Risk adjustment in colorectal cancer

1
Risk adjustment in colorectal cancer

• PP Tekkis, JD Poloniecki,
• MR Thompson, JD Stamatakis
• On behalf of the Association of Coloproctology
• of Great Britain and Ireland
• Risk stratification seminar
• 7th March 2003

2
aims

• risk-assessment tools
• ASA
• Colorectal-POSSUM
• setting standards of care
• audit benchmarking

3
clinical effectiveness

• quality of care
• expected outcomes
• prediction about what ought to happen
• observed outcomes
• case-mix adjustment
• comparative audit

4
clinical effectiveness algorithm


5
study population

• ACPGBI study 2002
• Colorectal-POSSUM study
• 15,873 patients
• colorectal cancer patients
• 88 hospitals
• CCF colorectal cancer study 2002
• 7000 patients
• 9 hospitals

6
American Society of Anaesthesiology grading (ASA)

• I No systemic disease
• II Mild / moderate systemic disease
• III Severe systemic disturbance
• IV Life threatening disease
• V Moribund patient

Dripps et al JAMA 1963
7
POSSUM scoring system
Copeland et al BJS 1991
8
Colorectal-POSSUM

• Physiological score
• Age
• Cardiac history
• Blood pressure
• Pulse
• Urea
• Haemoglobin

• Operative severity score
• Complexity
• Urgency
• Peritoneal soiling
• Cancer staging

Tekkis et al BJS 2002
9
potential users

• patient
• informed consent
• information at the public domain
• clinician
• cross-sectional analysis
• dynamic monitoring of surgical performance
• hospital
• self monitoring
• league tables

10
simple dataset
11
which risk factors

• morbidity co-morbidity
• ASA grade
• Colorectal POSSUM
• age
• mode of surgery
• cancer resection
• Dukes stage

12
risk is not uniform
13
non-hierarchical model
Which patient?
14
hierarchical model
Which patient? What hospital?
15
How is mortality affected by age?
16
Does mortality differ between hospitals?
17
ACPGBI ASA model
18
conversion of odds ratios to probability
UNIT j
19
Does it work in the UK?
20
Does it work outside the UK?
Fazio et al ASCRS 2003

• UK
• Urgent vs elective
• 21
• ASA II vs I
• 21
• ASA III vs I
• 51
• Dukes D vs A
• 21

• USA
• Urgent vs elective
• 21
• ASA II vs I
• 21
• ASA III vs I
• 51
• Dukes D vs A
• 2.51

21
which tool?

• ASA
• ROC curve 78.8
• widely used
• simple
• cheap
• operated patients only
• subjective
• stood the test of time

• CR-POSSUM
• ROC curve 86.6
• more complex
• more-expensive
• automated
• operated and non-operated patients
• objective
• additional validation

22
mortality control chart
Tekkis et al BMJ 2003
23
conclusions

• risk analysis part of clinical practice
• early outcomes operative mortality
• acute physiological derangement
• co-morbidity
• risk adjustment essential
• POSSUM based system

24
thank

• The Royal College of Surgeons of England
• Contributing consultants