AI assistance is heading your way: ChatGPT not half bad at describing aspects of primary care

No doubt many of you will have heard of OpenAI’s chatbot ChatGPT, which provides very human-sounding responses to questions and has been used for writing songs, essays, position statements and all kinds of other tasks.

The AI’s responses seem so real and lifelike because the models behind it were trained on vast amounts of data from the internet written by humans, including conversations.

AI as a concept has been around in the medical sphere for a while, radiology being one area where it is most advanced.

In radiology, the systems can be trained on a huge number of images with the corresponding reports and codes of those reports. It’s similar to pathology but, because radiology films have an orientation whereas pathology slides don’t, it’s much easier for the AI to recognise patterns in radiology.

AI has also been used a lot for pattern recognition in skin diagnosis, particularly melanoma.

Asked, “What are the implications of AI in healthcare?”, ChatGPT responded:

“AI has the potential to revolutionise healthcare by improving diagnostic accuracy, reducing medical errors, and increasing efficiency in various medical tasks. Some specific applications of AI in healthcare include:

• Medical imaging: AI-assisted diagnostic tools can analyse medical images and identify potential abnormalities or diseases with a high degree of accuracy.
• Electronic health records: AI-powered systems can help organise and analyse large amounts of patient data, making it easier for doctors to access and use this information.
• Personalised medicine: AI can be used to analyse a patient’s genetic information and other medical data to develop personalised treatment plans.
• Drug discovery: AI can help identify new potential drug targets and aid in the development of new drugs.
• Clinical decision support: AI-powered systems can help doctors make better and more informed decisions by analysing patient data and providing relevant information.”

ChatGPT is quite happy to respond to follow-up questions. “What about AI in primary care?”

It says: “AI has the potential to improve primary care by providing doctors with more accurate and timely information, reducing administrative tasks and allowing doctors to spend more time with patients. Some specific ways that AI is being used or could be used in primary care include:

• Virtual assistants: AI-powered virtual assistants can help patients schedule appointments, answer questions, and provide basic health information.
• Electronic health records: (Same answer as before.)
• Diagnostic assistance: AI can help doctors make diagnoses by analysing patient symptoms, medical history, and test results.
• Clinical decision support: (Same answer as before.)
• Chronic disease management: AI can help identify patients at risk for chronic diseases and provide personalised care plans to help prevent or manage these conditions.
• Triage: AI-powered triage systems can help identify patients with urgent needs and direct them to the appropriate level of care.”

There are concerns about the ethical implications, such as the potential for bias in algorithms and the risk of data breaches

AI is not meant to replace human doctors, but rather to assist them in their work and improve the overall efficiency and quality of care.

However, it is important to note there are concerns about the ethical implications of using AI in healthcare, including primary care, such as the potential for bias in algorithms and the risk of data breaches.

It’s important for healthcare providers and researchers to be transparent about their methods and results, and to be aware of these ethical issues when designing and implementing AI-based systems.

That being said, there seem to be some uses for this technology coming our way, and certainly some PHOs have used AI to analyse the databases they hold.

In general practice, very few datasets contain conversations held and subsequent actions, such as a prescription or a measurement.

These conversations could be recorded and the inputs we put into our practice management symptoms could be correlated with conversations. It would then be fascinating to try to do that and then get data to look out outcomes. This is all possible.

How about we ask the AI a clinical question. I hope it works better than older symptom-checkers that were still playing 20 questions after being told I had a fever, was vomiting, had a purple rash and was slipping in and out of consciousness!

Here we go…

“I have a cough and recurrent wheeze and get a bit tight in my chest. What should I do?”

Answer: “If you have a persistent cough and recurrent wheezing, as well as tightness in your chest, it is important to see a healthcare provider for an evaluation. These symptoms could be caused by a variety of conditions, such as asthma, chronic obstructive pulmonary disease (COPD), or an infection, and an accurate diagnosis is needed in order to provide appropriate treatment. Your healthcare provider may recommend medications, lifestyle changes, or other treatments to help manage your symptoms and improve your overall lung function. Additionally, it’s recommended to avoid smoking, exposure to pollution and allergens as much as you can.”

That seems sensible advice.

OK, let’s give it a health management question: “I have been diagnosed with asthma. I’ve been getting a bit wheezy at night, what should I do?”

Answer: “If you have been diagnosed with asthma and are experiencing wheezing at night, it’s important to follow your asthma action plan that was provided to you by your healthcare provider.

“This plan will outline the steps you should take to manage your asthma symptoms, including what medications to take and when...

“If you’re experiencing symptoms of asthma at night, you should take your ‘quick-relief’ or ‘rescue’ inhaler (such as albuterol) as directed by your healthcare provider. This medication can help to open up your airways and relieve wheezing and shortness of breath.

“Additionally, it’s recommended to avoid things that triggers asthma such as dust, smoke, pollution, pets, etc. If you’re not sure what triggers your asthma, you might want to discuss with your doctor.

“If your symptoms persist or worsen despite following your asthma action plan, it’s important to contact your healthcare provider for further evaluation. They may need to adjust your treatment plan or refer you to a specialist.”

We GPs, of course, offer a lot more – the human touch, the relationship, the trust. There are all kinds of limitations with AI, but these systems will get better. Combine them with spoken language input and output, and perhaps a visual avatar, and we could see some help coming for overworked GPs!

Richard Medlicott is a specialist GP based in Wellington