A pilot study on the capability of artificial intelligence in preparation of patients' educational materials for Indian public health issues.
Autor: | Mondal H; Department of Physiology, All India Institute of Medical Sciences, Deoghar, Jharkhand, India., Panigrahi M; Department of Pharmacology, Bhima Bhoi Medical College and Hospital, Balangir, Odisha, India., Mishra B; Department of Physiology, Sri Jagannath Medical College and Hospital, Puri, Odisha, India., Behera JK; Department of Physiology, Nagaland Institute of Medical Science and Research, Nagaland, India., Mondal S; Department of Physiology, Raiganj Government Medical College and Hospital, West Bengal, India. |
---|---|
Jazyk: | angličtina |
Zdroj: | Journal of family medicine and primary care [J Family Med Prim Care] 2023 Aug; Vol. 12 (8), pp. 1659-1662. Date of Electronic Publication: 2023 Aug 29. |
DOI: | 10.4103/jfmpc.jfmpc_262_23 |
Abstrakt: | Background: Patient education is an essential component of improving public health as it empowers individuals with the knowledge and skills necessary for making informed decisions about their health and well-being. Primary care physicians play a crucial role in patients' education as they are the first contact between the patients and the healthcare system. However, they may not get adequate time to prepare educational material for their patients. An artificial intelligence-based writer like ChatGPT can help write the material for physicians. Aim: This study aimed to ascertain the capability of ChatGPT for generating patients' educational materials for common public health issues in India. Materials and Methods: This observational study was conducted on the internet using the free research version of ChatGPT, a conversational artificial intelligence that can generate human-like text output. We conversed with the program with the question - "prepare a patients' education material for X in India." In the X, we used the following words or phrases - "air pollution," "malnutrition," "maternal and child health," "mental health," "noncommunicable diseases," "road traffic accidents," "tuberculosis," and "water-borne diseases." The textual response in the conversation was collected and stored for further analysis. The text was analyzed for readability, grammatical errors, and text similarity. Result: We generated a total of eight educational documents with a median of 26 (Q1-Q3: 21.5-34) sentences with a median of 349 (Q1-Q3: 329-450.5) words. The median Flesch Reading Ease Score was 48.2 (Q1-Q3: 39-50.65). It indicates that the text can be understood by a college student. The text was grammatically correct with very few (seven errors in 3415 words) errors. The text was very clear in the majority (8 out of 9) of documents with a median score of 85 (Q1-Q3: 82.5-85) in 100. The overall text similarity index was 18% (Q1-Q3: 7.5-26). Conclusion: The research version of the ChatGPT (January 30, 2023 version) is capable of generating patients' educational materials for common public health issues in India with a difficulty level ideal for college students with high grammatical accuracy. However, the text similarity should be checked before using it. Primary care physicians can take the help of ChatGPT for generating text for materials used for patients' education. Competing Interests: There are no conflicts of interest. (Copyright: © 2023 Journal of Family Medicine and Primary Care.) |
Databáze: | MEDLINE |
Externí odkaz: |