Scoring systems

Question 1

What type of scoring systems are available?

Answer 1

1. Disease specific scores (Child-Pugh, Wells, CURB 65, CHADS VASc, GRACE, Glasgow-Imrie)
2. Physiological and illness severity scores (APACHE II, SOFA, ICNARC, MODS, TISS)

Question 2

What are the features of an ideal scoring system?

Answer 2

1. Valid
2. Simple
3. Discriminatory (plot sensitivity/specificity curve to get AUROC - should be at least 0.7
4. Calibrated
5. Transferrable
6. Generateable from readily available data

Question 3

What are scoring systems used for?

Answer 3

1. To assess performance of ICU care and outcomes
2. To assess patient’s risk of mortality
3. To compare populations
4. To facilitate audit
5. To tailor treatment

Question 4

How can we compare different ICUs?

Answer 4

1. Outcome-based:
   - SMR
   - ICNARC case mix programme
   - CRBSI rate
   - VTE rate
   - MRSA rate
   - unit based bacteraemia
2. Process-based:
   - Audit participation
   - delirium screening
   - FAST HUG
   - early EN
   - hand hygiene compliance
   - VTE assessments.
3. Structure-based: GPICS standards (consultant cover, nurse:patient ratio, % days at full occupancy, MDT ward rounds, ICU follow-up)

Question 5

What is the standardised mortality rate?

Answer 5

SMR = observed mortality/predicted mortality

Problems with it include:

• doesn’t take into account pre-hospital or pre-ICU care discrepancies
• doesn’t take into account case-mix
• doesn’t take into account local baseline health and deprivation
• there may be varying application of scoring systems

Question 6

What is the CURB 65 score?

Answer 6

Confusion - new
Urea >7.0
RR >30
Blood pressure <90
Age >65

Each scores 1
Increased score associated with increased mortality (4 = 40%)
Score >3 is severe and likely needing ICU admission.