Background: With the advancement of technology and the provision of critical care, more critically ill patients are surviving the medical intensive care unit (ICU). Upon transfer to the general wards patients continue to face complex and ongoing medical issues that increase their risk of morbidity and mortality. This study aimed to determine patient characteristics as well as ICU and post-ICU (ward) care practices that are associated with poor short-term clinical outcomes.
Methods: A multi-site, retrospective chart review (1/1/2017 – 12/31/2017) of randomly selected patients who were admitted to the medical ICU from the emergency department and subsequently transferred to the medical ward. Exclusion criteria included patients who were admitted for acute stroke s/p TPA and those with DKA. Data extracted included patient demographics and comorbidities as well as markers of illness severity (mechanical ventilation, pressor support, midodrine use), presence of high-risk medications (benzodiazepines, antipsychotics, opioids) and care practices (indwelling Foley catheter, immobility, malnutrition). Outcome measures included: in-hospital mortality, functional decline, rapid response (RRT), ICU readmission, hospital length of stay (LOS), and hospital readmission. The Chi-Square test or Fisher’s exact were used to determine univariate associations between the variables and outcomes. A multiple logistic regression with backwards selection was used to model the outcomes.
Results: Of the 400 charts reviewed, 151 met inclusion criteria. Amongst this cohort, the median age was 64 (SD ± 19.13), 73% were female and 67% were White. The most frequent admission diagnoses to the ICU were sepsis (54%), respiratory failure (51%) and gastrointestinal bleeding (12%). With regard to patient factors, the Charlson Comorbidity Index was associated with in-hospital mortality (OR=1.6, 95% OR CI: 1.1 – 2.3, p = 0.0069) and functional decline (β=0.07, p =0.0425). Severity of illness indicators were associated with in-hospital mortality (midodrine use: OR = 7.5, 95% OR CI: 1.3 – 44.5, p = 0.0256), longer hospital LOS (mechanical ventilation: β=0.44, p=0.0024 and pressor support: β=0.57, p=0.0001) and RRTs (mechanical ventilation: OR = 12.9, OR 95% CI: 1.6 – 102.6, p = 0.0157). With regard to clinical practices, presence of indwelling Foley catheters was a significant predictor of readmission to ICU (OR = 6.0, 95% OR CI: 1.6 – 22.5, p = 0.0080). Most practices on the general wards, including immobility (bedrest), delirium (restraint use, new antipsychotic use, and constant observation), high risk medications (opioid and benzodiazepine use) and malnutrition (NPO orders), were not associated with short-term clinical outcomes.
Conclusions: While critically-ill patients are more frequently surviving the ICU, they are still at risk for significant morbidity and mortality on the medical wards. Hospitalists should be aware of patient factors as well as ICU severity of illness indicators that may be associated with poor in-hospital outcomes.