Differences in Utilization of Life Support and End-of-Life Care for Medical ICU Patients With Versus Without Cancer.

OBJECTIVES: To explore differences in the utilization of life support and end-of-life care between patients dying in the medical ICU with cancer compared with those without cancer. DESIGN: Retrospective review of 403 deaths or hospice transfers in the medical ICU from January 1, 2012, to June 30, 2013. SETTING: Urban tertiary care university hospital. PATIENTS: Consecutive medical ICU deaths or hospice transfers over an 18-month period. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: One hundred eighty-two patients (45%) had a diagnosis of active cancer and 221 (55%) did not. Despite similar severity of illness, there were significant differences in the use of life support and end-of-life care. Patients without cancer had longer medical ICU length of stay (median, 5 vs 4 d; p = 0.0495), used mechanical ventilation more often and for longer (83.7% vs 70.9%, p = 0.002; 4 vs 3 d, p = 0.017), and initiated dialysis more frequently (26.7% vs 14.8%; p = 0.0038). Patients without active cancer had family meetings later (median, 3 vs 2 d; p = 0.001), less frequent palliative care consultation (17.6% vs 32.4%; p = 0.0006), and took longer to transition to do not resuscitate or comfort care (median, 4 vs 3 d; p = 0.048). CONCLUSIONS: Among patients dying in the medical ICU, the diagnosis of active cancer influences the intensity of life support utilization and the quality of end-of-life care. Patients with active cancer use less life support and may receive better end-of-life care than similar patients without cancer. These differences are likely due to biases or misunderstandings about the trajectory of advanced nonmalignant disease among patients, families, and perhaps providers.

Incidence of Multidrug Resistant Infections in Emergency Department Patients with Suspected Sepsis.

BACKGROUND: Inappropriate antibiotic therapy in sepsis is associated with poor outcomes, clinicians often provide routine coverage for multidrug resistant (MDR) bacteria. However, these regimens may contribute to problems related to antibiotic overuse. To understand the incidence and related factors of multidrug resistant bacterial infections in ED patients with sepsis, we examined how often patients with sepsis in our emergency department had MDR infections. We also explored risk factors for, and outcomes from, MDR bacterial infections. METHODS: We reviewed records of patients presenting to our emergency department (ED) meeting criteria for severe sepsis or septic shock from March 2012 to July 2013. Patient demographics, comorbidities, preadmission location, and APACHE II scores were analyzed, as were clinical outcomes. RESULTS: A total of 191 episodes were examined. 108 (57%) cases were culture-positive, and of these, 23 (12.0%) had an MDR pathogen recovered. Among patients with positive cultures, MDR patients used mechanical ventilation more often 29% vs. 52% (P = 0.03) and had longer mean ICU and hospital length of stays: 4.0 vs 9.3 (P < 0.08) and 10.6 vs 20.8 (P = 0.01), respectively. We did not identify statistically significant predictors of MDR infection. CONCLUSIONS: The overall number of infections due to MDR bacteria was low, and MDR gram-negative infections were uncommon. The use of multiple empiric antibiotics for resistant gram-negative infections in the ED may be beneficial in only a small number of cases. Additionally, empiric coverage for vancomycin-resistant enterococci may need to be considered more often. Larger studies may help further elucidate the rates of MDR infections in ED patients, and identify specific risk factors to rationally guide empiric antibiotic treatment.

Increased Healthcare-Associated Infections in a Surgical Intensive Care Unit Related to Boarding Non-Surgical Patients.

Background: Hospital over-capacity often forces boarding patients outside of their designated intensive care unit (ICU). Anecdotal evidence suggested medical intensive care unit (MICU) patients boarding in the surgical intensive care unit (SICU) were responsible for increases in healthcare-associated infection (HAI) rates. We studied the effect of ICU boarding on rates of SICU HAIs. Methods: This single-center, retrospective two-year database study compared primary SICU patients (Home) to MICU patients boarding in the SICU (Boarders). Variables studied included age, gender, Acute Physiology and Chronic Health Evaluation III (APACHE III) scores, and comorbidities. Healthcare-associated infections included Clostridium difficile infection, catheter-associated urinary tract infections, central line-associated blood stream infection, and ventilator-associated pneumonia. Student t-test, Fisher exact testing, and a multivariable regression model were used to determine the significance of associations. Results: A total of 2,562 patients were included in the study; 328 (12.8%) were Boarders and 2,234 (87.2%) were Home. Univariable analysis demonstrated that Boarders were older (64.0 ± 16.9 vs. 60.2 ± 17.4), more severely ill (APACHE III score 70.5 ± 31.1 vs. 53.4 ± 21.9), more likely to have cirrhosis, coronary artery disease, and asthma/chronic obstructive pulmonary disease, but less likely to have hypertension. On univariable analysis boarding was associated with an increase HAI rate (19 HAI/1,000 patient days vs. 6.2, p < 0.001). Multivariable regression modeling demonstrated boarding status remained independently associated with HAI (odds ratio [OR] 1.83 95% confidence interval [CI] 1.02-3.27). Cost estimates demonstrated an additional cost of $83,303 per 1,000 patient days. Conclusion: The practice of hospital boarding is associated with development of HAI and increased hospital costs. Efforts at determining the cause of this increase and then reducing HAIs will improve patient care and help hospital budgets.