A study showing the usefulness of artificial intelligence in the field of healthcare

Almost every day, new predictions appear about the impact of artificial intelligence in all sectors in the future, and some of these predictions are negative, such as: eliminating millions of jobs, and some are positive, such as: helping us discover new facts about the universe.

Many predictions are based on non-existent types of AI technologies that are still in development, but recent studies have found that AI technologies currently in use in the healthcare sector can save billions of dollars annually.

A recent study conducted by Harvard University and management consulting firm McKinsey revealed that currently available artificial intelligence technologies could save up to $360 billion in health care in the United States annually, or 10% of annual spending on healthcare.

The study showed that cost-saving AI technologies in the healthcare sector can be used for administrative tasks such as: optimizing staffing levels as per patient demand, optimizing operating room schedules so that more surgeries can be performed, or managing patient data effectively.

It can also be used in clinical healthcare tasks, such as: predicting recurrence of a patient with the same disease, diagnosing radiographs, and rapidly monitoring epidemic diseases, in addition to promoting medical research and developing new drugs.

The study indicated that cost savings will occur without sacrificing quality and accessibility, as these opportunities can also lead to non-financial benefits such as: improved quality of healthcare, increased access, better patient experience, and greater satisfaction for clinicians.

The study confirmed that saving these costs can be achieved within the next five years, as expanding the use of artificial intelligence remains a challenge due to problems related to gaining confidence and implementation time.

Nikhil Sahni, partner at McKinsey and a participant in the study, said: “Everything we are talking about here; It is technologies that already exist today and are used in case studies that we have seen in our experience, we are not talking about fancy new use cases that no one has ever seen implement.”

Numerous uses of generative AI are now emerging in the healthcare sector, especially Language Large Models (LLMs). For example: Carbon Health is working on integrating an interactive chatbot based on the GPT-4 language model - the same one used In (ChatGPT) - in their EHR (Electronic Health Records) system. During the appointment; the bot will record the conversation between the patient and the provider, and create a summary of the visit that includes the patient's medical history, diagnostic codes, and any other notes the provider adds.

Statistics show that a robot like this can, on average, fill out a medical chart in less than four minutes compared to 16 minutes for a manual chart, and this technology can also give doctors more time to treat patients, as service providers usually spend about 34 minutes with each patient, but About half of the appointment is devoted to administrative work, such as: planning.

in addition to; Researchers at New York University Grossman School of Medicine have developed a large language model for analyzing electronic health records (NYUTron) that can do multiple tasks, including estimating the length of a patient's stay in the hospital, the risk of death, the probability of a patient's re-infection, among other factors.

The researchers found that the NYUTron model was able to predict about 80 percent of predictions that a patient would develop the same disease again within a month.

Thus, it is expected that artificial intelligence will make its way into every aspect of health care, but these technologies cannot completely replace service providers. Rather, they are auxiliary tools that save a lot of money and effort and improve the quality of service provided.