Intravenous Opioid Administration During Mechanical Ventilation and Use After Hospital Discharge.

Importance: Guidelines recommend an analgesia-first strategy for sedation during mechanical ventilation, but associations between opioids provided during mechanical ventilation and posthospitalization opioid-related outcomes are unclear. Objective: To evaluate associations between an intravenous opioid dose received during mechanical ventilation and postdischarge opioid-related outcomes in medical (nonsurgical) patients. Design, Setting, and Participants: This retrospective cohort study evaluated adults receiving mechanical ventilation lasting 24 hours or more for acute respiratory failure and surviving hospitalization. Participants from 21 Kaiser Permanente Northern California hospitals from January 1, 2012, to December 31, 2019, were included. Data were analyzed from October 1, 2020, to October 31, 2023. Exposures: Terciles of median daily intravenous fentanyl equivalents during mechanical ventilation. Main Outcomes and Measures: The primary outcome was the first filled opioid prescription in 1 year after discharge. Secondary outcomes included persistent opioid use and opioid-associated complications. Secondary analyses tested for interaction between opioid doses during mechanical ventilation, prior opioid use, and posthospitalization opioid use. Estimates were based on multivariable-adjusted time-to-event analyses, with death as a competing risk, and censored for hospice or palliative care referral, rehospitalization with receipt of opioid, or loss of Kaiser Permanente plan membership. Results: The study included 6746 patients across 21 hospitals (median age, 67 years [IQR, 57-76 years]; 53.0% male). Of the participants, 3114 (46.2%) filled an opioid prescription in the year prior to admission. The median daily fentanyl equivalent during mechanical ventilation was 200 µg (IQR, 40-1000 µg), with terciles of 0 to 67 µg, more than 67 to 700 µg, and more than 700 µg. Compared with patients who did not receive opioids during mechanical ventilation (n = 1013), a higher daily opioid dose was associated with opioid prescriptions in the year after discharge (n = 2942 outcomes; tercile 1: adjusted hazard ratio [AHR], 1.00 [95% CI, 0.85-1.17], tercile 2: AHR, 1.20 [95% CI, 1.03-1.40], and tercile 3: AHR, 1.25 [95% CI, 1.07-1.47]). Higher doses of opioids during mechanical ventilation were also associated with persistent opioid use after hospitalization (n = 1410 outcomes; tercile 3 vs no opioids: odds ratio, 1.44 [95% CI, 1.14-1.83]). No interaction was observed between opioid dose during mechanical ventilation, prior opioid use, and posthospitalization opioid use. Conclusions and Relevance: In this retrospective cohort study of patients receiving mechanical ventilation, opioids administered during mechanical ventilation were associated with opioid prescriptions following hospital discharge. Additional studies to evaluate risks and benefits of strategies using lower opioid doses are warranted.

ICU Staffing in the United States.

BACKGROUND: The last national estimates of US ICU physician staffing are 25 years old and lack information about interprofessional teams. RESEARCH QUESTION: How are US adult ICUs currently staffed? STUDY DESIGN AND METHODS: We conducted a cross-sectional survey (May 4, 2022-February 2, 2023) of adult ICU physicians (targeting nurse/physician leadership) contacted using 2020 American Hospital Association (AHA) database information and, secondarily, through professional organizations. The survey included questions about interprofessional ICU staffing availability and roles at steady state (pre-COVID-19). We linked survey data to hospital data in the AHA database to create weighted national estimates by extrapolating ICU staffing data to nonrespondent hospitals based on hospital characteristics. RESULTS: The cohort consisted of 596 adult ICUs (response rates: AHA contacts: 2.1%; professional organizations: unknown) with geographic diversity and size variability (median, 20 beds; interquartile range, 12-25); most cared for mixed populations (414 [69.5%]), yet medical (55 [9.2%]), surgical (70 [11.7%]), and specialty (57 [9.6%]) ICUs were well represented. A total of 554 (93.0%) had intensivists available, with intensivists covering all patients in 75.6% of these and onsite 24 h/d in one-half (53.3% weekdays; 51.8% weekends). Of all ICUs, 69.8% had physicians-in-training and 77.7% had nurse practitioners/physician assistants. For patients on mechanical ventilation, nurse to patient ratios were 1:2 in 89.6%. Clinical pharmacists were available in 92.6%, and respiratory therapists were available in 98.8%. We estimated 85.1% (95% CI, 85.7%-84.5%) of hospitals nationally had ICUs with intensivists, 51.6% (95% CI, 50.6%-52.5%) had physicians-in-training, 72.1% (95% CI, 71.3%-72.9%) had nurse practitioners/physician assistants, 98.5% (95% CI, 98.4%-98.7%) had respiratory therapists, and 86.9% (95% CI, 86.4%-87.4%) had clinical pharmacists. For patients on mechanical ventilation, 86.4% (95% CI, 85.8%-87.0%) used 1:2 nurses/patients. INTERPRETATION: Intensivist presence in adult US ICUs has greatly increased over 25 years. Intensivists, respiratory therapists, and clinical pharmacists are commonly available, and each nurse usually provides care for two patients on mechanical ventilation. However, team composition and workload vary.

Strengthening the use of artificial intelligence within healthcare delivery organizations: balancing regulatory compliance and patient safety.

OBJECTIVES: Surface the urgent dilemma that healthcare delivery organizations (HDOs) face navigating the US Food and Drug Administration (FDA) final guidance on the use of clinical decision support (CDS) software. MATERIALS AND METHODS: We use sepsis as a case study to highlight the patient safety and regulatory compliance tradeoffs that 6129 hospitals in the United States must navigate. RESULTS: Sepsis CDS remains in broad, routine use. There is no commercially available sepsis CDS system that is FDA cleared as a medical device. There is no public disclosure of an HDO turning off sepsis CDS due to regulatory compliance concerns. And there is no public disclosure of FDA enforcement action against an HDO for using sepsis CDS that is not cleared as a medical device. DISCUSSION AND CONCLUSION: We present multiple policy interventions that would relieve the current tension to enable HDOs to utilize artificial intelligence to improve patient care while also addressing FDA concerns about product safety, efficacy, and equity.

Video-Assisted vs Robotic-Assisted Lung Lobectomies for Operating Room Resource Utilization and Patient Outcomes.

Importance: With increased use of robots, there is an inadequate understanding of minimally invasive modalities' time costs. This study evaluates the operative durations of robotic-assisted vs video-assisted lung lobectomies. Objective: To compare resource utilization, specifically operative time, between video-assisted and robotic-assisted thoracoscopic lung lobectomies. Design, Setting, and Participants: This retrospective cohort study evaluated patients aged 18 to 90 years who underwent minimally invasive (robotic-assisted or video-assisted) lung lobectomy from January 1, 2020, to December 31, 2022, with 90 days' follow-up after surgery. The study included multicenter electronic health record data from 21 hospitals within an integrated health care system in Northern California. Thoracic surgery was regionalized to 4 centers with 14 board-certified general thoracic surgeons. Exposures: Robotic-assisted or video-assisted lung lobectomy. Main Outcomes and Measures: The primary outcome was operative duration (cut to close) in minutes. Secondary outcomes were length of stay, 30-day readmission, and 90-day mortality. Comparisons between video-assisted and robotic-assisted lobectomies were generated using the Wilcoxon rank sum test for continuous variables and the χ2 test for categorical variables. The average treatment effects were estimated with augmented inverse probability treatment weighting (AIPTW). Patient and surgeon covariates were adjusted for and included patient demographics, comorbidities, and case complexity (age, sex, race and ethnicity, neighborhood deprivation index, body mass index, Charlson Comorbidity Index score, nonelective hospitalizations, emergency department visits, a validated laboratory derangement score, a validated institutional comorbidity score, a surgeon-designated complexity indicator, and a procedural code count), and a primary surgeon-specific indicator. Results: The study included 1088 patients (median age, 70.1 years [IQR, 63.3-75.8 years]; 704 [64.7%] female), of whom 446 (41.0%) underwent robotic-assisted and 642 (59.0%) underwent video-assisted lobectomy. The median unadjusted operative duration was 172.0 minutes (IQR, 128.0-226.0 minutes). After AIPTW, there was less than a 10% difference in all covariates between groups, and operative duration was a median 20.6 minutes (95% CI, 12.9-28.2 minutes; P < .001) longer for robotic-assisted compared with video-assisted lobectomies. There was no difference in adjusted secondary patient outcomes, specifically for length of stay (0.3 days; 95% CI, -0.3 to 0.8 days; P = .11) or risk of 30-day readmission (adjusted odds ratio, 1.29; 95% CI, 0.84-1.98; P = .13). The unadjusted 90-day mortality rate (1.3% [n = 14]) was too low for the AIPTW modeling process. Conclusions and Relevance: In this cohort study, there was no difference in patient outcomes between modalities, but operative duration was longer in robotic-assisted compared with video-assisted lung lobectomy. Given that this elevated operative duration is additive when applied systematically, increased consideration of appropriate patient selection for robotic-assisted lung lobectomy is needed to improve resource utilization.

Predicting stroke risk after sepsis hospitalization with new-onset atrial fibrillation.

BACKGROUND: New-onset atrial fibrillation (AF) during sepsis is common, but models designed to stratify stroke risk excluded patients with secondary AF. We assessed the predictive validity of CHA2DS2VASc scores among patients with new-onset AF during sepsis and developed a novel stroke prediction model incorporating presepsis and intrasepsis characteristics. METHODS: We included patients ≥40 years old who survived hospitalizations with sepsis and new-onset AF across 21 Kaiser Permanente Northern California hospitals from January 1, 2011 to September 30, 2017. We calculated the area under the receiver operating curve (AUC) for CHA2DS2VASc scores to predict stroke or transient ischemic attack (TIA) within 1 year after a hospitalization with new-onset AF during sepsis using Fine-Gray models with death as competing risk. We similarly derived and validated a novel model using presepsis and intrasepsis characteristics associated with 1-year stroke/TIA risk. RESULTS: Among 82,748 adults hospitalized with sepsis, 3992 with new-onset AF (median age: 80 years, median CHA2DS2VASc of 4) survived to discharge, among whom 70 (2.1%) experienced stroke or TIA outcome and 1393 (41.0%) died within 1 year of sepsis. The CHA2DS2VASc score was not predictive of stroke risk after sepsis (AUC: 0.50, 95% confidence interval [CI]: 0.48-0.52). A newly derived model among 2555 (64%) patients in the derivation set and 1437 (36%) in the validation set included 13 variables and produced an AUC of 0.61 (0.49-0.73) in derivation and 0.54 (0.43-0.65) in validation. CONCLUSION: Current models do not accurately stratify risk of stroke following new-onset AF secondary to sepsis. New tools are required to guide anticoagulation decisions following new-onset AF in sepsis.

Content of Patient Electronic Messages to Physicians in a Large Integrated System.

This quality improvement study describes the content of electronic health record messages from patients to physicians in a large integrated health care system using natural language processing algorithms.

Automated CT Analysis of Body Composition as a Frailty Biomarker in Abdominal Surgery.

Importance: Prior studies demonstrated consistent associations of low skeletal muscle mass assessed on surgical planning scans with postoperative morbidity and mortality. The increasing availability of imaging artificial intelligence enables development of more comprehensive imaging biomarkers to objectively phenotype frailty in surgical patients. Objective: To evaluate the associations of body composition scores derived from multiple skeletal muscle and adipose tissue measurements from automated segmentation of computed tomography (CT) with the Hospital Frailty Risk Score (HFRS) and adverse outcomes after abdominal surgery. Design, Setting, and Participants: This retrospective cohort study used CT imaging and electronic health record data from a random sample of adults who underwent abdominal surgery at 20 medical centers within Kaiser Permanente Northern California from January 1, 2010, to December 31, 2020. Data were analyzed from April 1, 2022, to December 1, 2023. Exposure: Body composition derived from automated analysis of multislice abdominal CT scans. Main Outcomes and Measures: The primary outcome of the study was all-cause 30-day postdischarge readmission or postoperative mortality. The secondary outcome was 30-day postoperative morbidity among patients undergoing abdominal surgery who were sampled for reporting to the National Surgical Quality Improvement Program. Results: The study included 48 444 adults; mean [SD] age at surgery was 61 (17) years, and 51% were female. Using principal component analysis, 3 body composition scores were derived: body size, muscle quantity and quality, and distribution of adiposity. Higher muscle quantity and quality scores were inversely correlated (r = -0.42; 95% CI, -0.43 to -0.41) with the HFRS and associated with a reduced risk of 30-day readmission or mortality (quartile 4 vs quartile 1: relative risk, 0.61; 95% CI, 0.56-0.67) and 30-day postoperative morbidity (quartile 4 vs quartile 1: relative risk, 0.59; 95% CI, 0.52-0.67), independent of sex, age, comorbidities, body mass index, procedure characteristics, and the HFRS. In contrast to the muscle score, scores for body size and greater subcutaneous and intermuscular vs visceral adiposity had inconsistent associations with postsurgical outcomes and were attenuated and only associated with 30-day postoperative morbidity after adjustment for the HFRS. Conclusions and Relevance: In this study, higher muscle quantity and quality scores were correlated with frailty and associated with 30-day readmission and postoperative mortality and morbidity, whereas body size and adipose tissue distribution scores were not correlated with patient frailty and had inconsistent associations with surgical outcomes. The findings suggest that assessment of muscle quantity and quality on CT can provide an objective measure of patient frailty that would not otherwise be clinically apparent and that may complement existing risk stratification tools to identify patients at high risk of mortality, morbidity, and readmission.

Validation of a Multivariable Model to Predict Suicide Attempt in a Mental Health Intake Sample.

Importance: Given that suicide rates have been increasing over the past decade and the demand for mental health care is at an all-time high, targeted prevention efforts are needed to identify individuals seeking to initiate mental health outpatient services who are at high risk for suicide. Suicide prediction models have been developed using outpatient mental health encounters, but their performance among intake appointments has not been directly examined. Objective: To assess the performance of a predictive model of suicide attempts among individuals seeking to initiate an episode of outpatient mental health care. Design, Setting, and Participants: This prognostic study tested the performance of a previously developed machine learning model designed to predict suicide attempts within 90 days of any mental health outpatient visit. All mental health intake appointments scheduled between January 1, 2012, and April 1, 2022, at Kaiser Permanente Northern California, a large integrated health care delivery system serving over 4.5 million patients, were included. Data were extracted and analyzed from August 9, 2022, to July 31, 2023. Main Outcome and Measures: Suicide attempts (including completed suicides) within 90 days of the appointment, determined by diagnostic codes and government databases. All predictors were extracted from electronic health records. Results: The study included 1 623 232 scheduled appointments from 835 616 unique patients. There were 2800 scheduled appointments (0.17%) followed by a suicide attempt within 90 days. The mean (SD) age across appointments was 39.7 (15.8) years, and most appointments were for women (1 103 184 [68.0%]). The model had an area under the receiver operating characteristic curve of 0.77 (95% CI, 0.76-0.78), an area under the precision-recall curve of 0.02 (95% CI, 0.02-0.02), an expected calibration error of 0.0012 (95% CI, 0.0011-0.0013), and sensitivities of 37.2% (95% CI, 35.5%-38.9%) and 18.8% (95% CI, 17.3%-20.2%) at specificities of 95% and 99%, respectively. The 10% of appointments at the highest risk level accounted for 48.8% (95% CI, 47.0%-50.6%) of the appointments followed by a suicide attempt. Conclusions and Relevance: In this prognostic study involving mental health intakes, a previously developed machine learning model of suicide attempts showed good overall classification performance. Implementation research is needed to determine appropriate thresholds and interventions for applying the model in an intake setting to target high-risk cases in a manner that is acceptable to patients and clinicians.

Using Electronic Health Records to Improve HIV Preexposure Prophylaxis Care: A Randomized Trial.

BACKGROUND: Preexposure prophylaxis (PrEP) use remains limited and inequitable, and strategies are needed to improve PrEP provision in primary care. METHODS: We conducted a cluster randomized trial at Kaiser Permanente, San Francisco, to evaluate the effectiveness of a clinical decision support intervention guided by an electronic health record (EHR)-based HIV risk prediction model to improve PrEP provision. Primary care providers (PCPs) were randomized to usual care or intervention, with PCPs who provide care to people with HIV balanced between arms. PCPs in the intervention arm received an EHR-based staff message with prompts to discuss HIV prevention and PrEP before upcoming in-person or video visits with patients whose predicted 3-year HIV risk was above a prespecified threshold. The main study outcome was initiation of PrEP care within 90 days, defined as PrEP discussions, referrals, or prescription fills. RESULTS: One hundred twenty-one PCPs had 5051 appointments with eligible patients (2580 usual care; 2471 intervention). There was a nonsignificant increase in initiation of PrEP care in the intervention arm (6.0% vs 4.5%, HR 1.32, 95% CI: 0.84 to 2.1). There was a significant interaction by HIV provider status, with an intervention HR of 2.59 (95% CI: 1.30 to 5.16) for HIV providers and 0.89 (95% CI: 0.59 to 1.35) for non-HIV providers (P-interaction <0.001). CONCLUSION: An EHR-based intervention guided by an HIV risk prediction model substantially increased initiation of PrEP care among patients of PCPs who also care for people with HIV. Higher-intensity interventions may be needed to improve PrEP provision among PCPs less familiar with PrEP and HIV care.

Training the next generation of delivery science researchers: 10-year experience of a post-doctoral research fellowship program within an integrated care system.

Introduction: Learning health systems require a workforce of researchers trained in the methods of identifying and overcoming barriers to effective, evidence-based care. Most existing postdoctoral training programs, such as NIH-funded postdoctoral T32 awards, support basic and epidemiological science with very limited focus on rigorous delivery science methods for improving care. In this report, we present the 10-year experience of developing and implementing a Delivery Science postdoctoral fellowship embedded within an integrated health care delivery system. Methods: In 2012, the Kaiser Permanente Northern California Division of Research designed and implemented a 2-year postdoctoral Delivery Science Fellowship research training program to foster research expertise in identifying and addressing barriers to evidence-based care within health care delivery systems. Results: Since 2014, 20 fellows have completed the program. Ten fellows had PhD-level scientific training, and 10 fellows had clinical doctorates (eg, MD, RN/PhD, PharmD). Fellowship alumni have graduated to faculty research positions at academic institutions (9), and research or clinical organizations (4). Seven alumni now hold positions in Kaiser Permanente's clinical operations or medical group (7). Conclusions: This delivery science fellowship program has succeeded in training graduates to address delivery science problems from both research and operational perspectives. In the next 10 years, additional goals of the program will be to expand its reach (eg, by developing joint research training models in collaboration with clinical fellowships) and strengthen mechanisms to support transition from fellowship to the workforce, especially for researchers from underrepresented groups.

Heterogeneity of Benefit from Earlier Time-to-Antibiotics for Sepsis.

Rationale: Shorter time-to-antibiotics improves survival from sepsis, particularly among patients in shock. There may be other subgroups for whom faster antibiotics are particularly beneficial.Objectives: Identify patient characteristics associated with greater benefit from shorter time-to-antibiotics.Methods: Observational cohort study of patients hospitalized with community-onset sepsis at 173 hospitals and treated with antimicrobials within 12 hours. We used three approaches to evaluate heterogeneity of benefit from shorter time-to-antibiotics: 1) conditional average treatment effects of shorter (⩽3 h) versus longer (>3-12 h) time-to-antibiotics on 30-day mortality using multivariable Poisson regression; 2) causal forest to identify characteristics associated with greatest benefit from shorter time-to-antibiotics; and 3) logistic regression with time-to-antibiotics modeled as a spline.Measurements and Main Results: Among 273,255 patients with community-onset sepsis, 131,094 (48.0%) received antibiotics within 3 hours. In Poisson models, shorter time-to-antibiotics was associated with greater absolute mortality reduction among patients with metastatic cancer (5.0% [95% confidence interval; CI: 4.3-5.7] vs. 0.4% [95% CI: 0.2-0.6] for patients without cancer, P < 0.001); patients with shock (7.0% [95% CI: 5.8-8.2%] vs. 2.8% [95% CI: 2.7-3.5%] for patients without shock, P = 0.005); and patients with more acute organ dysfunctions (4.8% [95% CI: 3.9-5.6%] for three or more dysfunctions vs. 0.5% [95% CI: 0.3-0.8] for one dysfunction, P < 0.001). In causal forest, metastatic cancer and shock were associated with greatest benefit from shorter time-to-antibiotics. Spline analysis confirmed differential nonlinear associations of time-to-antibiotics with mortality in patients with metastatic cancer and shock.Conclusions: In patients with community-onset sepsis, the mortality benefit of shorter time-to-antibiotics varied by patient characteristics. These findings suggest that shorter time-to-antibiotics for sepsis is particularly important among patients with cancer and/or shock.

Clinical outcomes in hospitalized plasma and platelet transfusion recipients prior to and following widespread blood donor SARS-CoV-2 infection and vaccination.

BACKGROUND: The safety of transfusion of SARS-CoV-2 antibodies in high plasma volume blood components to recipients without COVID-19 is not established. We assessed whether transfusion of plasma or platelet products during periods of increasing prevalence of blood donor SARS-CoV-2 infection and vaccination was associated with changes in outcomes in hospitalized patients without COVID-19. METHODS: We conducted a retrospective cohort study of hospitalized adults who received plasma or platelet transfusions at 21 hospitals during pre-COVID-19 (3/1/2018-2/29/2020), COVID-19 pre-vaccine (3/1/2020-2/28/2021), and COVID-19 post-vaccine (3/1/2021-8/31/2022) study periods. We used multivariable logistic regression with generalized estimating equations to adjust for demographics and comorbidities to calculate odds ratios (ORs) and 95% confidence intervals (CIs). RESULTS: Among 21,750 hospitalizations of 18,584 transfusion recipients without COVID-19, there were 697 post-transfusion thrombotic events, and oxygen requirements were increased in 1751 hospitalizations. Intensive care unit length of stay (n = 11,683) was 3 days (interquartile range 1-5), hospital mortality occurred in 3223 (14.8%), and 30-day rehospitalization in 4144 (23.7%). Comparing the pre-COVID, pre-vaccine and post-vaccine study periods, there were no trends in thromboses (OR 0.9 [95% CI 0.8, 1.1]; p = .22) or oxygen requirements (OR 1.0 [95% CI 0.9, 1.1]; p = .41). In parallel, there were no trends across study periods for ICU length of stay (p = .83), adjusted hospital mortality (OR 1.0 [95% CI 0.9-1.0]; p = .36), or 30-day rehospitalization (p = .29). DISCUSSION: Transfusion of plasma and platelet blood components collected during the pre-vaccine and post-vaccine periods of the COVID-19 pandemic was not associated with increased adverse outcomes in transfusion recipients without COVID-19.

Impact of Reducing Time-to-Antibiotics on Sepsis Mortality, Antibiotic Use, and Adverse Events.

Rationale: Shorter time-to-antibiotics is lifesaving in sepsis, but programs to hasten antibiotic delivery may increase unnecessary antibiotic use and adverse events. Objectives: We sought to estimate both the benefits and harms of shortening time-to-antibiotics for sepsis. Methods: We conducted a simulation study using a cohort of 1,559,523 hospitalized patients admitted through the emergency department with meeting two or more systemic inflammatory response syndrome criteria (2013-2018). Reasons for hospitalization were classified as septic shock, sepsis, infection, antibiotics stopped early, and never treated (no antibiotics within 48 h). We simulated the impact of a 50% reduction in time-to-antibiotics for sepsis across 12 hospital scenarios defined by sepsis prevalence (low, medium, or high) and magnitude of "spillover" antibiotic prescribing to patients without infection (low, medium, high, or very high). Outcomes included mortality and adverse events potentially attributable to antibiotics (e.g., allergy, organ dysfunction, Clostridiodes difficile infection, and culture with multidrug-resistant organism). Results: A total of 933,458 (59.9%) hospitalized patients received antimicrobial therapy within 48 hours of presentation, including 38,572 (2.5%) with septic shock, 276,082 (17.7%) with sepsis, 370,705 (23.8%) with infection, and 248,099 (15.9%) with antibiotics stopped early. A total of 199,937 (12.8%) hospitalized patients experienced an adverse event; most commonly, acute liver injury (5.6%), new MDRO (3.5%), and Clostridiodes difficile infection (1.7%). Across the scenarios, a 50% reduction in time-to-antibiotics for sepsis was associated with a median of 1 to 180 additional antibiotic-treated patients and zero to seven additional adverse events per death averted from sepsis. Conclusions: The impacts of faster time-to-antibiotics for sepsis vary markedly across simulated hospital types. However, even in the worst-case scenario, new antibiotic-associated adverse events were rare.

Development and validation of a machine learning model using electronic health records to predict trauma- and stressor-related psychiatric disorders after hospitalization with sepsis.

A significant minority of individuals develop trauma- and stressor-related disorders (TSRD) after surviving sepsis, a life-threatening immune response to infections. Accurate prediction of risk for TSRD can facilitate targeted early intervention strategies, but many existing models rely on research measures that are impractical to incorporate to standard emergency department workflows. To increase the feasibility of implementation, we developed models that predict TSRD in the year after survival from sepsis using only electronic health records from the hospitalization (n = 217,122 hospitalizations from 2012-2015). The optimal model was evaluated in a temporally independent prospective test sample (n = 128,783 hospitalizations from 2016-2017), where patients in the highest-risk decile accounted for nearly one-third of TSRD cases. Our approach demonstrates that risk for TSRD after sepsis can be stratified without additional assessment burden on clinicians and patients, which increases the likelihood of model implementation in hospital settings.

Pharmacologic and Genetic Downregulation of Proprotein Convertase Subtilisin/Kexin Type 9 and Survival From Sepsis.

OBJECTIVES: Treatments that prevent sepsis complications are needed. Circulating lipid and protein assemblies-lipoproteins play critical roles in clearing pathogens from the bloodstream. We investigated whether early inhibition of proprotein convertase subtilisin/kexin type 9 (PCSK9) may accelerate bloodstream clearance of immunogenic bacterial lipids and improve sepsis outcomes. DESIGN: Genetic and clinical epidemiology, and experimental models. SETTING: Human genetics cohorts, secondary analysis of a phase 3 randomized clinical trial enrolling patients with cardiovascular disease (Evaluation of Cardiovascular Outcomes After an Acute Coronary Syndrome During Treatment With Alirocumab [ODYSSEY OUTCOMES]; NCT01663402), and experimental murine models of sepsis. PATIENTS OR SUBJECTS: Nine human cohorts with sepsis (total n = 12,514) were assessed for an association between sepsis mortality and PCSK9 loss-of-function (LOF) variants. Incident or fatal sepsis rates were evaluated among 18,884 participants in a post hoc analysis of ODYSSEY OUTCOMES. C57BI/6J mice were used in Pseudomonas aeruginosa and Staphylococcus aureus bacteremia sepsis models, and in lipopolysaccharide-induced animal models. INTERVENTIONS: Observational human cohort studies used genetic PCSK9 LOF variants as instrumental variables. ODYSSEY OUTCOMES participants were randomized to alirocumab or placebo. Mice were administered alirocumab, a PCSK9 inhibitor, at 5 mg/kg or 25 mg/kg subcutaneously, or isotype-matched control, 48 hours prior to the induction of bacterial sepsis. Mice did not receive other treatments for sepsis. MEASUREMENTS AND MAIN RESULTS: Across human cohort studies, the effect estimate for 28-day mortality after sepsis diagnosis associated with genetic PCSK9 LOF was odds ratio = 0.86 (95% CI, 0.67-1.10; p = 0.24). A significant association was present in antibiotic-treated patients. In ODYSSEY OUTCOMES, sepsis frequency and mortality were infrequent and did not significantly differ by group, although both were numerically lower with alirocumab vs. placebo (relative risk of death from sepsis for alirocumab vs. placebo, 0.62; 95% CI, 0.32-1.20; p = 0.15). Mice treated with alirocumab had lower endotoxin levels and improved survival. CONCLUSIONS: PCSK9 inhibition may improve clinical outcomes in sepsis in preventive, pretreatment settings.

Identifying Optimal Acute Care Comparators to Inform the Evaluation of an Advanced Care at Home Pilot Program.

BACKGROUND: Hospital at Home (H@H) programs-which seek to deliver acute care within a patient's home-have become more prevalent over time. However, existing literature exhibits heterogeneity in program structure, evaluation design, and target population size, making it difficult to draw generalizable conclusions to inform future H@H program design. OBJECTIVE: The objective of this work was to develop a quality improvement evaluation strategy for a H@H program-the Kaiser Permanente Advanced Care at Home (KPACAH) program in Northern California-leveraging electronic health record data, chart review, and patient surveys to compare KPACAH patients with inpatients in traditional hospital settings. METHODS: The authors developed a 3-step recruitment workflow that used electronic health record filtering tools to generate a daily list of potential comparators, a manual chart review of potentially eligible comparator patients to assess individual clinical and social criteria, and a phone interview with patients to affirm eligibility and interest from potential comparator patients. RESULTS: This workflow successfully identified and enrolled a population of 446 comparator patients in a 5-month period who exhibited similar demographics, reasons for hospitalization, comorbidity burden, and utilization measures to patients enrolled in the KPACAH program. CONCLUSION: These initial findings provide promise for a workflow that can facilitate the identification of similar inpatients hospitalized at traditional brick and mortar facilities to enhance outcomes evaluations for the H@H programs, as well as to identify the potential volume of enrollees as the program expands.

Predictive Modeling Using Artificial Intelligence and Machine Learning Algorithms on Electronic Health Record Data: Advantages and Challenges.

The rapid adoption of electronic health record (EHR) systems in US hospitals from 2008 to 2014 produced novel data elements for analysis. Concurrent innovations in computing architecture and machine learning (ML) algorithms have made rapid consumption of health data feasible and a powerful engine for clinical innovation. In critical care research, the net convergence of these trends has resulted in an exponential increase in outcome prediction research. In the following article, we explore the history of outcome prediction in the intensive care unit (ICU), the growing use of EHR data, and the rise of artificial intelligence and ML (AI) in critical care.

Comparison of Mammography AI Algorithms with a Clinical Risk Model for 5-year Breast Cancer Risk Prediction: An Observational Study.

Background Although several clinical breast cancer risk models are used to guide screening and prevention, they have only moderate discrimination. Purpose To compare selected existing mammography artificial intelligence (AI) algorithms and the Breast Cancer Surveillance Consortium (BCSC) risk model for prediction of 5-year risk. Materials and Methods This retrospective case-cohort study included data in women with a negative screening mammographic examination (no visible evidence of cancer) in 2016, who were followed until 2021 at Kaiser Permanente Northern California. Women with prior breast cancer or a highly penetrant gene mutation were excluded. Of the 324 009 eligible women, a random subcohort was selected, regardless of cancer status, to which all additional patients with breast cancer were added. The index screening mammographic examination was used as input for five AI algorithms to generate continuous scores that were compared with the BCSC clinical risk score. Risk estimates for incident breast cancer 0 to 5 years after the initial mammographic examination were calculated using a time-dependent area under the receiver operating characteristic curve (AUC). Results The subcohort included 13 628 patients, of whom 193 had incident cancer. Incident cancers in eligible patients (additional 4391 of 324 009) were also included. For incident cancers at 0 to 5 years, the time-dependent AUC for BCSC was 0.61 (95% CI: 0.60, 0.62). AI algorithms had higher time-dependent AUCs than did BCSC, ranging from 0.63 to 0.67 (Bonferroni-adjusted P < .0016). Time-dependent AUCs for combined BCSC and AI models were slightly higher than AI alone (AI with BCSC time-dependent AUC range, 0.66-0.68; Bonferroni-adjusted P < .0016). Conclusion When using a negative screening examination, AI algorithms performed better than the BCSC risk model for predicting breast cancer risk at 0 to 5 years. Combined AI and BCSC models further improved prediction. © RSNA, 2023 Supplemental material is available for this article.

Prediction of In-hospital Mortality Among Intensive Care Unit Patients Using Modified Daily Laboratory-based Acute Physiology Score, Version 2.

BACKGROUND: Mortality prediction for intensive care unit (ICU) patients frequently relies on single ICU admission acuity measures without accounting for subsequent clinical changes. OBJECTIVE: Evaluate novel models incorporating modified admission and daily, time-updating Laboratory-based Acute Physiology Score, version 2 (LAPS2) to predict in-hospital mortality among ICU patients. RESEARCH DESIGN: Retrospective cohort study. PATIENTS: ICU patients in 5 hospitals from October 2017 through September 2019. MEASURES: We used logistic regression, penalized logistic regression, and random forest models to predict in-hospital mortality within 30 days of ICU admission using admission LAPS2 alone in patient-level and patient-day-level models, or admission and daily LAPS2 at the patient-day level. Multivariable models included patient and admission characteristics. We performed internal-external validation using 4 hospitals for training and the fifth for validation, repeating analyses for each hospital as the validation set. We assessed performance using scaled Brier scores (SBS), c -statistics, and calibration plots. RESULTS: The cohort included 13,993 patients and 107,699 ICU days. Across validation hospitals, patient-day-level models including daily LAPS2 (SBS: 0.119-0.235; c -statistic: 0.772-0.878) consistently outperformed models with admission LAPS2 alone in patient-level (SBS: 0.109-0.175; c -statistic: 0.768-0.867) and patient-day-level (SBS: 0.064-0.153; c -statistic: 0.714-0.861) models. Across all predicted mortalities, daily models were better calibrated than models with admission LAPS2 alone. CONCLUSIONS: Patient-day-level models incorporating daily, time-updating LAPS2 to predict mortality among an ICU population performs as well or better than models incorporating modified admission LAPS2 alone. The use of daily LAPS2 may offer an improved tool for clinical prognostication and risk adjustment in research in this population.