Effects of ICU characters, human resources and workload to outcome indicators in Thai ICUs: the results of ICU-RESOURCE I study

Abstract

Objective: There was a lack of available data regarding intensive care unit (ICU) characteristics, human resources, workload, and ICU outcomes in Thai ICUs. Therefore, the objectives of the present study were to describe these parameters and to demonstrate the association of these predictors to ICU outcome indicators including crude mortality, ventilator days and ICU length of stay (ICU-LOS).

Material and method: Retrospective survey to 155 participated ICUs across Thailand. ICUs characters, physician and nurse staffing, patient density and ICU outcomes indicator at the year 2011 of monthly mortality, ventilator days and ICU length of stay were recorded. Multilevel mixed effect linear regression was used for cluster analysis. Statistical difference was defined as p-value < 0.05.

Results: The 132 ICUs (85.16%) were identified as open ICU or low physician staffing. The ICUs were directed or consulted by intensivists or critical care physicians in 53 ICUs and nearly half of them were located in ICUs at academic hospitals. The median value of average daily nurse to patient ratio (NPR) was 0.5 (Inter-quartile range, IQR 0.23). The median crude mortality was 13.92% (IQR 10.16). Median ventilator days and ICU-LOS were 5.31 (IQR 4.42) and 5.8 (IQR 3.0), respectively. A multilevel mixed model demonstrated crude mortality benefit in groups of closed ICU management or high physician staffing, academic ICUs, regular multidisciplinary round, ICU physician staffing availability and low patient density. Although the NPR did not demonstrate any benefit in crude mortality, a lower NPR (higher number of nurse staff) was associated with lower ventilator days.

Conclusion: Thai ICUs showed differences in administration systems. The outcome indicators of crude mortality, ventilator days and ICU-LOS were impacted by the ICU characteristics, human resources and ICU workload (Thai Clinical Trial Registry: TCTR-201200005).