Background
Hospital-acquired pressure injuries are a serious problem among critical care patients. Some can be prevented by using measures such as specialty beds, which are not feasible for every patient because of costs. However, decisions about which patient would benefit most from a specialty bed are difficult because results of existing tools to determine risk for pressure injury indicate that most critical care patients are at high risk.
Objective
To develop a model for predicting development of pressure injuries among surgical critical care patients.
Methods
Data from electronic health records were divided into training (67%) and testing (33%) data sets, and a model was developed by using a random forest algorithm via the R package “randomforest.”
Results
Among a sample of 6376 patients, hospital-acquired pressure injuries of stage 1 or greater (outcome variable 1) developed in 516 patients (8.1%) and injuries of stage 2 or greater (outcome variable 2) developed in 257 (4.0%). Random forest models were developed to predict stage 1 and greater and stage 2 and greater injuries by using the testing set to evaluate classifier performance. The area under the receiver operating characteristic curve for both models was 0.79.
Conclusion
This machine-learning approach differs from other available models because it does not require clinicians to input information into a tool (eg, the Braden Scale). Rather, it uses information readily available in electronic health records. Next steps include testing in an independent sample and then calibration to optimize specificity. (American Journal of Critical Care. 2018; 27:461–468)
IMPORTANCE:It is not know if hospital-level extracorporeal cardiopulmonary resuscitation (ECPR) case volume, or postcannulation clinical management associate with survival outcomes.
OBJECTIVES:To describe variation in postresuscitation management practices, and annual hospital-level case volume, for patients who receive ECPR and to determine associations between these management practices and hospital survival.DESIGN: Observational cohort study using case-mix adjusted survival analysis.
SETTING AND PARTICIPANTS:Adult patients greater than or equal to 18 years old who received ECPR from the Extracorporeal Life Support Organization Registry from 2008 to 2019.
MAIN OUTCOMES AND MEASURES:Generalized estimating equation logistic regression was used to determine factors associated with hospital survival, accounting for clustering by center. Factors analyzed included specific clinical management interventions after starting extracorporeal membrane oxygenation (ECMO) including coronary angiography, mechanical unloading of the left ventricle on ECMO (with additional placement of a peripheral ventricular assist device, intra-aortic balloon pump, or surgical vent), placement of an arterial perfusion catheter distal to the arterial return cannula (to mitigate leg ischemia); potentially modifiable on-ECMO hemodynamics (arterial pulsatility, mean arterial pressure, ECMO flow); plus hospital-level annual case volume for adult ECPR.
RESULTS:Case-mix adjusted patient-level management practices varied widely across individual hospitals. We analyzed 7,488 adults (29% survival); median age 55 (interquartile range, 44-64), 68% of whom were male. Adjusted hospital survival on ECMO was associated with mechanical unloading of the left ventricle (odds ratio [OR], 1.3; 95% CI, 1.08-1.55; p = 0.005), performance of coronary angiography (OR, 1.34; 95% CI, 1.11-1.61; p = 0.002), and placement of an arterial perfusion catheter distal to the return cannula (OR, 1.39; 95% CI, 1.05-1.84; p = 0.022). Survival varied by 44% across hospitals after case-mix adjustment and was higher at centers that perform more than 12 ECPR cases/yr (OR, 1.23; 95% CI, 1.04-1.45; p = 0.015) versus medium-and low-volume centers.
CONCLUSIONS AND RELEVANCE:Modifiable ECMO management strategies and annual case volume vary across hospitals, appear to be associated with survival and should be the focus of future research to test if these hypothesisgenerating associations are causal in nature.
This study evaluated oral medication adherence among adolescents and young adults (AYAs) with cancer during a trial of a smartphone-based medication reminder application (app). Methods: Twenty-three AYAs receiving at least one prescribed, scheduled oral medication related to their outpatient cancer treatment participated in this 12-week single-group interrupted time series longitudinal design study. Baseline oral medication adherence was monitored using electronic monitoring caps for 4 weeks. Participants then used a medication reminder app and continued to have their oral medication adherence monitored for 8 weeks. Participants completed an electronically administered weekly survey addressing perceived adherence and reasons for nonadherence. Results: Four adherence phenotypes were identified using visual graphical analysis of individual participants' weekly adherence: (1) high adherence during the preintervention and intervention periods (n = 13), (2) low preintervention adherence and improved adherence during the intervention period (n = 3), (3) low adherence during both periods (n = 6), and (4) high preintervention adherence and low adherence during the intervention period (n = 1). Growth curve models did not show significant changes in adherence by preintervention versus intervention trajectories (p > 0.05); however, the variance in adherence during the intervention narrowed for more highly adherent AYAs. ''Forgetfulness'' was the most frequently reported reason for nonadherence. Conclusion: Although overall adherence did not improve following use of the app, the variance decreased for more highly adherent participants. Additional or alternative interventions are needed for AYAs with persistently poor adherence. Assessment of adherence patterns may support individualized recommendation of tailored interventions.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.