Mentor

Gadd, Holly

Document Type

Dissertation

Publication Date

2020

Abstract

Objective: Invasive mechanical ventilation (IMV) is lifesaving and one of the most common interventions implemented in the intensive care unit (ICU). More than half of the patients in the ICU require IMV within the first 24 hours after ICU admission. This project aimed to evaluate and predict the mortality rate of hospitalized patients on IMV by examining their risk factors, such as patient demographic characteristics, disease status, social environment, and discharge status.

Methods: The National Inpatient Sample (NIS) database 2016 was used to identify patients requiring IMV. Mortality was the dependent variable. Independent variables had four major categories. First, patients’ demographic characteristics which contained two categories, non-modifiable risk factors (age, age groups, gender, and race), and socioeconomic status (income and primary expected payer). Second, disease status including principal physical diagnosis, duration of mechanical ventilation, and modifiable risk factors (malnutrition, obesity, ETOH abuse, nicotine abuse, and opioids abuse). Third, social environment which included admission status and hospital stratum. Admission status included weekdays versus weekend admission, indication of ED service, and elective versus non-elective services. The final category was discharge status, which included total cost of hospital service and length of stay. Total of 601 patients were selected.

Results: Four variables which included found to be independent predictors of mortality among patients on ventilators. These were age, pulmonary embolism (PE), Pneumonia (PNA), and indication of emergency department (ED). The logistic regression model was statistically significance, χ2(7) = 53.59, p = 0.00 (p < 0.005). The model explained 14.5% (Nagelkerke R2) of the variance in those who died during hospitalization and correctly classified 83.7% of cases. Sensitivity was 3.1%, specificity was 99.2%, positive predictive value was 42.9% and negative predictive value was 84.2%. Principal physical diagnosis of pulmonary embolism had 7.02 times (OR=7.02, CI =1.89-26.07, p = 0.00) higher odds to cause death during hospitalization than respiratory failure. Furthermore, principal physical diagnosis of pneumonia had 2.38 times (OR= 2.38, CI =1.15-4.92, p = 0.02) higher odds to cause death during hospitalization than respiratory failure. Increasing age was associated with an increased likelihood of death during hospitalization among patients on IMV (OR= 1.04, CI =1.03 – 1.06, p = 0.00). No indication of ED service was associated with a reduction (OR=0.53, CI =0.29 – 0.95, p = 0.03) in the likelihood of death during hospitalization compare to indication of ED service.

Conclusion: Age, age group, race, primary expected payer, principal physical diagnosis, length of IMV, and modifiable risk factors were significantly associated with in-hospital mortality among patients on IMV. Only age, PE, pneumonia, and indication of ED service could predict in-hospital mortality among patients on IMV. A clear understanding of these risk factors is integral for an appropriate and timely management and further to improve patients’ outcomes.

Download

Share

COinS