If data is the new oil, is ChatGPT the new refinery? Or is it the fire?

For data to be helpful in addressing healthcare’s many problems, then like oil, we need to refine it. Artificial Intelligence (AI) has potential in so many aspects of healthcare, from clinical care delivery to Research and Development, Public Health and policy development. While AI may not replace clinical decision-making, it could dramatically improve the quality of decision making as well as helping provide more empathetic care and patient engagement.

It is clear that for AI to advance and play a bigger role in healthcare will depend on numerous factors, most importantly, GE Healthcare’s recent study highlighted that the trust of providers, patients and health-care professionals in AI-based technologies is vital.

Whether we like it or not, generative AI is surely coming to healthcare, and it can serve as a powerful member of the healthcare team – If you’re ready to onboard it and understand its strengths and weaknesses.

Here are four things health providers can do to prepare their systems and care processes to derive the most benefit from LLMs while also avoiding the pitfalls:

1. Bring data together to enable rich insights

AI is only as good as the data it has access to. Every patient interaction and test result creates a new data point, meaning there is a mountain of data in health that, if harnessed correctly, could provide rich insights.

The first step in onboarding AI is to unify data and information collected across multiple systems. While that mountain may seem impossible, it isn’t.

While integrating and managing such a large volume of data can seem daunting, it’s entirely achievable. Health organisations should prioritise creating a strong, centralised data foundation where systems, applications, and services are interoperable and can share information securely and efficiently. This connected approach supports the safe and effective integration of AI and machine learning. It empowers clinicians and decision-makers with the insights they need to deliver better, more targeted care.

2. Be clear on your specific needs and the benefits you hope to gain.

While industries like financial services are ahead in adopting AI, it doesn’t mean solutions designed for their needs meet the specific needs of healthcare providers and systems.

Technology boffins aren’t immersed in our world and don’t innately understand how healthcare operates or the necessary guardrails to ensure patient safety and regulatory compliance. Missteps with this technology could compromise our high-trust healthcare relationships. To get solutions that will work for you and your patients, the tech must be designed with clear input from people having in-depth, first-hand knowledge and experience of the healthcare environment.

Starting with an intense focus on user-centred design, products such as our Amadeus Digital Care Record, enable organisations to gain the best benefit of software designed to not only satisfy but also to enhance their unique needs. It’s no surprise that in addition to meeting safety and regulatory requirements, AI for healthcare needs to be designed with that same intense end-user focus.

3. Facilitate patient autonomy, participation and empowerment.

Is your organisation ready to put patients in the driver’s seat of their healthcare? Do you have strategies in place to grow your capabilities around patient centered care? In today’s world, patient centric care is an absolute necessity and not a “bolt on” after the fact.

As an extended member of a patient’s care team, AI in partnership with other well established tools such as patient portals providing access to their complete care record, can significantly enhance patient autonomy. This, for example, by providing triage support, symptom checking, essential health advice and guidance.

Tools like Orion Health’s Virtuoso Digital Front Door solution enable patients to participate in most aspects of their healthcare and derive benefits as we introduce and expand the use of AI for better health outcomes.

4. Enhance data governance, security and privacy.

Data security is the foundation of trust between an organisation and its clients. Trust rings especially true for the health industry. Personal health information (PHI) is incredibly personal, at times highly sensitive, and the demands of regulators are rightfully so intense that data security and privacy are paramount.

Using de-identification tools, healthcare organisations can safely share anonymised patient data while remaining compliant with health privacy legislation including HIPAA and GDPR. This gives healthcare professionals enhanced opportunities to confidently advance healthcare outcomes while the patient’s privacy and trust remain intact.

AI, especially the LLM solutions such as ChatGPT are already impacting and driving change in the way healthcare operates. Wherever you are in the process of onboarding AI, having the right technology is critical to unlocking its potential and refining its use for your healthcare setting

MEET THE AUTHOR

Dr. Chris Hobson

Chief Medical Officer at Orion Health. Dr Hobson has 20 years’ experience as a physician working in primary and secondary care, including 10 years’ of experience in family medicine.

Connect with Dr Hobson:
Linkedin
Twitter