Prescription of potentially inappropriate medications after an intensive care unit stay for acute respiratory failure.

BACKGROUND: Among survivors of critical illness, prescription of potentially inappropriate medications (PIM) at hospital discharge is thought to be an important, modifiable patient safety concern. To date, there are little empirical data evaluating this issue. RESEARCH QUESTION: The objective of this study was to determine the frequency of PIM prescribed to survivors of acute respiratory failure (ARF) at hospital discharge and explore their association with readmissions or death within 90 days of hospital discharge. STUDY DESIGN AND METHODS: Prospective multicenter cohort study of ARF survivors admitted to ICUs and discharged home. Prospective of new PIMs with a high-adverse-effect profile ("high impact") at discharge was the primary exposure. Potential inappropriateness was determined by a structured consensus process using Screening Tool of Older Persons' Prescriptions-Screening Tool to Alert to Right Treatment, Beers' criteria, and clinical context of prescriptions by a multidisciplinary team. Covariate balancing propensity score was used for the primary analysis. RESULTS: Of the 195 Addressing Post Intensive Care Syndrome-01 (APICS-01) patients, 169 (87%) had ≥1 new medications prescribed at discharge, with 154 (91.1%) prescribed with one or more high-impact (HI) medications. Patients were prescribed a median of 5 [3-7] medications, of which 3 [1-4] were HI. Twenty percent of HI medications were potentially inappropriate. Medications with significant central nervous system side-effects were most prescribed potentially inappropriately. Forty-six (30%) patients experienced readmission or death within 90 days of hospital discharge. After adjusting for prespecified covariates, the association between prescription of potentially inappropriate HI medications and the composite primary outcome did not meet the prespecified threshold for statistical significance (risk ratio: 0.54; 0.26-1.13; p = 0.095) or with the constituent endpoints: readmission (risk ratio: 0.57, 0.27-1.11) or death (0.7, 0.05-9.32). CONCLUSION: At hospital discharge, most ARF survivors are prescribed medications with a high-adverse-effect profile and approximately one-fifth are potentially inappropriate. Although prescription of such medications was not associated with 90-day readmissions and mortality, these results highlight an area for additional investigation.

Clostridium difficile and other adverse events from overprescribed antibiotics for acute upper respiratory infection.

BACKGROUND: Guidelines widely recommend avoiding antibiotics for many acute upper respiratory infections (aURIs) to avert adverse events in the absence of likely benefit. However, the extent of harm from these antibiotics remains a subject of debate and could inform patient-centered decision-making. Prior estimates finding a number needed to harm (NNH) between 8 and 10 rely on patient-reported adverse events of any severity. In this analysis, we sought to estimate adverse events by only measuring comparatively severe events that require subsequent clinical evaluation. METHODS: We constructed a retrospective cohort, including 51 million patient encounters. Using logistic regression models, we determined the adjusted odds ratio (aOR) of clinically detectable adverse events following antibiotic use compared with events among unexposed individuals with aURIs. Our outcomes included candidiasis, diarrhea, Clostridium difficile infection (CDI), and a composite outcome. FINDINGS: From our analysis, 62.4% of the population received antibiotics in an aURI encounter. Observed adverse events in the antibiotic-exposed group were 54,279 and 46,936 for diarrhea and candidiasis, respectively, yielding an aOR of 1.24 and 1.61, and an NNH of 3,126 and 1,975. Observed events of CDI in the exposed group were 30,133, and aORs of isolated CDI and combined adverse events were 1.07 and 1.30, resulting in an NNH of 17,695 and 1,150, respectively. Females were more likely to be diagnosed with any adverse event. Overall antibiotics were found to result in 5.7 additional cases of CDI per 100,000 outpatient prescriptions following an upper respiratory tract infection. INTERPRETATION: Despite higher NNH than previous methods of analysis, we find substantial iatrogenic harm associated with prescribing antibiotics in aURIs.

Pediatric Heart Network Trial of Losartan vs. Atenolol in Children and Young Adults with Marfan Syndrome: Impact on Prescription Practices.

The Pediatric Heart Network (PHN) trial showed similar efficacy of ß-blockers (BB) and angiotensin receptor blockers (ARB) for aortic root dilation in Marfan syndrome, but the impact on prescription practices is unknown. We hypothesized BB and ARB prescriptions would increase after the trial results were published (2014). Prescription data (2007-2016) were obtained from outpatient encounters (IBM Marketscan) for Marfan syndrome patients (6 months-25 years old). Excluding 2014 as a washout period, we analyzed two intervals: 2007-2013 and 2015-2016. Medication categories included BB, ARB, angiotensin converting enzyme inhibitors (ACEI), combination (BB/ARB and/or BB/ACEI), and no drug. Interrupted time-series analysis assessed immediate level change after publication and change in slope for the trend pre- and post-publication. Odds ratios (OR) and 95% confidence intervals from logistic regressions and generalized estimating equation methods accounted for correlation of prescriptions within patients. In 1499 patients (age 14.1 ± 6.1 years, 59% female) seen 2007-2013, BB trended lower [OR 0.91 (0.89, 0.93), p < 0.001] and ARB trended higher [OR 1.12 (1.07, 1.18), p < 0.001], while combination, ACEI, and no drug remained stable. This trend persisted, but was not significant, for BB [OR 0.54 (0.27, 1.08), p = 0.37] and ARB [OR 1.91 (0.55, 6.69), p = 0.31] in 2015-2016. Combination, ACEI, and no drug remained similar. In short term follow-up, changes in prescription practices following publication of the PHN trial were not statistically significant. This may be due to a change seen prior to publication with early adoption of ARBs that was maintained after confirmation of their effectiveness.

Learning from past respiratory failure patients to triage COVID-19 patient ventilator needs: A multi-institutional study.

BACKGROUND: Unlike well-established diseases that base clinical care on randomized trials, past experiences, and training, prognosis in COVID19 relies on a weaker foundation. Knowledge from other respiratory failure diseases may inform clinical decisions in this novel disease. The objective was to predict 48-hour invasive mechanical ventilation (IMV) within 48 h in patients hospitalized with COVID-19 using COVID-like diseases (CLD). METHODS: This retrospective multicenter study trained machine learning (ML) models on patients hospitalized with CLD to predict IMV within 48 h in COVID-19 patients. CLD patients were identified using diagnosis codes for bacterial pneumonia, viral pneumonia, influenza, unspecified pneumonia and acute respiratory distress syndrome (ARDS), 2008-2019. A total of 16 cohorts were constructed, including any combinations of the four diseases plus an exploratory ARDS cohort, to determine the most appropriate cohort to use. Candidate predictors included demographic and clinical parameters that were previously associated with poor COVID-19 outcomes. Model development included the implementation of logistic regression and three ensemble tree-based algorithms: decision tree, AdaBoost, and XGBoost. Models were validated in hospitalized COVID-19 patients at two healthcare systems, March 2020-July 2020. ML models were trained on CLD patients at Stanford Hospital Alliance (SHA). Models were validated on hospitalized COVID-19 patients at both SHA and Intermountain Healthcare. RESULTS: CLD training data were obtained from SHA (n = 14,030), and validation data included 444 adult COVID-19 hospitalized patients from SHA (n = 185) and Intermountain (n = 259). XGBoost was the top-performing ML model, and among the 16 CLD training cohorts, the best model achieved an area under curve (AUC) of 0.883 in the validation set. In COVID-19 patients, the prediction models exhibited moderate discrimination performance, with the best models achieving an AUC of 0.77 at SHA and 0.65 at Intermountain. The model trained on all pneumonia and influenza cohorts had the best overall performance (SHA: positive predictive value (PPV) 0.29, negative predictive value (NPV) 0.97, positive likelihood ratio (PLR) 10.7; Intermountain: PPV, 0.23, NPV 0.97, PLR 10.3). We identified important factors associated with IMV that are not traditionally considered for respiratory diseases. CONCLUSIONS: The performance of prediction models derived from CLD for 48-hour IMV in patients hospitalized with COVID-19 demonstrate high specificity and can be used as a triage tool at point of care. Novel predictors of IMV identified in COVID-19 are often overlooked in clinical practice. Lessons learned from our approach may assist other research institutes seeking to build artificial intelligence technologies for novel or rare diseases with limited data for training and validation.

Cost-Effectiveness of Transitional Care Services After Hospitalization With Heart Failure.

Background: Patients with heart failure (HF) discharged from the hospital are at high risk for death and rehospitalization. Transitional care service interventions attempt to mitigate these risks. Objective: To assess the cost-effectiveness of 3 types of postdischarge HF transitional care services and standard care. Design: Decision analytic microsimulation model. Data Sources: Randomized controlled trials, clinical registries, cohort studies, Centers for Disease Control and Prevention life tables, Centers for Medicare & Medicaid Services data, and National Inpatient Sample (Healthcare Cost and Utilization Project) data. Target Population: Patients with HF who were aged 75 years at hospital discharge. Time Horizon: Lifetime. Perspective: Health care sector. Intervention: Disease management clinics, nurse home visits (NHVs), and nurse case management. Outcome Measures: Quality-adjusted life-years (QALYs), costs, net monetary benefits, and incremental cost-effectiveness ratios (ICERs). Results of Base-Case Analysis: All 3 transitional care interventions examined were more costly and effective than standard care, with NHVs dominating the other 2 interventions. Compared with standard care, NHVs increased QALYs (2.49 vs. 2.25) and costs ($81 327 vs. $76 705), resulting in an ICER of $19 570 per QALY gained. Results of Sensitivity Analysis: Results were largely insensitive to variations in in-hospital mortality, age at baseline, or costs of rehospitalization. Probabilistic sensitivity analysis confirmed that transitional care services were preferred over standard care in nearly all 10 000 samples, at willingness-to-pay thresholds of $50 000 or more per QALY gained. Limitation: Transitional care service designs and implementations are heterogeneous, leading to uncertainty about intervention effectiveness and costs when applied in particular settings. Conclusion: In older patients with HF, transitional care services are economically attractive, with NHVs being the most cost-effective strategy in many situations. Transitional care services should become the standard of care for postdischarge management of patients with HF. Primary Funding Source: Swiss National Science Foundation, Research Council of Norway, and an Intermountain-Stanford collaboration.

Learning From Past Respiratory Infections to Predict COVID-19 Outcomes: Retrospective Study.

BACKGROUND: For the clinical care of patients with well-established diseases, randomized trials, literature, and research are supplemented with clinical judgment to understand disease prognosis and inform treatment choices. In the void created by a lack of clinical experience with COVID-19, artificial intelligence (AI) may be an important tool to bolster clinical judgment and decision making. However, a lack of clinical data restricts the design and development of such AI tools, particularly in preparation for an impending crisis or pandemic. OBJECTIVE: This study aimed to develop and test the feasibility of a "patients-like-me" framework to predict the deterioration of patients with COVID-19 using a retrospective cohort of patients with similar respiratory diseases. METHODS: Our framework used COVID-19-like cohorts to design and train AI models that were then validated on the COVID-19 population. The COVID-19-like cohorts included patients diagnosed with bacterial pneumonia, viral pneumonia, unspecified pneumonia, influenza, and acute respiratory distress syndrome (ARDS) at an academic medical center from 2008 to 2019. In total, 15 training cohorts were created using different combinations of the COVID-19-like cohorts with the ARDS cohort for exploratory purposes. In this study, two machine learning models were developed: one to predict invasive mechanical ventilation (IMV) within 48 hours for each hospitalized day, and one to predict all-cause mortality at the time of admission. Model performance was assessed using the area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, positive predictive value, and negative predictive value. We established model interpretability by calculating SHapley Additive exPlanations (SHAP) scores to identify important features. RESULTS: Compared to the COVID-19-like cohorts (n=16,509), the patients hospitalized with COVID-19 (n=159) were significantly younger, with a higher proportion of patients of Hispanic ethnicity, a lower proportion of patients with smoking history, and fewer patients with comorbidities (P<.001). Patients with COVID-19 had a lower IMV rate (15.1 versus 23.2, P=.02) and shorter time to IMV (2.9 versus 4.1 days, P<.001) compared to the COVID-19-like patients. In the COVID-19-like training data, the top models achieved excellent performance (AUROC>0.90). Validating in the COVID-19 cohort, the top-performing model for predicting IMV was the XGBoost model (AUROC=0.826) trained on the viral pneumonia cohort. Similarly, the XGBoost model trained on all 4 COVID-19-like cohorts without ARDS achieved the best performance (AUROC=0.928) in predicting mortality. Important predictors included demographic information (age), vital signs (oxygen saturation), and laboratory values (white blood cell count, cardiac troponin, albumin, etc). Our models had class imbalance, which resulted in high negative predictive values and low positive predictive values. CONCLUSIONS: We provided a feasible framework for modeling patient deterioration using existing data and AI technology to address data limitations during the onset of a novel, rapidly changing pandemic.

Breast cancer screening of pregnant and breastfeeding women with BRCA mutations.

Screening recommendations for women with BRCA mutations include annual breast MRI starting at age 25, with annual mammogram added at age 30. The median age of childbearing in the US is age 28, therefore many BRCA mutation carriers will be pregnant or breastfeeding during the time when intensive screening is most important to manage their increased breast cancer risk. Despite this critical overlap, there is little evidence to guide clinicians on the appropriate screening for women with BRCA mutations during pregnancy or breastfeeding. Hormonal shifts that occur during pregnancy, the postpartum period, and breastfeeding result in changes to the breasts that may further complicate the sensitivity and specificity of screening modalities. We explore the safety and efficacy of available breast cancer screening modalities, including clinical breast exam, mammogram, breast MRI, and ultrasound among women with BRCA mutations who are pregnant or breastfeeding, providing recommendations from the most current published literature and expert opinion.

Two new hemoglobin variants: Hb Tallahassee [α3(A1)Ser→Tyr; HBA2: c.11C>A] and Hb madison-NC [ß119(GH2)Gly→Ser; HBB: c.358G>A].

Of the 1570 reported hemoglobin (Hb) variants detected to date, 390 are α2-globin chain (some variants have yet to be identified by DNA analyses and are therefore presumed) and 827 are the result of mutations of the ß-globin chain. Due to their location on the Hb structure, only a minority of these variants result in a clinical phenotype; most are silent and are detected during routine surveillance, are found incidentally during other disease-related investigations or following newborn screening programs. In this report we discuss phenotype/genotype and molecular characteristics of two new Hb variants, both of which were clinically silent. One is an α2-globin chain variant located at codon 3 [α3(A1)Ser→Tyr; HBA2: c.11C > A] named Hb Tallahassee and the other is a ß-globin chain variant located at codon 119 [ß119(GH2)Gly→Ser; HBB: c.358G > A] called Hb Madison-NC.

Rare occurrence of Hb Lepore-Baltimore in African Americans: molecular characteristics and variations of Hb Lepores.

Hb Lepore is the hybrid hemoglobin (Hb) composed of two alpha-globin chains and two deltabeta hybrid chains and is associated with the clinical findings of thalassemia minor in its heterozygous form. Hb Lepore can be found in many ethnic groups, commonly in southern European countries, but rarely in African Americans. The first Hb Lepore case in an African-American individual was named Hb Lepore-The Bronx (Hb Lepore-Boston). Hb Lepore-Washington-Boston and Hb Lepore-Baltimore with a breakpoint of (delta50Ser/beta86Ala) were later reported. In this paper, we describe an Hb Lepore-Baltimore (delta68Leu/beta84Thr) deltabeta-fusion gene with a different breakpoint detected for the first time in an African-American female. We have used state-of-the-art technology, combining protein- and DNA-based methods, in the analysis of the hybrid hemoglobin and discuss its molecular characteristics.