Prior authorization requirements in the office-based laboratory setting are administratively inefficient and threaten timeliness of care.

OBJECTIVE: The objective of this study was to investigate the administrative and clinical impacts of prior authorization (PA) processes in the office-based laboratory (OBL) setting. METHODS: This single-institution, retrospective analysis studied all OBL PAs pursued between January 2018 and March 2022. Case, PA, and coding information was obtained from the practice's scheduling database. RESULTS: Over the study period, 1854 OBL cases were scheduled; 8% (n = 146) required PA. Of these, 75% (n = 110) were for lower extremity arterial interventions, 19% (n = 27) were for deep venous interventions, and 6% (n = 9) were for other interventions. Of 146 PAs, 19% (n = 27) were initially denied but 74.1% (n = 7) of these were overturned on appeal. Deep venous procedures were initially denied, at 43.8% (n = 14), more often than were arterial procedures, at 11.8% (n = 13). Of 146 requested procedures, 4% (n = 6) were delayed due to pending PA determination by a mean 14.2 ± 18.3 working days. An additional 6% (n = 8) of procedures were performed in the interest of time before final determination. Of the seven terminally denied procedures, 57% (n = 4) were performed at cost to the practice based on clinical judgment. CONCLUSIONS: Using PA appeals mechanisms, while administratively onerous, resulted in the overturning of most initial denials.

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.

Venous ablation procedures by provider type, including advanced practice providers.

OBJECTIVE: The necessary training and certification of providers performing venous ablation has become a topic of debate in recent years. As venous interventions have shifted away from the hospital, the diversity of provider backgrounds has increased. We aimed to characterize superficial venous ablation practice patterns associated with different provider types. METHODS: We analyzed Medicare Fee-For-Service data from 2010 through 2018. Procedures were identified by their Current Procedural Terminology code and included radiofrequency ablation, endovenous laser ablation, chemical adhesive ablation (ie, VenaSeal; Medtronic, Inc), and mechanochemical ablation. These procedures were correlated with the practitioner type to identify provider-specific trends. RESULTS: Between 2010 and 2018, the number of ablation procedures increased by 107% from 114,197 to 236,558 per year (P < .001). Most procedures were performed by surgeons without vascular board certification (28.7%; 95% confidence interval [CI], 28.7%-28.8%), followed by vascular surgeons (27.1%; 95% CI, 27.0%-27.2%). Traditionally noninterventional specialties, which exclude surgeons, cardiologists, and interventional radiologists, accounted for 14.1% (95% CI, 14.1%-14.2%), and APPs accounted for 3.5% (95% CI, 3.4%-3.5%) of all ablation procedures during the study period. The total number of ablations increased by 9.7% annually (95% CI, 9.7%-9.8%), whereas procedures performed by APPs increased by 62.0% annually (95% CI, 61.6%-62.4%). There were significant differences between specialties in the use of nonthermal ablation modalities: APPs had the highest affinity for nonthermal ablation (odds ratio [OR], 2.60; 95% CI, 2.51-2.69). Cardiologists were also more likely to use nonthermal ablation (OR, 1.62; 95% CI, 1.59-1.66). Similarly, the uptake of new nonthermal technology (ie, chemical adhesives) was greatest among APPs (OR, 3.57; 95% CI, 3.43-3.70) and cardiologists (OR, 1.86; 95% CI, 1.81-1.91). Vascular surgeons were less likely to use nonthermal modalities (OR, 0.95; 95% CI, 0.92-0.97), including new nonthermal technology in the first year of availability (OR, 0.93; 95% CI, 0.90-0.95). CONCLUSIONS: The use of venous procedures has increased rapidly during the past decade, particularly as endovenous ablations have been performed by a wider practitioner base, including APPs and noninterventionalists. Practice patterns differ by provider type, with APPs and cardiologists skewing more toward nonthermal modalities, including more rapid uptake of new nonthermal technology. Provider-specific biases for specific ablation modalities might reflect differences in training, skill set, the need for capital equipment, clinical privileges, or reimbursement. These data could help to inform training paradigms, the allocation of resources, and evaluation of appropriateness in a real-world setting.

Safety and efficacy of endovenous ablation in patients with a history of deep vein thrombosis.

OBJECTIVE: Endovenous ablation is the standard of care for patients with symptomatic superficial venous insufficiency. For patients with a history of deep vein thrombosis (DVT), concern exists for an increased risk of postprocedural complications, particularly venous thromboembolism. The objective of this study was to evaluate the safety and efficacy of endovenous thermal ablation in patients with a history of DVT. METHODS: The national Vascular Quality Initiative Varicose Vein Registry was queried for superficial venous procedures performed from January 2014 to July 2021. Limbs treated with radiofrequency or laser ablation were compared between patients with and without a DVT history. The primary safety end point was incident DVT or endothermal heat-induced thrombosis (EHIT) II-IV in the treated limb at <3 months of follow-up. The secondary safety end points included any proximal thrombus extension (ie, EHIT I-IV), major bleeding, hematoma, pulmonary embolism, and death due to the procedure. The primary efficacy end point was technical failure (ie, recanalization at <1 week of follow-up). Secondary efficacy end points included the risk of recanalization over time and the postprocedural change in quality-of-life measures. Outcomes stratified by preoperative use of anticoagulation (AC) were also compared among those with prior DVT. RESULTS: Among 33,892 endovenous thermal ablations performed on 23,572 individual patients aged 13 to 90 years, 1698 patients (7.2%) had a history of DVT. Patients with prior DVT were older (P < .001), had a higher body mass index (P < .001), were more likely to be male at birth (P < .001) and Black/African American (P < .001), and had greater CEAP classifications (P < .001). A history of DVT conferred a higher risk of new DVT (1.4% vs 0.8%; P = .03), proximal thrombus extension (2.3% vs 1.6%; P = .045), and bleeding (0.2% vs 0.04%; P = .03). EHIT II-IV, pulmonary embolism, and hematoma risk did not differ by DVT history (P = NS). No deaths from treatment occurred in either group. Continuing preoperative AC in patients with prior DVT did not change the risk of any complications after endovenous ablation (P = NS) but did confer an increased hematoma risk among all endovenous thermal ablations and surgeries (P = .001). Technical failure was similar between groups (2.0% vs 1.2%; P = .07), although a history of DVT conferred an increased recanalization risk over time (hazard ratio, 1.90; 95% confidence interval, 1.46, 2.46; P < .001). The groups had comparable improvements in postprocedural venous clinical severity scores and Heaviness, Aching, Swelling, Throbbing, and Itching scores (P = NS). CONCLUSIONS: Endovenous thermal ablation for patients with a history of DVT was effective. However, appropriate patient counseling regarding a heightened DVT risk, albeit still low, is critical. The decision to continue or withhold AC preoperatively should be tailored on a case-by-case basis.

Evaluating Large Language Models in Extracting Cognitive Exam Dates and Scores.

Importance: Large language models (LLMs) are crucial for medical tasks. Ensuring their reliability is vital to avoid false results. Our study assesses two state-of-the-art LLMs (ChatGPT and LlaMA-2) for extracting clinical information, focusing on cognitive tests like MMSE and CDR. Objective: Evaluate ChatGPT and LlaMA-2 performance in extracting MMSE and CDR scores, including their associated dates. Methods: Our data consisted of 135,307 clinical notes (Jan 12th, 2010 to May 24th, 2023) mentioning MMSE, CDR, or MoCA. After applying inclusion criteria 34,465 notes remained, of which 765 underwent ChatGPT (GPT-4) and LlaMA-2, and 22 experts reviewed the responses. ChatGPT successfully extracted MMSE and CDR instances with dates from 742 notes. We used 20 notes for fine-tuning and training the reviewers. The remaining 722 were assigned to reviewers, with 309 each assigned to two reviewers simultaneously. Inter-rater-agreement (Fleiss' Kappa), precision, recall, true/false negative rates, and accuracy were calculated. Our study follows TRIPOD reporting guidelines for model validation. Results: For MMSE information extraction, ChatGPT (vs. LlaMA-2) achieved accuracy of 83% (vs. 66.4%), sensitivity of 89.7% (vs. 69.9%), true-negative rates of 96% (vs 60.0%), and precision of 82.7% (vs 62.2%). For CDR the results were lower overall, with accuracy of 87.1% (vs. 74.5%), sensitivity of 84.3% (vs. 39.7%), true-negative rates of 99.8% (98.4%), and precision of 48.3% (vs. 16.1%). We qualitatively evaluated the MMSE errors of ChatGPT and LlaMA-2 on double-reviewed notes. LlaMA-2 errors included 27 cases of total hallucination, 19 cases of reporting other scores instead of MMSE, 25 missed scores, and 23 cases of reporting only the wrong date. In comparison, ChatGPT's errors included only 3 cases of total hallucination, 17 cases of wrong test reported instead of MMSE, and 19 cases of reporting a wrong date. Conclusions: In this diagnostic/prognostic study of ChatGPT and LlaMA-2 for extracting cognitive exam dates and scores from clinical notes, ChatGPT exhibited high accuracy, with better performance compared to LlaMA-2. The use of LLMs could benefit dementia research and clinical care, by identifying eligible patients for treatments initialization or clinical trial enrollments. Rigorous evaluation of LLMs is crucial to understanding their capabilities and limitations.

Temporal trends in surgical implants for in situ fixation of stable slipped capital femoral epiphysis.

Stable slipped capital femoral epiphysis (SCFE) is a pediatric hip disorder managed with in situ fixation of the proximal femoral epiphysis, otherwise known as 'pinning the hip'. The objective of this study was to characterize how the choice of implant for in situ fixation of stable SCFE has changed over time. A systematic review of publications concerning in situ fixation of stable SCFE from January 1993 to November 2021 was conducted. The change in the type, diameter, and number of implants used in publications over time and the age of their respective patient cohorts was evaluated. A total of 207 articles met inclusion criteria. There was an increase in publications using cannulated screws over time (P = 0.0113). As the yearly percentage of publications using threadless non-cannulated implants decreased (P = 0.0309), the percentage using cannulated screws increased (P = 0.0047). Single-implant fixation also increased (P = 0.0409). While there was no difference in the rate of increase of implants < 7 mm or ≥ 7 mm in diameter (P = 0.299), patients with larger-diameter implants were, on average, older than patients with smaller-diameter implants (P = 0.0462). In general, the age of patients undergoing in situ fixation of stable SCFE has not changed (P = 0.595). Irrespective of patient-specific considerations, single cannulated screws have become the implant of choice for in situ fixation of stable SCFE. There has not been a consensus on the optimal implant diameter; instead, patient-specific considerations are of paramount importance in this decision.

Creation and validation of 3D-printed head molds for stereotaxic injections of neonatal mouse brains.

BACKGROUND: An increasing number of rodent model systems use injection of DNA or viral constructs in the neonatal brain. However, approaches for reliable positioning and stereotaxic injection at this developmental stage are limited, typically relying on handheld positioning or molds that must be re-aligned for use in a given laboratory. NEW METHOD: A complete protocol and open-source software pipeline for generating 3D-printed head molds derived from a CT scan of a neonatal mouse head cast, together with a universal adapter that can be placed on a standard stereotaxic stage. RESULTS: A series of test injections with adenovirus encoding red fluorescent protein, or Fluorogold, were conducted using original clay molds and newly generated 3D printed molds. Several metrics were used to compare spread and localization of targeted injections. COMPARISON WITH EXISTING METHODS: The new method of head mold generation gave comparable results to the field standard, but also allowed the rapid generation of additional copies of each head mold with standardized positioning of the head each time. CONCLUSIONS: This 3D printing pipeline can be used to efficiently develop a series of head molds with standardized injection coordinates across multiple laboratories. More broadly, this pipeline can easily be adapted to other perinatal ages or species.