Item response theory as a feature selection and interpretation tool in the context of machine learning.

Optimizing the number and utility of features to use in a classification analysis has been the subject of many research studies. Most current models use end-classifications as part of the feature reduction process, leading to circularity in the methodology. The approach demonstrated in the present research uses item response theory (IRT) to select features independent of the end-classification results without the biased accuracies that this circularity engenders. Dichotomous and polytomous IRT models were used to analyze 30 histological breast cancer features from 569 patients using the Wisconsin Diagnostic Breast Cancer data set. Based on their characteristics, three features were selected for use in a machine learning classifier. For comparison purposes, two machine learning-based feature selection protocols were run-recursive feature elimination (RFE) and ridge regression-and the three features selected from these analyses were also used in the subsequent learning classifier. Classification results demonstrated that all three selection processes performed comparably. The non-biased nature of the IRT protocol and information provided about the specific characteristics of the features as to why they are of use in classification help to shed light on understanding which attributes of features make them suitable for use in a machine learning context.

Impact of Adherence to Ibrutinib on Clinical Outcomes in Real-World Patients With Chronic Lymphocytic Leukemia.

BACKGROUND: Chronic lymphocytic leukemia (CLL) is a B-cell neoplasm with clonal expansion of small lymphocytes. Ibrutinib, an irreversible inhibitor of Bruton tyrosine kinase (BTK), is a first-line treatment option, and recent data suggest that strict adherence is directly related to clinical outcomes. OBJECTIVES: The primary objective of this study was to quantify ibrutinib adherence rates in real-world patients with CLL on ibrutinib; secondary outcomes included progression-free survival and overall survival. METHODS: This retrospective study included subjects who were treated at a large academic medical center over approximately 5 years. Subjects were at least 18 years, diagnosed with CLL or small lymphocytic lymphoma, and treated with ibrutinib monotherapy for at least 6 months. Adherence was quantified using the medication possession ratio (MPR), which is the ratio of the sum of days' supply of medication in a period over the number of days in that period, and was based on fill history from the medical center's specialty pharmacy. RESULTS: For the 32 subjects in this study, the mean ibrutinib adherence rate was 91.7% (range, 84.4%-100%). Only 3 subjects had disease progression, and 1 death was recorded while on therapy (all with MPR < 95%); therefore, analyses of clinical outcomes were unable to be assessed due to a low number of events. There were no statistically significant differences in rates of adherence based on baseline characteristics and adverse drug events. CONCLUSION: In patients with CLL treated with ibrutinib, mean adherence was 91.7%, which is lower than rates seen in clinical trials.

Ultrasound-assisted catheter-directed thrombolysis versus systemic anticoagulation alone for submassive pulmonary embolism.

Pulmonary embolism (PE) is a significant contributor to morbidity and mortality in the United States. Catheter-directed, ultrasound-assisted thrombolysis (USAT) uses high-frequency, low-energy ultrasound waves to disaggregate uncrosslinked fibrin fibers and increase thrombus penetration of a locally delivered thrombolytic to treat an acute PE. The purpose of this study is to compare the efficacy and safety of catheter-directed USAT versus systemic anticoagulation alone in submassive PE. This was a single-center, retrospective study of patients with a diagnosis of acute submassive PE from April 4, 2014 to May 1, 2019 at a large, academic medical center. Subjects were split into two different groups based on treatment with either USAT with systemic anticoagulation or systemic anticoagulation alone. The primary outcome was the incidence of severe or life-threatening GUSTO bleeding within 72 h or until hospital discharge if sooner. A total of 130 subjects were included (n = 40 in the USAT group and n = 90 in systemic anticoagulation alone group). Significantly fewer subjects in the USAT group had an active diagnosis of cancer at the time of presentation (7.5% vs 28.9%, p = 0.006). There was no difference in severe or life-threatening GUSTO bleeding or any component of the GUSTO bleeding definitions. Administration of USAT with systemic anticoagulation was well-tolerated when compared to systemic anticoagulation alone, but bias may have led to selection of patients for USAT with a lower bleeding risk and higher functional status at baseline.

Multicenter validation of the Quality of Trauma Care Patient-Reported Experience Measure (QTAC-PREM).

BACKGROUND: Incorporating patient and family perspectives into injury care quality assessment is a necessary part of comprehensive quality improvement. However, tools to measure patient and family perspectives of injury care are lacking. Therefore, our objective was to assess the psychometric properties of the Quality of Trauma Care Patient-Reported Experience Measure (QTAC-PREM), the first measure developed to assess patient experiences with overall injury care. METHODS: We conducted a prospective multicenter cohort study of adult injury patients recruited from three trauma centers. Patients or surrogates completed an acute care survey measure in the hospital and a post-acute care survey measure after hospital discharge. RESULTS: Four hundred participants (78%) completed the acute care measure, and 207 (59%) completed the post-acute care measure. We identified three subscales on the acute measure and two subscales on the post-acute measure. All subscales and items had evidence of construct validity. Four subscales had good internal consistency, and three were independent predictors of participants' overall ratings of injury care quality. The majority of items demonstrated suitable test-retest reliability. Comparison of QTAC-PREM scores with those of an existing patient experience tool, the Hospital version of the Consumer Assessment of Healthcare Providers and Systems (HCAHPS), demonstrated evidence of appropriate divergent and convergent validity. CONCLUSION: This study demonstrates that the QTAC-PREM is feasible to implement at trauma centers and provides evidence of validity and reliability. The tool may be useful to incorporate patient perspectives into trauma care quality measurement and improvement.

Facilitating a just and trusting culture.

PURPOSE: The purpose of this paper is to identify managerial and organizational characteristics and behaviors that facilitate the fostering of a just and trusting culture within the healthcare system. DESIGN/METHODOLOGY/APPROACH: Two studies were conducted. The initial qualitative one was used to identify themes based on interviews with health care workers that facilitate a just and trusting culture. The quantitative one used a policy-capturing design to determine which factors were most likely to predict outcomes of manager and organizational trust. FINDINGS: The factors of violation type (ability vs integrity), providing an explanation or not, blame vs no blame by manager, and blame vs no blame by organization were all significant predictors of perceptions of trust. RESEARCH LIMITATIONS/IMPLICATIONS: Limitations to the generalizability of findings included both a small and non-representative sample from one health care region. PRACTICAL IMPLICATIONS: The present findings can be useful in developing training systems for managers and organizational executive teams for managing medical error events in a manner that will help develop a just and trusting culture. SOCIAL IMPLICATIONS: A just and trusting culture should enhance the likelihood of reporting medical errors. Improved reporting, in turn, should enhance patient safety. ORIGINALITY/VALUE: This is the first field study experimentally manipulating aspects of organizational trust within the health care sector. The use of policy-capturing is a unique feature that sheds light into the decision-making of health care workers as to the efficaciousness of particular managerial and organizational characteristics that impact a just and trusting culture.

Determinants of adverse events in hospitals--the potential role of patient safety culture.

This study investigated two contributing factors in predicting adverse events in hospital settings. We approached this issue using the statistical procedure of hierarchical linear modeling to test for multilevel relationships. We found that the resource intensity of the presenting case was related to the severity level of negative incidents in hospital settings in a large metropolitan center. More important, we found that a more positive culture of patient safety within hospital units was related to lower incident severity. More than 8,000 admissions within 40 different units in three hospitals were included in the analyses.