appointments and in investigations ordered for lack of an explanation. The laboratory pays in the scientist's time whenever the case is reopened, and in the hour spent reconstructing what was done last time from a report and a memory. Each budget sees its own slice and nobody sees the whole, so the sum that would justify the effort to close the case is never computed.
The incentives follow the tariff. A test has a price and a turnaround target; an open case has neither. A service is measured on tests reported, so it reports tests, and the case that would benefit most from a day of a scientist's attention competes with new referrals that count.
The open cases are not even listed. A service that cannot enumerate its unresolved cases, with their ages and their last activity, cannot estimate what they cost or where effort would pay. The backlog is a feeling rather than a number, and a feeling does not get funded. When a commissioner asks how many unresolved cases the service holds, the honest answer in most services is that someone would have to go through the files.
The investigations ordered in the absence of a diagnosis are rarely attributed to its absence. The muscle biopsy at six is recorded as a muscle biopsy. That it would not have been done if the exome at three had been interpreted with the clinical question in view, or reanalysed at four when the gene was published, appears in no record.
How Tulparus handles it
Every unresolved case is a visible object with a state, an age, a last activity and a list of open items. A service can see, at any moment, how many cases are open, how long each has been so, what each is waiting for, and which have had no activity for a year. The backlog becomes a list that can be sorted, costed and worked.
Reanalysis is triggered by what changed and costed by what it would take. When a knowledge release touches a case, the queue entry carries the specific change and the specific candidate, so the scientist's time goes to the cases where something has moved. A service can report how many cases a release touched, how many were reopened and how many closed, which is the number a commissioner can fund.
The next step is proposed with its cost and what it would resolve, so that the cheap informative action is visible beside the expensive uninformative one, and the service can show that a repeat examination or a relative's sample was proposed before a biopsy. The record of what was proposed, done and declined is kept with the case, so that when the question is asked later, why the biopsy, the answer is in the record and not in anyone's memory.
For the boy above, the case would have shown an open item at four, with the publication that created it, when the gene was published, a reanalysis proposed with the reason, and, had it been acted on, an end to the investigations that followed. The accounts would still show four tests. The case would show why there were not more.
What you can do tomorrow
Counting is the first step, and a service can count without new tools. The numbers below are the ones that have moved funding decisions in services that produced them.
- List your unresolved cases with the date of last activity. The number and the ages are the argument for funding reanalysis, and most services have never produced them.
- For ten open cases, add up the investigations and visits since the negative report. Present the figure to whoever funds the service.
- Give reanalysis a tariff internally, even a nominal one, so that time spent on open cases is counted as work rather than as a favour.
- When a case is finally solved, record what the diagnosis made unnecessary. The avoided costs are the case for doing it sooner next time.
Sources
- Stark Z, Schofield D, Alam K, et al. Prospective comparison of the cost-effectiveness of clinical whole-exome sequencing with that of usual care overwhelmingly supports early use and reimbursement. Genetics in Medicine, 2017.
- Schofield D, Alam K, Douglas L, et al. Cost-effectiveness of massively parallel sequencing for diagnosis of paediatric muscle diseases. npj Genomic Medicine, 2017.
- Farnaes L, Hildreth A, Sweeney NM, et al. Rapid whole-genome sequencing decreases infant morbidity and cost of hospitalization. npj Genomic Medicine, 2018.
- Ewans LJ, Schofield D, Shrestha R, et al. Whole-exome sequencing reanalysis at 12 months boosts diagnosis and is cost-effective when applied early in Mendelian disorders. Genetics in Medicine, 2018.
- 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.