vanderbilt sports medicine how to practice and teach ebm chapter 3 may 3, 2006
TRANSCRIPT
Vanderbilt Sports Medicine
How to practice and teach EBMHow to practice and teach EBMChapter 3Chapter 3
May 3, 2006May 3, 2006
Diagnosis and ScreeningDiagnosis and Screening
3 questions to consider about tests3 questions to consider about tests– Is the evidence about accuracy of a test Is the evidence about accuracy of a test
validvalid– Does evidence show test can distinguish Does evidence show test can distinguish
patients who do/don’t have disorderpatients who do/don’t have disorder– How can I apply test to a specific patientHow can I apply test to a specific patient
If the evidence is valid, is it If the evidence is valid, is it importantimportant– Ie: clinically worthwhileIe: clinically worthwhile
Validity of evidenceValidity of evidenceMeasurement (most important)Measurement (most important)– Was there an independent, blind Was there an independent, blind
comparison with a reference gold comparison with a reference gold standardstandard
RepresentativeRepresentative– Appropriate spectrum of patientsAppropriate spectrum of patients
AscertainmentAscertainment– Was reference standard ascertained Was reference standard ascertained
regardless of diagnostic test resultregardless of diagnostic test result
ImportanceImportance
Does evidence demonstrate ability of Does evidence demonstrate ability of test to distinguish patients w and w/o test to distinguish patients w and w/o disorderdisorder
SensitivitySensitivity– pts with disorder who have positive testpts with disorder who have positive test
SpecificitySpecificity– Pts without disorder who have negative Pts without disorder who have negative
testtest
ImportanceImportance
Likelihood ratio - LR(+)Likelihood ratio - LR(+)
Prob of positive test in presence of disorderProb of positive test in presence of disorder
Prob of positive test in absence of disorderProb of positive test in absence of disorder– LR+ = sensitivity/(1-specificity)LR+ = sensitivity/(1-specificity)– Positive test is more likely in pt w disorderPositive test is more likely in pt w disorder
LR(-) = (1-sensitivity)/specificityLR(-) = (1-sensitivity)/specificity– Negative test more likely in absence of Negative test more likely in absence of
disorddisord
ImportanceImportanceSnNoutSnNout– Test with high sensitivity (Sn)Test with high sensitivity (Sn)– Negative result (N)Negative result (N)– Rules OUT diagnosisRules OUT diagnosis
SpPinSpPin– Test with high specificity (Sp)Test with high specificity (Sp)– Positive result (P)Positive result (P)– Rules IN diagnosisRules IN diagnosis
Applying test to a patientApplying test to a patient
Is test available, affordable, Is test available, affordable, accurate, and precise accurate, and precise in our settingin our setting– Some tests have higher LR in florid Some tests have higher LR in florid
disease and lower LR in milddisease and lower LR in mild– As patients are referred to specialists, As patients are referred to specialists,
some tests loose power because some tests loose power because patients displaying symptoms will patients displaying symptoms will include more false-positivesinclude more false-positives
Applying test to a patientApplying test to a patientCan we generate an estimate of pre-test Can we generate an estimate of pre-test probabilityprobability– From experience, prevalence studies, practice From experience, prevalence studies, practice
databases, pre-test probability studiesdatabases, pre-test probability studies
Does post-test probability affect mgmt and Does post-test probability affect mgmt and help patienthelp patient– Does result cross threshold that causes us to Does result cross threshold that causes us to
stop testingstop testing(-) test that r/o likelihood of diagnosis(-) test that r/o likelihood of diagnosis
(+) test that r/in diagnosis(+) test that r/in diagnosis
– If not, perform additional tests to pursue If not, perform additional tests to pursue diagnosisdiagnosis
Test treatment thresholdsTest treatment thresholds
May not cross threshold until several May not cross threshold until several tests are performedtests are performed
If tests are independent, can chain If tests are independent, can chain LR together for each test to increase LR together for each test to increase overall likelihood ratiooverall likelihood ratio– Ex: LR test 1 (0.06) x LR test 2 (13) x LR Ex: LR test 1 (0.06) x LR test 2 (13) x LR
test 3 (11) = 9.13test 3 (11) = 9.13– Cvt LR to prob: 9.13/10.13 = 90% probCvt LR to prob: 9.13/10.13 = 90% prob
More than any single test aloneMore than any single test alone
Multilevel LRMultilevel LR
Expressing LR for two levels (+/-) is Expressing LR for two levels (+/-) is helpful, but giving LR for multiple helpful, but giving LR for multiple levels allows more distinctionlevels allows more distinction– Ex: 5 levels from extremely negative – Ex: 5 levels from extremely negative –
mod negative – neutral – mod pos – extr mod negative – neutral – mod pos – extr pospos
– Extremely positive more definitive than Extremely positive more definitive than if only had two levels (pos/neg), so if only had two levels (pos/neg), so result is more usefulresult is more useful
Multiple TestsMultiple Tests
Multiple tests grouped togetherMultiple tests grouped together– Considered a “cluster”Considered a “cluster”– Combination of results helps make Combination of results helps make
diagnosisdiagnosis– When validated in second independent When validated in second independent
groupgroup““Clinical prediction guideline”Clinical prediction guideline”
Critically Appraised Topic (CAT)Critically Appraised Topic (CAT)
Standardized, one page summary of Standardized, one page summary of evidence on one topicevidence on one topic– Shows bottom line resultShows bottom line result– Show scenario and clinical questionShow scenario and clinical question– Show studies and gold standardShow studies and gold standard– Presents evidence and resultsPresents evidence and results– ReferencesReferences
Screening/Case FindingScreening/Case Finding
ScreeningScreening– Making early diagnosis of pre-Making early diagnosis of pre-
symptomatic disease among well symptomatic disease among well general publicgeneral public
Case FindingCase Finding– Making early diagnosis of pre-Making early diagnosis of pre-
symptomatic disease in patients who symptomatic disease in patients who presented for an presented for an unrelatedunrelated disorder disorder
Screening/Case FindingScreening/Case FindingImplicitly recruit pts with promise that Implicitly recruit pts with promise that they will live longer or better if they are they will live longer or better if they are testedtested
Requires that evidence shows pts are Requires that evidence shows pts are better off in long run with early diagnosisbetter off in long run with early diagnosis– Labels all pts (may be harmful)Labels all pts (may be harmful)– False + test is only harmful (no benefit)False + test is only harmful (no benefit)– Test must not only be accurate, but treatment Test must not only be accurate, but treatment
must be efficaciousmust be efficacious
Screening/Case FindingScreening/Case FindingDoes RCT evidence show early diagnosis Does RCT evidence show early diagnosis leads to improved survival/quality of lifeleads to improved survival/quality of life– RCT type 1 – screen vs no screenRCT type 1 – screen vs no screen– RCT type 2 – (+) screen test – treat vs no treatRCT type 2 – (+) screen test – treat vs no treat
Will early diag pts cooper with treatmentWill early diag pts cooper with treatment– If not, no benefitIf not, no benefit
How do benefits/harms compareHow do benefits/harms compareDoes frequency/severity of disorder Does frequency/severity of disorder warrant the effortwarrant the effort
Screening/Case FindingScreening/Case Finding
BiasBias– Early detection always appears to improve Early detection always appears to improve
survival (may in fact not) - “pt sicker longer”survival (may in fact not) - “pt sicker longer”– Pts who volunteer for testing are more cooper Pts who volunteer for testing are more cooper
with health advicewith health advice
Have better outcomes anywayHave better outcomes anyway– Early tests ID pts with slower/more benign dzEarly tests ID pts with slower/more benign dz– Is follow-up adequateIs follow-up adequate
Early detection is not always a good thing!Early detection is not always a good thing!