Potential safety concerns in use of the artificial intelligence chatbot ‘ChatGPT’ for perioperative patient communication

Tung, Joshua Yi Min; Lim, Daniel Yan Zheng; Sng, Gerald

doi:10.1111/bju.16042

Cited by 11 publications

(7 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We asked 60 questions across 6 topics to reduce stochasticity implications. 4,5,14 Our methodology only included 1 iteration per question and was completed on a single date to reduce confounding. We focused on UCF as it is utilized nationally, widely trusted, and includes several sexual health conditions.…”

Section: Discussionmentioning

confidence: 99%

“…1,2 As technology companies continue to innovate artificial intelligence (AI) applications, this trend will increase. 3-5 Modern large language models, namely OpenAI’s ChatGPT, allow free text entry and generate novel responses. 6 Of note, these responses are often long and convoluted.…”

mentioning

confidence: 99%

“…5,6,12-14 Particularly worrisome is AI hallucination, where misinformation may arise from subjective, emotion-laden questioning. 3,4…”

mentioning

confidence: 99%

“…Education regarding men’s health was traditionally obtained from provider-created materials, particularly the Urology Care Foundation™ (UCF), which is the primary patient-facing service of the AUA. 4,19 As AI becomes increasingly popular amongst patients, we must understand its strengths and limitations. Here, we aimed to evaluate the utility of ChatGPT in improving accessibility and quality of educational information within men’s health, comparing responses with gold-standard UCF resources.…”

mentioning

confidence: 99%

See 3 more Smart Citations

Comparison of ChatGPT and Traditional Patient Education Materials for Men’s Health

Shah,

Ghosh,

Hochberg

et al. 2024

Urology Practice

View full text Add to dashboard Cite

Introduction:ChatGPT is an artificial intelligence platform available to patients seeking medical advice. Traditionally, urology patients consulted official provider-created materials, particularly the Urology Care FoundationÔ (UCF). Today, men increasingly go online due to the rising costs of health care and the stigma surrounding sexual health. Online health information is largely inaccessible to laypersons as it exceeds the recommended American sixth to eighth grade reading level. We conducted a comparative assessment of patient education materials generated by ChatGPT vs UCF regarding men's health conditions.Methods: All 6 UCF men's health resources were identified. ChatGPT responses were generated using patient questions obtained from UCF. Adjusted ChatGPT responses were generated by prompting, "Explain it to me like I am in sixth grade." Textual analysis was performed using sentence, word, syllable, and complex word count. Six validated formulae were used for readability analysis. Two physicians independently scored responses for accuracy, comprehensiveness, and understandability. Statistical analysis involved Wilcoxon matched-pairs test.Results: ChatGPT responses were longer and more complex. Both UCF and ChatGPT failed official readability standards, although ChatGPT performed significantly worse across all 6 topics (all P < .001). Conversely, adjusted ChatGPT readability typically surpassed UCF, even meeting the recommended level for 2 topics. Qualitatively, UCF and ChatGPT had comparable accuracy, although ChatGPT had better comprehensiveness and worse understandability.Conclusions: When comparing readability, ChatGPT-generated education is less accessible than provider-written content, although neither meets the recommended level. Our analysis indicates that specific artificial intelligence prompts can simplify educational materials to meet national standards and accommodate individual literacy.

show abstract

Section: Discussionmentioning

confidence: 99%

mentioning

confidence: 99%

“…5,6,12-14 Particularly worrisome is AI hallucination, where misinformation may arise from subjective, emotion-laden questioning. 3,4…”

mentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Comparison of ChatGPT and Traditional Patient Education Materials for Men’s Health

Shah,

Ghosh,

Hochberg

et al. 2024

Urology Practice

View full text Add to dashboard Cite

show abstract

“…The interpretation of this standard in cases involving LLMs will determine potential physician liability, with increasing acceptance and FDA authorization potentially affecting the outcome. 50 Tung et al 51 recently conducted an assessment of ChatGPT's perioperative patient communication. Although the chatbot was able to respond with empathy, reassurance, and accuracy in discerning minor vs major postoperative complications, some of the information presented was incorrect but presented with the authority of factual truth.…”

Section: Medical Educationmentioning

confidence: 99%

Academic Surgery in the Era of Large Language Models

Rengers,

Thiels,

Salehinejad

2024

JAMA Surg

View full text Add to dashboard Cite

ImportanceThis review aims to assess the benefits and risks of implementing large language model (LLM) solutions in an academic surgical setting.ObservationsThe integration of LLMs and artificial intelligence (AI) into surgical practice has generated international attention with the emergence of OpenAI’s ChatGPT and Google’s Bard. From an administrative standpoint, LLMs have the potential to revolutionize academic practices by reducing administrative burdens and improving efficiency. LLMs have the potential to facilitate surgical research by increasing writing efficiency, building predictive models, and aiding in large dataset analysis. From a clinical standpoint, LLMs can enhance efficiency by triaging patient concerns and generating automated responses. However, challenges exist, such as the need for improved LLM generalization performance, validating content, and addressing ethical concerns. In addition, patient privacy, potential bias in training, and legal responsibility are important considerations that require attention. Research and precautionary measures are necessary to ensure safe and unbiased use of LLMs in surgery.Conclusions and RelevanceAlthough limitations exist, LLMs hold promise for enhancing surgical efficiency while still prioritizing patient care. The authors recommend that the academic surgical community further investigate the potential applications of LLMs while being cautious about potential harms.

show abstract

“Incorporating large language models into academic neurosurgery: embracing the new era”

Aamir,

Hafsa

2024

Neurosurg Rev

View full text Add to dashboard Cite

Potential safety concerns in use of the artificial intelligence chatbot ‘ChatGPT’ for perioperative patient communication

Cited by 11 publications

References 8 publications

Comparison of ChatGPT and Traditional Patient Education Materials for Men’s Health

Comparison of ChatGPT and Traditional Patient Education Materials for Men’s Health

Academic Surgery in the Era of Large Language Models

“Incorporating large language models into academic neurosurgery: embracing the new era”

Contact Info

Product

Resources

About