EBLM Pres Final

8/3/2019 EBLM Pres Final

1/101


2/101

"THE THREE MAIN TASKS OF THE

CLINICIAN AREDIAGNOSIS, PROGNOSIS, ANDTREATMENT.

OF THESE DIAGNOSIS IS BY FAR THE

MOST IMPORTANT, FOR UPON IT THESUCCESS OF THE OTHER TWO

DEPENDS."

RYLE J.A.The natural history of disease 2nd ed.

Oxford University Press, 1948


3/101

What I will be sharing with youToday?


4/101

1.What is evidence based laboratory medicine?

2.What are the components of EBLM?

3. How to ask a question?

4. How to acquire information?

5.How to analyze the information?

6.How to apply the information?

7.Critics view of EBLM.


5/101

What is Evidence basedMedicine ?


6/101


7/101

EBLM

Conscientious explicit and judicious use of current best

evidence in Laboratory medicine for making well

informed decision


8/101

Individual expertiseBest external evidence

Patients values

&expectation

EBLM

COMPONENTS OF EBLM


9/101

Why evidence basedMedicine?


10/101

Increased innovation New technologies

Greater knowledge New treatments & Diagnostics

Increased workload More patient visits

More spending Salary and other costs

Patient expectation More knowledge from internet

Legal aspects


11/101


12/101

What are the justification for an evidencebased medicine?

Constant requirement for information

Constant addition of new information

Limited time availability

The poor quality of access to good

information


13/101

Limited number and poor quality of studies linking testResults to patients benefits.

The poor perception of the value of diagnostic tests.

The ever increasing demand for tests.

The disconnected approach to resource allocation.

Silo budgeting

What is particular to laboratory medicine?


14/101

How to practice ?

1. Identification of question

2. Track down the best evidence

3. Critical assessment of the best evidence.

4. Implementation of best practice.

5. Evaluate

ASK

Acquire

Appraise

ACT

AUDIT


15/101

Elements of EBLM


16/101

Convert a clinical situation into a searchable,(and hopefully answerable) question using

PICOPATIENT

INTERVENTION

COMPARISON

OUTCOME


17/101

atient or Problem

ntervention

omparison

utcome

Patient refers tothe person

presenting with theproblem, or moresimply, to theproblem itself. Bothconcepts areimportant insearching.


18/101

atient or Problem

ntervention

omparison

utcome

Interventionrefers to theaction taken in

response to theproblem. This isoften a drug or

surgicalprocedure, but itcan take many

forms


19/101

atient orProblem

ntervention

omparison

utcome

Comparison refersto the benchmark

against which theintervention ismeasured. Often it

refers to anothertreatment, notreatment, or aplacebo.


20/101

atient or Problem

ntervention

omparison

utcome

Outcomerefers to the

anticipatedresult of theintervention.


21/101

How to apply this for EBLM?


22/101

QUESTIONS TO BE ASKED

CARO QUESTIONS

C: Case What are the patient characteristics,conditions, symptoms, demographics ?

A: Assay Which procedure or strategy is considered ?

R:Reference

What is the standard procedure, thecomparator ?

O:Outcome

What is the interest, the diagnostic validity? Sensitivity, specificity, predictive values,prognosis ?


23/101

Types of question


24/101

Type I : Regarding diagnostic accuracy of the test

1.Patients presenting to the emergency departmentWith shortness of breath.

2.How well does N terminal pro B type natriureticpeptide

4. Predict heart failure as assessed by

3. The cardiac ejection fraction measured byEchocardiography


25/101

Type II : Related to the value of test in improving

Patients outcomes.

1. Patient admitted to the hospital for treatment of

heart failure.

2. How well does the use of N terminal Pro B typeNatriuretic peptide as a guide to therapy.

3. Improve the length of hospital stay and the rateOf subsequent readmission for heart failure ?


26/101

How to Acquire evidence ?


27/101

In laboratory medicine an alternative toClinical trail is Diagnostic accuracy studies.

The best design for diagnostic accuracy

Studies is a prospective cohort study with a

Blinded comparison of the performance of

Experimental test and that of an appropriate

Gold standard test in a spectrum of patients

Suspected to having the disease in question.


28/101

An important goal of studies of diagnostics

test is to Determine whether the new test

adds information to that known from patient

observation or other investigations


29/101

How to start a search ?


30/101

Computer system

Clinical Evidence or PIER (UpToDate)

ACP Journal Club, InfoPOEMS, Dynamed

Cochrane Library, PubMED ClinicalQueries, BMJUpdates, guidelines

Original Studies

OR SUMsearch or TRIP

How to seek evidence-based information


31/101

Choosing Resources

ForegroundBackground

Rare

Co

mmon

Unfiltered

Database(e.g. MEDLINE)

Filtered/Pre-appraised

Evidence

Textbooks


32/101

Where to search ?


33/101

It is best to start the search with looking forExternal evidence based guidelines that can beAdapted.

The search for evidence usually starts in databasesSuch as the Cochrane Library which contains high qualitySystematic reviews or meta analysis.


34/101

If a search is not successful in the secondaryLiterature one can look for primary reports in the

Medline.Use Pub Med for the search of Medline.

The best single search term for laboratory test

Is sensitivity .

However the word diagnostic test, DiagnosisDiagnostic use combined with the corresponding

Clinical condition ( eg: Chronic renal failure)andFinally the name of the test ( eg: Soluble transferrinReceptor.


35/101

Determine the level of evidence of the primaryStudies and reviews.

The highest level of evidence is a good quality wellConducted systematic review or meta analysis of RCTfor testing patient related outcomes.

( PSA for Screening Prostate cancer )

Prospective cohort studies for Diagnostic accuracy studies.

( Total PSA Vs the free PSA / Total PSA in the diagnosisOf prostate cancer )


36/101

What and Why do we choose

a systemic review?


37/101

Systematic SearchingSystematic Reviews


38/101

DefinitionsReview articles

A broad overview of a topic,similar to a textbookchapter.

Often covers multiple,background aspects of a

disease such as natural history,etiology, epidemiology, signs &symptoms, diagnosis,treatment, and prognosis.

The article summarizes theresults from many otherprimary studies.

The studies to summarize are

chosen at the discretion of theauthor.


39/101

DefinitionsReview articles

A broad overview of a topic,similar to a textbookchapter.

Studies are chosen using a

standardized protocol to minimizeselection bias.

Systematic Review

A type of review article thatfocuses on a focused clinicalquestion


40/101

Definitions

Review articles

A broad overview of a topic,similar to a textbookchapter.Systematic Review

A type of review article thatfocuses on a focused clinicalquestion

Meta-analysis

A type of systematic review in whichthe numerical results from individualstudies are mathematically combinedto give a single, overall estimate of

treatment effect.


41/101

Definitions

Review articles

Systematic Review

Meta-analysis

A systematic review can be thought of as aresearch project done on the medical literatureitself.

Instead of human beings acting as subjects, thesubjects of a systematic review are individual

RCTs


42/101

Finding Systematic Reviews

Produces high quality systematic reviews

Managed by the Cochrane Collaboration

A not-for-profit international organizationand one of the initial developers of

systematic reviews Available through the HSLIC web site.


43/101

Finding Systematic Reviews Pub Med Clinical Queries

They are accessed from the "Clinical Queries"link on the blue side bar of the PubMed homepage.


44/101


45/101

How to critically appraise an

Evidence?

E l


46/101

Essential ConceptsThree concepts are essential to understanding the

critical appraisal of systematic reviews. These are:

Publication bias. Publication bias is one of thefactors that systematic reviews attempt to avoid byselecting studies in a systematic way.

Heterogeneity. Heterogeneity is a statisticalmeasure of the difference between the results fromdifferent studies. The less heterogeneous resultsare, the easier it becomes to estimate overalleffect.


47/101

HOW TO DETECTHETEROGENICITY?


48/101

Forrest Plots


49/101

D'Souza, A. L et al. BMJ 2002;324:1361

Effect of probiotics on the risk of antibiotic associated diarrhoea


50/101

Forest plots. These graphicaldisplays show study data in a waythat makes it easy to see similaritiesand differences between studies.


51/101

Look at the title of the forest plot,

the intervention, outcome effectmeasure of the investigation and thescale

Th l b l ll h h i d


52/101

The label tells you what the comparison andoutcome of interest are



53/101

Scale measuring treatment effect.Take care when reading labels!



54/101

The names on the left are theauthors of the primary studiesincluded in the MA


55/101

Each study has an ID (author)



56/101

Treatment effect sizes for each study(plus 95% CI)



57/101

The small squares represent theresults of the individual trial results

The size of each square representsthe weight given to each study inthe meta-analysis

Horizontal lines are confidence intervals


58/101

Horizontal lines are confidence intervalsDiamond shape is pooled effectHorizontal width of diamond is confidence interval



59/101

The vertical line represents the line of

no effect, i.e. where there is no

statistically significant difference

between the treatment/intervention

group and the control group


60/101

The vertical line in middle is the line of no effectFor ratios this is 1, for means this is 0



61/101

Pooled Se = 0.71Heterogeneity p


62/101

Average men having an average meal


63/101

How to detect Bias?

F l l


64/101

64

Funnel plots

A funnel plot is a scatter plot oftreatment effect against a measure

of study size.


65/101

65

Funnel Plots

attempt to detect bias in studyselection

results of each study plotted againstsample size

what should we expect?


66/101

66

Why Funnel?

precision in the estimation of the truetreatment effect increases as the sample sizeincreases.

Small studies scatter more widely at thebottom of the graph

In the absence of bias the plot shouldresemble a symmetrical inverted funnel


67/101

67

Funnel Plot

Favors Treatment Favors Control

Odds Ratio


68/101

68

Funnel Plot


Odds Ratio


69/101

69

Funnel Plot


Odds Ratio


70/101

70

Funnel Plot


Odds Ratio


71/101

71


72/101

72


73/101

73

Publication Bias

Asymmetrical appearance ofthe funnel plot with a gap in abottom corner of the graph

Drawbacks to systematic


74/101

Drawbacks to systematicreviews/meta-analyses

Can be done badly 2 systematic reviews on same topic can have

different conclusions

Inappropriate aggregation of studies A meta-analysis is only as good as the

papers included Tend to look at broad questions that

may not be immediately applicable toindividual patients


75/101

How to rate or gradethe evidence?

Quality of primary studies and reviews


76/101

Q y f p y

Rating of the level of evidence of individual articles

1a

1b

II

III

IV

Meta analysis or systematic review based on atleast several level 1b studies

Diagnostic trial or outcome study of good quality

Diagnostic trial or outcome study of medium qualityInsufficient patients or other trials( Case control or other designs)

Descriptive studies , case reports etc

Statement of committees, opinion of experts,not systematic

Rating of the strength of the evidence supporting


77/101

g g pp gGuidelines recommendations

A

B

C

D

Supported by at least by two independentStudies of level 1b or one review of 1a

Supported by at least two independent studiesof level II

Not supported by sufficient studies of level

I of II

Advices of experts

Compile an evidence table


78/101

Compile an evidence table

1.Publication details of the individual studies.

2. Study design

3.Spectrum of patient and patient setting.

4.Prevalence of the condition.

5.Diagnostic test used of compared.

6.Out come measured.

7.Effects measured including measures of diagnosticaccuracy.


79/101

8. Comments on specific issues raised by the study.( biases)

9.Quality rating and level of evidence of the study.

Make the judgment based on:


80/101

Make the judgment based on:

1.Quality of the evidence :

The extent to which the studys design, conduct,And analysis have minimized selection, measurement andConfounding bias.

2.Quantity of evidence:

The number of studies that have evaluated the given

Topic and the sample size of each study.

3. Consistency of the evidence.

Meta-analysis Software


81/101

Meta analysis Software Free

RevMan [ReviewManager] Meta-Analyst Epi Meta Easy MA

Meta-Test Meta-Stat

Commercial Comprehensive Meta-

analysis

Meta-Win WEasy MA

General stats packages Stata SAS

S-Plushttp://www.prw.le.ac.uk/epidemio/personal/ajs22/meta/

Diagnostic accuracy studies allow the


82/101

Diagnostic accuracy studies allow the

calculation of various statistics that

provide an indication of "test

performance" how good the index test is

At detecting the target condition.

Whiting et al. in: BMC Medical ResearchMethodology 2003http://www.biomedcentral.com/1471-2288/3/25

Do we need a detailed statistical and epidemiological


83/101

Do we need a detailed statistical and epidemiologicalskills To practice EBLM ?

No

Then what is needed ?

Critical appraisal skill

Competent understanding of the strengths and weaknessof systemic Reviews and meta analysis

The laboratory personnel must direct more effect todemonstrate the impact of laboratory tests on a greatervariety of clinical outcomes.

DIAGNOSIS WORKSHEET


84/101

Are the Results of This Diagnostic Study Valid?

Was there an independent, blind comparisonwith a reference (gold) standard ofdiagnosis?

Was the diagnostic test evaluated in an

appropriate spectrum of patients (likethose in whom it would be used inpractice)?

Was the reference standard appliedregardless of the diagnostic test result?

Was the test (or cluster of tests) validatedin a second, independent group ofpatients?

Can We Apply This Valid, Important Evidence


85/101

pp y , pAbout a Diagnostic Test in Caring for Our Patient?

Is the diagnostic test available, affordable,accurate, and precise in our setting?

Can we generate a clinically sensible estimate ofour patients pre-test probability (from

personal experience, prevalence statistics,practice databases, or primary studies)?

Will the resulting post-test probabilities affectour management and help our patient?

*Could it move acrosis a test-treatment

threshold? *Would our patient be a willingpartner in carrying it out?

Would the consequences of the test help ourpatient?

STARD (Standards for reporting diagnostic


86/101

D ( n r f r r p r ng gnaccuracy) - a checklist

Introduction Diagnostic accuracy between tests or across

patient groupsProbands Demographic description, inclusion and exclusion

criteria, symptoms, data collection criteria.

Study design Time frame, number and group of probands, timeof measurements, treatment of probands

Referencestandard

Description of standard and rationale forcomparison.

Test method Technical, analytical specifications (linearity, cut-off levels, uncertainty, bias, etc)

Statisticalmethods Methods for reporting diagnostic validities,comparisons between groups, test reproducibility

Results Cross tabulaton of results (reference, test),analytical and diagnostic acuracy between groupsof probands, ROC-curves, Box-Whiskers plot.

Conclusion Clinical application P. M. Bossuyt et al. 2003

Evidence of performance designed to facilitate decisionki


87/101

Decisions

Cost effectiveness

Organizational impact

Clinical impact

Diagnostic TherapeuticOutcome

Diagnostic performance

Technical performance

making


88/101

How to act and Modify ?


89/101

Test Question Result Action Outcome

TroponinI

Has the patienthad a MI

7.2g/L Decide to admit,

Intensive care

Decreasedmorbidity &mortality

BNP Is this

breathlesspatientsufferingfromHeart

failure

56ng/L Seek alternative

diagnosticmethod

Avoid

incorrectdiagnosis &treatment

HbA1C Is this patientcomplying withtreatmentprotocol

10.6%( Nochangein ayear

Consider changingTreatment,closer monitoringand freq visit

Persistentlyhigh value hasincreased riskofcomplications

Promises of EBLM


90/101

It ties clinical practices to scientific standards ofevidence

Able to draw upon the objective experience of manyresearchers working with accepted scientific

standards of evidence

EBLM should also promote greater uniformity

Evaluate implementing cost cutting measures

EBM should provide a scientific basis for the constructionof public policy

Critics


91/101

Standard guidelines Disincentives of individual innovation

Becomes more like cook book medicine

Lower standards by deskilling practitioners

Instead to clinical judgment practitioners will beencouraged to Use protocols

Incapable of operating effectively in diverse situation


92/101

Is the highest level of evidence always the strongest

Recommendation ?

NO


93/101

Highest level of evidence may not provide the

Strongest recommendations in some local contest.

The evidence must be supplemented with considered

Judgment of the potential clinical benefits and harms

Patients preferences

The organizational and economic impact of testing.

h l h


94/101

In patients presenting with complaints with symptomsOf tongue and mouth the prevalence of Vit B12

Deficiency in only 8%.

The relatively low cost of Testing for B12 deficiencyAnd availability of effective treatment may counterBalance the low probability of this cause.

Might lead to recommendation of B12 testing inOne community .

But not so in another community because the relativeCosts may be different.


95/101

An example where patients choices are consideredIs the triple test used for antenatal screening ofDowns screening.

The consequences of positive screening test isAmniocentesis which may harm the fetus.

And in positive cases an abortion may be required.

BUT


96/101

BUT

Good professionals should treat guidelines more as options.

As True standards and professional organizations do not

enforce adherence.

Change in health care is possible with guidelines.

Its creation and Implementation reflects the collaborative

nature of health care.

Future


97/101

Establish a culture of EBLM

How ?

Change the pattern of JournalClub start from the Residents

Evaluating a systemic review orEven journal can be even a part

Of MD evaluation

Critical appraisal checklists


98/101

pp

CASP (Critical Skills Appraisal Programme)

http://www.phru.nhs.uk/casp/critical_appraisal_tools.htm

JAMA Users Guides to the Medical Literature

http://www.cche.net/usersguides/main.asp

Crombie I (1996) The Pocket Guide to CriticalAppraisal, BMJ Books, London

Greenhalgh T (2001) How to Read a Paper, BMJ Books,London

BestBETs CA database http://www.bestbets.org/cgi-bin/browse.pl?~show=appraisal
http://www.phru.nhs.uk/casp/critical_appraisal_tools.htmhttp://www.cche.net/usersguides/main.asphttp://www.bestbets.org/cgi-bin/browse.pl?~show=appraisalhttp://www.bestbets.org/cgi-bin/browse.pl?~show=appraisalhttp://www.bestbets.org/cgi-bin/browse.pl?~show=appraisalhttp://www.bestbets.org/cgi-bin/browse.pl?~show=appraisalhttp://www.cche.net/usersguides/main.asphttp://www.phru.nhs.uk/casp/critical_appraisal_tools.htm


99/101

There are different checklists for different studyDesigns at:

1.The centre for Evidence Based Medicine( WWW.cebm.net)

2.Casp International network( WWW.caspinternational.org.uk )

3. Centre for Health Evidence( WWW.cche.net )

Impact of EBLM
http://www.cebm.net/http://www.caspinternational.org.uk/http://www.cche.net/http://www.cche.net/http://www.caspinternational.org.uk/http://www.cebm.net/


100/101

Impact of EBLM


101/101

THANK YOU

For your patient listening

EBLM Pres Final

Documents