Contact isolation of infected or colonised hospitalised patients is instrumental to interrupting multidrug-resistant organism (MDRO) cross-transmission. Many studies suggest an increased rate of adverse events associated with isolation. We aimed to compare isolated to non-isolated patients in intensive care units (ICUs) for the occurrence of adverse events and medical errors.
Methods: We used the large database of the Iatroref III study that included consecutive patients from three ICUs to compare the occurrence of pre-defined medical errors and adverse events among isolated vs. non-isolated patients. A subdistribution hazard regression model with careful adjustment on confounding factors was used to assess the effect of patient isolation on the occurrence of medical errors and adverse events.
Results: Two centres of the Iatroref III study were eligible, an 18-bed and a 10-bed ICU (nurse-to-bed ratio 2.8 and 2.5, respectively), with a total of 1,221 patients. After exclusion of the neutropenic and graft transplant patients, a total of 170 isolated patients were compared to 980 non-isolated patients. Errors in insulin administration and anticoagulant prescription were more frequent in isolated patients. Adverse events such as hypo- or hyperglycaemia, thromboembolic events, haemorrhage, and MDRO ventilator-associated pneumonia (VAP) were also more frequent with isolation. After careful adjustment of confounders, errors in anticoagulant prescription [subdistribution hazard ratio (sHR) = 1.7, p = 0.04], hypoglycaemia (sHR = 1.5, p = 0.01), hyperglycaemia (sHR = 1.5, p = 0.004), and MDRO VAP (sHR = 2.1, p = 0.001) remain more frequent in isolated patients.
Conclusion: Contact isolation of ICU patients is associated with an increased rate of some medical errors and adverse events, including non-infectious ones.