Predicting healthcare-associated infections, length of stay, and mortality with the nursing intensity of care index.

OBJECTIVES: The objectives of this study were (1) to develop and validate a simulation model to estimate daily probabilities of healthcare-associated infections (HAIs), length of stay (LOS), and mortality using time varying patient- and unit-level factors including staffing adequacy and (2) to examine whether HAI incidence varies with staffing adequacy. SETTING: The study was conducted at 2 tertiary- and quaternary-care hospitals, a pediatric acute care hospital, and a community hospital within a single New York City healthcare network. PATIENTS: All patients discharged from 2012 through 2016 (N = 562,435). METHODS: We developed a non-Markovian simulation to estimate daily conditional probabilities of bloodstream, urinary tract, surgical site, and Clostridioides difficile infection, pneumonia, length of stay, and mortality. Staffing adequacy was modeled based on total nurse staffing (care supply) and the Nursing Intensity of Care Index (care demand). We compared model performance with logistic regression, and we generated case studies to illustrate daily changes in infection risk. We also described infection incidence by unit-level staffing and patient care demand on the day of infection. RESULTS: Most model estimates fell within 95% confidence intervals of actual outcomes. The predictive power of the simulation model exceeded that of logistic regression (area under the curve [AUC], 0.852 and 0.816, respectively). HAI incidence was greatest when staffing was lowest and nursing care intensity was highest. CONCLUSIONS: This model has potential clinical utility for identifying modifiable conditions in real time, such as low staffing coupled with high care demand.

Novel Strategies for Predicting Healthcare-Associated Infections at Admission: Implications for Nursing Care.

BACKGROUND: Accurate, real-time models to predict hospital adverse events could facilitate timely and targeted interventions to improve patient outcomes. Advances in computing enable the use of supervised machine learning (SML) techniques to predict hospital-onset infections. OBJECTIVES: The purpose of this study was to trial SML methods to predict urinary tract infections (UTIs) during inpatient hospitalization at the time of admission. METHODS: In a large cohort of adult hospitalizations in three New York City acute care facilities (N = 897,344), we used two SML methods-neural networks and decision trees-to predict having a hospital-onset UTI using data available and accessible on the first day of admission at healthcare facilities in the United States. RESULTS: Performance for both neural network and decision tree models were superior compared to logistic regression methods. The decision tree model had a higher sensitivity compared to neural network, but a lower specificity. DISCUSSION: SML methods show potential for automated accurate UTI risk stratification using electronic data routinely available at admission; this could relieve nurses from the burden of having to complete and document additional risk assessment forms in the electronic medical record. Future studies should pilot and test interventions linked to the risk stratification results, such as short nursing educational modules or alerts triggered for high-risk patients.