Risky business: healthcare professionals, their clinical decisions

Risky business: healthcare
professionals, their clinical
decisions and can we do
anything to improve them?
Dr Carl Thompson, Editor: Evidence Based Nursing
(ebn.bmj.com); Senior Lecturer, Health Sciences.
“It’s the decisions, stupid.”
Good idea or bad idea
• Weekly exercise programme for
staff?
• Bicycle safety education programme
for kids?
• Teenage driver education
programme?
• Lying babies prone for sleeping?
Bad news – uncertainty is
irreducible (Eddy, 1996)
• Technical (these can be reduced)
– Lack of information for Dx, Px, Rx
– Speed of growth of knowledge
– ‘natural variability’ (between clinicians and patients)
• Personal (always present)
– Patient preferences unknown
– Adequacy of surrogate decision making
– Hospitals organised around familiar techniques not
patients
• Conceptual (ditto)
– Which patient gets what?
– Application of general criteria to individual patients
– Past experiences to present patients?
Does any of this matter?
•
•
•
•
•
11% of admissions result in adverse
events, 50% of these down to error
1 million patients suffer iatrogenic
harm,
1000 die per annum
Between 7 and 8.4 additional bed days
per adverse event
Mandatory reporting has sensitivity of
around 5%
(NAO 2005, NPSA 2002, Akbari and Sheldon 2006)
The good news. Information
Behaviour is…
1. Think number between 10 and 20
2. Add the digits together (e.g. 13 = 1+3 =
4)
3. Subtract from the first number you
thought of
4. Subtract 5
5. Convert to a letter (e.g. 1=A, 2=B etc…)
6. Listen to Carl…
Anatomy of an evidence
based decision
expertise
preferences
resources
evidence
Intuition and uncertainty
• Main mode of decision making in
healthcare
– Deny it! (uncertainty)
– Heuristics (cognitive shortcuts and mottoes):
“if there is any chance of (the disease), the
(procedure) should be performed… if but one patient
is saved, the effort is worthwhile… costs should not
be considered in decisions about individual
patients… and when in doubt do it”
– Errors – base rate neglect; sample size
insensitivity; overweight irrelevant information
(> with more detail).
What do we know? EBP
• Limited time for info (17-20 minutes/week)
• Lots of decisions (circa 50 per shift on
MAU)
• CPD/education rather than practice
drivers
• Limited appraisal skills
• Everyone hates numbers
• Finite kinds of clinical questions once you
learn how to handle uncertainty
(Thompson et al. 2001, 2002, 2004, 2005;
Ely et al 2001; Cogdill 2004; McCaughan 2005, Ely 1997, Wyatt 2005)
What can we do: the evidence
Brian Haynes, R Evid Based Med 2006;11:162-164
Copyright ©2006 BMJ Publishing Group Ltd.
What can we do?
Decision making
• Get sexy with CPD!
– cognitive “forcing” and training
– Structuring decisions (PICO, decision analysis)
• Feedback on decisions*
• Make uncertainty, values and prior beliefs
explicit
• Recognise that decision support makes us
more not less professional
• Manage time
Mews – “simple decision
support”
Better or worse (ROC)?
Mews >4
.69
Sensitivity
P (judge+ | risk+)
>11 .65
>8 .61
>4 .57
1 – specificity (False positive rate)
Time – “detecting signals”
One more…
• Nutrition in acute stroke
– Screening and assessment using
validated tools
– Change management (multifacted,
targetted and substantive)
– 2 x 200 patients in a before-after
(quasi) experiment; 11 medical care of
elderly wards; MDTs
EBP what’s it all for – 1
• 9% more patients swallow function
screened <24 hours
• 17% increase in nutritional risk
screening <24 hours
• 11% reduction in patients for whom
nutrition decision making was
deferred >5 days post admission
EBP what’s it all for – 2
• Days without food dropped from 9 to
4
• 39% lesss chest infection, aspiration
pneumonia or sepsis
– or if you prefer… 20 fewer patients
– (33 before vs 13 after, p < 0.01)
Perry and McClaren, EBN 2003 6:68-71
Summary (the serious one)
• Know something about the decisions
and uncertainties you face and use this
knowledge (5S)
• Ask yourself “what if?”
• Accept decision support if its there
• Keep track of what happens as a result
of your decisions
• Find time from somewhere
“Even if all you have is a horse’s arse of
a decision you can learn something…”