Machine Learning-Based Early Warning Systems for Acute Care Utilization During Systemic Therapy for Cancer.

BACKGROUND: Emergency department visits and hospitalizations frequently occur during systemic therapy for cancer. We developed and evaluated a longitudinal warning system for acute care use. METHODS: Using a retrospective population-based cohort of patients who started intravenous systemic therapy for nonhematologic cancers between July 1, 2014, and June 30, 2020, we randomly separated patients into cohorts for model training, hyperparameter tuning and model selection, and system testing. Predictive features included static features, such as demographics, cancer type, and treatment regimens, and dynamic features, such as patient-reported symptoms and laboratory values. The longitudinal warning system predicted the probability of acute care utilization within 30 days after each treatment session. Machine learning systems were developed in the training and tuning cohorts and evaluated in the testing cohort. Sensitivity analyses considered feature importance, other acute care endpoints, and performance within subgroups. RESULTS: The cohort included 105,129 patients who received 1,216,385 treatment sessions. Acute care followed 182,444 (15.0%) treatments within 30 days. The ensemble model achieved an area under the receiver operating characteristic curve of 0.742 (95% CI, 0.739-0.745) and was well calibrated in the test cohort. Important predictive features included prior acute care use, treatment regimen, and laboratory tests. If the system was set to alarm approximately once every 15 treatments, 25.5% of acute care events would be preceded by an alarm, and 47.4% of patients would experience acute care after an alarm. The system underestimated risk for some treatment regimens and potentially underserved populations such as females and non-English speakers. CONCLUSIONS: Machine learning warning systems can detect patients at risk for acute care utilization, which can aid in preventive intervention and facilitate tailored treatment. Future research should address potential biases and prospectively evaluate impact after system deployment.

Association Between Familiarity of the Surgeon-Anesthesiologist Dyad and Postoperative Patient Outcomes for Complex Gastrointestinal Cancer Surgery.

Importance: The surgeon-anesthesiologist teamwork and relationship is crucial to good patient outcomes. Familiarity among work team members is associated with enhanced success in multiple fields but rarely studied in the operating room. Objective: To examine the association between surgeon-anesthesiologist dyad familiarity-as the number of times working together-with short-term postoperative outcomes for complex gastrointestinal cancer surgery. Design, Setting, and Participants: This population-based retrospective cohort study based in Ontario, Canada, included adults undergoing esophagectomy, pancreatectomy, and hepatectomy for cancer from 2007 through 2018. The data were analyzed January 1, 2007, through December 21, 2018. Exposures: Dyad familiarity captured as the annual volume of procedures of interest done by the surgeon-anesthesiologist dyad in the 4 years before the index surgery. Main Outcomes and Measures: Ninety-day major morbidity (any Clavien-Dindo grade 3 to 5). The association between exposure and outcome was examined using multivariable logistic regression. Results: Seven thousand eight hundred ninety-three patients with a median age of 65 years (66.3% men) were included. They were cared for by 737 anesthesiologists and 163 surgeons who were also included. The median surgeon-anesthesiologist dyad volume was 1 (range, 0-12.2) procedures per year. Ninety-day major morbidity occurred in 43.0% of patients. There was a linear association between dyad volume and 90-day major morbidity. After adjustment, the annual dyad volume was independently associated with lower odds of 90-day major morbidity, with an odds ratio of 0.95 (95% CI, 0.92-0.98; P = .01) for each incremental procedure per year, per dyad. The results did not change when examining 30-day major morbidity. Conclusions and Relevance: Among adults undergoing complex gastrointestinal cancer surgery, increasing familiarity of the surgeon-anesthesiologist dyad was associated with improved short-term patient outcomes. For each additional time that a unique surgeon-anesthesiologist dyad worked together, the odds of 90-day major morbidity decreased by 5%. These findings support organizing perioperative care to increase the familiarity of surgeon-anesthesiologist dyads.

Impact of Antibiotic Exposure Before Immune Checkpoint Inhibitor Treatment on Overall Survival in Older Adults With Cancer: A Population-Based Study.

PURPOSE: Antibiotic exposure before immune checkpoint inhibitor (ICI) treatment can negatively affect outcomes through alteration in the gut microbiome, but large-scale evaluations are lacking. We performed a population-level retrospective cohort study to evaluate the impact of antibiotic exposure before starting ICI on overall survival (OS). PATIENT AND METHODS: Patients with cancer, age 65 years or older, who initiated treatment with ICIs between June 2012 and October 2018 in Ontario, Canada, were identified using systemic therapy administration data. The cohort was deterministically linked to other health care databases to obtain covariates and antibiotic prescription claim data at both 1 year and 60 days before ICI therapy. Multivariable Cox models evaluated the association between exposure and OS. RESULTS: Among the 2,737 patients with cancer who received ICIs, 59% and 19% of patients received antibiotics 1 year and 60 days before ICI therapy, respectively. Median OS was 306 days. Any antibiotic exposure within 1 year before ICI was associated with worse OS (adjusted hazard ratio [aHR], 1.12; 95% CI, 1.12 to 1.23; P = .03). In antibiotic class analysis, exposure to fluoroquinolones within 1 year (aHR, 1.26; 95% CI, 1.13 to 1.40; P < .001) or 60 days before ICI (aHR, 1.20; 95% CI, 0.99 to 1.45; P = .06) was associated with worse OS, with a dose effect seen on the basis of total weeks of exposure over 1 year (aHR, 1.07 per week; 95% CI, 1.03 to 1.11; P < .001) and 60 days (aHR, 1.12 per week; 95% CI, 1.03 to 1.23; P = .01). CONCLUSION: In this population-level study, exposure to antibiotics and specifically fluoroquinolones before ICI therapy was observed to be associated with worse OS among older adults with cancer. Interventions aimed at altering the gut microbiome to boost immunogenicity may help improve outcomes for patients receiving ICIs with prior antibiotic exposure.

Variation in Anesthesiology Provider-Volume for Complex Gastrointestinal Cancer Surgery: A Population-Based Study.

OBJECTIVE: Examine between-hospital and between-anesthesiologist variation in anesthesiology provider-volume (PV) and delivery of high-volume anesthesiology care. BACKGROUND: Better outcomes for anesthesiologists with higher PV of complex gastrointestinal cancer surgery have been reported. The factors linking anesthesiology practice and organization to volume are unknown. METHODS: We identified patients undergoing elective esophagectomy, hepatectomy, and pancreatectomy using linked administrative health data sets (2007-2018). Anesthesiology PV was the annual number of procedures done by the primary anesthesiologist in the 2 years before the index surgery. High-volume anesthesiology was PV>6 procedures/year. Funnel plots to described variation in anesthesiology PV and delivery of high-volume care. Hierarchical regression models examined between-anesthesiologist and between-hospital variation in delivery of high-volume care use with variance partition coefficients (VPCs) and median odds ratios (MORs). RESULTS: Among 7893 patients cared for at 17 hospitals, funnel plots showed variation in anesthesiology PV (median ranging from 1.5, interquartile range: 1-2 to 11.5, interquartile range: 8-16) and delivery of HV care (ranging from 0% to 87%) across hospitals. After adjustment, 32% (VPC 0.32) and 16% (VPC: 0.16) of the variation were attributable to between-anesthesiologist and between-hospital differences, respectively. This translated to an anesthesiologist MOR of 4.81 (95% CI, 3.27-10.3) and hospital MOR of 3.04 (95% CI, 2.14-7.77). CONCLUSIONS: Substantial variation in anesthesiology PV and delivery of high-volume anesthesiology care existed across hospitals. The anesthesiologist and the hospital were key determinants of the variation in high-volume anesthesiology care delivery. This suggests that targeting anesthesiology structures of care could reduce variation and improve delivery of high-volume anesthesiology care.

The Association Between Hospital High-volume Anesthesiology Care and Patient Outcomes for Complex Gastrointestinal Cancer Surgery: A Population-based Study.

OBJECTIVE: To examine the association of between hospital rates of high-volume anesthesiology care and of postoperative major morbidity. BACKGROUND: Individual anesthesiology volume has been associated with individual patient outcomes for complex gastrointestinal cancer surgery. However, whether hospital-level anesthesiology care, where changes can be made, influences the outcomes of patients cared at this hospital is unknown. METHODS: We conducted a population-based retrospective cohort study of adults undergoing esophagectomy, pancreatectomy, or hepatectomy for cancer from 2007 to 2018. The exposure was hospital-level adjusted rate of high-volume anesthesiology care. The outcome was hospital-level adjusted rate of 90-day major morbidity (Clavien-Dindo grade 3-5). Scatterplots visualized the relationship between each hospital's adjusted rates of high-volume anesthesiology and major morbidity. Analyses at the hospital-year level examined the association with multivariable Poisson regression. RESULTS: For 7893 patients at 17 hospitals, the rates of high-volume anesthesiology varied from 0% to 87.6%, and of major morbidity from 38.2% to 45.4%. The scatter plot revealed a weak inverse relationship between hospital rates of high-volume anesthesiology and of major morbidity (Pearson: -0.23). The adjusted hospital rate of high-volume anesthesiology was independently associated with the adjusted hospital rate of major morbidity (rate ratio: 0.96; 95% CI, 0.95-0.98; P <0.001 for each 10% increase in the high-volume rate). CONCLUSIONS: Hospitals that provided high-volume anesthesiology care to a higher proportion of patients were associated with lower rates of 90-day major morbidity. For each additional 10% patients receiving care by a high-volume anesthesiologist at a given hospital, there was an associated reduction of 4% in that hospital's rate of major morbidity.