Skip to Main Content
Page banner

Submitted
Being Assessed
Responded
Thanks for your vote
This voting milestone has now ended, and the idea has progressed to a private milestone
Return to idea list
Thanks for approving the idea
This milestone has now ended, and the idea has progressed to a private milestone
Return to idea list

Automated reporting and interpretation of COVID-19 results using machine learning AI- Increasing throughput capabilities, whilst maintaining quality standards.

E: A connection is helpful

We have a machine learning system (pcr.ai) for automated interpretation, monitoring and reporting of COVID tests which can be used to link PoC devices to central laboratories.  In under three minutes, results from any thermocycler are available for the patient. Patented AI ensures accuracy of results and standardisation ensures results from different equipment and labs can be compared - all in real-time. Complete QC monitoring further ensures reliability of results.

Pcr.ai enables sample collection, extraction and amplification to take place at PoC with the data being interpreted and monitored automatically with notifications of errors being sent to a central laboratory managed by experienced Virologists. 

Remote access for people who are off-site is simple as pcr.ai is based in the cloud.  This also enables near-real-time disease data to be available to third parties such as government agencies, other labs, etc  for epidemiology, trend studies or other analyses.  

In-built functionality tracks deviations in controls before they become an issue, data can be stored centrally and fed into multiple reporting systems as needed.

Pcr.ai is already in use at Kings College Hospital - with strong endorsements from their clinical and laboratory teams.

Pcr.ai can enable teams with limited clinical pcr experience to run tests in a safer and more integrated manner and furthermore can help experienced teams to expand more quickly and easily, without additional hiring for data analysis, review and reporting.

Validation

Our system has been validated by a lab based in Glasgow and Gartnavel NHS Hospital - the study was published in the JCV - https://doi.org/10.1016/j.jcv.2019.08.005 . They concluded that ‘...PCR.Ai (from diagnostics.ai) is a highly accurate time-saving tool that reduces complexity of qPCR analysis and hence the need for specialists and hands-on time. It demonstrated capabilities to enable us to get results out more quickly with lower costs and less risk of errors.’

At Kings College Hospital, the system is a critical part of the infrastructure enabling them to process 500-1000 Covid tests a day, they plan to expand to up to 5-10,000/day in the coming weeks.

 
 
 

How quickly could this be deployed and what are the dependencies?

Pcr.ai takes a few hours to days setup per assay and laboratory. 


 

 

 

What is the likely production volume?

Pcr.ai can process at least 300,000 samples a day and can be scaled to handle millions of tests.

 
 
 

What are the risks and barriers to using this at scale?

The system is reliant on correct labelling of samples and controls otherwise can suffer from ‘garbage in, garbage out’ issues (where not detected due to control failures etc).


Additional redundancy and capacity would be added as usage grows to ensure multiple cloud failover points.


 

 

 

Who are you already partnering with on this?

We are working closely with the team at KCH NHS to enable their provision of service at scale.

 
 
 
edited on Apr 14, 2020 by Simon Bengen
You will need to login to post a comment

Manjinder Virk Apr 16, 2020

Really simple to use, has helped increase our throughput and handle the QC monitoring tremendously.

Reply 0

Ken Garner Apr 20, 2020

Hi Simon, I have recent experience of a similar solution to this but in a remote patient monitoring platform where real time data streaming in from wearables was risk stratified to trigger alerts and provide population level trends. (It successfully predicted a winter flu outbreak) I would like to learn more, could we chat sometime today or tomorrow?

Reply 0

Simon Bengen Apr 20, 2020

Hi Ken - sure happy to chat. Tomorrow would be better - how can I contact you?

Reply 0

Ken Garner Apr 20, 2020

Hi Simon, Could you ping me at ken@researchandinvest.com?
Tomorrow is a bit manic but if you could let me know what time(s) work for you I could re-schedule to suit your availability.
Cheers
Ken

Reply 0

Bev Matthews Apr 23, 2020

The idea has been progressed to the next milestone.

Reply 0

Bev Matthews Apr 24, 2020

Status label added: E

Reply 0

Bev Matthews Jun 15, 2020

Hello Simon, thank you for your submission. Your solution has now been forwarded to the relevant team within the testing programme. They will be in touch with you directly if they require any further information. Thanks Bev

Reply 0

Bev Matthews Jun 15, 2020

The idea has been progressed to the next milestone.

Reply 0

Share