Artificial Intelligence (AI) has come to stay, and it already plays a role in many areas of our lives. How will Artificial Intelligence affect medicine?
What most users are familiar with at present, are primarily AI tools based on LLM (large language models). These tools are already being used in medicine, especially for learning purposes and in ways that improve daily management.
The role of Artificial Intelligence in medicine
For purposes of learning, there are websites and applications based on AI that make it easier to search for information, and to find results that more accurately and precisely answer our questions. In addition, AI makes it possible to glean information from multiple sources at the same time, so that the doctor does not need to spend time reviewing individual websites or journals. A multitude of scientific journals can be consulted in a single search.
These generative models of AI are increasingly better trained to answer questions about complex clinical cases. This can give us insight regarding surgical approach options, or surgical techniques, and can help us decide if a pathology does, in fact, require surgery or if it would be better to continue with conservative treatment.
It is true that AI tools have not yet reached the level of an experienced specialist when it comes to deciding which surgical technique is best for a specific case. Most likely, this is because there may not be consensus on the point in the current medical literature. Medical professionals do not agree always agree with each other, and many times at congresses, we argue for or against a certain option. We should not forget that, in this type of decision, the surgeon’s own experience and expertise with a specific technique is fundamental, something that these AI models are currently not able to reflect or interpret.
How AI can be used by doctors and patients
Patients can search for information in AI applications, but may still find the answers confusing. First, because although they can ask the application to answer the question in simple language, the complexity of the subject matter, and the number of factors that influence a decision, make it difficult to trust that a decision as important as whether or not we should have surgery, or what technique should be used, should be delegated to an AI model.
The role of AI in diagnostics
AI is making great strides in helping with diagnosis. The results of various tests have been published showing that AI is able to interpret hundreds or thousands of X-rays with excellent results and very little error, and of course, in much less time than it would take a single human. In the same way, we can expect AI to help doctors with anatomical pathology diagnosis, making it safer and more efficient.
In daily practice, we can use AI to help to answer patients’ questions or to research answers to clinical challenges. On more than one occasion, when a patient who was taking ten different medications and experiencing side effects, I have used AI to clarify which of the medications was the likely cause. And in complex clinical cases, I have also used AI by entering a list of the various disparate symptoms that a patient was suffering to find out what possible diagnosis it could suggest based on the information provided.
Analysis of clinical data
Thanks to the advantage AI offers for analyzing large amounts of data, there are now monitoring and alert systems for patients who wear devices ("wearables") that send clinical information to hospitals and doctors. If we have a tool capable of efficiently analyzing a large amount of data, then it makes sense to record a large amount of data.
Other benefits of AI that we are already beginning to see, are in the area of management. Doctors spend countless hours writing reports, ordering tests, and writing down what patients tell to us. Time spent on these tasks has taken away time that could have been dedicated to the more human side of healthcare and a higher quality of care for our patients. The time consumed in these tasks is also cited as one of the main causes of burnout or exhaustion, and leaving the profession. We believe that AI-based applications will keep getting better, and will rapidly become more widely used, so that doctors will have more time, and be able to spend better quality time with our patients.
In addition, thanks to automated data collection, we can look forward to the information in a patient’s medical record being better and more accurate. And in the not too distant future, AI should be able to propose treatment options and answers to patients’ questions based on all the published literature, and based on evidence. No doubt soon, we can expect to find proposals for more personalized treatments. Likewise, we will be able to propose improved designs for implants, to better fit the anatomy and biomechanics of our patients. And in the operating room, we will be able to plan trajectories for the placement of the implants with the utmost precision, as we already do at Instituto Clavel with the Mazor X robotic arm.
The human factor of a specialist physician cannot be replaced
At least for now, and for the near future, the proposals, answers, analyses, and summaries provided by AI must be still be validated by a specialist medical professional. It is important to remember that, in medicine, there is no room for error and our duty is to propose the safest treatment. We know that AI models also make mistakes, and the information they provide may include biases that do not necessarily benefit the patient.
And finally, there are still issues to resolve regarding confidentiality and data protection. We are dealing with sensitive information, everyone’s medical information, and it must be treated with the utmost security.
Dr. Pablo Clavel
Neurosurgeon and Director of Instituto Clavel
Categories: Innovation