As part of a nationwide quality improvement program, the standardised mortality ratio (SMR) of your Intensive Care Unit was compared to other similar Intensive Care Units using a funnel plot.
You are ICU “A” (filled red dot)
a) What does the graph show about your ICU “A”? (20% marks)
b) Explain how the SMR is calculated. (20% marks)
c) Give the causes of an increased SMR. (60% marks)
The SMR of ICU A is above the upper 99% CI indicating the SMR is significantly higher than similar
hospitals. Your ICU has significantly more deaths than expected compared to similar hospitals.
The overall SMR for the group is less than 1 and the SMR for ICU A is less than 1
SMR = O/E
O= observed number of deaths, E = expected number of deaths
E is derived from the average of the sample/ population.
Usually a risk adjustment model is used to calculate and account for severity of illness.
- Can be “apparent” or “real”.
- Data quality
- Incomplete or errors in data submission causing underestimated expected risk
- Widely different casemix of this ICU compared to others.
- Statistical model (risk adjustment) may no longer well calibrated
- True increase in mortality which can be due to
- i. Factors internal to ICU: very high occupancy, poor processes,, inadequate staffing
- ii. Factors external to ICU; problems in services that are high users of ICU e.g. surgery, system
Pass rate: 27%
Highest mark: 6.5
Additional Examiners’ Comments:
Many candidates showed a significant knowledge gap relating to this commonly used quality indicator with insufficient details and structure in their answers.