Title Place Date. Centre for Evidence-Based Medicine What I’ve done / do/don’t do l Done: I’ve...

Title

Place

Date

Centre for Evidence-Based Medicine

What I’ve done / do/don’t do

Done: I’ve gotten out of date and retrained in Internal Medicine twice

Do: I run an in-patient General Medicine service (all comers) at a UK DGH:» 208 admissions last month» strive to use evidence at the bedside

Don’t: I’ve cancelled my journal subscriptions (and give away the JCI and BMJ)


The Problems:

We need evidence (about the accuracy of diagnostic tests, the power of prognostic markers, the comparative efficacy and safety of interventions, etc.) about 5 times for every in-patient (and twice for every 3 out-patients).

We get less than a third of it


The Problems:

To keep up to date in Internal Medicine, I need to read 17 articles a day, 365 days a year

Need to read Don’t Nor does anyone else


Median minutes/week spent reading about my

patients:

Self-reports at 17 Grand Rounds: Medical Students: 90 minutes House Officers (PGY1): 0 (up to 70%=none) SHOs (PGY2-4): 20 (up to 15%=none) Registrars: 45 (up to 40%=none) Sr. Registrars 30 (up to 15%=none) Consultants:

» Grad. Post 1975: 45 (up to 30%=none)» Grad. Pre 1975: 30 (up to 40%=none)


Performance deteriorates, too

Determinants of the clinical decision to treat some, but not other, hypertensives:

1 Level of blood pressure.2 Patient’s age.3 The physician’s year of graduation from

medical school.4 The amount of target-organ damage.


No wonder, then, that CME is growing

Big, and getting huge. Usually instructionally (fact) oriented. Several randomised trials have shown

that it does not improve clinical performance.


Three solutions

Clinical performance can keep up to date:1 by learning how to practice evidence-

based medicine ourselves.2 by seeking and applying evidence-based

medical summaries generated by others.3 by applying evidence-based strategies

for changing our clinical behaviour.


When did EBM begin ?

Certainly in post-revolutionary Paris.

Arguably in B.C China.

Some late-comers named it in 1992.


What evidence-based medicine is:

The practice of EBM is the integration of individual clinical expertise

with the best available external clinical evidence

from systematic research.and

patient’s values and expectations


I.Individual Clinical Expertise:

Clinical skills and clinical judgement Vital for determining whether the

evidence (or guideline) applies to the individual patient at all and, if so, how


II. Best External Evidence:

From real clinical research amongintact patients.

Has a short doubling-time (10 years). Replaces currently accepted diagnostic

tests and treatments with new ones that are more powerful, more accurate, more efficacious, and safer.


III. Patients’ Values & Expectations

Have always played a central role in determining whether and which interventions take place

We’re getting better at quantifying and integrating them


What EBM is not:

EBM is not cook-book medicine» evidence needs extrapolation to my

patient’s unique biology and values EBM is not cost-cutting medicine

» when efficacy for my patient is paramount, costs may rise, not fall


Evidence-Based Medicine:The Practice

When caring for patients creates the need for information:

1 Translation to an answerable question (patient/manoeuvre/outcome).

2 Efficient track-down of the best evidence » secondary (pre-appraised) sources

e.g., Cochrane; E-B Journals» primary literature


Evidence-Based Medicine:The Practice

3 Critical appraisal of the evidence for its validity and clinical applicability generation of a 1-page summary.

4 Integration of that critical appraisal with clinical expertise and the patient’s unique biology and beliefs action.

5 Evaluation of one’s performance.


We needn’t always carry out all 5 steps to provide

E-B Care

Asking an answerable question. SearchingSearching for the best evidence. Critically-appraisingappraising the evidence. Integrating the evidence with our

expertise and our patient’s unique biology and values

evaluating our performance


We’ve identified 3 different modes of practice

“Searching & appraising”» provides E-B care, but is expensive in time and

resources “Searching only”

» much, quicker, and if carried out among E-B resources, can provide E-B care

“Replicating” the practice of experts» quickest, but may not distinguish evidence-

based from ego-based recommendations


Even fully EB-trained clinicians work in all 3

modes

“Searching & appraising” mode for the problems I encounter daily.

“Searching only” mode among E-B resources for problems I encounter once a month.

“Replicating” the practice of experts mode for problems I encounter once a decade(and crossing my fingers!).


Patients can benefit

Even if <10% of clinicians are capable of practicing in the “searching & appraising” mode (5% of GPs)

As long as most of them practice in a “searching” mode within high-quality evidence sources (70-80% of GPs):» Cochrane Library, E-B Journals, E-B

Guidelines, etc


Three solutions

Clinical performance can keep up to date:1 by learning how to practice evidence-

based medicine ourselves.2 by seeking and applying evidence-based

medical summaries generated by others.3 by applying evidence-based strategies

for changing our clinical behaviour.


Information required within seconds

Systematic reviews, periodically updated, of randomised trials of the effects of health care (from all sources, and in all languages):

The Cochrane Collaboration.

Cochrane Systematic Reviews (522; another 500 in preparation)

Database of Abstracts of Reviews of Effectiveness (1895)

Registry of Randomised Controlled Trials (218,355)


Information required within seconds

CD-Evidence-based journals of 2º publication:

screen 50-70 clinical journals per week for clinical articles that pass critical appraisal quality filters conclusions likely to be true.

select the subset that are clinically relevant.

summarise as “more-informative” abstracts.

add commentaries from clinical experts.

introduce with declarative titles.


2. Seeking and Applying EBM generated by others

Evidence-Based Medicine is published in: English French German Italian Portuguese Spanish



New Evidence-based journals of 2º publication: E-B Cardiovascular Medicine E-B Health Policy & Management E-B Nursing E-B Mental Health

And as new departments in 1º journals.



E-B Textbooks: E-B Pain Relief E-B Cardiology

includes icons for levels of evidence “E-B On-Call”

includes > 1300 CATs


Can you really practice EBM?

Is there any “E” for EBM ?


Conventional Wisdom

“only about 15% of medical interventions are supported by solid scientific evidence” (BMJ Editorial)


Even on the U.S. Talk-Shows: (“Health Outrage of

the Week”)

“..... this would put 80 to 90 per cent of accepted medical procedures in this country under the heading of quackery!”


Problems with Conventional Wisdom

uses clinical manoeuvres, rather than patients, as the denominator.

tends to focus on high-technology, “big ticket” items.

relies on simple literature searches that miss over half of the most rigorous types of evaluations.

conducted from armchairs.


Performed an empirical study on a busy in-patient

service

on the general medicine in-patient service of the Nuffield Department of Medicine at the Oxford-Radcliffe NHS Hospital Trust (“The John Radcliffe”)

all our admissions arise from urgent referral from local GPs or via the Emergency Room


The Protocol

At the time of discharge, death, or month’s end, each patient was reviewed and consensus reached on:

The primary diagnosis: the disease, syndrome or condition most

responsible for the patient’s admission to hospital


The Protocol (cont.)

The Primary Intervention the treatment or other manoeuvre that

constituted our most important attempt to cure, alleviate, or care for the primary diagnosis

traced into the literature to determine its basis in evidence

– the Consultant’s “Instant Resource Book”– bibliographic data base searches


Primary Interventions were Classified by Level:

Evidence from Randomised Control Trials (better yet: systematic reviews of all relevant, high-quality RCTs)

Convincing non-experimental evidence (unnecessary & unethical to randomise)

Interventions without substantial evidence


Conclusions from E-B oriented General Medicine:

82% of our patients received evidence-based care.

treatments for 53% were justified by RCTs or systematic reviews of RCTs.

Of 28 relevant RCTs and SRs, 21 were accessible within seconds.

treatments for 29% were justified by convincing non-experimental evidence


Evidence from RCTs (53%)

36% had Cardiovascular diagnoses:

» Ischaemic heart disease 17%» Heart failure 6%» Arrhythmia 2%» Thromboembolism 3%» Cerebrovascular 8%


Evidence from RCTs (53%)

7% had taken poison 5% received chemotherapy or analgesia

for cancer 3 % had gastrointestinal disorders 2% had obstructive airways disease


Convincing non-experimental evidence

(29%)

Infections 15% Cardiac disorders 7% Miscellany (non-compliance, drug

reactions, bowel or bladder neck obstruction, dehydration, micturition syncope) 7%


Interventions without substantial evidence

(18%)

Specific symptomatic and supportive care for mild poisoning, non-cardiac chest pain, viral (non-herpetic) meningitis, terminal CNS disease, confusion, and food poisoning.


Better Outcomes for Patients When EBM Is

Practised

E-B practise vs. Outcome in stroke (US): When cared for by E-B neurologists,

patients were 44% more likely to receive warfarin, and much more likely to be placed in a stroke care unit,

And were 22% less likely to die in the next 90 days. (Mitchell et al: stroke 1996;27:1937-43)


Centres for Evidence-Based Surgery

E-B General/Vascular Unit in Liverpool:» 95% received evidence-based Rx

24% Level 1 71% Level 2

E-B Paediatric Unit in Liverpool:» 77% received evidence-based Rx

11% Level 1 66% Level 2


Worse Outcomes for Patients When EBM Is Not

Practised:

In a city-wide study of E-B practise vs. Outcome in carotid stenosis:

Generated E-B indications for endarterectomy and reviewed 291 pts.

Found the surgical indications:» Appropriate in 33%»Questionable in 49%»Inappropriate in 18%


Worse Outcomes for Patients When EBM Is Not

Practised

Stroke or death within the next 30 days: Expected (if left alone): 0.5% Expected (if properly selected and

operated): 1.5% Observed among operated patients (2/3

operated for questionable or inappropriate reasons): >5%

Wong et al. Stroke 1997;28: 891-8.


Evidence-Based Ambulatory Paediatrics

54% of manoeuvres were evidence-based (“experts” had predicted <20%)» 77% of diagnostic manoeuvres» 67% of treatments» 59% of health promotion


Centres for Evidence-Based

Psychiatry

In-Patients (Oxford)» 67% treated on the basis of RCTs

Out-Patient» >80% received evidence-based Rx


Evidence-Based General Practice

122 consecutive consultations in a suburban (Leeds, UK) practice.

81% evidence-based:» 31% based on RCTs or overviews» 50% based on convincing non-experimental

evidence» 19% without substantial evidence

(Gill et al, BMJ 1996;312:819-21)


Can we get evidence to the bedside?

Need it within seconds if it is to be incorporated into busy clinical rounds

Our initial attempts to bring the best evidence to a busy clinical team caring for 200+ admissions per month


Searching for Evidence in the Month Before the Cart:

Expected searches = 98 Identified searching needs = 72 Only 19 searches (26%) carried out.


Contents of the Cart:

Infra-red simultaneous stethoscope with 12 remote receivers.

Physical diagnosis text book and reprints (JAMA Rational Clinical Exam).

Notebook computer, computer projector, and pop-out screen.

Rapid printer.


Contents of the Cart (cont):Library Round-Trip = 7 min

125 summaries (1-3 pp) of evidence previously appraised and summarised by Side A teams (in the form of “Redbook” entries or

Critically-Appraised Topics : “CATs”).

Access Time to the “bottom line” = 12 sec.Access Time to the “bottom line” = 12 sec.

CAN FIND THE CAT IN 12 SECONDSMIS WITH HYPERGLYCAEMIA BENEFIT FROM INTENSIVE INSULIN THERAPY.

Clinical Bottom Line:Treating 9 hyperglycaemic MI patients iwth intensive insulin => 3 months will prevent oneadditional death over the next 3.4 years.

Appraised by: Sackett; 24 October 1997The Study: Non-blinded concealed randomised controlled trial with intention-to-treat.Swedish patients admitted with MI in the prior 24 hours with blood glucose >11 mmol/l with orwithout prior known diabetes. 50% thrombolysed; by discharge 80% given aspirin, 70% givenbeta-blockers, and 31% given ACE-inhibitors.Control group (N = 314; 314 analysed): Routine MI care (including aspirin and beta-blockers) but no (extra) insulin unless "clinically indicated" (43%, 45% and 49% on insulin at discharge, 3months, and 1 year)).Experimental group (N = 306; 306 analysed): Routine MI care plus glucose+insulin infusion for =>24 hours and qid insulin for =>3 months (87%, 80% and 72% on insulin at discharge, 3months, and 1 year).The Evidence:Outcome Time to

OutcomeCER EER RRR ARR NNT

death (allpatients)

3.4 years 0.439 0.333 24% 0.106 9

95%ConfidenceIntervals:

7% to 41% 0.030 to 0.182 5 to 34

Comments: 1. Benefit greatest in low risk patients (< 70 y/o, no prior MI, no CHF, no digitalis Rx) notpreviously on insulin (RRR 46%; ARR 0.15; NNT 7).2. PTCA and CABG done in 5% and 11% of controls and in 4% and 11% of intensive insulinpatients.3. Glucose 11.7 and 9.6 at 24 hours; 9 and 8.2 at discharge.4. 97% of deaths were cardiovascular, and most of the mortality benefit was seen afterdischarge. Expiry date: October 1998References: Malmberg K for the DIGAMI Study Group: Prospective randomised study of intensive insulintreatment on long term survival after acute myocardial infarction in patients with diabetesmellitus. BMJ 1997;314:1512-5.

CAN OBTAIN THE BOTTOM LINE IN 2 SECONDS:

MIS WITH HYPERGLYCAEMIA BENEFIT FROM INTENSIVE INSULINTHERAPY.

Clinical Bottom Line: Treating 9 hyperglycaemic MI patients with intensiveinsulin => 3 months will prevent one additional deathover the next 3.4 years.

CAN READ THE EVIDENCE IN 2 MINUTES

The Study: Non-blinded concealed randomised controlled trial with intention-to-treat.Swedish patients admitted with MI in the prior 24 hours with blood glucose >11 mmol/l with or without priorknown diabetes. 50% thrombolysed; by discharge 80% given aspirin, 70% given beta-blockers, and 31%given ACE-inhibitors.Control group (N = 314; 314 analysed): Routine MI care (including aspirin and beta-blockers) but no (extra) insulin unless "clinically indicated" (43%, 45% and 49% on insulin at discharge, 3 months, and 1 year)).Experimental group (N = 306; 306 analysed): Routine MI care plus glucose+insulin infusion for =>24 hours and qid insulin for =>3 months (87%, 80% and 72% on insulin at discharge, 3 months, and 1 year).

The Evidence:Outcome Time to

OutcomeCER EER RRR ARR NNT

death (allpatients)

3.4years

0.439 0.333 24% 0.106 9

95%ConfidenceIntervals:

7% to41%

0.030 to 0.182 5 to 34

CAN STUDY THE ORIGINAL EVIDENCE FOR HOURS

Reference: Malmberg K for the DIGAMI Study Group: Prospective randomisedstudy of intensive insulin treatment on long term survival after acutemyocardial infarction in patients with diabetes mellitus. BMJ1997;314:1512-5.


Contents of the Cart (cont):Library Round-Trip = 7 min

CD of Best EvidenceAccess Time to the “bottom line” = 26 sec.Access Time to the “bottom line” = 26 sec. CD of WinSPIRS (5-year clinical subsets)

Access Time to useful abstract = 90 secAccess Time to useful abstract = 90 sec. . (so used for filling Educational Rx after (so used for filling Educational Rx after rounds)rounds)

CD of the Cochrane Library (used for filling Educational Rx after rounds)(used for filling Educational Rx after rounds)


Usefulness of the Cart:

81% of searches were for evidence that could affect diagnostic and/or treatment decisions.

90% of these searches were successful in finding useful evidence.

*


Of the successful searches (from the perspective of the most junior responsible

team member):

52% confirmed diagnostic and/or management decisions

23% led to changes in existing decisions

25% led to additional decisions


Searching for Evidence in a 3-day period after the Cart:

Expected searches = 10 Identified searching needs = 41 Only 5 searches (12%) carried out.


Can we get evidence to the bedside?

Yes, and it will improve patient care. But can we provide it in a less

cumbersome form?


EBM and Purchasing

In harmony:

When we clinicians stop doing things that are useless or harmful

When we use just-as-good but less expensive treatments, carers, and sites for care.


What we could save in Oxford

by switching from:

LASIX frusemide: £ 90,000

simvastatin cerivastatin: £ 500,000

TENORMIN atenolol: £ 700,000

diclofenac ibuprofen: £ 1,000,000

Total: £ 2,290,000 how many hips would these savings

purchase?


EBM and Purchasing

Still in harmony:

When we spend now to save later.


EBM and Purchasing

In grudging collaboration:

Waiting lists, once we understand the opportunity costs of shortening them:» it’s not about money» it’s about what else we won’t be able to do

if we shorten them


EBM and Purchasing

In conflict:

When we identify so strongly with a dying patient’s short-term goals that we use resources that we know would “add more QALYs” if used for other patients.


EBM and E-B Guidelines

EBM integrates evidence, expertise, and the unique biology and values of individual patients.

Local EB Provision ought to integrate evidence, expertise, and the unique biology and values of the local scene.



The best evidence comes from systematic reviews (such as Cochrane) and/or E-B journals of 2º publication:» Much more likely (than personal search and

critical appraisal) to be true » Saves the clinician’s precious (scarce!) time

Avoids error and duplication of effort



But NO systematic review can (or should try to) identify the “4 B’s:» Burden» Barriers» Behaviours» Balance

They can ONLY be determined at the local (or even patient) level


1. Burden

The burden of illness, disability, and untimely death that would occur if the evidence were NOT applied

the consequences of doing nothing


2. Barriers

Patient-values & preferences Geography Economics Administration/Organisation Tradition “Expert” opinion


3. Behaviours

The behaviours required from providers and patients if the evidence is applied.

All that guidelines can do is specify the former!


4. Balance

The opportunity cost of applying this guideline rather than some other one.


Killer B’s

Burden: too small to warrant action. Barriers: ultimately down to patients’

values. Behaviours: may not be achievable. Balance: may favour another guideline

over this one.


Two monumental wastes of time and energy

First, national/international evidence-summarising groups prescribing how patients everywhere should be treated.

Their expertise: predicting the health consequences if you do treat.

Their ignorance: the local B’s, and whether killer B’s are operating.


Two monumental wastes of time and energy

Second, local groups attempting to systematically review the evidence.

Their expertise: identifying the local B’s and eliminating the killer B’s

Their ignorance: searching for all relevant evidence; Chinese; performing tests for heterogeneity.


Applying a study result to my patient

Never interested in “generalising”

Am interested in a special form of extrapolation: particularising


Extrapolating (particularising) to my

individual patient:

First and foremost: Is my patient so different from those in the trial that its results can make no contribution to my treatment decision?

if no contribution, I restart my search if it could help, I need to integrate the

evidence with my clinical expertise and my patient’s unique biology and values...


To add Clinical Expertise and Patient’s Biology &

Values:

What is my patient’s RISK ?» of the event the treatment strives to prevent?» of the side-effect of treatment?

What is my pt’s RESPONSIVENESS? What is the treatment’s FEASIBILITY in

my practice/setting? What are my patient’s VALUES ?


To add Clinical Expertise and Patient’s Biology &

Values:

I begin by considering Risk and Responsiveness for the event I hope to prevent with the treatment:

The report gives me (or I can calculate) an Absolute Risk Reduction [ARR] for the average patient in the trial.

ARR = probability that Rx will help the average patient.


For example, Warfarin in nonvalvular atrial

fibrillation:

After 1.8 years of follow-up in an RCT: Control Event Rate (placebo) = 4.3% Exper. Event Rate (warfarin) = 0.9% so, for the average patient in the trial,

the probability of being helped, or Absolute Risk Reduction = (CER - EER) = 3.4% ACPJC

1993;118:42


How can I adjust that ARR for my pt’s Risk and Responsiveness?

Could try to do this in absolute terms:» my Patient’s Expected Event Rate: PEER» and multiply that by the RRR» and factor in my Patient’s expected

responsiveness Clinicians are not very accurate at

estimating absolute Risk and Responsiveness



Clinicians are pretty good at estimating their patient’s relative Risk and Responsiveness

So, I express them as decimal fractions:» f~risk (if at three times the risk, f~risk = 3)» f~resp (if only half as responsive [e.g., low

compliance], f~resp = 0.5)



probability that Rx will help my patient = ARR x f~risk x f~resp

If ARR is 3.4% and I judge that their f~risk is 3 and that their f~resp is 0.5 then the probability that warfarin will

help my patient = 3.4% x 3 x 0.5 = 5.1%


Must also consider the probability that I will do

harm:

In the case of warfarin: serious bleeding (requiring transfusion) from the g-i tract, or into the urine, soft tissues or oropharynx.

Absolute Risk Increase = 3% at 1 yr, so ARI estimated to be 5% in 1.8 years

ACPJC 1994;120:52


…and adjust the probability of harm for my

patient

Again, can express my clinical judgement in relative terms: f~harm

Given my patient’s age, I judge their f~harm to be doubled: 2

then the probability that Rx will harm my patient = ARI x f~harm =

5% x 2 = 10%


Can now begin to estimate the Likelihood of Help vs.

Harm

Probability of help: ARR (embolus) x f~risk x f~resp = 5.1%

Probability of harm: ARI (haemorrhage) x f~harm = 10%

My patient’s Likelihood of Being Helped vs. Harmed [LHH] is: (5.1% to 10%) or 2 to 1 against warfarin!

…or is it ?


The LHH has to include my patient’s values

I need to take into account my patient’s views (“preferences,” “utilities”) about the relative severity:

» of the bleed I might cause

» to the embolus I hope to prevent Expressed in relative terms = s~

» if the bleed is half as bad as the embolus, then s~ = 0.5


On in-patient services in Oxford and Toronto:

When Dr. Sharon Straus has described a typical embolic stroke (with its residual disability) and typical moderate bleed (brief hospitalisation and transfusion but no permanent disability):

for most of her patients, a bleed is only 1/5th as bad as a stroke

so the s~ is 0.2


So the LHH becomes:

{ARR for embolus} x {f~risk} x {f~resp} vs.

{ARI for bleed} x {f-harm} x {s~}

3.4% x 3 x 0.5 = 5.1% vs. 5% x 2 x 0.2 = 2%

LHH = 5.1 to 2 or 2.5 to1 » (I am more than twice as likely to help than harm my

patient if they accept my offer of Rx)


We can work out the LHH for most patients <6

minutes

To be feasible on our service: has to be “do-able” in 3 minutes.


Reactions from our patients

All are grateful that their values/opinions are being sought

1/3 want to see the calculations, perhaps change their value for s~, and make up their own minds.

1/3 adopt the LHH as presented. 1/3 say “Whatever you tell me, doctor!”

Title Place Date. Centre for Evidence-Based Medicine What I’ve done / do/don’t do l Done: I’ve...

Documents

Transcript of Title Place Date. Centre for Evidence-Based Medicine What I’ve done / do/don’t do l Done: I’ve...