Document

Report
Translating evidence into
practice
Peter Morley
Royal Melbourne Hospital, University of Melbourne
Plan of attack
• Patterns of research (eg. ICU)
• Reasons people doubt the literature
• Reasons people don’t want to change
ANZICS Clinical trials group
• Crit Care Resusc 2007; 9: 198–204
• Australian and New Zealand Intensive Care
Society Clinical Trials Group (ANZICS-CTG).
• Established in 1994, the ANZICS-CTG has
published a number of highimpact studies of
increasing complexity and size
• trial of low-dose dopamine for patients with early
acute renal failure (n=328, Lancet 2000)
• the Saline Albumin Fluid Evaluation (SAFE) study
(n=6997, NEJM 2004)
• Medical Early Response Intervention and Therapy
(MERIT) study (a cluster randomised trial in 23
hospitals, Lancet 2005)
• collaborations between Australian, New Zealand
and Canadian investigators in RCTs include
– VASST (low dose vasopressin in septic shock),
– DECRA (decompressive craniectomy in traumatic brain
injury)
– NICE-SUGAR (glucose control regimens in ICU), and
– PRO- TECT (unfractionated versus low molecular
weight heparin for venous thromboembolism
prophylaxis)
•
•
•
•
barriers to patient recruitment,
obtaining funding in a timely manner
consent issues specific to the critically ill
altered pharmacokinetics associated with critical
illness
Problems arose!
Confounders
“. . . there are known unknowns,
and there are unknown unknowns . . .”
Observational studies
•
•
•
•
•
Often data collected for “other” purposes
Associate factor with particular outcome
Try to account for “known knowns”
Try to account for “known unknowns”!
BUT . . .
The Effectiveness of Right Heart
Catheterization in the Initial Care of
Critically Ill Patients.
Connors JAMA 1996; 276(11) 889-97
• Observational study
• Compared group using PA catheters with
those who didn’t
• Sicker patients but proportionately more
died!
The Effectiveness of Right Heart
Catheterization in the Initial Care of
Critically Ill Patients.
Connors JAMA 1996; 276(11) 889-97
• “Sensitivity analysis suggested that a missing
covariate would have to increase the risk of death
6-fold and the risk of RHC 6-fold for a true
beneficial effect of RHC to be misrepresented as
harmful.”
• “In this observational study of critically ill patients,
after adjustment for treatment selection bias, RHC
was associated with increased mortality and
increased utilization of resources”.
Patient selection?
No more ICU research?
I know your son is dying, and
you are having trouble coping
with the news, the grieving,
the distress, but . . .
• I would like to put him in a trial where we
will toss a coin, to see whether we will
remove half his skull . . .
• We are doing everything possible, but by
chance we will see if we can do something
more . . .
Ethics?
ARDSnet major issues
• Consent via surrogate
• Is it appropriate to compare two extremes of
practice, or better to use usual care or
average values for control
ARDSnet
• In 2003, after investigators of the Acute
Respiratory Distress Syndrome Network
from San Francisco were publicly accused
of doing research without requiring
adequate consent, a definition for a legal
delegate was introduced in the Californian
state law.
New guidelines were required
for critical care research in
USA
PAC-Man study
• intended to assess the usefulness, or not, of the
pulmonary artery catheter as a monitoring tool for
acutely ill patients.
• Only 2.6%of patients could consent before
randomization.
• Relatives granted consent in 81%of cases and
refused it in only 0.6%.
• Patients who regained cognitive capacity
afterwards gave retrospective consent in nearly all
cases (93%). Only 3% refused.
• Curr OpinCrit Care13:122–125. ß2007
Methodologic limitations?
• Crit Care Med 2007 Vol. 35, No. 2
• In the meta-analyses included in their study,
the investigators observed that 69% had
major flaws (OQAQ score < 5) and the
• mean OQAQ score was only 3.3 (95%
• confidence interval, 3.0 –3.6)
What about COI?
The control group?
Amiodarone in VF:
ARR EST
• Amiodarone in the out-of-hospital
Resuscitation of REfractory Sustained
ventricular Tachyarrhythmias (ARREST trial)
– 504 adult non-traumatic arrest, King County,
Washington still in VF/VT after 3 or more shocks
– 246 received 300mg amiodarone rapid IV infusion,
v 258 received rapid IV diluent (detergent)
– 44% with amiodarone survived to hospital admission (v
35%)
– 1% better hospital discharge rate (N.S.)
– Eligible but not enrolled, survival = amiodarone!
Conflicting results?
“However, there were more deaths in the metoprolol group
than in the placebo group (129 [3·1%] vs 97 [2·3%] patients;
1·33, 1·03–1·74; p=0·0317)”.
Diagnostic and prognostic
studies?
Studies of diagnostic tests
• “Test” = examination finding/investigation
• Starting point is initial “test” on patients
– Compare “test” result with known outcome (“gold
standard”)
– Develop threshold result (to alter Mx)
= Clinical Decision Rule (CDR)
• Better = confirm result in multiple centers
Bozeman 1996: (o)esophageal
bulb
Tracheal
intubation
Yes
No
Oeoph
bulb
totals
Totals
Reinflate 294
0
294
Not
3
reinflate
297
19
22
19
316
Diagnosis outcomes
• Sensitivity
– = 294/294 = 1 (or 100%)
– 95% CI = 0.988 to 1.000
• Specificity
– = 19/22 = 0.864 (or 86.4%)
– 95% CI = 0.651 to 0.971
Diagnosis outcomes
• Likelihood ratio for a positive test
– = sens/(1-spec) = 1/(1-0.864) = 7.33
– 95% CI = 2.56 to 20.99
• Likelihood ratio for a negative test
– = (1-sens)/spec = (1-1)/0.864 = 0
Translation into practice
What rhythm is our process
in?
Threshold for practice
change
• “Portfolio”
– Totality of evidence
• Quality & level
– Magnitude of effect
• NNT
• Population at risk
– Simplicity of training &
implementation
• Necessity of
change
– Bretylium
– Monophasic
– Pauses in
compressions
– Problems with training
& implementation
Critical care specialists
understanding of the literature
“Many interventions that have not been tested by
RCT were believed to have been tested;
conversely, some interventions actually tested by
RCT were not mentioned. Few interventions used
in the ICU have actually been shown by RCT to
have a positive effect on outcome”.
Are my patients so different??
"The Surviving Sepsis Campaign: International
guidelines for management of severe sepsis
and septic shock: 2008. An assessment by the
Australian and New Zealand Intensive Care
Society. "
P Hicks et al. Anaesthesia and Intensive Care
36.2 (2008): 149-51
ANZICS surviving sepsis
• Early goal directed therapy
– Limitations of single RCT. Planning repeat trial: ARISE
• Glucose control <8.3
– No data! Planning trial: NICE SUGAR
• Stress dose steroid
– Adjusted analysis from subgroup of negative RCT
• rhAPC
– Limitations with single positive study
• Noradrenaline or dopamine
– Widespread use of adrenaline (not shown inferior)
Should we use therapeutic
hypothermia after cardiac
arrests?
How are we doing with
hypothermia?
•
•
•
•
Clearly aware of our limitations
Really keen to improve
Admitting we may be slow to change
Promising to do better
EVER!!!
• Curr Opin Crit Care 2003, 9:321–325
• “the most cost-effective opportunity to
improve patient outcomes over the next
quarter century will likely come not from
discovering new therapies but from
discovering how to deliver therapies that
are known to be effective”
Plan of attack
•
•
•
•
•
Introduction to Intensive Care Research
Progress over 30 years
Successes
Problems
The future
Evidence Based Medicine
• = the conscientious, explicit and judicious
use of current best evidence in making
decisions about individual patients
BMJ 2003;327:1459–61
How Best to Ventilate?: Trial Design and
Patient Safety in Studies of the Acute
Respiratory Distress Syndrome
Steinbrook, Robert. Crit Care Med 348(14), 2003, 1393-1401
• The NHLBI established the ARDS Network
in 1994 for the conduct of clinical trials. It
now comprises 20 academic medical
centers in the United States and Canada.
So far, the total amount that has been spent
for the network and its trials is $37.4 million.
• No useful information from studies??
Large RCTs (usually asking a
single, and possibly already
answered question) are
expensive
We should also focus on other information
I Civil
Some possible mechanisms
•
•
•
•
•
•
•
•
•
Conservative resuscitation
Acceptance of permissive hypotension
FAST scanning
rapid ED assessment
early CT scanning
Damage control
brain oriented intensive care
DVT prophylaxis
early rehab
IHI (bundles)
http://www.ihi.org/ihicomponents
• Key features of Ventilator Bundle :
– Elevation of the Head of the Bed
– Daily "Sedation Vacations" and Assessment of
Readiness to Extubate
– Peptic Ulcer Disease Prophylaxis
– Deep Venous Thrombosis Prophylaxis
Get your act together!
• Norway
– OOHCA, admitted to ED, cardiac cause
• Standardised post-resuscitation care
– Haemodynamics, oxygenation, ventilation
– PCI, Hypothermia, sedation, metabolic (eg.
glucose)
• Doubling favorable neurologic outcome
– 26 to 56%
ACLS training works!
Moretti et al. Resus 2007
Strategies for
implementing change
•
•
•
•
•
•
•
Educational materials
Conferences, courses
Interactive small group meetings
Use of opinion leaders
Feedback on performance
Computers
Mass media campaigns
Grol R. Lancet 2003;362:1225-30

similar documents