stat 13, intro. to statistical methods for the life and health ...frederic/13/f17/13day08.pdfmean...
TRANSCRIPT
Stat 13, Intro. to Statistical Methods for the Life and Health Sciences.
1. Observational studies and experiments. 2. Experiments and aspirin example. 3. Random sampling, random assignment, and blocking. 4. Blinding. 5. Portacaval shunt example. 6. Coverage, non-response bias, adherer bias, and clofibrate example. 7. More about confounding factors.
Finish reading chapter 4. http://www.stat.ucla.edu/~frederic/13/F17 .HW3 is due Tue Nov 7. 4.CE.10, 5.3.28, 6.1.17, and 6.3.14. In 5.3.28d, use the theory-based formula. You do not need to use an applet.The midterm is Tue Nov 7. There is no lecture Thu Nov 9.
1
HW3 is due Tue Nov 7. 4.CE.10, 5.3.28, 6.1.17, and 6.3.14.
4.CE.10 starts out "Studies have shown that children in the U.S. who have been spanked have a significantly lower IQ score on average...."
5.3.28 starts out "Recall the data from the Physicians' Health Study: Of the 11,034 physicians who took the placebo ...."
6.1.17 starts out "The graph below displays the distribution of word lengths ...."
6.3.14 starts out "In an article titled 'Unilateral Nostril Breathing Influences Lateralized Cognitive Performance' that appeared ...."
2
1. ObservationalstudiesversusExperiments
Section4.2
ObservationalStudiesvs.Experiments
• Inanobservationalstudy,thesubjectsthemselvesdecidetheirleveloftheexplanatoryvariable.• Forexample,theresearchersdidn’tcontrolwhichchildrensleptwithanightlightonornot.• Observationalstudiesalwayshavepotentialconfoundingvariablespresentandthesemaypreventusfromdeterminingcauseandeffect.
Nightlightsandnear-sightedness
• Couldparents’eyesightbeanotherexplanation?• Maybeparentswithpooreyesighttendtousemorelighttomakeiteasiertonavigatetheroomatnightandparentswithpooreyesightalsotendtohavechildrenwithpooreyesight.• Nowwehaveathirdvariableofparents’eyesight• Parents’eyesightisconsideredaconfoundingvariable.• Otherpossibleconfounders?Wealth?Books?Computers?
ConfoundingVariables• Aconfoundingvariableisassociatedwithboththeexplanatoryvariableandtheresponsevariable.• Wesayitisconfoundingbecauseitseffectsontheresponsecannotbeseparatedfromthoseoftheexplanatoryvariable.• Becauseofthis,wecan’tdrawcauseandeffectconclusionswhenconfoundingvariablesarepresent.
ConfoundingVariables• Sinceconfoundingvariablescanbepresentinobservationalstudies,wecan neverconcludecausationfromasingleobservationalstudy.• Thisdoesn’tmeantheexplanatoryvariableisn’tinfluencingtheresponsevariable.Associationmaynotimplycausation,butcanbeaprettybighint.• SometimesthroughaseriesofcarefullydoneobservationalstudieseveryplausibleconfounderhasbeenaccountedforandthenpeopleconcludeXcausesY.Butitisneverasconclusiveaswithexperimentalresults.•
2.Experimentsandaspirinexample.
• Inanexperiment,theresearcherssettheleveloftheexplanatoryvariableforeachsubject.• Theselevelsmaycorrespondtoatreatmentandcontrol.• Welldesignedexperimentscancontrolforconfoundingvariablesbymakingthetreatmentandcontrolgroupssimilarexceptforwhattheexperimentermanipulates.
Aspirinexample.
Physicians’HealthStudyI (studyaspirin’saffectonreducingheartattacks.
• Startedin1982with22,071malephysicians.
• Thephysicianswererandomlyassignedintooneoftwogroups.
• Halftooka325mgaspirineveryotherdayandhalftookaplacebo.
Results• Intendedtogountil1995,theaspirinstudywasstoppedin1988afterfindingsignificantresults.• 189(1.7%)heartattacksoccurredintheplacebogroupand104(0.9%)intheaspiringroup.(45%reductioninheartattacksfortheaspiringroup.)• Whataboutconfoundingvariables?Couldtheaspiringroupbedifferentthantheplacebogroupinsomeotherways?• Didtheyhaveabetterdiet?• Didtheyexercisemore?• Weretheygeneticallylesslikelytohaveheartattacks?• Weretheyyounger?
TheBigIdea• Confoundingvariablesareoftencircumventedinexperimentsduetotherandomassignmentofsubjectstotreatmentgroups.• Randomlyassigningpeopletogroupstendstobalanceoutallothervariablesbetweenthegroups.• Soconfoundingvariables,includingonestheresearchersdidn'tanticipate,shouldberoughlyequalizedbetweenthetwogroupsandthereforeshouldnotbeconfounding.• Thus,causeandeffectconclusionsaresometimespossibleinexperimentsthroughrandomassignment.
Randomsamplingandrandomassignment.• Withobservationalstudiesorexperiments,randomsamplingisoftendone.Thispossiblyallowsustomakeinferencesfromthesampletothepopulationwherethesamplewasdrawn.• Withexperiments,randomassignmentisdone.Thispossibleallowsustoconcludecausation.
• ThePhysician’sHealthStudyusedrandomassignment.Diditalsouserandomsampling?• No,hardlyanyexperimentsuserandomsampling,butgettheirsubjectsinotherways.• ThePhysician’sHealthStudysentoutinvitationlettersandquestionnairestoall261,248malephysiciansbetween40and84yearsofagewholivedintheUnitedStates.• Ofthe59,285whowerewillingtoparticipateinthetrial,26,062weretoldtheycouldnotbecauseofsomemedicalconditionorcurrentmedicaltreatment.• So,itisquestionnable towhatpopulationwecangeneralizetheresultthataspirinhelpsreducetheriskofheartattacks.
ArticleBaselineDemographicsAfterRandomAssignment
Parameter Placebo(n=129)
Uceris(n=128)
Meanage,years(range) 39.9(12–68) 37.6(13–66)Men 77(59.7) 70(54.7)Women 52(40.3) 58(45.3)Meandiseaseduration(yrs) 6.3 5.5Duration≤1 year,n(%) 23(17.8) 28(21.9)Duration>5 years,n(%) 51(39.5) 44(34.4)Proctosigmoiditis 64(49.6) 58(45.3)Left-sidedcolitis 44(34.1) 37(28.9)MeanbaselineUCDAIscore 6.2 6.5MeanbaselineEIscore 6.6 6.5Priormesalazine use 75(58.1) 66(51.6)Priorsulfasalazineuse 28(21.7) 33(25.8)
Sandborn WJ,TravisS,MoroL,JonesR,Gautille T,Bagin R,HuangM,YeungP,BallardED2nd Once-dailybudesonideMMX®extended-releasetabletsinduceremissioninpatientswithmildtomoderateulcerativecolitis:resultsfromtheCOREIstudy. Gastroenterology 2012Nov;143(5):1218-26
BlockingandRandomAssignment• Thegoalinrandomassignmentistomakethetwogroupsassimilaraspossibleinallwaysotherthanthetreatment.• Sometimethereareknownconfoundersandyoucanblockon(controlfor)thesevariables.• Forexample,ifoursubjectsconsistof60%femalesand40%males,wecanforcethetreatmentandcontrolgroupstobe60%femaleand40%male,usingamatchedpairdesign.• Blockingmakessensewhenthereareknownconfoundersyouwanttocontrolfor.Butrandomassignmentmakesthemsimilarevenintermsofunknownconfounders.
3.Blinding.Eveninexperiments,thetreatmentandcontrolgroupscanbedifferentinwaysotherthantheexplanatoryvariable.Thisisespeciallytruewhentheresponsevariableissomewhatsubjective.Painisanexample.Onestudyfoundthat1/4ofpatientssufferingfrompost-operativepain,whengivenaplacebo(justapillofsugarandwater)claimedtheyexperienced"significantpromptpainrelief".
3.Blinding.Peoplemightnotbeabletojudgetheirownlevelsofpainverywell,andmaybeinfluencedbythebeliefthattheyhavetakenaneffectivetreatment.Thusinanexperimentwithsucharesponsevariable,researchersshouldensurethesubjectdoesnotknowwhetherheorshereceivedthetreatmentorthecontrol.Thisiscalledblinding.Inadouble-blind experiment,neitherthesubjectnortheresearcherrecordingtheresponsevariableknowstheleveloftheexplanatoryvariableforeachsubject(i.e.treatmentorcontrol).
4.Portacaval shuntexample.Thefollowingexampleshowstheimportanceofdoingarandomizedcontrolledexperiment.Theportacaval shuntisamedicalprocedureaimedatcurbingbleedingtodeathinpatientswithcirrhosisoftheliver.Thefollowingtablesummarizes51studiesontheportacaval shunt.Thepoorlydesignedstudieswereveryenthusiasticaboutthesurgery,whilethecarefullydesignedstudiesprovethatthesurgeryislargelyineffective.
DegreeofenthusiasmDesign High Moderate NoneNocontrols 24 7 1Controls,butnotrandomized 10 3 2Randomizedcontrolled 0 1 3
4.Portacaval shuntexample.Whydidthepoorlydesignedstudiescometothewrongconclusion?Alikelyexplanationisthatinthestudieswherepatientswerenotrandomlyassignedtothetreatmentorcontrolgroup,byandlargethehealthierpatientsweregiventhesurgery.Thisalonecouldexplainwhythetreatmentgroupoutlivedthecontrolgroupinthesestudies.
DegreeofenthusiasmDesign High Moderate NoneNocontrols 24 7 1Controls,butnotrandomized 10 3 2Randomizedcontrolled 0 1 3
5.Moreproblemswithstudies,andClofibrateexample.
Surveysareobservational.• Coverageisacommonissue.Coverageistheextenttowhichthepeopleyousampledfromrepresenttheoverallpopulation.Asurveyatafancyresearchhospitalinawealthyneighborhoodmayyieldpatiencewithhigherincomes,highereducation,etc.• Non-responsebiasisanothercommonproblem.Poorcoveragemeansthepeoplegettingthesurveydonotrepresentthegeneralpopulation.Non-responsebiasmeansthatoutofthepeopleyougavethesurveyto,thepeopleactuallyfillingitoutandsubmittingitaredifferentfromthepeoplewhodidnot.• Sameexactissuesinwebsurveys.
5.Moreproblemswithstudies,andClofibrateexample.
Non-responsebiasissimilartoadhererbias,inexperiments.Adrugcalledclofibrate wastestedon3,892middle-agedmenwithhearttrouble.Itwassupposedtopreventheartattacks.1,103assignedatrandomtotakeclofibrate,2,789toplacebo(lactose)group.Subjectswerefollowedfor5years.Isthisanexperimentoranobservationalstudy?
Clofibrate patientswhodiedduringfollowupadherers 15%non-adherers 25%total 20%
5.Moreproblemswithstudies,andClofibrateexample.
Non-responsebiasissimilartoadhererbias,inexperiments.Adrugcalledclofibrate wastestedon3,892middle-agedmenwithhearttrouble.Itwassupposedtopreventheartattacks.1,103assignedatrandomtotakeclofibrate,2,789toplacebo(lactose)group.Subjectswerefollowedfor5years.Isthisanexperimentoranobservationalstudy?
Itisanexperiment.DoesClofibrate work?Clofibrate patientswhodiedduringfollowup
adherers 15%non-adherers 25%total 20%
Clofibrate patientswhodiedduringfollowupadherers 15%non-adherers 25%total 20%
--------------------------------------------------------------------------Placebo
adherers 15%nonadherers 28%total 21%
Thosewhotookclofibrate didmuchbetterthanthosewhodidn'tkeeptakingclofibrate.Doesthismeanclofibrate works?
Clofibrate patientswhodiedduringfollowupadherers 15%non-adherers 25%total 20%
--------------------------------------------------------------------------Placebo
adherers 15%nonadherers 28%total 21%
Thosewhoadheredtoplaceboalsodidmuchbetterthanthosewhostoppedadhering.
Clofibrate patientswhodiedduringfollowupadherers 15%non-adherers 25%total 20%
--------------------------------------------------------------------------Placebo
adherers 15%nonadherers 28%total 21%
Allinalltherewaslittledifferencebetweenthetwogroups.
Clofibrate patientswhodiedduringfollowupadherers 15%non-adherers 25%total 20%
--------------------------------------------------------------------------Placebo
adherers 15%nonadherers 28%total 21%
Adherersdidbetterthannon-adherers,notbecauseofclofibrate,butbecausetheywerehealthieringeneral.Why?
Clofibrate patientswhodiedduringfollowupadherers 15%non-adherers 25%total 20%
--------------------------------------------------------------------------Placebo
adherers 15%nonadherers 28%total 21%
Adherersdidbetterthannon-adherers,notbecauseofclofibrate,butbecausetheywerehealthieringeneral.Why?• adherersarethetypetoengageinhealthierbehavior.• sickpatientsarelesslikelytoadhere.
6.Moreaboutconfoundingfactors.• Byaconfoundingfactor,wemeananalternativeexplanationthatcouldexplaintheapparentrelationshipbetweenthetwovariables,eveniftheyarenotcausallyrelated.Typicallythisisdonebyfindinganotherdifferencebetweenthetreatmentandcontrolgroup.Forinstance,differentstudieshaveexaminedsmokersandnon-smokersandhavefoundthatsmokershavehigherratesoflivercancer.Oneexplanationwouldbethatsmokingcauseslivercancer.Butisthereanyother,alternativeexplanation?• Onealternativewouldbethatthesmokerstendtodrinkmorealcohol,anditisthealcohol,notthesmoking,thatcauseslivercancer.
6.Moreaboutconfoundingfactors.• Anotherplausibleexplanationisthatthesmokersareprobablyolderonaveragethanthenon-smokers,andolderpeoplearemoreatriskforallsortsofcancerthanyoungerpeople.• Anothermightbethatsmokersengageinotherunhealthyactivitiesmorethannon-smokers.• Notethatifonesaidthat“smokingmakesyouwanttodrinkalcoholwhichcauseslivercancer,”thatwouldnotbeavalidconfoundingfactor,sinceinthatexplanation,smokingeffectiveiscausallyrelatedtolivercancerrisk.
6.Moreaboutconfoundingfactors.• Aconfoundingfactormustbeplausiblylinkedtoboththeexplanatoryandresponsevariables.Soforinstancesaying“perhapsahigherproportionofthesmokersaremen”wouldnotbeaveryconvincingconfoundingfactor,unlessyouhavesomereasontothinkgenderisstronglylinkedtolivercancer.• Anotherexample:left-handednessandageatdeath.PsychologistsDianeHalpernandStanleyCoren lookedat1,000deathrecordsofthosewhodiedinSouthernCaliforniainthelate1980sandearly1990sandcontactedrelativestoseeifthedeceasedwererighthanded orlefthanded.Theyfoundthattheaverageagesatdeathofthelefthanded was66,andfortherighthanded itwas75.Theirresultswerepublishedinprestigiousscientificjournals,NatureandtheNewEnglandJournalofMedicine.
6.Moreaboutconfoundingfactors.Allsortsofcausalconclusionsweremadeabouthowthisshowsthatthestressofbeinglefthanded inourrighthanded worldleadstoprematuredeath.
6.Moreaboutconfoundingfactors.• Isthisanobservationalstudyoranexperiment?
6.Moreaboutconfoundingfactors.• Isthisanobservationalstudyoranexperiment?Itisanobservationalstudy.• Arethereplausibleconfoundingfactorsyoucanthinkof?
6.Moreaboutconfoundingfactors.• Aconfoundingfactoristheageofthetwopopulationsingeneral.Leftiesinthe1980swereonaverageyoungerthanrighties.Manyoldleftieswereconvertedtorightiesatinfancy,intheearly20thcentury,butthispracticehassubsided.Thusinthe1980sand1990s,therewererelativelyfewoldleftiesbutmanyyoungleftiesintheoverallpopulation.Thisaloneexplainsthediscrepancy.