Comparison of artificial intelligence large language model chatbots in answering frequently asked questions in anaesthesia.

Background: Patients are increasingly using artificial intelligence (AI) chatbots to seek answers to medical queries. Methods: Ten frequently asked questions in anaesthesia were posed to three AI chatbots: ChatGPT4 (OpenAI), Bard (Google), and Bing Chat (Microsoft). Each chatbot's answers were evaluated in a randomised, blinded order by five residency programme directors from 15 medical institutions in the USA. Three medical content quality categories (accuracy, comprehensiveness, safety) and three communication quality categories (understandability, empathy/respect, and ethics) were scored between 1 and 5 (1 representing worst, 5 representing best). Results: ChatGPT4 and Bard outperformed Bing Chat (median [inter-quartile range] scores: 4 [3-4], 4 [3-4], and 3 [2-4], respectively; P<0.001 with all metrics combined). All AI chatbots performed poorly in accuracy (score of ≥4 by 58%, 48%, and 36% of experts for ChatGPT4, Bard, and Bing Chat, respectively), comprehensiveness (score ≥4 by 42%, 30%, and 12% of experts for ChatGPT4, Bard, and Bing Chat, respectively), and safety (score ≥4 by 50%, 40%, and 28% of experts for ChatGPT4, Bard, and Bing Chat, respectively). Notably, answers from ChatGPT4, Bard, and Bing Chat differed statistically in comprehensiveness (ChatGPT4, 3 [2-4] vs Bing Chat, 2 [2-3], P<0.001; and Bard 3 [2-4] vs Bing Chat, 2 [2-3], P=0.002). All large language model chatbots performed well with no statistical difference for understandability (P=0.24), empathy (P=0.032), and ethics (P=0.465). Conclusions: In answering anaesthesia patient frequently asked questions, the chatbots perform well on communication metrics but are suboptimal for medical content metrics. Overall, ChatGPT4 and Bard were comparable to each other, both outperforming Bing Chat.

Brief Assessment of Patient Phenotype to Explain Variability in Postsurgical Pain and Opioid Consumption after Cesarean Delivery: Performance of a Novel Brief Questionnaire Compared to Long Questionnaires.

BACKGROUND: Understanding factors that explain why some women experience greater postoperative pain and consume more opioids after cesarean delivery is crucial to building an evidence base for personalized prevention. Comprehensive psychosocial assessment with validated questionnaires in the preoperative period can be time-consuming. A three-item questionnaire has shown promise as a simpler tool to be integrated into clinical practice, but its brevity may limit the ability to explain heterogeneity in psychosocial pain modulators among individuals. This study compared the explanatory ability of three models: (1) the 3-item questionnaire, (2) a 58-item questionnaire (long) including validated questionnaires (e.g., Brief Pain Inventory, Patient Reported Outcome Measurement Information System [PROMIS]) plus the 3-item questionnaire, and (3) a novel 19-item questionnaire (brief) assessing several psychosocial factors plus the 3-item questionnaire. Additionally, this study explored the utility of adding a pragmatic quantitative sensory test to models. METHODS: In this prospective, observational study, 545 women undergoing cesarean delivery completed questionnaires presurgery. Pain during local anesthetic skin wheal before spinal placement served as a pragmatic quantitative sensory test. Postoperatively, pain and opioid consumption were assessed. Linear regression analysis assessed model fit and the association of model items with pain and opioid consumption during the 48 h after surgery. RESULTS: A modest amount of variability was explained by each of the three models for postoperative pain and opioid consumption. Both the brief and long questionnaire models performed better than the three-item questionnaire but were themselves statistically indistinguishable. Items that were independently associated with pain and opioid consumption included anticipated postsurgical pain medication requirement, surgical anxiety, poor sleep, pre-existing pain, and catastrophic thinking about pain. The quantitative sensory test was itself independently associated with pain across models but only modestly improved models for postoperative pain. CONCLUSIONS: The brief questionnaire may be more clinically feasible than longer validated questionnaires, while still performing better and integrating a more comprehensive psychosocial assessment than the three-item questionnaire.

Frequency and Risk Factors for Difficult Intubation in Women Undergoing General Anesthesia for Cesarean Delivery: A Multicenter Retrospective Cohort Analysis.

BACKGROUND: Estimates for the incidence of difficult intubation in the obstetric population vary widely, although previous studies reporting rates of difficult intubation in obstetrics are older and limited by smaller samples. The goals of this study were to provide a contemporary estimate of the frequency of difficult and failed intubation in women undergoing general anesthesia for cesarean delivery and to elucidate risk factors for difficult intubation in women undergoing general anesthesia for cesarean delivery. METHODS: This is a multicenter, retrospective cohort study utilizing the Multicenter Perioperative Outcomes Group database. The study population included women aged 15 to 44 yr undergoing general anesthesia for cesarean delivery between 2004 and 2019 at 1 of 45 medical centers. Coprimary outcomes included the frequencies of difficult and failed intubation. Difficult intubation was defined as Cormack-Lehane view of 3 or greater, three or more intubation attempts, rescue fiberoptic intubation, rescue supraglottic airway, or surgical airway. Failed intubation was defined as any attempt at intubation without successful endotracheal tube placement. The rates of difficult and failed intubation were assessed. Several patient demographic, anatomical, and obstetric factors were evaluated for potential associations with difficult intubation. RESULTS: This study identified 14,748 cases of cesarean delivery performed under general anesthesia. There were 295 cases of difficult intubation, with a frequency of 1:49 (95% CI, 1:55 to 1:44; n = 14,531). There were 18 cases of failed intubation, with a frequency of 1:808 (95% CI, 1:1,276 to 1:511; n = 14,537). Factors with the highest point estimates for the odds of difficult intubation included increased body mass index, Mallampati score III or IV, small hyoid-to-mentum distance, limited jaw protrusion, limited mouth opening, and cervical spine limitations. CONCLUSIONS: In this large, multicenter, contemporary study of more than 14,000 general anesthetics for cesarean delivery, an overall risk of difficult intubation of 1:49 and a risk of failed intubation of 1:808 were observed. Most risk factors for difficult intubation were nonobstetric in nature. These data demonstrate that difficult intubation in obstetrics remains an ongoing concern.

Anesthetic Management of Von Willebrand Disease in Pregnancy: A Retrospective Analysis of a Large Case Series.

BACKGROUND: Although Von Willebrand disease (vWD) is the most common heritable bleeding disorder, there are limited reports regarding the safety of neuraxial anesthesia in the obstetric population and no definitive guidelines specifying recommended pretreatment or therapies for patients with vWD. The aim of this study is to describe the anesthetic management of pregnant patients with vWD at a large tertiary-care center. METHODS: In this retrospective analysis, the study population was identified from vWD patients evaluated by our high-risk obstetric anesthesia consultation service and by diagnosis codes from our institutional research database registry. We manually reviewed records of patients with vWD in pregnancy who delivered at our institution between January 1, 2000 and January 1, 2019 for demographic characteristics, circumstances of vWD diagnosis, history of bleeding, laboratory studies, and overall management of vWD. Anesthetic management of vWD was at the discretion of individual providers, based on multidisciplinary consensus and the specific circumstances of each patient's disease and obstetric presentation. RESULTS: We identified 106 deliveries among 71 individual vWD patients. Of the unique patients, 54 had vWD type 1, 6 had vWD type 2, and 11 had vWD type unknown. Forty-three cases (40.6%) were cesarean deliveries. Neuraxial techniques were used in 94 of 106 deliveries (88.7%). Treatment with desmopressin or Von Willebrand factor/factor VIII concentrate before neuraxial anesthesia occurred in 27 of 94 neuraxial anesthetics (28.7%). Eleven deliveries (10.4%) were complicated by postpartum hemorrhage (PPH), defined as estimated blood loss of ≥1000 mL. There were no noted adverse anesthetic outcomes (0 of 106; 0% [95% confidence interval, 0-3.4]), including neuraxial hematoma or thromboembolic events. CONCLUSIONS: In this large case series, the majority of vWD patients received neuraxial anesthesia for labor and delivery, with no noted adverse events. This suggests that neuraxial anesthesia can be safely performed with the peripartum management that we describe. Pretreatment was dictated by the type and severity of vWD. Multidisciplinary planning is important to optimize the coagulation status of patients with vWD and facilitate options for analgesia and anesthesia.

Patient characteristics associated with SARS-CoV-2 infection in parturients admitted for labour and delivery in Massachusetts during the spring 2020 surge: A prospective cohort study.

BACKGROUND: While studies from large cities affected by coronavirus disease 2019 (COVID-19) have reported on the prevalence of SARS-CoV-2 in the context of universal testing during admission for delivery, the patient demographic, social and clinical factors associated with SARS-CoV-2 infection in pregnant women are not fully understood. OBJECTIVE: To evaluate the epidemiological factors associated with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection in women admitted for labour and delivery, in the context of universal screening at four Boston-area hospitals. METHODS: In this prospective cohort study, we reviewed the health records of all women admitted for labour and delivery at four hospitals from the largest health system in Massachusetts between 19 April 2020 and 27 June 2020. We calculated the risk of SARS-CoV-2 infection, including asymptomatic infection. We calculated associations between SARS-CoV-2 infection and demographic and clinical characteristics. RESULTS: A total of 93 patients (3.2%, 95% confidence interval 2.5, 3.8) tested positive for SARS-CoV-2 infection on admission for labour and delivery out of 2945 patients included in the analysis; 80 (86.0%) of the patients who tested positive were asymptomatic at the time of testing. Factors associated with SARS-CoV-2 infection included the following: younger age, obesity, African American or Hispanic race/ethnicity, residence in heavily affected communities (as measured in cases reported per capita), presence of a household member with known SARS-CoV-2 infection, non-health care essential worker occupation and MassHealth or Medicaid insurance compared to commercial insurance. 93.8% of patients testing positive for SARS-CoV-2 on admission had one or more identifiable factors associated with disease acquisition. CONCLUSIONS: In this large sample of deliveries during the height of the surge in infections during the spring of 2020, SARS-CoV-2 infection was largely concentrated in patients with distinct demographic characteristics, those largely from disadvantaged communities. Racial disparities seen in pregnancy persist with respect to SARS-CoV-2 infection.

The Most Influential Publications in Obstetric Anesthesiology, 1998-2017: Utilizing the Delphi Method for Expert Consensus.

BACKGROUND: There have been many advances in obstetric anesthesiology in the past 2 decades. We sought to create a list of highly influential publications in the field using the Delphi method among a group of obstetric anesthesiology experts to create an important educational, clinical, and research resource. METHODS: Experts in the field, defined as obstetric anesthesiologists selected to present the Gerard W. Ostheimer Lecture at the Society for Obstetric Anesthesia and Perinatology (SOAP) annual meeting within the past 20 years, were recruited to participate. The Delphi technique was used by administering 3 rounds of surveys. Participants were initially asked to identify the highly influential publications from the year they presented the Ostheimer lecture, in addition to the most influential publications from the time period overall. Highly influential publications were defined as those that changed traditional views, invoked meaningful practices, catalyzed additional research, and fostered ideas or practices that had durability over time. After each round of surveys, responses were collected and used as choices for subsequent surveys with the goal of obtaining group consensus. RESULTS: We determined expert consensus on 22 highly influential publications from 1998 to 2017. The focus of these publications ranged from disease entities, interventions, treatment methodologies, and complications. CONCLUSIONS: Key themes in the publications chosen included the reduction of maternal morbidity and mortality and refinements in the analgesic and anesthetic management of labor and delivery.

Introduction of a Novel System for Quantitating Blood Loss After Vaginal Delivery: A Retrospective Interrupted Time Series Analysis With Concurrent Control Group.

BACKGROUND: Postpartum hemorrhage (PPH) is a leading cause of maternal mortality worldwide. While volumetric and gravimetric blood loss measurement techniques have been shown to increase PPH detection compared to visual estimation of blood loss (vBL), a combination of these methods has not been evaluated. The primary aim of this study was to compare the change in odds of PPH detection after vaginal delivery (VD; the intervention series) versus cesarean delivery (CD; the control series) before versus after implementation of a device (Triton L&D, Gauss Surgical, Inc, Los Altos, CA) that combines gravimetric and volumetric estimation of blood loss (tBL). METHODS: After institutional review board (IRB) approval, patients were identified who had a VD or CD at our institution 3 months before and 3 months after the implementation of a tBL device for VD. The levels and trends in odds of PPH detection (the primary outcome, defined as blood loss ≥500 mL for VD and ≥1000 mL for CD) were compared within and between delivery methods before and after implementation of the device. Secondary outcomes were blood loss ≥1000 mL, total blood loss, secondary uterotonic use, and a composite outcome related to PPH management (transfusion, vasopressor administration, and/or surgical procedures). A post hoc subgroup analysis compared nadir hematocrit, hematocrit reduction ≥10%, and difference between vBL or tBL and calculated blood loss (cBL) between VDs before and after introduction of the device. All outcome comparisons were performed using segmented regression with inverse probability of treatment weighting. RESULTS: The weighted odds ratio (wOR) of PPH detection post- versus preimplementation of the device was 2.74 (95% confidence interval [CI], 1.39-5.41; P = .004) for VDs versus 1.43 (95% CI, 0.72-2.85; P = .304) for CDs (P for interaction .190). No difference was detected in the secondary outcomes for VDs between time periods. In the subgroup of VD patients who had postdelivery hematocrits measured, the mean difference between vBL or tBL and cBL was smaller in the tBL (mean ± standard deviation [SD]: -237 ± 522 mL) versus vBL group (-600 ± 596 mL; weighted difference in means [95% CI]: 349 mL [13-684]; P = .042). CONCLUSIONS: In this interrupted time series analysis, the odds of PPH detection after VD increased after implementation of the tBL device. However, a statistical difference was not detected between the increase in PPH detection post- versus preimplementation of the device in VDs versus CDs.

Trends in Postpartum Hemorrhage in the United States From 2010 to 2014.

Postpartum hemorrhage (PPH) is a leading cause of morbidity and mortality in the United States; its prevalence increased during the 1990s-2000s. The purpose of this study was to reevaluate trends in PPH using the National Inpatient Sample. From 2010 to 2014, the prevalence of PPH increased from 2.9% (95% confidence interval [CI], 2.7%-3.1%) to 3.2% (95% CI, 3.1%-3.3%) of deliveries. Adjusting for PPH risk factors did not substantially attenuate this trend. Among patients with PPH, there was a decline in associated coagulopathy, acute respiratory failure, and maternal death, but an increase in sepsis and acute renal failure. Continued focus on PPH management is warranted.

Toxic Ingestion of Acetaminophen and Acetylsalicylic Acid in a Parturient at 33 Weeks Gestation: A Case Report.

The anesthetic management of toxic ingestion during pregnancy requires concomitant concerns for both mother and fetus. We describe the management of a parturient at 33 weeks of gestation after a suicide attempt by ingestion of acetaminophen (APAP) and acetylsalicylic acid (ASA). Timing of toxin ingestion must be determined, prompt antidote administration prioritized, and hepatotoxicity-induced hematologic impairment anticipated. Fetal exposure to toxins must also be minimized. The use of point-of-care rotational thromboelastometry in conjunction with standard coagulation testing in such cases facilitates consideration of neuraxial anesthesia and determination of risk for postpartum hemorrhage.