Page 139 - Hospital Authority Convention 2018

Service Enhancement Presentations

F4.5 Clinical Safety and Quality Services II 16:15 Room 421

Intensive Care Unit Outcomes Monitoring and Improvement Programme
Ho CM, Yan WW, Chan KC, Shum HP, Wong WT, Lam KW, Lai KY, Ma J, Leung CM, Chan KM, Tam MN
Coordinating Committee in Intensive Care, Hospital Authority, Hong Kong
Introduction HOSPITAL AUTHORITY CONVENTION 2018
Adult intensive care is an integral service for patients with life-threatening critical illnesses. In the past, adult intensive care
units (ICUs) in Hong Kong adopted an external model for the performance review. Intensive Care Unit Outcomes Monitoring
and Improvement Programme (ICUOMP), commissioned by Co-Ordinating Committee (COC) in Intensive Care, was launched
in 2015. Refined risk-adjusted models were created to benchmark the performance of intensive care units in HK.

Objectives
(1) To develop a reliable local contemporary clinical audit related to critically ill patients in Hong Kong ICUs; (2) to measure
and strengthen the quality of ICU services; and (3) improve the strategic planning of ICU services.

Methodology
Since 2015, all admissions to the 15 adult ICUs under Hospital Authority have been screened. Some admissions were
excluded following internationally accepted criteria. Diagnosis and physiological data within the first 24 hours of admission
were collected. Data validity was checked using random sampling by independent ICU specialists since 2016. Mortality and
length of stay (LOS) of patients were modelled, by independent academic biostatisticians, with generalised linear mixed
model.

Results Monday, 7 May 2018
Data from 27,844 admissions were collected after exclusion; 12,394 and 12,731 admissions, from 2015 and 2016 respectively,
were analysed. The main findings were as follows: (1) There was an annual increase of 4% in ICU admissions. (2) The
crude hospital mortality rate was 16.4%. (3) An outlier was identified in the mortality model of 2015. Upon feedback on
this information, the unit reviewed their situation and the anomaly was reverted in 2016. (4) Another outlier was identified in
the mortality model of 2016. Feedback has been provided and we await the data of 2017. (5) In the post-hoc analysis, ICU
performance was proven correlated with off-hour intensivists staffing, the number of doctors in the unit, the total ICU LOS
provided per doctor and per nurse. (6) Patient’s mortality also showed association with out-of-office-hour ICU discharge. (7)
There were significant variations in the length of stay among the ICUs. Exploration of the variation may improve the efficiency
of ICU resources.

Conclusion
ICUOMP is an important and effective audit to identify areas of improvement in clinical practice and subsequent patient
outcome.

137