To develop a healthcare chatbot service (AI-guided bot) that conducts real-time conversations using large language models to provide accurate health information to patients. To provide accurate and specialized medical responses, we integrated several cancer practice guidelines. The size of the integrated meta-dataset was 1.17 million tokens. The integrated and classified metadata were extracted, transformed into text, segmented to specific character lengths, and vectorized using the embedding model. The AI-guide bot was implemented using Python 3.9. To enhance the scalability and incorporate the integrated dataset, we combined the AI-guide bot with OpenAI and the LangChain framework. To generate user-friendly conversations, a language model was developed based on Chat-Generative Pretrained Transformer (ChatGPT), an interactive conversational chatbot powered by GPT-3.5. The AI-guide bot was implemented using ChatGPT3.5 from Sep. 2023 to Jan. 2024. The AI-guide bot allowed users to select their desired cancer type and language for conversational interactions. The AI-guided bot was designed to expand its capabilities to encompass multiple major cancer types. The performance of the AI-guide bot responses was 90.98 ± 4.02 (obtained by summing up the Likert scores). The AI-guide bot can provide medical information quickly and accurately to patients with cancer who are concerned about their health.

Lee JW, et al. J Korean Med Sci. 2019 Dec;35(2):e15. https://doi.org/10.3346/jkms.2020.35.e15

Purpose: To investigate the association between statin use and stomach cancer incidence in individuals with hypercholesterolemia. Materials and methods: To examine the cumulative effect of statins, we defined a statin user as one who used statins during 2002–2003 at baseline. Statin users were further classified into high and low users according to the medication possession rate. Statin non-users consisted of participants who had never used statins during the entire period of 2002–2015, despite having hypercholesterolemia (total cholesterol level ≥250 mg/dL at baseline). Ultimately, 17,737 statin users and 13,412 statin non-users were used in the analysis. We performed survival analyses, considering the diagnosis of stomach cancer as an event of interest. Results: Median follow-up duration was 12.9 years. The cumulative incidence rates of stomach cancer were lowest in high users (1.90% in men and 0.98% in women). Compared to non-users, hazard ratios (95% confidential intervals) for stomach cancer of low users and high users were 0.953 (0.755–1.203) and 0.526 (0.399–0.693) in men and 0.629 (0.457–0.865) and 0.370 (0.256–0.535) in women, respectively, after adjusting for possible confounders. Conclusions: We observed an inverse association between statin use and stomach cancer incidence in participants with hypercholesterolemia.

Aging has become a global problem, and the interest in healthy aging is growing. Healthy aging involves a focus on the maintenance of the function and well-being of elderly adults, rather than a specific disease. Thus, the management of frailty, which is an accumulated decline in function, is important for healthy aging. The adaptation method was used to develop clinical practice guidelines on frailty management that are applicable in primary care settings. The guidelines were developed in three phases: preparation (organization of committees and establishment of the scope of development), literature screening and evaluation (selection of the clinical practice guidelines to be adapted and evaluation of the guidelines using the Korean Appraisal of Guidelines for Research and Evaluation II tool), and confirmation of recommendations (three rounds of Delphi consensus and internal and external reviews). A total of 16 recommendations (five recommendations for diagnosis and assessment, 11 recommendations for intervention of frailty) were made through the guideline development process. These clinical practice guidelines provide overall guidance on the identification, evaluation, intervention, and monitoring of frailty, making them applicable in primary care settings. As aging and “healthy aging” become more and more important, these guidelines are also expected to increase in clinical usefulness.

OBJECTIVES: Opioids are prescribed to treat moderate to severe pain. We investigated recent trends in opioid (morphine, oxycodone, fentanyl, and hydromorphone) prescriptions using data from the Korean National Health Insurance Service-National Sample Cohort between 2002 and 2015.METHODS: The morphine milligram equivalent (MME) was calculated to standardize the relative potency of opioids. The number (cases) or amount (MME) of annual opioid prescriptions per 10,000 registrants was computed to analyze trends in opioid prescriptions after age standardization. Joinpoint regression analysis was conducted to calculate the annual percentage change and average annual percentage change (AAPC).RESULTS: The number (cases) of prescriptions per 10,000 registrants increased from 0.07 in 2002 to 41.23 in 2015 (AAPC, 76.0%; 95% confidence interval [CI], 61.6 to 91.7). The MME per 10,000 registrants increased from 15.06 in 2002 to 40,727.80 in 2015 (AAPC, 103.0%; 95% CI, 78.2 to 131.3). The highest AAPC of prescriptions and MME per 10,000 registrants were observed in the elderly (60-69 years) and in patients treated at general hospitals. Fentanyl prescriptions increased most rapidly among the 4 opioids.CONCLUSIONS: Consumption of opioids greatly increased in Korea over the 14-year study period.