UNIT VII EPIDEMIOLOGICAL METHODS PDF.pdf

Epidemiological Methods

1. EpidemiologicalMethods
1Burhan Uddin, Karachi

2. LearningObjectives
Atthecompletionof thisunitlearners willbe ableto:
1.DiscusstheDescriptivein termofTime,place and
Person.
2.DiscusstheAnalyticaland basicConcepts ofCross
Sectional
3. DescribetheIntervention/ Experimentalstudy
2Burhan Uddin, Karachi

3. Epidemiology
Epidemiologyisthestudyofthedistribution
anddeterminantsofhealth-relatedstatesor
events(includingdisease),
and
theapplicationofthisstudytothe control of
diseasesand other healthproblems.
(WHO2017)
3Burhan Uddin, Karachi

4. StudyDesign/Method
A studydesignisa specificplanor protocolfor
conductingthestudy,whichallowsthe
investigatortotranslatetheconceptual
hypothesisintoanoperationalone.
4Burhan Uddin, Karachi

5. EpidemiologyMethods
Experimental
(RCTs)
Observational
Analytical Descriptive
Case-ControlCohort
+cross-sectional&ecologic
5Burhan Uddin, Karachi

6. EpidemiologyMethods
•Descriptivestudies
–Examinepatternsofdisease
•Analyticalstudies
–Studiesofsuspectedcausesofdiseases
•Experimentalstudies
–Comparetreatmentmodalities
6Burhan Uddin, Karachi

7. EpidemiologyMethods
7Burhan Uddin, Karachi

8. HierarchyofEpidemiologicStudyDesign
Tower&Spector,2007 8Burhan Uddin, Karachi

9. ObservationalStudies
(nocontroloverthecircumstances)
-Descriptive:Mostbasic demographicstudies
*CaseReport
*CaseSeries
*Crosssectional
*Ecological/Correlationstudy
-Analytical:Comparativestudiestestinganhypothesis
*Cross-sectional
(asnapshot;noideaoncause-and-effectrelationship)
*Cohort
(prospective;cause-and-effectrelationshipcanbeinferred)
* Case-control
(retrospective;cause-and-effectrelationshipcanbeinferred)
9Burhan Uddin, Karachi

10. EpidemiologyMethods
10Burhan Uddin, Karachi

11. CaseReport
•What?
•Theprofileofasingle
patientisreportedin
detailbyoneormore
clinicians
•Example
•In1961,apublished
casereportofa 40
year-old women who
developed pulmonary
embolismafter
beginninguseoral
contraceptive
11Burhan Uddin, Karachi

12. CaseSeries
•What?
•Anindividualcase
reportthathasbeen
expandedtoincludea
numberofpatients
withagiven disease
•Example
•InLosAngeles,five
young homosexuals
men, previously
healthy,were
diagnosedwith
pneumocystcariini
pneumoniain
6-monthsperiod
12Burhan Uddin, Karachi

13. EcologicalorCorrelation
•EcologicalStudies
–wholepopulationis the unitofanalysis
–relationshipbetweenexposureandoutcomeat the
individuallevelismissing(incompletedesign)
–ecologicalfallacy
13Burhan Uddin, Karachi

14. AnalyticalStudies
(comparativestudiestestingan hypothesis)
•Cohort(prospective)
–Begins with an exposure (smokers and non-
smokers)
•Case-control(retrospective)
–Begins with outcome (cancer cases and healthy
controls)
14Burhan Uddin, Karachi

15. Population
Peoplewithout
disease
Exposed
Notexposed
Disease
Nodisease
Disease
Nodisease
CohortStudies
15Burhan Uddin, Karachi

16. time
Study beginshere
Study
population
freeof
disease
Factor
present
Factor
absent
disease
nodisease
disease
nodisease
present
future
16

17. AdvantagesofCohortStudies
Canestablishpopulation-basedincidence
Accuraterelativerisk(riskratio)estimation
Can examinerareexposures(asbestos>lungcancer)
Temporalrelationshipcanbeinferred(prospectivedesign)
Time-to-eventanalysis ispossible
Canbeused whererandomizationisnotpossible
Magnitudeofariskfactor’seffectcanbequantified
Selectionandinformationbiases aredecreased
Multipleoutcomescanbestudied
smoking> lungcancer,COPD,larynxcancer)
17Burhan Uddin, Karachi

18. DisadvantagesofCohortStudies
Lengthyandexpensive
Mayrequireverylargesamples
Notsuitableforrare diseases
Notsuitablefordiseaseswithlong-latency
Unexpected environmental changes may influence the
association
Non-response,migration and loss-to-follow-upbiases
Sampling, ascertainmentandobserverbiasesare stillpossible
18Burhan Uddin, Karachi

19. time
Exposure Studystarts
Disease
occurrence
ProspectiveCohortStudy
ExposureStudystarts
Disease
occurrence
time
EPIET19Burhan Uddin, Karachi

20. Retrospectivecohortstudies
Exposure
time
Disease
occurrence
Studystarts
EPIET20Burhan Uddin, Karachi

21. Population
Cases
Controls
Exposed
Case-ControlStudies
Notexposed
Exposed
Notexposed
21Burhan Uddin, Karachi

22. Case-ControlStudies
Schulz&Grimes,2002
22Burhan Uddin, Karachi

23. AdvantagesofCase-ControlStudies
Cheap,easyandquickstudies
Multipleexposurescanbeexamined
Rarediseasesanddiseaseswithlonglatencycan
bestudied
Suitablewhenrandomizationisunethical
(alcoholandpregnancyoutcome)
23Burhan Uddin, Karachi

24. Disadvantagesof Case-ControlStudies
Caseandcontrolselectiontroublesome
Subjecttobias(selection,recall,misclassification)
Directincidenceestimationisnotpossible
Temporalrelationshipis notclear
Multipleoutcomescannotbestudied
If theincidenceofexposureishigh,itisdifficultto
showthedifferencebetweencases andcontrols
Reversecausationisaproblemininterpretation
24Burhan Uddin, Karachi

25. Application Exercise:
Case/ControlStudy
•Describe a case/control study on the
relationshipbetweenchildhoodobesity,
smokinghistory,andoccurrenceof
hypertensioninmiddle-aged men.
•Whatresearchquestioncanweanswer?
25Burhan Uddin, Karachi

26. Cross-SectionalStudies
•Measurement ofriskandoutcome atthe
sametime.
Riskfactor
Outcome
26

27. Cross-SectionalDesign
•Theonlystudy
capableofcalculating
prevalence.
–Proportion of the
populationwith
theoutcomeatany
point intime.
27Burhan Uddin, Karachi

28. ApplicationExercise:
Cross-SectionalStudy
•Designacross-sectionalstudythatexamines
therelationshipbetweendietarysodiumand
hypertensioninmiddle-agedmen.
•Whatresearchquestioncanweanswer?
28Burhan Uddin, Karachi

29. AdvantagesofCross-SectionalStudies
•Cheap andquickstudies.
•Dataisfrequentlyavailablethroughcurrent
recordsorstatistics.
•Idealforgeneratingnewhypothesis.
29Burhan Uddin, Karachi

30. DisadvantagesofCross-SectionalStudies
Theimportanceof therelationshipbetween the
causeand theeffectcannotbedetermined.
•Temporalweakness:
–Cannotdetermineifcauseprecededtheeffector
theeffectwasresponsibleforthecause.
–Therulesofcontributorycause cannotbe fulfilled.
30Burhan Uddin, Karachi

31. Typeof
Study
Alternative
Name
ClinicalTrials
Unitof
Study
PatientsRandomisedControlled
Trials
FieldTrials
CommunityTrials
Community
InterventionStudies
HealthyPeople
Communities
31Burhan Uddin, Karachi

32. Types oftrials
BlindedNotblinded
Randomised Notrandomised
ControlledNotcontrolled
Trial
32Burhan Uddin, Karachi

33. time
Studybeginshere(baselinepoint)
Study
population
outcome
Intervention
nooutcome
outcome
Control
nooutcome
baseline
future
RANDOMIZATION
33Burhan Uddin, Karachi

34. Intervention Randomisation
Control Blinding
34Burhan Uddin, Karachi

35. ExperimentalStudiesareUseful forEvaluating
•Newdrugorothertreatmentfordisease
•Newmedical/healthcaretechnology
•Methodsofprevention
•Methodsofhealthpromotion
•Newhealthprotection policies
•Programsforscreeninganddiagnosis
•Methodsofprovidinghealthcare
•Newhealthcarepolicies
35Burhan Uddin, Karachi

36. EthicalConsiderationsin
ExperimentalStudies
•Isproposedtreatment safe?
•Forthe sakeoftrial, cana treatmentethicallybe
withheld?
•Whatpatientsmaybebroughtintotrialand
allocatedrandomlytotreatments?
•Is itethicaltouseaplaceboor dummytreatment?
•Isit properforthetrialtobein anywaymasked?
AdaptedfromHill(1977)
36Burhan Uddin, Karachi

37. Advantages(I)
–the “goldstandard”ofresearchdesigns.
They thus provide the most convincing
evidenceofrelationshipbetweenexposure
andeffect.Example:
•trialsofhormonereplacementtherapyin
menopausalwomenfoundnoprotection
forheartdisease,contradictingfindings
ofpriorobservationalstudies
37Burhan Uddin, Karachi

38. Advantages (II)
Best evidencestudy design
Noselectionbias(usingblinding)
Controllingforpossibleconfounders
ComparableGroups(usingrandomization)
38Burhan Uddin, Karachi

39. Disadvantages
•Isthe mostexpensivestudydesignin termsof
money,time,andnumberofpatients.
–Issuesof patientattritionandcompliancemay
invalidatetheresults.
–Can be problematicforethicalreasons.
•Useofplacebo
•Harmoutweighingbenefits
•Zerotolerancefor some exposures
39Burhan Uddin, Karachi

40. 40Burhan Uddin, Karachi

41. 41Burhan Uddin, Karachi