Health Tech Awards Finalists 2021: Best Use of AI and Automation Tools 

A popular category, that grows each year, here we take a look at the finalists in the Best Use of AI and Automation Tools category.

We explore the programmes at Portsmouth Hospitals University NHS Trust, OncoHost, Cognetivity, Owkin, Sensyne Health, Al Rehab Ltd, UCLH and Diaceutics.

Delve into the finalists here – which would you pick as your winner?


This entry focuses on the in-house development of a ‘Hospital Oxygen Usage Monitoring Application’ to monitor the flow rate through the hospital’s vacuum insulated evaporator (main oxygen tank) to provide an overview of daily usage, as COVID-19 case increased and oxygen use increased.

The trust said: “As the number of patients requiring oxygen due to coronavirus increased, so did the demand of the flow through the VIE. Monitoring this flow is difficult and can only be done at the VIE itself if the correct monitoring equipment has been installed. Due to the size of our hospital and our multi-site operation and complex VIE install we devised a different solution.”

The team created  a web interface to monitor the amount of liquid oxygen in the VIE at certain time intervals, so they knew when it needed refilling.

The automated solution uses open-source tools and their methodology and code is available on GitHub. The team noted the process:

  • Robotic process automation is used via an open-source project TagUI to go to the website and grab the VIE fill data every day.
  • These data are then manipulated in an app to calculate flow rate and compare it to the previous days.
  • An automated email is sent then sent out daily to the estates team, reporting team and clinical managers so there is an awareness of the current usage compared to threshold.
  • The results are then graphically displayed in the app (

The app has been running for 15 months as an entirely automated solution and has allowed the trust to confidently manage hospitals oxygen supplies throughout both waves of the pandemic. The code can be found at

Portsmouth Hospitals University NHS Trust 

This entry focus on automating ambulance handovers, by using automation to reduce the time spent processing information.

The trust said: “Handovers at Portsmouth Hospitals University NHS Trust have been transformed through the introduction of ‘Bob’, an Intelligent Automation digital worker that automates transfer of data from South Coast Ambulance Services software into the Emergency Department system.

Members of the IT team spent time shadowing the ED administration department and began mapping the patient record workflow and pathways from ambulance to triage. Through this mapping exercise the concept of a digital worker (‘Bob’) was employed to assist with the reduction of manual patient documentation uploads.

The entry states: “With 137 patients on average arriving at the hospital by ambulance, patient information transfers between SCAS and ED can take up to three minutes per patient. Further time is then accumulated when ambulance holds occur, leaving ED staff having to track document uploads in increased periods of demand. Driving the reduction of manual data transference between the two systems, allows staff to prioritise patient care – a clear priority in this implementation.”

‘Bob’ was piloted during select periods of 2020 and 2021, before being activated 24/7. Since going live with ‘Bob’ in April 2021, the ED team has:

  • Uploaded 115 documents on average a day, resulting in IA performing 82% of uploaded patient information documentation
  • Supported the increase in information being uploaded before the doctor has seen the patient, from 69.3% to 77.6%. This in turn has led to increased patient safety
  • Saved an average of 1 minute by automating the uploading of documentation. Releasing on average 2.3 hours back to ED admin staff per day, improving patient flow

Phillip Kenney, Deputy CIO, commented: “There are so many opportunities to improve the workforce efficiencies through the use of IA. Improving care we deliver, improving moral of our colleagues who are having to complete manual repetitive tasks due to silos of information, we have a really exciting programme ahead of us.”


The next entry in this category focuses on the use of AI to understand patients’ unique response to treatment and overcome one of the major obstacles in clinical oncology today – resistance to therapy.

Following over a decade of academic research on the body’s biological response to anti-cancer treatment led by Yuval Shaked, Ph.D., a professor at Israel’s Technion Institute of Technology, OncoHost was founded in 2017.

OncoHost developed PROphet®, a precision oncology diagnostic platform that combines advanced AI with proprietary proteomic analysis to analyse a broad range of select proteins in a series of patient blood samples. The sophisticated technology uses machine learning algorithms to characterise, analyse and anticipate a patient’s response to treatment through a simple blood test.

To date, the system is showing very high accuracy ( in predicting how melanoma and non-small cell lung cancer (NSCLC) patients respond to various therapies. Currently, clinical trials are taking place in the United States, the UK and Israel at multiple research sites, including Yale Cancer Center, Fox Chase Cancer Center, Rutgers Cancer Institute of New Jersey, Jefferson Health and Hadassah Medical Center (Israel), and, most recently, the UK’s National Health Services (NHS) (


Next up, is Cognetivity, an integrated cognitive assessment platform enabled by InterSystems IRIS for Health.

The company is reinventing early dementia detection though its scalable ‘Integrated Cognitive Assessment’ test (ICA) that uses advances in neuroscience, AI and the power of InterSytems IRIS for Health cloud data management platform, and is now set for its commercial roll-out.

The ICA test is based on humans’ strong reaction to animal stimuli, and the ability of a healthy brain to process images of animals in less than 200ms. ICA is a rapid animal/non-animal visual categorisation test, engaging brain areas affected in pre-symptomatic stages of Alzheimer’s and detecting subtle impairments in information processing speed, aiming to detect the earliest signs of disease before the onset of memory loss symptoms.

The clinical effectiveness of the ICA platform is significantly enhanced by the data management and integration capabilities of InterSystems’ highly advanced IRIS for Health data management platform.

Cognetivity’s ICA platform is approved by the UK’s Medicines and Healthcare Products Regulatory Agency and has been deployed with the specialist NHS mental health organisation North Staffordshire Combined Healthcare NHS Trust since September 2020. It is also in use in primary care in Sunderland.

The entry said: “The ICA is a major advance in early dementia detection, helping society deal with one of the most significant healthcare challenges it will face for decades to come. Its simplicity removes language, cultural or educational barriers and bias, whilst also removing the ‘learning effect’ – i.e. the difficulty in ascertaining whether patients’ health is improving, or whether they are ‘learning’ the elements of the test. This enables the solution to achieve adoption almost anywhere in the world, making detection faster, cheaper and more reliable.”

Al Rehab Ltd.

The next entry in this category focuses on a scalable and cost-effective solution to unmet rehabilitation needs.

The entry notes there are 255,000 falls related emergency admission per annum, and a workforce short of 4,000 physiotherapists.

To address this, AI Rehab has developed Slider, a solution combining artificial intelligence with a suite of smart products. The products included an app, with augmented reality gamification, and a remote monitoring wearable medical device. Data obtained from the testing of Slider at the UK National Physical Laboratory and publicly available data from the US National Institute of Health Osteoarthritis Initiative is being used to train the deep learning algorithm.

Slider is designed to encourage and motivate patients to do pre and postop hip and knee exercises and to measure compliance and progress. The app incorporates alerts for remote notification to the patients rehab team when the patient’s progress deviates from the pathway and PROMS to standardise outcome measures and monitor physiotherapy results. The gamification is designed to improve patient compliance. The AI detects outliers and patterns. Slider builds a unique patient profile by combining information about the patient, severity of arthritis, surgery data (if applicable) and movement data. It uses the profile to optimise the exercise plan, timing of surgery and discharge destination, and it refines the profile iteratively.

The company highlights the benefits of the system:

  • Supports the optimisation of personalised and risk stratified rehab programmes
  • Addresses unmet need resulting from physiotherapy shortages
  • Puts power in the hands of the patient and improves compliance with evidence-based physiotherapy programmes
  • Setting agnostic, providing continuity and consistency throughout the patient journey
  • Standardises the quality of rehab service provision across the NHS
  • Provides surgeons, GPs and physiotherapists with objective measures of progress and early warning of deterioration, so they can intervene early and prevent further deterioration or falls
  • Useful research tool, providing objective measures of the impact of changes in practice, new technologies and/or new drugs
  • Secure cloud capability enables scalability

AI Rehab has worked with Imperial College to develop the prototype and Lancaster University and the University of Central Lancashire (UCLan) to test and bring the solution to market. Slider has been evaluated at hospitals across the UK. AI Rehab has access to NHS provider’s and commissioner’s through UCLan’s Health MATTERS programme and the Idea to Scale-Up Programme. AI Rehab is also working with a world leading market access and health economic agency, Genesyze Ltd., on a global market entry plan and a blueprint for rapid adoption.


Owkin has developed a deep learning model to predict RNA-Seq expression of tumours from whole slide images.

The new genomic analysis tool (HE2RNA) predicts gene expression from histology whole-slide images. It facilitates patient diagnosis and improves the prediction of response to treatment and survival outcomes. The omani said “it is interpretable by nature and can help pathologists predict genes involved in cancer development, also tumor status and response to therapies.”

The entry states: “HE2RNA, is a deep-learning algorithm specifically customised for the prediction of gene expression from WSI. For training our model, we collected WSIs and their corresponding RNA-Seq data from The Cancer Genome Atlas (TCGA) public database. We then investigated how HE2RNA could be used to generate heatmaps for a spatial visualisation of any gene expression. Finally, we show how the internal representation (transcriptomic representation) learned by the model can improve the prediction of a specific molecular phenotype such as microsatellite instability.

“HE2RNA robustly and consistently predicted subsets of genes expressed in different cancer types, including genes involved in immune cell activation status and immune cell signaling. We hypothesise that the algorithm can recognise immune cells and correlate their presence with the expression of a subset of protein-coding genes (such as CQ1B). Major breakthroughs in cancer therapy are driven by discoveries of treatment targets for immunotherapy in many types of cancer. HE2RNA could be a useful tool for pathologists and oncologists to consistently quantify immune infiltration or guide treatment decision in the context of immunotherapy.”

Sensyne Health and Chelsea and Westminster Hospital NHS Foundation Trust

Sensyne Health utilised de-identified and anonymised patient datasets to support the clinical decision-making process for COVID-19 patients with the SYNE-COV prediction algorithm for Chelsea and Westminster Hospital NHS Foundation Trust.

The tool, a COVID-19 patient outcome prediction algorithm, in collaboration with Chelsea and Westminster Hospital NHS Foundation Trust, helps clinicians proactively manage patients by utilising machine learning and artificial intelligence to predict patient outcomes and inform near real time clinical decision making.

The entry cited “the volume of data points generated for patients in hospital has increased exponentially, and the breadth increases as the disease progresses. For example, one patient in ICU is expected to produce 20,000 – 30,000 points of data per day, including vital signs, respiratory tests and laboratory results. During the pandemic, early intervention was essential, not only to understand the likely impact on intensive care resources but more importantly to save lives.”

The hospital and Sensyne Health worked together to develop the AI-based digital solution, that provides clinicians with an individual patient risk score to aid critical clinical assessment of three potential COVID outcomes when a patient enters hospital:

  • Admission to intensive care
  • Invasive mechanical ventilation
  • In-hospital mortality

Diaceutics – Diagnostic Deductive Pathways

Diaceutics’ entry into this award category focuses on the application of machine learning and AI to identify the best possible testing journey for patients using DDPs® (Diagnostic Deductive Pathways).

The company says that DXRX – The Diagnostic Network® – is the world’s first diagnostic commercialisation platform for precision medicine, integrating multiple pipelines of real-world diagnostic testing data from a global network of laboratories. Through machine learning and the standardisation of millions of aggregated de-identified patient testing events, Diaceutics says it can identify the best possible testing journey or “Deductive Diagnostic Pathway (DDP®)” for patients at disease level, providing the industry with a guide to ‘getting the right medicines to the right patients, faster’.

Diaceutics launched DXRX – The Diagnostic Network® in 2020 to provide a ‘digital, scalable solution’ for precision medicine diagnostics in line with therapy launches, by integrating multiple pipelines of ‘real-world diagnostic testing data’ with a ‘vibrant diagnostics marketplace’. The company combines data from multiple sources including laboratory result data, diagnostic profiling meta data and CMS and commercial claims data into its DXRX platform, providing access to more than 365 million de-identified patient records globally, in 53 countries.

With 476 potential precision medicine therapies now in late phase of development and over 30 per cent of all FDA approvals between 2018-2020 being for Precision Medicines, the pipeline of therapies requiring a companion diagnostic solution is rapidly growing. Diaceutics has embarked on addressing this significant and pressing issue by building its data enabled platform, which has been designed to get ‘every patient the right treatment at the right time’ by identifying and solving hurdles in a therapy’s testing ecosystem ahead of therapy launch. To date, the company has leveraged deep disease level data analytics and implementation solutions to improve the diagnostic testing infrastructure for over 600 projects with 39 of the world’s leading pharmaceutical companies.

Diaceutics also says that the labeling of clinical data allows data-driven insights, providing a cleaner and easier dataset to work with, as well as allowing the creation of groupings and filters. By using a standardised six-step approach for better data, testing and treatment, it ensures that anyone working with the data has the same starting point with same patient cohort and methods. This means data can be analysed and aggregated at a high level quickly and efficiently, meaning ‘more patients get the right treatment at the right time enabled by DXRX’.