Measuring Outcomes in audit and research

1
(No Transcript)
2
Measuring Outcomes in audit and research
3
ICU Mortality

• Most common outcome measure
• death common in ICU
• Unequivocal endpoint
• Easy to measure
• can compare with predicted outcome
• we wish to design trials to improve mortality

4
ICU mortality Whats wrong with using this as an
outcome measure?

• This is the topic of my talk
• I will talk about the problems of using this
  measure in research and audit separately
• the problems in research and audit are very
  different
• Useful to compare and contrast audit and research

5
Clinical ICU Research

• We are trying to understand the effect of a
  treatment
• Does treatment X benefit the patient?
• Does treatment X lead to less patients dying in
  ICU? Seems like a good question to ask

6
What is wrong with ICU mortality as our end-point
in research?

• Its not a patient centred outcome
• Doesnt tell you what happened next
• Patient may die soon after
• Many of our patients leave ICU in organ failure
• Patient may be so frail after ICU they have no
  quality of life
• Im going to focus on long-term survival as an
  alternative measure

7
Why look at long-term outcome?

• Its all the cardiologists fault!

AIR Study Ramipril after MI
8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
Why might it be difficult to prove long-term
benefit in a sepsis trial?

• Cardiology patient may only have IHD
• Sort the blocked artery and LV function better
  for ever
• Patient continues treatment long term
• Why is a patient septic? may have underlying
  problems disease ICU cant cure
• Did patient die of sepsis or an underlying
  disease?
• ICU intervention stops when patient leaves ICU

12
patient dies of MRSA pneumonia

• Elderly patient
• Severe MS
• Bed bound
• In hospital for 5 months
• Could any sepsis super drug have got this patient
  home?

13
Oesophagectomy for cancer

• Hospital Mortality 5-10
• Comorbidities and operative complications predict
  hospital mortality
• 1 year mortality around 50
• Tumour stage predicts 1 year mortality
• What we do in ICU is never going to
• improve 1 year mortality

14
Patients leave ICU frail or with organ failure

• 30 of patients leaving ICU have a respiratory
  SOFA score of 3
• i.e. p02 9 kPa on 35 02
• ATICS Study involving over 2000 patients in
  Scottish ICUs
• May be ICU style treatment has to continue post
  ICU to show improvement in long-term outcome

15
Clinical Audit

• What does this mean?

16
SMR comparisons

• Leads to reassurance / complacency
• Leads to mediocre care
• Unfair comparisons
• Most importantly
• Doesnt create any feed back for health care
  workers delivering care
• No mechanism for improving care or identifying
  problems

17
(No Transcript)
18
Survival doesnt guarantee the treatment was good!

• Failure of many non catastrophic processes does
  not lead to death or severe injury within hours
  of the failure
• Example hand-washing
• This is due to biological resilience
• These processes tend to have poor compliance
• In contrast disconnection from a ventilator for
  10 minutes

19
What should Clinical Audit Do?

• Audit cant prove a treatment works
• Audit asks are we giving our patients good
  treatment?
• It can establish whether we are delivering a
  treatment in a reliable/safe/timely manner
• Feed back and results to healthcare providers in
  a timely fashion
• Create an audit loop
• This is known as Process Audit

20
Why is process audit important?

• missing link in Evidence Based Medicine
• What evidence do we have that we are delivering
  good care?
• Large study from USA 53.5 of acute patients
  received best evidence care McGlynn, et al NEJM
  2003 348 2635-2645
• Idealistic striving for best care

21
Summary

• ICU Research needs to look beyond ICU survival to
  long term survival and QOL but this will be
  challenging
• ICU Audit needs to stop relying on ICU survival
  and SMR as evidence of good care and embrace
  process audit

22
Anyone still awake?