The UK has agreed to settle a lawsuit over how it selected an IT contract for coronavirus testing at its Lighthouse labs.
The BBC understands that the settlement will cost the government up to £2m.
British company Diagnostics AI claimed it lost out to a European rival UgenTec despite spotting some positive coronavirus cases its rival missed.
It sued the government over the decision, claiming the selection process was “unfair and unlawful”.
Lighthouse labs are a UK-wide network of specialist coronavirus laboratories managed by the government and run by private firms. When the labs were set up, companies pitched to analyse the test results.
The dispute was due to be played out in court. It would have meant a public examination of the accuracy and speed of the testing system, at a time when it has come under serious criticism.
But the government has decided to settle the case and will pay Diagnostics AI compensation and most of its legal fees.
However, despite agreeing to the payout, the government has refuted the claims made by Diagnostics AI, saying they are “inaccurate”.
“The tests are reliable and effective, the laboratories that undertake them have been reviewed and assessed by experts and the percentage of false negatives or positives is tiny,” said a Department of Health and Social Care spokesperson.
“This was a commercial dispute over a software contract where a number of factors were considered before it was awarded, which is still subject to final agreement over costs.”
As the contract was worth more than £1m, the BBC understands the settlement including legal costs could amount to around £2m.
Dispute over contract decision
Swabs are taken from people at testing sites or home tests and treated with a chemical process that produces a graph. The software is used to determine whether the graphs show the sample was positive or negative for coronavirus.
Diagnostics AI claimed UgenTec’s analysis of a trial run of 2,000 samples was flawed. In some cases, it claimed UgenTec found negative coronavirus results, when the results were actually positive or inconclusive.
“The system that they ultimately went with and decided to pay for missed around 50 out of 800 positive [results], so that’s around one in 15, or so, one in 16 – to be precise – positives,” Diagnostics AI’s chief executive Aron Cohen told the BBC.
“Obviously when that translates to hundreds of thousands of samples a day, that’s potentially thousands of missed positives going out every day. So that was really worrying for us.”
UgenTec in return claimed that no patients were affected at all as it was a trial run.
“We provide crucial covid interpretation services to the Lighthouse Labs to help them manage the vast amounts of data they generate. These claims are inaccurate and misleading,” UgenTec’s chief executive Steven Verhoeven told the BBC.
“None of these samples refer to actual results given to patients or the public and to imply any public health impact is wrong. Live tests were not being supported by our software at the time which was in the process of being implemented. As illustrated by independent tests, we have every confidence in our software and the services we provide.”
‘A commercial dispute’
Two non-profit companies owned and funded by the government were also sued by Diagnostics AI – namely UK Biocentre and Medicines Delivery Catapult (MDC), which ran the process to decide which company to use.
Court papers show that between 31 March and 14 April, Diagnostics AI repeatedly requested information about exactly what services were required and how their bid would be evaluated.
Diagnostics AI say it never received the information it asked for. This is refuted by UK Biocentre, which says both providers were given the same information.
When the two bids were being considered in early April, the UK was facing what Boris Johnson had called a “moment of national emergency”.
In such urgent circumstances, the law does make provision for the government to buy services without a competitive process, if certain conditions are met.
However, it is understood that both Diagnostics AI and Ugentec had been recommended to UK Biocentre, and so a decision was made to evaluate both offers.
Diagnostics AI says this process was unfair and flawed, but UK Biocentre insists it was fair to both bidders.
A spokesperson for UK Biocentre said: “The allegations are groundless; this was a commercial dispute. The software in question is being used widely in the Lighthouse Laboratories, in some NHS laboratories and abroad.
“External quality assurance has confirmed that the polymerase chain reaction (PCR) testing in the Lighthouse Laboratories, of which the automated diagnostic software forms part, is performing well.”
A spokesperson for MDC also provided the BBC with a statement: “The full results of evaluation identified UgenTec as a safe and quality provider, able to deliver in high volumes, and with a comprehensive support system in place. It has performed superbly over the past six months, analysing over eight million test results for the nation. The litigation was purely a commercial dispute.”
The BBC understands that investigations were carried out into the claims made by Diagnostics AI, but concluded that concerns over the safety of UgenTec’s software were unsubstantiated.
However Mr Cohen disagrees: “The government is paying out a lot of money. And they’re paying this out, you know, to avoid it at least in part, to avoid having to have these issues aired in court, and to have discussions over the accuracy of the testing.”