Generative artificial intelligence chatbots may provide appropriate informational responses to common vascular surgery questions by patients

OBJECTIVES

Generative artificial intelligence (AI) has emerged as a promising tool to engage with patients. The objective of this study was to assess the quality of AI responses to common patient questions regarding vascular surgery disease processes.

METHODS

OpenAI's ChatGPT-3.5 and Google Bard were queried with 24 mock patient questions spanning seven vascular surgery disease domains. Six experienced vascular surgery faculty at a tertiary academic center independently graded AI responses on their accuracy (rated 1-4 from completely inaccurate to completely accurate), completeness (rated 1-4 from totally incomplete to totally complete), and appropriateness (binary). Responses were also evaluated with three readability scales.

RESULTS

ChatGPT responses were rated, on average, more accurate than Bard responses (3.08 ± 0.33 vs 2.82 ± 0.40, p < .01). ChatGPT responses were scored, on average, more complete than Bard responses (2.98 ± 0.34 vs 2.62 ± 0.36, p < .01). Most ChatGPT responses (75.0%, n = 18) and almost half of Bard responses (45.8%, n = 11) were unanimously deemed appropriate. Almost one-third of Bard responses (29.2%, n = 7) were deemed inappropriate by at least two reviewers (29.2%), and two Bard responses (8.4%) were considered inappropriate by the majority. The mean Flesch Reading Ease, Flesch-Kincaid Grade Level, and Gunning Fog Index of ChatGPT responses were 29.4 ± 10.8, 14.5 ± 2.2, and 17.7 ± 3.1, respectively, indicating that responses were readable with a post-secondary education. Bard's mean readability scores were 58.9 ± 10.5, 8.2 ± 1.7, and 11.0 ± 2.0, respectively, indicating that responses were readable with a high-school education (p < .0001 for three metrics). ChatGPT's mean response length (332 ± 79 words) was higher than Bard's mean response length (183 ± 53 words, p < .001). There was no difference in the accuracy, completeness, readability, or response length of ChatGPT or Bard between disease domains (p > .05 for all analyses).

CONCLUSIONS

AI offers a novel means of educating patients that avoids the inundation of information from "Dr Google" and the time barriers of physician-patient encounters. ChatGPT provides largely valid, though imperfect, responses to myriad patient questions at the expense of readability. While Bard responses are more readable and concise, their quality is poorer. Further research is warranted to better understand failure points for large language models in vascular surgery patient education.

Natural history of internal carotid artery stenosis progression

OBJECTIVE

The aim of this study was to investigate the natural history of internal carotid artery (ICA) stenosis progression.

METHODS

This single-institution retrospective cohort study analyzed patients diagnosed with ICA stenosis of 50% or greater on duplex ultrasound from 2015 to 2022. Subjects were drawn from our institutional Intersocietal Accreditation Commission-accredited noninterventional vascular laboratory database. Primary outcomes were incidences of disease progression, and stroke or revascularization after index study. Progression was defined as an increase in stenosis classification category. Imaging, demographic, and clinical data was obtained from our institutional electronic medical record via a database mining query. Cases were analyzed at the patient and artery levels, with severity corresponding to the greatest degree of ICA stenosis on index and follow-up studies.

RESULTS

Of 577 arteries in 467 patients, mean cohort age was 73.5 ± 8.9 years at the time of the index study, and 45.0% (n = 210) were female. Patients were followed with duplex ultrasound for a mean of 42.2 ± 22.7 months. Of 577 arteries, 65.5% (n = 378) at the index imaging study had moderate (50%-69%) stenosis, 23.7% (n = 137) had severe (70%-99%) stenosis, and 10.7% (n = 62) were occluded. These three groups had significant differences in age, hypertension, hyperlipidemia prevalence, and proportion on best medical therapy. Of the 467-patient cohort, 56.5% (n = 264) were on best medical therapy, defined as smoking cessation, treatment with an antiplatelet agent, statin, and antihypertensive and glycemic agents as indicated. Mean time to progression for affected arteries was 28.0 ± 20.5 months. Of those arteries with nonocclusive disease at diagnosis, 21.3% (n = 123) progressed in their level of stenosis. Older age, diabetes, and a history of vasculitis were associated with stenosis progression, whereas antiplatelet agent use trended towards decreased progression rates. Of the 467 patients, 5.6% (n = 26) developed symptoms; of those, 38.5% (n = 10) had ischemic strokes, 26.9% (n = 7) had hemispheric transient ischemic attacks, 11.5% (n = 3) had amaurosis fugax, and 23.1% (n = 6) had other symptoms. A history of head and neck cancer was positively associated with symptom development. Of 577 affected arteries, 16.6% (n = 96) underwent intervention; 81% (n = 78) of interventions were for asymptomatic disease and 19% (n = 18) were for symptomatic disease. No patient-level factors were associated with risk of intervention.

CONCLUSIONS

A significant number of carotid stenosis patients experience progression of disease. Physicians should consider long-term surveillance on all patients with carotid disease, with increased attention paid to those with risk factors for progression, particularly those with diabetes and a history of vasculitis.

Endovenous laser ablation: A comprehensive review

OBJECTIVE

To provide an evidence-based overview of endovenous laser ablation and describe its role as an effective and durable technique for the management of superficial venous insufficiency.

METHODS

The published literature on the treatment of varicose veins using endovenous laser ablation was reviewed. The literature search focused on the history of endovenous laser ablation, its safety and durability, known complications, and differences in outcomes based on the iterations of fiber type and laser wavelength.

RESULTS

Treatment safety and efficacy of endovenous laser ablation appear to be based on the amount of energy administered over a defined distance, or the linear endovenous energy density. The ideal linear endovenous energy density varies with the laser wavelength and fiber-type. Post-operative pain and bruising may be reduced by the use of higher wavelength fibers or the use of radial or jacket-tip fibers as compared to bare-tip fibers. The incidence of endothermal heat-induced thrombosis remains low and has declined with increasing experience. Reports have demonstrated a greater than 90% technical success rate with saphenous endovenous laser ablation, long-term durability of ablation, and commensurate improvement in quality of life.

CONCLUSIONS

Endovenous laser ablation is a safe and durable treatment option for the management of incompetent superficial and perforator veins of the lower extremities. As an endothermal technology, it remains a key component of the standard of care for the treatment of chronic venous insufficiency.

Determining If Sex Bias Exists in Human Surgical Clinical Research

Importance

Sex is a variable that is poorly controlled for in clinical research.

Objectives

To determine if sex bias exists in human surgical clinical research, to determine if data are reported and analyzed using sex as an independent variable, and to identify specialties in which the greatest and least sex biases exist.

Design, Setting, and Participants

For this bibliometric analysis, data were abstracted from 1303 original peer-reviewed articles published from January 1, 2011, through December 31, 2012, in 5 surgery journals.

Main Outcomes and Measures

Study type, location, number and sex of participants, degree of sex matching of included participants, and inclusion of sex-based reporting, statistical analysis, and discussion of data.

Results

Of 2347 articles reviewed, 1668 (71.1%) included human participants. After excluding 365 articles, 1303 remained: 17 (1.3%) included males only, 41 (3.1%) included females only, 1020 (78.3%) included males and females, and 225 (17.3%) did not document the sex of the participants. Although female participants represent more than 50% (n = 57 688 606) of the total number (115 377 213) included, considerable variability existed with the number of male (46 111 818), female (58 805 665), and unspecified (10 459 730) participants included among the journals, between US domestic and international studies, and between single vs multicenter studies. For articles included in the study, 38.1% (497 of 1303) reported these data by sex, 33.2% (432 of 1303) analyzed these data by sex, and 22.9% (299 of 1303) included a discussion of sex-based results. Sex matching of the included participants in the research overall was poor, with 45.2% (589 of 1303) of the studies matching the inclusion of both sexes by 50%. During analysis of the different surgical specialties, a wide variation in sex-based inclusion, matching, and data reporting existed, with colorectal surgery having the best matching of male and female participants and cardiac surgery having the worst.

Conclusions and Relevance

Sex bias exists in human surgical clinical research. Few studies included men and women equally, less than one-third performed data analysis by sex, and there was wide variation in inclusion and matching of the sexes among the specialties and the journals reviewed. Because clinical research is the foundation for evidence-based medicine, it is imperative that this disparity be addressed so that therapies benefit both sexes.

Patient controlled analgesia and GI dysfunction

A prospective randomized study of 300 total joint replacement patients was done comparing the gastrointestinal (GI) side effects of morphine, meperidine, and nalbuphine given by patient controlled analgesia (PCA). All three groups of patients developed some form of GI dysfunction, but no significant difference was found among the three narcotics used. It was found that normal GI assessment in the initial postoperative period does not guarantee that GI function will remain normal. This finding indicates a need for extended nursing assessment of the GI status of the postoperative PCA patient.