BACKGROUND:Compared to standard labor epidurals, placement of a dural puncture epidural (DPE) has been shown to improve labor analgesia efficacy, speed onset of analgesia, and decrease failure rates. However, there is a concern of increased post-dural puncture headache (PDPH) risk. A 25-gauge or larger spinal needle is typically needed for a DPE to have improved efficacy over a standard epidural. A prior meta-analysis comparing combined spinal-epidurals (CSEs) with epidurals did not identify a difference in PDPH rates, but many of the included randomized controlled trials (RCTs) used smaller diameter (27-29 gauge) spinal needles. This systematic review and meta-analysis of RCTs evaluates headache frequency following CSE or DPE labor analgesia techniques with 24- or 25-gauge spinal needles. METHODS:Medline, EMBASE, Scopus, Database of Abstracts of Reviews of Effects, Web of Science Core Collection, and Cochrane databases were searched on May 7, 2024, for RCTs comparing either DPE or CSE techniques using 24- or 25-gauge spinal needles with epidural labor analgesia. The primary outcome was headache. The risk-of-bias 2 tool was used to assess bias. The odds ratio (OR) was the chosen effect size using binomial family Bayesian estimation. The reference category for estimation was the conventional epidural technique group. RESULTS:Encompassing 3278 patients, 1765 (53.8%) received labor epidurals, 971 (29.6%) received DPEs, and 542 (16.5%) received CSEs. The rate of headaches in both groups was low, with headache rates of CSE/DPE and epidural groups (0.59% vs 0.34%), respectively. Headache was a secondary outcome in all the RCTs. In 10 of the 16 RCTs (62.5%), no patients of 1428 (0%; 95% credible interval [CI], 0.00-0.25) reported a headache. A total of nine (0.6%) headaches were reported in the DPE/CSE patients (denominator 1513) and six (0.3%) headaches were reported in the traditional epidural patients (denominator 1765). The posterior distribution of the treatment effect ( θ ) showed that the log OR of 0.35 favored the control group as having fewer events, but the number of events was rare, and the 95% CI was wide (-0.49 to 1.22). CONCLUSIONS:Headache was a rare event in both groups. Although these results are reassuring that either technique is clinically safe, they should be interpreted cautiously. A large RCT examining PDPH as the primary outcome with long-term follow-up is needed to more precisely determine if placement of a DPE or CSE with a 24- or 25-gauge spinal needle increases the risk of PDPH.

BACKGROUND:Air pollution, especially particle pollution, is increasingly recognized as a potential perioperative risk factor, yet modeling environmental exposures in surgical cohorts remains methodologically underdeveloped. We demonstrate a Bayesian hierarchical framework to quantify probabilistic associations between preoperative fine particulate matter (PM2.5) exposure and postoperative complications, highlighting its interpretability and flexibility for clinical environmental epidemiology. METHODS:We conducted a single center, retrospective cohort study using data from 49,615 surgical patients in Utah who underwent elective surgical procedures from 2016 to 2018. Patients' addresses were geocoded and linked to daily Census-tract level PM2.5 estimates. The exposure variable was defined as the maximum PM2.5 concentrations in the 7 days prior to surgery. The binary outcome was a composite of postoperative complications: pneumonia, surgical site infection, urinary tract infection, sepsis, stroke, myocardial infarction, or thromboembolic event. A hierarchical Bayesians regression model with weakly informative priors was used adjusting for age, sex, season, neighborhood disadvantage, and the Elixhauser index of comorbidities with census tract as a group (random) effect. We present posterior estimates with credible intervals, highlight model transparency and sensitivity, and discuss contrasts with standard frequentist methods. RESULTS:Postoperative complications were associated in a dose-dependent manner with higher concentrations of PM2.5 exposure. We found a relative increase of 8.2% in the odds of complications (OR = 1.082) for every 10.ug/m3 increase in the highest single-day 24-h PM2.5 exposure during the 7 days prior to surgery. For an increase in PM2.5 from 1 to 30 ug/m3, the odds of complication rose to over 27% (95% CI: 4%-55%). The results were robust across prior choices and model specifications. We report full posterior distributions and highlight advantages of Bayesian modeling for uncertainty quantification and clinical interpretability. CONCLUSIONS:This case study demonstrates the application of hierarchical Bayesian modeling to quantify the probabilistic associations between preoperative PM2.5 exposure and postoperative complications, highlighting transparent risk estimation and uncertainty characterization that may inform the design of future multicenter perioperative environmental studies. EDITORIAL COMMENT:Using Bayesian statistical analysis, the authors demonstrate a dose-dependent risk for postoperative complications in patients exposed to air polluted with fine particulate matter with a size of less than 2.5 μm.

Background:Social determinants of health continue to drive persistent disparities in perioperative care. Our team has previously demonstrated racial and socioeconomic disparities in perioperative processes, notably in the administration of antiemetic prophylaxis, in several large perioperative registries. Given how neighborhoods are socially segregated in the United States, we examined geospatial clustering of perioperative antiemetic disparities. Objective:The study aimed to determine whether disparities in perioperative antiemetic prophylaxis exhibit geographic clustering based on neighborhood-level disadvantage and whether patients from disadvantaged communities are more likely to be undertreated after adjusting for individual postoperative nausea and vomiting risk. Methods:We conducted a retrospective cohort study of anesthetic records from the University of Utah Hospital involving 19,477 patients who met the inclusion criteria. We geocoded patient home addresses and combined them with the census block group-level neighborhood disadvantage, a composite index from the National Neighborhood Data Archive. We stratified our patients by antiemetic risk score and calculated the number of antiemetic interventions. We used Poisson spatial scan statistics, implemented in SaTScan (Information Management Services, Inc), to detect geographic clusters of undertreatment. Results:We identified 1 significant cluster (P<.001) of undertreated perioperative antiemetic prophylaxis cases. The relative risk of the whole cluster was 1.44, implying that patients within the cluster were 1.44 times more likely to receive fewer antiemetics after controlling for antiemetic risk. Patients from more disadvantaged neighborhoods were more likely to receive below-median antiemetic prophylaxis after controlling for risk. Conclusions:To our knowledge, this is the first geospatial cluster analysis of perioperative process disparities; we leveraged innovative geostatistical methods and identified a spatially defined, geographic cluster of patients whose home address census-tract level neighborhood deprivation index predicted disparities in risk-adjusted antiemetic prophylaxis.

BACKGROUND:Arterial blood gas analysis (ABGA) is the reference standard for the diagnosis of respiratory failure (RF) and metabolic disturbance (MD), but peripheral venous blood gas analysis (PVBGA) is increasingly being used for the estimation of carbon dioxide, pH, and other variables in the context of acutely unwell adults presenting to hospitals and emergency departments. OBJECTIVES:The primary objective of this review is to evaluate the performance of PVBGA by comparing it with the reference standard ABGA, which is assumed to be error-free for the diagnosis of (1) respiratory failure, (2) hypercarbia, and (3) metabolic disturbance (the three target conditions) in adults. The secondary objective is to evaluate the performance of the index test to diagnose nine specific subtypes of respiratory failure and metabolic disturbance. The definitions for these additional conditions are determined by changes to one or more of the following: pH (acidity), pO2 (partial pressure of oxygen), pCO2 (partial pressure of carbon dioxide), HCO3 (bicarbonate), as stated in the Methods section of this review (target conditions). We aimed to explore the following covariates: participant demographics (e.g. age, weight, and sex); participant comorbidities (e.g. chronic lung disease, chest wall deformity, and central nervous system disorder such as spinal cord injury); and the indication for blood gas sampling (e.g. shortness of breath, critical illness, resuscitation, trauma, or whilst under general anaesthesia). SEARCH METHODS:On 10 July 2024, we searched the electronic databases MEDLINE, EMBASE, CINAHL, and LILACS. We also manually searched 19 respiratory and critical care journals, and we searched ClinicalTrials.gov for ongoing trials. SELECTION CRITERIA:We considered consecutive series studies and case-control studies that directly compared the index test PVBGA to the reference standard ABGA for adults over the age of 16 years. The included studies contained data for any one of the target conditions of respiratory failure and metabolic disturbance, as determined by individual changes to pO2 (partial pressure of oxygen), pCO2 (partial pressure of carbon dioxide), pH (acidity), and HCO3 (bicarbonate) concentration. Studies that only provided mean values for summed data were ineligible for inclusion. However, we invited authors of such studies to provide individual patient data for inclusion in this systematic review. There are nine studies awaiting classification. DATA COLLECTION AND ANALYSIS:Two authors independently evaluated the quality of the relevant studies and extracted data from them. We conducted a quality assessment using the QUADAS-2 tool. Our statistical analysis used 2 x 2 tables for the positive and negative results of each test. We estimated a bivariate meta-analysis of sensitivity and specificity. MAIN RESULTS:We included six studies (919 participants) in our quantitative analysis. All studies were at high risk of bias due to one or more of the following factors: patient selection, since it was unclear if consecutive patients were included or where they were located; index test, with poor reporting of cut-offs; flow and timing domain because the fraction of inspired oxygen was frequently not stated and any difference between the collection of the VBGA and the ABGA could introduce bias. Respiratory failure For the diagnosis of respiratory failure of any type, when using PVBGA, the estimated summary sensitivity (Sn) was 97.6% (95% credible interval (CI) 94.1 to 99.4) and the estimated summary specificity (Sp) was 36.9% (95% CI 17.1 to 60.1) (6 studies, 805 participants of whom 291 (36%) were diagnosed with respiratory failure by ABGA; sensitivity: low-certainty evidence; specificity: very low certainty evidence). Isolated hypercarbia For the diagnosis of isolated hypercarbia (regardless of oxygen level), when using PVBGA, the estimated summary Sn was 97.1% (95% CI 93.3 to 99.2); the estimated summary Sp value was 53.9% (95% CI 39.8 to 66.7) (6 studies with 805 participants, 269 (33%) with ABGA confirmation; low-certainty evidence). Other findings Results for metabolic disturbance and our secondary target conditions are presented in the full review. AUTHORS' CONCLUSIONS:Very limited data suggest PVBGA performs poorly as a diagnostic test for respiratory failure compared to the reference standard of ABGA. The index test PVBGA was highly sensitive for the diagnosis of respiratory failure and isolated hypercarbia, but its specificity was poor for these two primary target conditions. The high sensitivity means PVBGA may have a useful role as a "rule out test" for respiratory failure and isolated hypercarbia; however, the high false-positive rates make the clinical interpretation of a positive test difficult. Moreover, we are uncertain regarding these estimates because we have only low to very low certainty about the evidence. Further studies that use (ABGA) established thresholds for the diagnosis of each target condition are needed.

Prevention of explicit memory function for intraoperative events (i.e., awareness with recall [AWR]) is a key deliverable for every general anaesthetic. Development of monitoring technologies, drugs, and anaesthetic techniques to achieve this objective is an important priority of contemporary anaesthesia research. The volatile anaesthetic concentrations associated with lack of memory function in the unstimulated state (i.e. MAC-amnesia) are not well known. This pilot study aimed to estimate MAC-amnesia using a variation of the continuous recognition memory task (CRMT). We hypothesised that MAC-amnesia would be slightly lower than MAC-awake.1

center dot PURPOSE: To explore the patterns of anesthesia use and their determinants during vitreoretinal (VR) surgeries in academic and community hospitals across the US, using data from the Multicenter Perioperative Outcomes Group center dot DESIGN: A retrospective, multicenter, cohort study. center dot METHODS: We queried the MPOG database of 107,066 patients undergoing VR surgeries. Patients ( >18 years) undergoing VR surgery with monitored anesthesia care (MAC) or general anesthesia (GA) from January 1, 2015 to December 31, 2021 were included. Patient-level, casebased, and institutional-level covariates were collected. We performed multivariable mixed-effects models to determine predictors of anesthesia type use. The primary outcome was the type of anesthesia (MAC or GA) used during VR surgeries. As a secondary outcome, MAC cases were further classified based on the additional use of sedation into MAC with or without sedation. center dot RESULTS: We found that 67.45% of VR surgery cases received MAC, and 73.63% of institutions administered MAC to more than half of cases. Random effect modeling revealed that 47.76% of the variation in MAC use was attributed to institutions. A trend toward increased use of MAC with increasing age was observed. Patients diagnosed with chronic pulmonary disease, liver disease, or a history of drug abuse were less likely to receive MAC. Conversely, we found that patients with reported alcohol abuse disorder, diabetes with complications, and those with American Society of Anesthesiologists (ASA) physical status of 4 (vs. 1, 2, or 3) were more likely to use MAC. Compared to non-complex VR surgeries, there was a notably decreased likelihood of MAC use in complex PPV ( P = .004), PPV + scleral buckle (SB) for retinal detachment ( P < .0001), and primary SB surgery center dot CONCLUSIONS: Approximately 2/3 of VR anesthesia is under MAC, but GA is still preferred for SBs, complex vitrectomy, and younger patients. We show that large interinstitutional variation for using MAC in practice exists. (Am J Ophthalmol 2024;267: 30-40. (c) 2024 Elsevier Inc. All rights are reserved, including those for text and data mining, AI training, and similar technologies.)

Background:While exposure to fine particulate matter air pollution (PM 2.5 ) is known to cause adverse health effects, its impact on postoperative outcomes in US adults remains understudied. Perioperative exposure to PM 2.5 may induce inflammation that interacts insidiously with the surgical stress response, leading to higher postoperative complications. Methods:We conducted a single center, retrospective cohort study using data from 49,615 surgical patients living along Utah's Wasatch Front and who underwent elective surgical procedures at a single academic medical center from 2016-2018. Patients' addresses were geocoded and linked to daily Census-tract level PM 2.5 estimates. We hypothesized that elevated PM 2.5 concentrations in the week prior to surgery would be associated with an increase in a bundle of major postoperative complications. A hierarchical Bayesians regression model was fit adjusting for age, sex, season, neighborhood disadvantage, and the Elixhauser index of comorbidities. Results:Postoperative complications increased in a dose-dependent manner with higher concentrations of PM 2.5 exposure, with a relative increase of 8% in the odds of complications (OR=1.082) for every 10ug/m 3 increase in the highest single-day 24-hr PM 2.5 exposure during the 7 days prior to surgery. For a 30 fold increase in PM 2.5 (1 ug/m 3 to 30ug/m 3 ) the odds of complication rose to over 27% (95%CI: 4%-55%). The association persisted after controlling for comorbidities and confounders; our inferences were robust to modeling choices and sensitivity analysis. Conclusions:In this large Utah cohort, exposure to elevated PM 2.5 concentrations in the week before surgery was associated with a dose-dependent increase in postoperative complications, suggesting a potential impact of air pollution on surgical outcomes. These findings merit replication in larger datasets to identify populations at risk and define the interaction and impact of different pollutants. PM 2.5 exposure is a potential perioperative risk factor and, given the unmitigated air pollution in urban areas, a global health concern.

Background: Limited data exist to guide oxygen administration during one-lung ventilation for thoracic surgery. We hypothesised that high intraoperative inspired oxygen fraction during lung resection surgery requiring one-lung ventilation is independently associated with postoperative pulmonary complications (PPCs). Methods: We performed this retrospective multicentre study using two integrated perioperative databases (Multicenter Perioperative Outcomes Group and Society of Thoracic Surgeons General Thoracic Surgery Database) to study adult thoracic surgical procedures using one-lung ventilation. The primary outcome was a composite of PPCs (atelectasis, acute respiratory distress syndrome, pneumonia, respiratory failure, reintubation, and prolonged ventilation >48 h). The exposure of interest was high inspired oxygen fraction (FiO(2)), defined by area under the curve of a FiO(2) threshold > 80%. Univariate analysis and logistic regression modelling assessed the association between intraoperative FiO(2) and PPCs. Results: Across four US medical centres, 141/2733 (5.2%) procedures conducted in 2716 patients (55% female; mean age 66 yr) resulted in PPCs. FiO(2) was univariately associated with PPCs (adjusted OR [aOR]: 1.17, 95% confidence interval [CI]: 1.04-1.33, P=0.012). Logistic regression modelling showed that duration of one-lung ventilation (aOR: 1.20, 95% CI: 1.03-1.41, P=0.022), but not the time-weighted average FiO(2) (aOR: 1.01, 95% CI: 1.00-1.02, P=0.165), was associated with PPCs. Conclusions: Our results do not support limiting the inspired oxygen fraction for the purpose of reducing postoperative pulmonary complications in thoracic surgery involving one-lung ventilation.

Intraoperative hypoglycemia is presumed to be rare, but generalizable multicentre incidence and risk factor data for adult patients are lacking. We used a multicentre registry to characterize adults with intraoperative hypoglycemia and hypothesized that intraoperative insulin administration would be associated with hypoglycemia. We conducted a cross-sectional retrospective multicentre cohort study. We searched the Multicenter Perioperative Outcomes Group registry to identify adult patients with intraoperative hypoglycemia (glucose < 3.3 mmol·L−1 [< 60 mg·dL−1]) from 1 January 2015 to 31 December 2019. We evaluated characteristics of patients with intraoperative glucose measurements and with intraoperative hypoglycemia. Of 516,045 patients with intraoperative glucose measurements, 3,900 (0.76

BACKGROUND:Acute kidney injury (AKI) is a frequent yet understudied postoperative total joint arthroplasty complication. This study aimed to describe cardiometabolic disease co-occurrence using latent class analysis, and associated postoperative AKI risk. METHODS:This retrospective analysis examined patients ≥18 years old undergoing primary total knee or hip arthroplasties within the US Multicenter Perioperative Outcomes Group of hospitals from 2008 to 2019. AKI was defined using modified Kidney Disease: Improving Global Outcomes (KDIGO) criteria. Latent classes were constructed from eight cardiometabolic diseases including hypertension, diabetes, and coronary artery disease, excluding obesity. A mixed-effects logistic regression model was constructed for the outcome of any AKI and the exposure of interaction between latent class and obesity status adjusting for preoperative and intraoperative covariates. RESULTS:Of 81 639 cases, 4007 (4.9%) developed AKI. Patients with AKI were more commonly older and non-Hispanic Black, with more significant comorbidity. A latent class model selected three groups of cardiometabolic patterning, labelled 'hypertension only' (n=37 223), 'metabolic syndrome (MetS)' (n=36 503), and 'MetS+cardiovascular disease (CVD)' (n=7913). After adjustment, latent class/obesity interaction groups had differential risk of AKI compared with those in 'hypertension only'/non-obese. Those 'hypertension only'/obese had 1.7-fold increased odds of AKI (95% confidence interval [CI]: 1.5-2.0). Compared with 'hypertension only'/non-obese, those 'MetS+CVD'/obese had the highest odds of AKI (odds ratio 3.1, 95% CI: 2.6-3.7), whereas 'MetS+CVD'/non-obese had 2.2 times the odds of AKI (95% CI: 1.8-2.7; model area under the curve 0.76). CONCLUSIONS:The risk of postoperative AKI varies widely between patients. The current study suggests that the co-occurrence of metabolic conditions (diabetes mellitus, hypertension), with or without obesity, is a more important risk factor for acute kidney injury than individual comorbid diseases.

King Cholera and the first geospatial analysis (GA) by an anesthesiologist King Cholera had wreaked havoc on the health of Londoners for generations, but the outbreak on August 31, 1854 was extreme even by a society accustomed to mass death. Within 3 days, 127 people had died, and within a week, a majority of the population had fled the area. John Snow, considered a pioneer of anesthesia for administering Chloroform to Queen Victoria during the birth of Prince Leopold in 1853, was skeptical of the dominant miasma (airborne) theory of disease spread for Cholera.1 With the help of what would later become known as a Voronoi Diagram,2 Snow was able to visualize geospatial proximity, namely that patients who had consumed water from the Broad Street pump eventually contracted Cholera, while those who consumed water from different sources did not contract the disease. Presented with this map to show clusters of disease, the local authorities disabled the pump (Fig. 1).Figure 1: John Snow Map on mode of communication of cholera. This figure is taken from the map of the book “On the Mode of Communication of Cholera” by John Snow, published in 1854 C.F. Cheffins, Lith, Southampton. Use is in public domain. This is the first example in the literature of a dot map used to display density of cases in a geographic context and is a foundational map of medical geography and epidemiology.We leave it to historians to dispute whether or not the famous map of the outbreak was created until after the outbreak.3 Regardless, the reasoning and approach to the epidemic employed by Snow became foundational work for epidemiology and. GA Snow linked employment with certain companies to worsening of disease transmission, and by identifying geographic clusters of mortality, linked the Cholera outbreak to a polluted water source,4 an approach similar to the focus of this manuscript. The cross-disciplinary expertise, required for GA, common in the days before medical specialization and enabled by the relative paucity of development in various fields compared with today, can be reborn with the advent of Electronic Health Records (EHR) systems that make the tools of medical geography accessible to a wide variety of health care researchers and clinical subspecialties. As the story of John Snow demonstrates, the combination of GA and specialty-based medical knowledge (eg, Anesthesiology and Surgery), can serve as a method to derive insights into socioeconomic and environmental drivers of perioperative care and may lead to the design of appropriate interventions to ameliorate disparities in outcomes. Social context as key driver of health and health care disparities The focus of this manuscript is how context, location, and geography impact health and health care processes.5 Social determinants of health (SDOH) are the fundamental causes of disease, where primary causes manifest eventually as medical conditions. The central SDOH mediate the disease pathway all along; actionable mechanisms of disparity invite targeted countermeasures, once they are clearly identified and exposed. For example, food deserts, the absence of parks, and racial stress all contribute to obesity, a risk factor for obstructive sleep apnea. Night shifts with insufficient rest may lead to a car accident; language barriers and poor health literacy may hinder the provision of regional anesthesia; excessive opioid administration may trigger a respiratory arrest. In a second example, poor vision secondary to a cataract could be corrected with cataract surgery, but lack of social capital and health insurance prevent a surgical intervention. The poor vision compounds the social isolation of the patient, further enhancing cognitive decline and frailty. When a fall eventually leads to a hip fracture and hospital admission, the underlying fundamental causes of disease were poverty, lack of health insurance, social isolation, but the admission diagnosis code will be an unfortunate accident. In Table 1, a fictitious rural patient tells a harrowing story illustrating how geography, socioeconomic status, and ethnicity impact health care access and care, with the SDOHs of health tabulated as key drivers of health care processes and outcomes, with potential remedial action targeting concrete mechanisms. Our fictitious narrative of the tremendous hardships faced by rural, poor, underserved and minoritized populations is corroborated by quantitative and qualitative studies.6–8 An example of how SDOH impacts disease outcomes is in the treatment of melanoma—the skin cancer with the highest mortality in the United States. In fact, the largest Black-White disparities in cancer survival is in melanoma—an absolute survival difference of 22%.9 Retrospective database analysis have shown that older patients, non-White patients and those on Medicare/Medicaid are less likely to be offered a sentinel lymph node biopsy procedure as part of their surgical management.10,11 The sentinel lymph node biopsy is the most important predictor of prognosis for patients with melanoma and facilitates decision about adjuvant treatment as well as intensity of surveillance and therefore omitting it as part of treatment for patients in whom it is indicated is a great disservice. Patients who live in rural communities present with late-stage melanoma compared with their urban counterparts, face significant economic, financial and emotional hardships as care for complex cancers are usually hundreds of miles away from their work and local network of friends.6,12,13 The social and geographic circumstances we are born into, drive perioperative outcomes. Parental income and school district predicts scholastic achievement better than the teacher quality and is a better predictor of longevity than a person’s genetic code. Local pollution may trigger obstructive pulmonary disease.14 Occupational hazards can trigger acute and cause chronic conditions, contingent also on local mitigation strategies and safety precautions taken, which in turn depend on legal and social context. The recent US Supreme Court decision on Roe v Wade, with accompanying in state-level policy responses to the legality of abortion services, underlines the importance of legal and social context for health services offered and the circumstances under which care is offered or denied. Redlining, “cherry picking” and other racist and xenophobic policies illustrate that race and class both matter for access to quality care, and that context impacts the care we receive and the resulting outcomes.15–18 In summary, social circumstances are fundamental causes of some diseases.19 Social circumstances are distal causes of disease in contrast to the more proximal causes traditionally the focus of medical research and teaching. For example, myocardial infarct may be due to atherosclerosis, in turn the result of uncontrolled diabetes mellitus, hypertension, and hyperlipidemia, as the proximal causes. SDOH—the conditions we are born into, learn and live under, work and live in, and age in—are the distal and upstream drivers (through poverty, health literacy, insurance status, income, race/ethnicity) of the proximal causes (diabetes, obesity, hypertension)19 SDOH explains why some children have worse postsurgical outcomes after controlling for comorbidities.20 It also explains why some persons face barriers to access needed chronic pain services,21 the particular need of geriatric communities,22 or why some Black children receive less pain medication for acute appendicitis.23 SDOH facilitates postoperative recovery for some patients but fail to rescue others,24 who fall between the cracks of social networks and medical care. Table 1 - A fictitious narrative illustrating mechanisms of cancer disparities. Social determinant of health Mechanism Countermeasure Lack of transportation, high cost of travel, centralized care Reduced access to quality tertiary care Satellite clinics, public transport, decentralized health care Health literacy, social capital Delayed recognition of signs and symptoms Health curriculum, routine primary care physician visits Trust in health care system, culturally congruent care Compliance with primary prevention and best practices Community nurses, health system confidence building Poor working conditions, inflexible work hours Unable to attend to health and care needs Social policy and workers unions. family and sick leave of absence Poverty, lack of insurance, high health care cost Health care affordability Universal health insurance, living wages, social justice Unsafe working conditions Occupational exposure Occupational safety Environmental hazards Airborne exposure Environmental policy This fictitious narrative illustrates mechanism how social determinants of health (SDOH) can lead to disparities in rural cancer care, and what countermeasure might target such mechanism, [with SDOH in brackets in the text] 6–8:“Traveling back and forth for my cancer treatment is very challenging. [Lack of transportation, high cost of travel, centralized care] My name is Talamo. I live with my wife and two kids and work in one of the mines in Utah, [Unsafe working conditions, Environmental hazards] while my wife looks after the kids.Five years ago, I started having trouble breathing, and I didn’t think much of it until the dizziness began and the headaches began. [Health literacy] I needed to see a doctor, but I don’t get paid if I don’t work. [Inflexible work hours] My wife convinced me to go to a local clinic 40-minute drive away, [Centralized care] but it’s always hard taking off work and even worse getting there. [High cost of travel] Our area has no bus routes, and my car has been faulty for over a week now. [Lack of transportation], Thankfully, my buddy agreed to drive me to the clinic and back since he was off work. [High cost of travel]Three weeks was the earliest appointment time. [Social capital] While there, the doctor took my blood for tests and asked me several questions but couldn’t tell me what was wrong, or at least I did not understand what she was saying. [Health literacy] She spent very little time with me, no wonder she could not figure it out. [Culturally congruent care] I was given pills for the headache and told to get plenty of rest and drink lots of water. My grandmother died after she was seen in the same clinic. I decided not to take the pills. [Trust in healthcare system] The next day, I returned to work; I had no choice. [Inflexible work hours] But the headaches became more frequent and the dizziness almost unbearable. I could go to the ER an hour and a half away, but I can’t afford the bill. [Poverty] I have insurance, but I can’t afford the copay. [Lack of insurance, high healthcare cost] My wife’s grandfather works as a traditional healer in our small town. I saw him and he had a very long talk with me. He advised some life changes and a special tea, which made me feel somewhat better. [Culturally congruent care]Eventually, I went back to the same clinic six months later, and the doctor did a chest x-ray. He found a lump in my lungs but could not explain it precisely. This all did not make sense. [Health literacy] I was told to see a specialist a four hours’ drive from me, [Centralized care] but I don’t know what to do. [Lack of transportation, high cost of travel] I have a wife and kids to take care of. Time passed. My headache has worsened; some days, I could barely get out of bed, and now I can barely see from my right eye. [Health literacy] I woke up one morning coughing hard, struggling to breathe, and later found myself in an emergency room two hours away from our home [Centralized care] and a buddy of mine beside me. I had fainted. The doctor ran some tests and a CT scan of my chest. He also said a lump is growing on my chest, making it hard for me to breathe. He was also concerned about my other symptoms and thought this could be cancer that had spread to my brain, causing headaches, dizziness, and now a loss of vision in my right eye. I wept the entire night.” Dimensions of SDOH We propose to conceptualize SDOH in 3 dimensions, pertaining to (1) the identity (REAL for Race, Ethnicity, Affinity, and Language),25,26 (2) the social standing (SES) and (3) circumstances related to location and geography (GEO).17 Race and class both matter for health care outcomes, and so does geography.27 This article focuses on the third dimension, geography and location, but we need to consider interactions between REAL, SES, and GEO. The different dimension of SDOH (REAL, SES, GEO) can conspire insidiously to compound disparities in access to health, health care processes and outcomes: For example, poverty (SES) compounds barriers to access care in rural areas due to transportation cost.28,29 Provider racism (REAL) against a Black parturient may be accentuated by community health beliefs and health illiteracy (SES).30 Both examples also illustrate how SDOH at the individual person level (Black race, or poverty) may interact with family or community level SDOH (community health beliefs or absence of health services in rural areas).31 This conceptualization can be rendered in a polar diagram with 3 axis, REAL, SES, and GEO (Fig. 2); REAL stands for Race, Ethnicity, Affinity and Language, characteristics pertaining to identity; SES refers to socioeconomic characteristics like wealth, social or legal status, scholastic achievement, income; finally, geospatial pertains to geographic and location characteristics, for example the home, the neighborhood, the built environment, and other location characteristics. The individual is in center of the polar diagram. However, characteristics of an individual’s family, neighborhood, community, county, workplace, the state or nation they live in influence their health, access to care, health care processes and outcomes. These levels are depicted in progressively more peripheral perimeters from the individual. SDOH can be organized according to their axis and their proximity to the individual. For example, a person’s ethnicity is in the center of the polar diagram of SDOH on the REAL axis, and through xenophobia or racisms at the hand of clinicians can lead to barriers to optimal care. A food desert would be a characteristic of the built social environment of the neighborhood GEO axis and a bit further out on the polar diagram.17Figure 2: Polar dimensions of social determinants of health. We organize social determinants of health (SDOH) in this figure by spatial level and in 3 axis. SDOH concern (1) identity: REAL (Race, Ethnicity, Affinity, and Language), (2) socioeconomic status: SES (income, social capital, health literacy, etc.), and (2) the geographic domain: GEO (geographic factors, eg, food desert, availability of public transport, spatial accessibility of medical services). SDOH can act at different spatial levels, pertaining to the individual, their family, community, neighborhood, county, state, and nation. Individual mechanisms can be placed in a polar diagram to illustrate intersectionality and interaction between mechanisms leading to perioperative process and outcome disparities. Pollution is an example for a geographic SDOH acting at the state or community level. The impact of health literacy (SES factor) may span the personal, family and community level. Racism (REAL factor) may act at different spatial scales by different mechanisms: interpersonal racism drives disparities at the person-level, for example, when a clinician neglects a Black patient; structural racism may act at the state or community level, for example, through apartheid. Food desert or poor Public Transport are SDOH somewhere between the GEO and SES axis, acting more at the community than the personal level.The impact geospatial factors on perioperative access and process disparities Perioperative health care disparities can concern access to care, perioperative care processes or subsequent postoperative outcomes. Each approach (focus on access vs. process vs. outcome), has advantages and disadvantages. Below, we pick 2 examples of perioperative process disparities to illustrate the power of GA to explore actionable mechanism with a view to suggesting potential remedial action. The impact of US-census tract level SDOH on equitable antiemetic prophylaxis is the focus of the first case study,32 and the impact of state level policy on access to cataract surgery is illustrated in the second case study. SDOH impact the trajectory of cataract patients. The development of cataracts is driven by exposure to sun and professional exposure. Health literacy, social capital and connections facilitate the recognition of visual problems and a timely diagnosis, but these can be hindered by language barriers. Insurance status will drive access to ophthalmology services and the scheduling of cataract surgery. Lack of transport, social support, and poorly controlled comorbidities dues to lack of primary care can interfere with successful completion of cataract surgery and recovery. The goal is to understand processes leading to disparities with the same granular detail as any cancer pathophysiology. Such granular understanding of cause and effect, mechanism and pathways leading to perioperative disparity, would allow to target and test concrete countermeasures in a framework of continuous quality improvement as already practiced by other specialties.33,34 The added value of anesthesiologists as perioperative physicians would come from improving health equity through original investigations of mechanism of perioperative disparity embedded in continuous quality improvement efforts. GA and the boundaries of geography We begin by explaining GA as defined by the Environmental Systems Research Institute Inc. (ESRI): “The process of examining the locations, attributes, and relationships of features in spatial data through overlay and other analytical techniques in order to address a question or gain useful knowledge.” Spatial analysis extracts or creates new information from spatial data.35GA has been used in a variety of different fields, from urban planning to business development to social sciences and public health. What makes GA distinct from standard statistics is the use of spatial information within a relational database, where statistical analyses are dependent on the use of information that corresponds to a location in space, most usually corresponding to a precise latitude and longitude on planet Earth, and or to the relation of areas in space, notably polygons that represent specific areas such as a Census Tract. This point cannot be emphasized enough to the non-Geography audience. The use of any type of data with a “where” component requires special care and approaches to understand them properly and use them precisely in data analysis and interpretation. This is explained with reference to the case of the water crisis in Flint, Michigan, below. Geocoding and the perils of the Zip Code: The Water Crisis of Flint, Michigan Geocoding is a tool used extensively by Geospatial Information Systems (GIS) practitioners and is defined as the transformation of a textual address field into a latitude and longitude within a geographic information system36 that is a location on the earth’s surface. Geocoding can appear relatively straightforward, but it is important to emphasize that shortcuts in the process and a lack of understanding of GIS themselves can lead to classification bias. Typically, a GIS will use an address locator database to facilitate matching of textual addresses to known places on the earth’s surface. This gives varying levels of accuracy, which within the GIS are well described and tagged in the results of the geocoding. The use of Business Analyst extension from ESRI ArcGIS is one such toolset.37 The first attempt by the GIS is to match the address exactly based on street number, which then if that fails, it attempts to locate the address within a range of street numbers. If that is also not a match, most systems will then default to matching based on the street itself and attempt to find a median location in the street to which it can tag the address. Attempts are also made by the GIS to correct for common spelling errors, which are then given corresponding accuracy scores. Finally, if no matches are found through the above processes, a match based on Zip Code is attempted, with a location being assigned in the Zip Code centroid. Further attempts can also be made at the municipality level, though those are usually of limited utility depending on the intended application of the data. Special attention should be paid to choosing the right geographic projection for the data set, especially when concerning largely spread data sets (eg, multiple states/countries). In addition, matching to Zip Code (postal code) alone can lead to numerous problematic results. This is seen in one popular method of zip-code matching, which is called crosswalk Zip Code matching. This is a process by which a Zip Code from an address file is matched to a “Cross-walk” index file that has associations of Zip Codes with US Census Tracts. This process is problematic for several reasons, which are detailed further below. In summary, the greatest shortcoming is that Zip Codes can often occur across the boundaries of census tracts leading to misclassification bias in the results. Zip Codes have become mainstream within the popular health science media for describing the influence of geography on health of populations. The oft-cited catchphrase that “Zip Code is more important than genetic code”38 has seen a steady rise an adoption, and this has had the unfortunate side effect of leading many researchers without geography training to utilize Zip Codes as a unit of analysis for geocoding of patient data sets. It is the opinion of the authors that this method should only be used as a last resort; we will explain several examples of when this has become problematic. What are Zip Codes? Within the US, Zip Codes are designations of walking paths for delivery of mail created by the US Postal Service in the 1960s.39 During this time in the United States, discrimination by geography was commonplace and often codified, leading to phenomenon such as redlining40 where minorities were denied access to housing in predominately white neighborhoods, among other practices during the age of Jim Crow. The result is an inherent set of bias in the designation of Zip Codes and in the creation of their often-changing boundaries. Instead, Zip Code Tabulation Areas (ZCTAs) are utilized by many geographers to approximate the boundaries of Zip Codes. The issue with ZCTAs is that they also change periodically, as do Zip Codes with changes in populations and new housing developments, and so they can be an unreliable source of boundaries when comparing populations before and after a boundary change.41 The issues with Zip Codes stand in contrast to the design of US Census geographic units. Census Bureau designed their analysis units (Census Block, Census Tract, etc.) with geography in mind from the start and with a basis in the characteristics of the populations in the defined area. Contrast this with Zip Codes which are based on the convenience of mail sorting and delivery. In addition, the Census Bureau has standard processes to evaluate census tracts and other geographies on a regular basis, and utilized a well-grounded, transparent and scientific process.42 Census geographies are also much more amenable to associations based on location. That is, because they are designed to have “like within like,” if one is to geocode an address and it is found within a specific Census Tract, associating population characteristics from Census to individual characteristics for the purposes of a population health study (ie, associating social determinants data with an individual address) is statistically more valid than using Zip Codes. This incongruity is illustrated in Figure 3 below.Figure 3: Comparison of Zip Code and Census tract boundaries. This figure overlays Zip Code Tabulation Area (ZCTA) boundaries on top of Census Tract boundaries. Underlying this is Neighborhood Socioeconomic Disadvantage (NSD) which is detailed later in the text, showing red as most disadvantaged and blue as least disadvantaged. Note the different boundaries, as the ZCTA aggregates and, had that been used as the geography boundary, would have obscured heterogeneity that is evident in the Census Tracts. At the top of the figure, Zip Code 84103 is an example of this, with areas of high social disadvantage mixed together with areas of very low (blue) disadvantage within the same Zip Code.Bronx, a borough of New York City is home to the wealthiest and poorest who live only a few blocks apart. Often, Zip Codes can include both wealthy and disadvantaged neighborhoods within the same ZCTAs, while the US Census will attempt to prevent this type of grouping from occurring. The result of using ZCTA is a further potential for bias. The issue of Zip Codes is most painfully and obviously illustrated by the example of Flint, Michigan. Around 2015 Flint, Michigan began having water quality issues with its municipal tap water. Local physicians began noting elevated blood lead levels among area children, and reported this to the Michigan Department of Health and Human Services.43 The state examined the data and initially declared that there was no statistical association between the new service of Flint River water supply. There was a flaw in the data, however, as the state utilized area Zip Codes to examine the association of blood lead levels. Independent researchers recognized this and published their own study demonstrating that there was in fact an association.44 The independent research team recognized that the state had utilized Zip Codes, and upon examining a map of Flint and its water system the research team understood that the city water system was conterminous with the city boundaries. However, the Zip Codes crossed the City of Flint boundaries and included nearby towns. When the state examined blood lead levels of children by Zip Code, they included approximately 1/3 of addresses that were outside the boundaries of the Flint water system. This resulted in a massive misclassification bias within their data set, and lead them to concluding, erroneously, that there was no statistically significant elevation in blood lead levels. Fortunately, persistent community and medical community action, in collaboration with skilled geographers, were able to demonstrate that the new water system was poisoning the children in the community.43 The erosion of trust in the city and state officials has led residents to maintain a persistent level of suspicion with regards to anything said by local government officials,45 undermining the crucial trust in public health authorities, similar to other trauma experienced by minority communities, for example, the US Public Health Services (USPHS) Syphilis Study at Tuskegee. Zip Codes can have their role in GA, and they are sometimes the only information available to health research teams. However, as the above examples illustrate, they must be utilized with caution and a recognition of their limitations. Leveraging geospatial SDOH to improve perioperative health equity Continuous quality improvement framework Geospatial SDOH can be leveraged to inform public health policy, to improve equitable clinical practice processes for populations at risk, and for health systems science and health care disparity research. Obviously, in the framework of continuous quality improvement, these aforementioned activities are integrated in a process of collecting, analyzing & using data to improve the quality of health services for marginalized populations and ensure more equitable outcomes through process improvement for specific populations at risk, on an ongoing basis.46 Integration of SDOH and clinical data First, we detail how geospatial SDOH can be integrated into electronic health systems and perioperative health registries to leverage the neighborhood level information for health equity research and for equitable clinical care.47 Second, we illustrate with the 2 example use cases the power, promise, and pitfalls of GA for perioperative health systems science.48 Overview of geocoding of SDOH Geocoding SDOH from patient addresses follows this process: Starting from the patient home address at the time of service, we first perform geocoding in order to affix the textual address with a place on Earth corresponding to a specific latitude and longitude. This is then utilized in a GIS software package to match the latitude and longitude to a specific census geographic boundary.42 In terms of the National Neighborhood Data Archive (NaNDA),49 the boundary of choice is the US-Census Tract. The NaNDA database has a variety of socioeconomic (and geographic, such as park access) data sets, all coded at the US-Census Tract and ZCTA level. Matching takes place by the GIS examining each the boundaries of each census tract and determining which tract (polygon, in geographic speak) the geocode falls within. This is how the matching of NaNDA (or any geographic information) takes places within the GIS. As NaNDA data is encoded in Census Tracts (as well as ZCTA), any NaNDA data set can be matched to specific geocoded addresses. This data is then exported from the GIS as a flat file, associating the NaNDA data with each individual patient record and allowing for further statistical analysis. Likewise, data can remain in the GIS for further geostatistical analysis. Recall also, that any set of data that is geographic in nature, or has been coded into a GIS, can be associated with an individual geocode. Once this is complete, one can perform a GA to test hypotheses about social circumstances as key

Background There is insufficient prospective evidence regarding the relationship between surgical experience and prolonged opioid use and pain. The authors investigated the association of patient characteristics, surgical procedure, and perioperative anesthetic course with postoperative opioid consumption and pain 3 months postsurgery. The authors hypothesized that patient characteristics and intraoperative factors predict opioid consumption and pain 3 months postsurgery. Methods Eleven U.S. and one European institution enrolled patients scheduled for spine, open thoracic, knee, hip, or abdominal surgery, or mastectomy, in this multicenter, prospective observational study. Preoperative and postoperative data were collected using patient surveys and electronic medical records. Intraoperative data were collected from the Multicenter Perioperative Outcomes Group database. The association between postoperative opioid consumption and surgical site pain at 3 months, elicited from a telephone survey conducted at 3 months postoperatively, and demographics, psychosocial scores, pain scores, pain management, and case characteristics, was analyzed. Results Between September and October 2017, 3,505 surgical procedures met inclusion criteria. A total of 1,093 cases were included; 413 patients were lost to follow-up, leaving 680 (64%) for outcome analysis. Preoperatively, 135 (20%) patients were taking opioids. Three months postsurgery, 96 (14%) patients were taking opioids, including 23 patients (4%) who had not taken opioids preoperatively. A total of 177 patients (27%) reported surgical site pain, including 45 (13%) patients who had not reported pain preoperatively. The adjusted odds ratio for 3-month opioid use was 18.6 (credible interval, 10.3 to 34.5) for patients who had taken opioids preoperatively. The adjusted odds ratio for 3-month surgical site pain was 2.58 (1.45 to 4.4), 4.1 (1.73 to 8.9), and 2.75 (1.39 to 5.0) for patients who had site pain preoperatively, knee replacement, or spine surgery, respectively. Conclusions Preoperative opioid use was the strongest predictor of opioid use 3 months postsurgery. None of the other variables showed clinically significant association with opioid use at 3 months after surgery. Editor’s Perspective What We Already Know about This Topic What This Article Tells Us That Is New

Background: Anaesthesiologists are at high risk for exposure to SARS-CoV-2. We describe the development, implementation and efficacy of a high-fidelity simulation course for anaesthesiologists in response to this threat. Methods: This is an observational study of preparedness conferred by a simulation course. The following four knowledge gaps were identified: i) personal protective equipment (PPE) selection and donning/doffing; ii) operating room and anaesthesia machine turnover; iii) routine and difficult airway management; and iv) emergent airway management and Advanced Cardiac Life Support (ACLS) protocols. Four simulations were developed and offered to faculty, residents and Certified Registered Nurse Anaesthetists (CRNAs) at an anaesthesiology department. A survey estimated prior knowledge and assessed knowledge gained after the course. Results: Prior to the simulations, 27% and 26% of the participants estimated ability, respectively, to don and doff PPE, compared to 99% of the participants who demonstrated knowledge after the simulations: donning, odds ratio (OR) = infinity (19.6, infinity); doffing, OR = infinity (19.9, infinity). Prior to the simulations, 28% of the participants estimated knowledge specific to airway management; while after the simulations, 84% of the participants identified correct principles of airway management (OR = 9.1 [4.3, 20.2]). Prior to the simulations, 15% of the participants estimated knowledge of principles guiding emergency airway management and/or ACLS protocols, compared to 94% of the participants who correctly identified those principles after the simulations (OR = infinity [21.9, infinity]). Only 18% of clinicians estimated that they were competent prior to the simulations, while 89% of clinicians demonstrated competency on the written survey after the simulations (OR = 37 [10.8, 212]). Conclusion: Our results suggest that this simulation course effectively prepared anaesthesiology faculty, residents and staff to select appropriate PPE, and don and doff it, as well as apply best practices during airway management and ACLS protocols.

Background: Opioid analgesics continue to be prescribed after ambulatory surgery despite untoward adverse effects, risk of overdose, and association with substance use disorder. Purpose/Hypothesis: The purpose was to investigate the use of a novel system to provide scheduled and simultaneous dosing of acetaminophen, celecoxib, and pregabalin after anterior cruciate ligament reconstruction (ACLR). It was hypothesized that this system would markedly reduce pain and opioid use compared with existing best practice. Study Design: Randomized controlled trial; Level of evidence, 1. Methods: Included were 100 patients scheduled for elective, primary ACLR using allograft or hamstring tendon autograft. Selection criteria included age between 18 and 65 years and weight between 65 and 120 kg. Exclusion criteria were a known allergy to any drug used in the study or the use of opioid analgesics before surgery. Patients in the intervention group received a blister pack with scheduled, simultaneous doses of acetaminophen, celecoxib, and pregabalin; patients were also given oxycodone 5 mg as needed for breakthrough pain. Patients in the control group were prescribed ibuprofen and oxycodone 5 mg/acetaminophen 325 mg as needed for pain. The primary outcome measure was pain. Secondary outcomes were nausea, itching, and daily oxycodone use. Patients were asked to quantify their average pain at rest, nausea, and itching on an 11-point verbal scale (from 0 to 10). These data were recorded for 6 days during daily telephone contacts with patients after hospital discharge. Results: Cumulative results for 6 days showed significantly lower values in the intervention group compared with the control group for pain (median [interquartile range], 28 [14-35] vs 35 [28-41], respectively; P = .009) and oxycodone use (median [interquartile range] number of tablets, 0 [0-2] vs 8 [1.25-16], respectively; P < .001). Based on these data, the upper tolerance limits for the number of oxycodone tablets required by 90% of patients in the intervention and control groups were 8 tablets and 30 tablets, respectively. Cumulative results for nausea and itching were also significantly lower for the intervention group. Most patients in the intervention group used no opioids during recovery. Conclusion: Simultaneous dosing of 3 nonopioid analgesics resulted in reduced postoperative pain and markedly lower opioid use. Registration: NCT04015908 ( ClinicalTrials.gov identifier).

The relationship between intraoperative physiology and postoperative stroke is incompletely understood. Preliminary data suggest that either hypo- or hypercapnia coupled with reduced cerebrovascular inflow (e.g., due to hypotension) can lead to ischemia. This study tested the hypothesis that the combination of intraoperative hypotension and either hypo- or hypercarbia is associated with postoperative ischemic stroke.We conducted a retrospective, case-control study via the Multicenter Perioperative Outcomes Group. Noncardiac, nonintracranial, and nonmajor vascular surgical cases (18 yr or older) were extracted from five major academic centers between January 2004 and December 2015. Ischemic stroke cases were identified via manual chart review and matched to controls (1:4). Time and reduction below key mean arterial blood pressure thresholds (less than 55 mmHg, less than 60 mmHg, less than 65 mmHg) and outside of specific end-tidal carbon dioxide thresholds (30 mmHg or less, 35 mmHg or less, 45 mmHg or greater) were calculated based on total area under the curve. The association between stroke and total area under the curve values was then tested while adjusting for relevant confounders.In total, 1,244,881 cases were analyzed. Among the cases that screened positive for stroke (n = 1,702), 126 were confirmed and successfully matched with 500 corresponding controls. Total area under the curve was significantly associated with stroke for all thresholds tested, with the strongest combination observed with mean arterial pressure less than 55 mmHg (adjusted odds ratio per 10 mmHg-min, 1.17 [95% CI, 1.10 to 1.23], P < 0.0001) and end-tidal carbon dioxide 45 mmHg or greater (adjusted odds ratio per 10 mmHg-min, 1.11 [95% CI, 1.10 to 1.11], P < 0.0001). There was no interaction effect observed between blood pressure and carbon dioxide.Intraoperative hypotension and carbon dioxide dysregulation may each independently increase postoperative stroke risk.

Background The automated acquisition of intraoperative patient temperature data via temperature probes leads to the possibility of producing a number of artifacts related to probe positioning that may impact these probes’ utility for observational research. Objective We sought to compare the performance of two de novo algorithms for filtering such artifacts. Methods In this observational retrospective study, the intraoperative temperature data of adults who received general anesthesia for noncardiac surgery were extracted from the Multicenter Perioperative Outcomes Group registry. Two algorithms were developed and then compared to the reference standard—anesthesiologists’ manual artifact detection process. Algorithm 1 (a slope-based algorithm) was based on the linear curve fit of 3 adjacent temperature data points. Algorithm 2 (an interval-based algorithm) assessed for time gaps between contiguous temperature recordings. Sensitivity and specificity values for artifact detection were calculated for each algorithm, as were mean temperatures and areas under the curve for hypothermia (temperatures below 36 °C) for each patient, after artifact removal via each methodology. Results A total of 27,683 temperature readings from 200 anesthetic records were analyzed. The overall agreement among the anesthesiologists was 92.1%. Both algorithms had high specificity but moderate sensitivity (specificity: 99.02% for algorithm 1 vs 99.54% for algorithm 2; sensitivity: 49.13% for algorithm 1 vs 37.72% for algorithm 2; F-score: 0.65 for algorithm 1 vs 0.55 for algorithm 2). The areas under the curve for time × hypothermic temperature and the mean temperatures recorded for each case after artifact removal were similar between the algorithms and the anesthesiologists. Conclusions The tested algorithms provide an automated way to filter intraoperative temperature artifacts that closely approximates manual sorting by anesthesiologists. Our study provides evidence demonstrating the efficacy of highly generalizable artifact reduction algorithms that can be readily used by observational studies that rely on automated intraoperative data acquisition.

Background:High airway driving pressure is associated with adverse outcomes in critically ill patients receiving mechanical ventilation, but large multicentre studies investigating airway driving pressure during major surgery are lacking. We hypothesised that increased driving pressure is associated with postoperative pulmonary complications in patients undergoing major abdominal surgery. Methods:In this preregistered multicentre retrospective observational cohort study, the authors reviewed major abdominal surgical procedures in 11 hospitals from 2004 to 2018. The primary outcome was a composite of postoperative pulmonary complications, defined as postoperative pneumonia, unplanned tracheal intubation, or prolonged mechanical ventilation for more than 48 h. Associations between intraoperative dynamic driving pressure and outcomes, adjusted for patient and procedural factors, were evaluated. Results:Among 14 218 qualifying cases, 389 (2.7%) experienced postoperative pulmonary complications. After adjustment, the mean dynamic driving pressure was associated with postoperative pulmonary complications (adjusted odds ratio for every 1 cm H2O increase: 1.04; 95% confidence interval [CI], 1.02-1.06; P<0.001). Neither tidal volume nor PEEP was associated with postoperative pulmonary complications. Increased BMI, shorter height, and female sex were predictors for higher dynamic driving pressure (β=0.35, 95% CI 0.32-0.39, P<0.001; β=-0.01, 95% CI -0.02 to 0.00, P=0.005; and β=0.74, 95% CI 0.63-0.86, P<0.001, respectively). Conclusions:Dynamic airway driving pressure, but not tidal volume or PEEP, is associated with postoperative pulmonary complications in models controlling for a large number of risk predictors and covariates. Such models are capable of risk prediction applicable to individual patients.