Carcinoma, Hepatocellular
Conditions
Keywords
ChatGPT, Patient Education
Brief summary
Liver cancer is a leading cause of cancer-related deaths in Taiwan, with its onset linked to factors like chronic liver conditions, cirrhosis, and genetic predispositions. According to the Barcelona Clinic Liver Cancer (BCLC) classification, early-stage liver cancer is demarcated by stages 0 to A. Upon such diagnosis, both patients and their families often have numerous questions and concerns, ranging from treatment choices to long-term outcomes. The research proposes a GPT-3.5-based chatbot to assist these patients by providing timely, personalized information, aiming to enrich their understanding of the disease and improve communication between patients and health professionals. The research methodology employs a Randomized Controlled Trial (RCT) design, dividing participants into a control cohort receiving standard patient education routine and an experimental cohort receiving both the AI chatbot and traditional education routine. The comparative analysis of these cohorts will determine the effectiveness of the AI intervention in improving patients' health literacy and satisfaction.
Detailed description
Liver cancer is the second most common cause of cancer-related deaths in Taiwan. Various factors play a role in its development, such as chronic liver conditions, cirrhosis, viral infections, alcohol intake, obesity, diabetes, and genetic predispositions, among others. Based on the Barcelona Clinic Liver Cancer (BCLC) system, early-stage liver cancer falls within stages 0 to A. When faced with an early-stage liver cancer diagnosis, patients and their relatives frequently express concerns. These may range from the potential effects of the disease on daily living, evaluating treatment options, potential side effects, costs involved, the chances of recurrence, and survival rates, to the care required after the treatment. Addressing these worries often requires extensive explanations and time for the patients to process the information. The research proposes using a chatbot built upon the GPT-3.5 language model developed by OpenAI for patient education services. Such a chatbot would aid early-stage liver cancer patients navigate the complexities of obtaining relevant information. As an artificial intelligence technology, the chatbot can offer timely, personalized information and psychological support. By responding to patients' inquiries, the chatbot can provide a thorough understanding of basic liver cancer knowledge, its causes, and treatment approaches, thereby facilitating a deeper comprehension of the early stages of liver cancer and its treatment regimen. Patients and their relatives can comprehend their condition and treatment plans, enhancing their conversations with medical staff and promoting a harmonious doctor-patient relationship. The research uses a Randomized Controlled Trial (RCT) methodology, dividing patients into a control group undergoing the conventional patient education routine, and an experimental group that leverages both the chatbot and traditional education. By comparing selected outcomes between the two groups, the experiment's effectiveness will be determined.
Interventions
BEHAVIORALChatGPT
Patients receive additional education using a GPT-3.5-based educational robot on top of the traditional education.
BEHAVIORALpatient education with traditional methods.
Patients receive standard traditional education procedures.
Sponsors
Taipei Veterans General Hospital, Taiwan
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
HEALTH_SERVICES_RESEARCH
Masking
SINGLE (Subject)
Masking description
The research uses a Randomized Controlled Trial (RCT) methodology, dividing patients into a control group undergoing the conventional patient education routine, and an experimental group that leverages both the chatbot and traditional education. By comparing selected outcomes between the two groups, the experiment's effectiveness will be determined.
Intervention model description
To compare the educational effectiveness of a chatbot integrated with health education information to traditional health education methods. This comparison encompassed aspects such as the patient's health literacy and clinical satisfaction. Based on the findings, recommendations and improvements would be proposed to promote the application and development of large language models in the medical field.
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Patients with early-stage hepatocellular carcinoma from both gastroenterology and general surgery outpatient departments were included. Early-stage hepatocellular carcinoma is defined based on the Barcelona Clinic Liver Cancer (BCLC) staging as stages 0 to A.
Exclusion criteria
* Patients under the age of 18 or those currently undergoing treatment for other cancers.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Health literacy score of patients | 1 weeks to 1 month | Primarily measured using the Liver Cancer Knowledge Scale. The scale consists of 20 questions with options including correct, incorrect, and unsure, with 14 correct answers and 6 incorrect ones (questions 4, 7, 11, 15, 18, 19). Each correct answer scores 5 points, while incorrect or unsure answers score 0 points. The score range is from 0 to 100, with a total score of 100 points. |
| Satisfaction score with medical care | 1 weeks to 1 month | It mainly includes satisfaction with traditional health education and AI-based health education tools. The Likert scale assessed the score, which offers options ranging from very dissatisfied (1) to very satisfied (5). |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Degree of patient anxiety | 1 weeks to 1 month | Measured using The GAD-7 questionnaire, a scale designed to assess anxiety levels. |
Countries
Taiwan
Outcome results
None listed