A new study suggests that ChatGPT, an artificial intelligence (AI) chatbot, may be of great help to patients with cirrhosis or hepatocellular carcinoma (HCC) and their clinicians. It provides easily understandable data about the disease which can be used for various purposes. ChatGPT can supply precise, reproducible answers to commonly asked questions from patients concerning cirrhosis and hepatocellular carcinoma. Even though the majority of responses offered by ChatGPT are considered accurate by clinic specialists, most of them also termed it as “correct but inadequate”.
Apart from just providing the required data, ChatGPT also offers empathetic and practical advice to both patients and caregivers. However, its utility clearly reduces in the absence of tailored recommendations, something that the researchers unearthed while studying it. According to Brennan Spiegel, MD, with Cedars-Sinai, Los Angeles, California, ChatGPT, while it has limited capabilities, can help empower patients and improve healthcare literacy for different populations.
ChatGPT was developed by OpenAI as a natural language processing tool which makes it possible for people to talk to an AI bot and receive detailed answers when posed with a question. Even though the technology is being utilized in the field of medicine, the research by Cedars-Sinai to study and compare ChatGPT’s performance with that of physicians is among the firsts of its kind.
During the study, two liver transplant specialists independently graded the responses given by ChatGPT to 164 questions relevant to cirrhosis and HCC, slotted under five different categories — basic knowledge, diagnosis, treatment, lifestyle, and preventive medicine. ChatGPT answered 77% of the questions correctly, with 91 of them being highly accurate according to the specialists. Interestingly, ChatGPT was found to generate the most accurate responses Basic Knowledge, Lifestyle and Treatment and less so in the fields of Diagnosis and Preventive Medicine.
The researchers also analysed the chatbot’s responses against two published questionnaires for cirrhosis quality measures recommended by the American Association for the Study of Liver Diseases. For these questions also, ChatGPT offered correct responses 77% of the times. However, it failed to specify the decision-making cutoffs and treatment durations. With respect to regional guidelines such as HCC screening criteria, ChatGPT is still behind the curve.
OpenAI’s ChatGPT then proved itself suitable for providing “practical and multifaceted” advice to both patients and caregivers. It can also help clinicians by offering a framework to draft tailored questions. With the high proportion of confirmed accurate responses and development opportunities in the future, physicians can take help of ChatGPT and save time drafting patient responses. Not to forget, ChatGPT also helps in empowering the patients by providing them with reliable, accurate health information that is easy to comprehend.
OpenAI is a leading artificial intelligence company that has been in operation since 2015. Their machine learning methods, which focus on AI projects, make use of the latest technologies to enable machines to think. Human-level performance on certain tasks can be achieved with their products, allowing new possibilities for businesses and research institutes.
Brennan Spiegel, MD is a renowned clinician from Cedars-Sinai, Los Angeles, California. He is the lead study investigator in the ChatGPT project. His work majorly focuses on empowering patients and improving healthcare literacy for people from challenged backgrounds, those suffering from cirrhosis, and liver cancer. He believes that OpenAI’s chatbot can help facilitate a more efficient physician workflow and with minimal revisions, can help them address patient queries in a better way.
