At the heart of AI’s application in healthcare lies massive volumes of data. For all the hype around AI and Gen AI, the technology is only as good as the data available.

Having access to data sources, the integration of the data (both structured and unstructured) and data governance are all critical factors in ensuring AI can be impactful. This allows us to scale these solutions to truly make a positive impact on care.

For example, many healthcare organisations are currently implementing electronic patient record systems (ERP) as part of large-scale digitisation programs. This rich source of organisational-specific data, combined with data collated within an integrated healthcare system, means we will have access to a wider spectrum of data that moves away from organisational need to patient need.

The question is, will we know what to do with it? Is it of good quality? Do we have enough data scientists to analyse it? 

There are many use cases where AI can be beneficial to healthcare. But before we invest limited funds and resources into another large-scale program, proof of concepts and studies are needed.

And the good news is DXC Technology is working with several healthcare organisations to do exactly that. 

For example, working with Singapore General Hospital (SGH) and the national health agency Synapxe, DXC developed an AI system to tackle pneumonia and combat growing antibiotic resistance. The AI model, which processes symptom data from patients, helps distinguish between viral and bacterial infections, ensuring antibiotics are only prescribed when necessary.

Technology’s ability to sift through vast data to detect infection patterns is key in deciding whether antibiotics are required.

The AI system, called Augmented Intelligence in Infectious Diseases, was trained on data including clinical symptoms, responses to infection, X-rays and vital signs from about 8,000 SGH patients.

Further studies will assess the AI system's impact on antibiotic use in clinical settings, with future expansion plans to include other common infections.

Another example is a PoC that my colleagues conducted when they teamed up with a large medical group in California to look at the early detection and improved tracking of breast cancer.

This proof of concept required a limited amount of funding and a much lower utilisation of resources, leading to a quicker and cheaper study. However, its significance was its use of historical data pulled from the EPR system to assist in the early detection and diagnosis of breast cancer.

By using historical data, the PoC’s aim was to determine if AI could assist in the early detection and diagnosis of breast cancer to significantly improve outcomes for patients, by successfully identifying false positives and false negatives in breast cancer radiology results. 

Though AI was born nearly seven decades ago, its partnership with healthcare is just beginning—and collaboration and responsible implementation will shape the outcome.   

Remember, comprehensive patient data is the lifeblood of AI. That's why healthcare organisations must insist that they have access to patient data, or they risk missing out on the benefits AI can bring to clinicians, caregivers and the patients that they serve.  

• Learn more about healthcare and life sciences 
• Life sciences companies can speed regulatory approval with GenAI
• Optimise clinical service desks to enhance physician support 
  
• Barts Health uses robotics to transform HR operations