Development and internal validation of machine learning-based models and external validation of existing risk scores for outcome prediction in patients with ischaemic stroke.

Aims: We developed new machine learning (ML) models and externally validated existing statistical models [ischaemic stroke predictive risk score (iScore) and totalled health risks in vascular events (THRIVE) scores] for predicting the composite of recurrent stroke or all-cause mortality at 90 days and at 3 years after hospitalization for first acute ischaemic stroke (AIS). Methods and results: In adults hospitalized with AIS from January 2005 to November 2016, with follow-up until November 2019, we developed three ML models [random forest (RF), support vector machine (SVM), and extreme gradient boosting (XGBOOST)] and externally validated the iScore and THRIVE scores for predicting the composite outcomes after AIS hospitalization, using data from 721 patients and 90 potential predictor variables. At 90 days and 3 years, 11 and 34% of patients, respectively, reached the composite outcome. For the 90-day prediction, the area under the receiver operating characteristic curve (AUC) was 0.779 for RF, 0.771 for SVM, 0.772 for XGBOOST, 0.720 for iScore, and 0.664 for THRIVE. For 3-year prediction, the AUC was 0.743 for RF, 0.777 for SVM, 0.773 for XGBOOST, 0.710 for iScore, and 0.675 for THRIVE. Conclusion: The study provided three ML-based predictive models that achieved good discrimination and clinical usefulness in outcome prediction after AIS and broadened the application of the iScore and THRIVE scoring system for long-term outcome prediction. Our findings warrant comparative analyses of ML and existing statistical method-based risk prediction tools for outcome prediction after AIS in new data sets.

An artificial intelligence approach for predicting death or organ failure after hospitalization for COVID-19: development of a novel risk prediction tool and comparisons with ISARIC-4C, CURB-65, qSOFA, and MEWS scoring systems.

BACKGROUND: We applied machine learning (ML) algorithms to generate a risk prediction tool [Collaboration for Risk Evaluation in COVID-19 (CORE-COVID-19)] for predicting the composite of 30-day endotracheal intubation, intravenous administration of vasopressors, or death after COVID-19 hospitalization and compared it with the existing risk scores. METHODS: This is a retrospective study of adults hospitalized with COVID-19 from March 2020 to February 2021. Patients, each with 92 variables, and one composite outcome underwent feature selection process to identify the most predictive variables. Selected variables were modeled to build four ML algorithms (artificial neural network, support vector machine, gradient boosting machine, and Logistic regression) and an ensemble model to generate a CORE-COVID-19 model to predict the composite outcome and compared with existing risk prediction scores. The net benefit for clinical use of each model was assessed by decision curve analysis. RESULTS: Of 1796 patients, 278 (15%) patients reached primary outcome. Six most predictive features were identified. Four ML algorithms achieved comparable discrimination (P > 0.827) with c-statistics ranged 0.849-0.856, calibration slopes 0.911-1.173, and Hosmer-Lemeshow P > 0.141 in validation dataset. These 6-variable fitted CORE-COVID-19 model revealed a c-statistic of 0.880, which was significantly (P < 0.04) higher than ISARIC-4C (0.751), CURB-65 (0.735), qSOFA (0.676), and MEWS (0.674) for outcome prediction. The net benefit of the CORE-COVID-19 model was greater than that of the existing risk scores. CONCLUSION: The CORE-COVID-19 model accurately assigned 88% of patients who potentially progressed to 30-day composite events and revealed improved performance over existing risk scores, indicating its potential utility in clinical practice.

Characteristics, Treatment Patterns, and Clinical Outcomes After Heart Failure Hospitalizations During the COVID-19 Pandemic, March to October 2020.

OBJECTIVE: To compare clinical characteristics, treatment patterns, and 30-day all-cause readmission and mortality between patients hospitalized for heart failure (HF) before and during the coronavirus disease 2019 (COVID-19) pandemic. PATIENTS AND METHODS: The study was conducted at 16 hospitals across 3 geographically dispersed US states. The study included 6769 adults (mean age, 74 years; 56% [5033 of 8989] men) with cumulative 8989 HF hospitalizations: 2341 hospitalizations during the COVID-19 pandemic (March 1 through October 30, 2020) and 6648 in the pre-COVID-19 (October 1, 2018, through February 28, 2020) comparator group. We used Poisson regression, Kaplan-Meier estimates, multivariable logistic, and Cox regression analysis to determine whether prespecified study outcomes varied by time frames. RESULTS: The adjusted 30-day readmission rate decreased from 13.1% (872 of 6648) in the pre-COVID-19 period to 10.0% (234 of 2341) in the COVID-19 pandemic period (relative risk reduction, 23%; hazard ratio, 0.77; 95% CI, 0.66 to 0.89). Conversely, all-cause mortality increased from 9.7% (645 of 6648) in the pre-COVID-19 period to 11.3% (264 of 2341) in the COVID-19 pandemic period (relative risk increase, 16%; number of admissions needed for one additional death, 62.5; hazard ratio, 1.19; 95% CI, 1.02 to 1.39). Despite significant differences in rates of index hospitalization, readmission, and mortality across the study time frames, the disease severity, HF subtypes, and treatment patterns remained unchanged (P>0.05). CONCLUSION: The findings of this large tristate multicenter cohort study of HF hospitalizations suggest lower rates of index hospitalizations and 30-day readmissions but higher incidence of 30-day mortality with broadly similar use of HF medication, surgical interventions, and devices during the COVID-19 pandemic compared with the pre-COVID-19 time frame.

Hyperlipidaemia and mortality among patients hospitalised with pneumonia: retrospective cohort and propensity score matched study.

OBJECTIVE: To characterise the potential association of hyperlipidaemia (HLP) versus no HLP with all-cause mortality among patients hospitalised for pneumonia. DESIGN: Propensity score matched retrospective study. PARTICIPANTS: The study cohort consisted of consecutive 8553 adults hospitalised at a large academic centre with a discharge diagnosis of pneumonia from 1996 through 2015, followed until death or end of the study period, 17 August 2017. OUTCOMES: The outcome was HR for mortality at 28 days and in the long term in patients with pneumonia with concurrent HLP compared with those with no HLP. We first constructed multivariable Cox proportional regression models to estimate the association between concurrent HLP versus no HLP and mortality after pneumonia hospitalisation for the entire cohort. We then identified 1879 patients with pneumonia with concurrent HLP and propensity score matched in a 1:1 ratio to 1879 patients with no HLP to minimise the imbalance from measured covariates for further analysis. RESULTS: Among 8553 unmatched patients with pneumonia, concurrent HLP versus no HLP was independently associated with lower mortality at 28 days (HR 0.52, 95% CI 0.41 to 0.66) and at a median follow-up of 3.9 years (HR 0.75, 95% CI 0.70 to 0.80). The risk difference in mortality was consistent between 1879 propensity score matched pairs both at 28 days (HR 0.65, 95% CI 0.49 to 0.86) and at a median follow-up of 4 years (HR 0.88, 95% CI 0.81 to 0.96). In the subgroup of patients with clinically measured low-density lipoprotein cholesterol (LDL-C), graded inverse associations between LDL-C levels and mortality were found in both unmatched and matched cohorts. CONCLUSIONS: Among hospitalised patients with pneumonia, a diagnosis of HLP is protective against both short-term and long-term risk of death after adjustment for other major contributors to mortality in both unmatched and propensity score matched cohorts. These findings should be further investigated.

Initial and subsequent 3-year cost after hospitalization for first acute ischemic stroke and intracerebral hemorrhage.

AIMS: To examine 1) the major drivers of index hospitalization and 3-year post-acute follow-up care, 2) cost for rehabilitation and homecare, and 3) indirect cost from lost productivity after acute ischemic stroke (AIS) and intracerebral hemorrhage (ICH). METHODS: Retrospective study of adults hospitalized with AIS (n = 811) and ICH (N = 145) between 2003 and 2014. Direct costs standardized to Medicare reimbursement rates were captured for hospitalization and 3-year follow-up or death. Adjusted cost estimates were assessed using generalized linear modeling with gamma distribution. Costs for rehabilitation, home healthcare, and lost productivity were assessed using sets of cost captured through literature review. RESULTS: Calculated as mean cost per person: hospitalization $18,154 for AIS and $24,077 for ICH; monthly 3-year aggregate $5138 for AIS and $8172 for ICH; 3-year inpatient rehabilitation $4185 for AIS and $4196 for ICH; homecare $19,728 for AIS and $14,487 for ICH; indirect cost from lost productivity $77,078 for AIS and $56,601 for ICH. Age < 55 years, being non-white, and stroke severity were strongly associated with greater hospitalization cost for AIS and ICH. Hyperlipidemia incurred lower while cancer, coronary artery disease, asthma/chronic obstructive pulmonary disease, heart failure, and anemia incurred higher 3-year aggregate cost for AIS. Cancer and diabetes mellitus incurred higher 3-year aggregate cost for ICH. CONCLUSIONS: We provide estimates of direct and indirect costs incurred for acute and continuing post-acute care through a 3-year follow-up period after first-ever AIS and ICH with important comparisons for predictors between index hospitalization and 3-year post-stroke costs.

Reining in Sternal Wound Infections: The Achilles' Heel of Bilateral Internal Thoracic Artery Grafting.

Background: Although the survival advantage of bilateral internal thoracic artery grafting (BITA) is well known in patients undergoing coronary artery bypass grafting (CABG), this technique has not been widely adopted. This is mainly because of the increased risk of deep sternal wound infections (DSWI) associated with its use. However, in recent years the overall risk of DSWI has decreased. This is mainly because of strategies that have been adopted to decrease the risk of these infections in patients undergoing CABG. Conclusion: In this review we identified DSWI preventive strategies and described them in detail so that their use by surgeons can be increased. This would minimize the risk of DSWI after BITA grafting and maximize the use of this highly effective surgical technique.

Minimally Invasive Approaches to Surgical Aortic Valve Replacement: A Meta-Analysis.

BACKGROUND: Limited data exist studying the outcomes of the 2 minimally invasive aortic valve replacement (AVR) strategies-mini-sternotomy (AVR-st) and right anterior thoracotomy (AVR-th). We conducted an indirect meta-analysis to compare the outcomes of these minimally invasive approaches with each other and with conventional AVR (cAVR). METHODS: We Searched Medline, PubMed, Embase, and Web of Science in December 2017 for studies comparing AVR-st, AVR-th, and cAVR. Clinical outcomes were compared between cohorts with inverse weighted random effects modeling. Endpoints studied included hospital mortality, stroke, atrial fibrillation, cardiopulmonary bypass (CPB) time, and length of stay. RESULTS: A total of 19 studies (>10,000 pooled patients) met the inclusion criteria. Mortality (p = 0.06) and stroke (p = 0.15) were comparable between minimally invasive and conventional AVR. CPB times were longer with AVR-th versus cAVR (12.4 minutes [range, 5 to 19]; p < 0.01). In the AVR-th cohort, CPB duration was weakly inversely related to study size (p = 0.06). Atrial fibrillation was much less after AVR-th (odds ratio 0.47 [0.35 to 0.63]; p < 0.001). Hospital stay was significantly lower after minimally invasive surgery (0.8 [0.4 to 1.3] days; p < 0.01). AVR-th patients were dismissed 2.1 (1.6 to 2.7) days earlier than cAVR patients. CONCLUSIONS: Minimally invasive approaches to AVR yield excellent outcomes in high-volume centers. They reduce hospital stay and incidence of postoperative atrial fibrillation, and therefore should be considered in patients undergoing AVR. The operative approach should be selected according to surgeon's technical expertise and what is best for specific patient profile, however.

Diagnosis of coronary artery fistulas: clinical aspects and brief review of the literature.

An 86-year-old woman with history of hypertension, diabetes, hyperlipidemia, and coronary artery disease was admitted with new-onset congestive heart failure and pneumonia. She underwent coronary angiography after suffering a non-ST elevation myocardial infarction. There was severe coronary artery disease in addition to coronary artery-pulmonary fistulas involving proximal right coronary artery and a branch of left main coronary artery. Coronary artery fistula (CAF) is an abnormal communication between one or more coronary arteries and great vessels or a cardiac chamber. We reviewed 15 cases of CAFs published in PubMed and studied the clinical features of CAFs.

An exploratory study of the use of advance directives by US oncologists.

PURPOSE: This study sought to determine the rate of advance directive completion among US oncologists and factors influencing such a decision. METHODS: We surveyed 7590 members of the American Society of Clinical Oncology using a web-based questionnaire. RESULTS: The response rate was 8.1%. Most respondents (59%) had completed at least 1 document: 9% living will, 9% power of attorney for health care, and 41% both. Respondents who were older, men, married, with children, working in the community setting, radiation oncologists, and practicing general oncology were more likely than their counterparts to have an advance directive. Among those who had one, 95% and 36% had discussed their wishes with their loved ones and health care providers, respectively. Factors including experience at work, spouse, children, family, and religion had the most influence on respondents' decision. The majority of those without an advance directive reported either no reason or lack of time. Those who had them were more likely to report having a comprehensive review of their wishes with those closest to them, being more knowledgeable, having more routine discussions with their patients, and being more comfortable helping their patients complete one. CONCLUSION: Only about half of US oncologists who responded to our survey have completed an advance directive.