The Cross-Sectional Study

1 The Cross-Sectional Study :Investigating Prevalence and AssociationRonald A. ThistedDepartments of Health Studies and StatisticsThe University of ChicagoCRTP Track I Seminar, Autumn, 2006 Lecture the structure of the Cross-Sectional Study design, the advantages and disadvantages of this design, the kinds of questions that can be addressed usingcross- sectional Clinical Scenario: Are Kidney Stones and HypertensionConnected?The Cross-Sectional StudyThe LogicThe StructurePros and ConsConducting a Cross-Sectional StudySteps in building a Cross-Sectional studyPractical IssuesDiscussion of clinical exampleThe data setThe questionsConfounding and interactionData analysis issuesSuggested ReadingOutlineA Clinical Scenario: Are Kidney Stones and HypertensionConnected?The Cross-Sectional StudyThe LogicThe StructurePros and ConsConducting a Cross-Sectional StudySteps in building a Cross-Sectional studyPractical IssuesDiscussion of clinical exampleThe data setThe questionsConfounding and interactionData analysis issuesSuggested ReadingAre Kidney Stones and Hypertension Connected?

2 BackgroundIWomen age 34 59 with history of kidney stones more likely toalso have prior diagnosis of hypertension. (Madore, 1998)IMen age 40 75 with stone history: same direction ofassociation, but weaker. (Madore, 1998b)IMultiple studies report relationship between BP and stones;wide variability in size of Gap: What is really going on?Hypotheses:ISome subgroups may be more susceptible to increased BPwhen stones are present than of previous results may be due to differences inrepresentation of high-risk of BP and stones varies with (a) sex and (b) to test these hypotheses?INeed to be able to identify and Study subgroups of to avoid selection by stone status or blood pressure (soclinic-based samples are probably out)IWould like to say something about subgroups as they exist inthe adult by Gillen, Coe, and Worcester using NHANES III:TheThird National Health and Nutrition Examination :ISome subgroups may be more susceptible to increased BPwhen stones are present than of previous results may be due to differences inrepresentation of high-risk of BP and stones varies with (a) sex and (b) to test these hypotheses?

3 INeed to be able to identify and Study subgroups of to avoid selection by stone status or blood pressure (soclinic-based samples are probably out)IWould like to say something about subgroups as they exist inthe adult by Gillen, Coe, and Worcester using NHANES III:TheThird National Health and Nutrition Examination :ISome subgroups may be more susceptible to increased BPwhen stones are present than of previous results may be due to differences inrepresentation of high-risk of BP and stones varies with (a) sex and (b) to test these hypotheses?INeed to be able to identify and Study subgroups of to avoid selection by stone status or blood pressure (soclinic-based samples are probably out)IWould like to say something about subgroups as they exist inthe adult by Gillen, Coe, and Worcester using NHANES III:TheThird National Health and Nutrition Examination Clinical Scenario: Are Kidney Stones and HypertensionConnected?

4 The Cross-Sectional StudyThe LogicThe StructurePros and ConsConducting a Cross-Sectional StudySteps in building a Cross-Sectional studyPractical IssuesDiscussion of clinical exampleThe data setThe questionsConfounding and interactionData analysis issuesSuggested ReadingThe Logic of Cross-Sectional StudiesILooks at a slice of the population at a single point in the selected sample is appropriately selected, composition ofthe sample reflects that in the random sampleICluster sampleIStratified random samplesIMulti-stage sampleIPerform pre-defined measurements and include questionnaire/survey questionsIWhat can we do with this sample? We can fraction of the population has a particularcharacteristic?[History of kidney stones? Diagnosed hypertension?SBP>140?] is the correlation between an exposure and an outcome? [Relationship of kidney stone history to HTN history]The Structure of a Cross-Sectional Study [People with disease/outcome][People without disease/outcome]BeginCompareNowRisk Factor +Risk Factor -Risk Factor -Risk Factor + Study PopulationThe Structure of a Cross-Sectional Study , ContinuedBeginCompare[People with risk factor][People without risk factor]NowOutcome +Outcome -Outcome -Outcome + Study PopulationPros and ConsIAdvantagesICheaper/easier than longitudinal Study : no follow-up required!

5 IAfford good control over the measurement/ascertainmentprocessICan maximize completeness of key data (compared toretrospective Study )IHave greater control over precision of estimates in subgroups(stratified sampling)IOften can be accomplished assecondary data analysis,that is,data collected by someone else (possibly for another purpose)Pros and Cons, ContinuedIDisadvantagesIIn secondary data analysis, no control over purpose, choice, ormethod of data collectionICannot tell us about causal relationships (only correlation)IGeneralizability limited by sampled population, populationdefinitionISample size requirements may be very large (especially whenlooking at rare outcomes or exposures)IPotential for selection Length-biased sampling results from the fact thatindividuals with long courses of a disease are more likely to bethe ones identified as prevalent cases than people with coursesof short Clinical Scenario: Are Kidney Stones and HypertensionConnected?

6 The Cross-Sectional StudyThe LogicThe StructurePros and ConsConducting a Cross-Sectional StudySteps in building a Cross-Sectional studyPractical IssuesDiscussion of clinical exampleThe data setThe questionsConfounding and interactionData analysis issuesSuggested ReadingHow to conduct a Cross-Sectional studyIIdentify (and define) population of interest[Adults aged 17 90 with knowledge of stone hx].IDefine outcomes[Previous diagnosis of HTN; SBP].IDefine exposures (for correlational analysis)[Lifetime history of kidney stones]ICreate data collection instruments [Surveys, interviews, physical measurement procedures: SBP]Data collection forms are very useful forIStandardizationIProtocol adherenceIInsure consistent ascertainment[Training of staff]How to conduct a Cross-Sectional Study , ContinuedISample from population appropriately[Multistage sample of households].

7 IObtain consent, then measure IAnalyze using appropriate statistical methodsIMay require special techniques to account for sampling adjustment forconfoundersis usually necessary; weare (usually) after relationships that remain after adjusting forother factors[Age, race, sex are differentially associated with both stoneformation and blood pressure]OR..Ask to conduct a Cross-Sectional Study , ContinuedISample from population appropriately[Multistage sample of households].IObtain consent, then measure IAnalyze using appropriate statistical methodsIMay require special techniques to account for sampling adjustment forconfoundersis usually necessary; weare (usually) after relationships that remain after adjusting forother factors[Age, race, sex are differentially associated with both stoneformation and blood pressure]OR..Ask Issues for the Primary cross -SectionalistINon responseIRepresentativenessILogistics issues: cluster sampling, household contactIDefining eligibility (target population)IDefining measures in advance; respondent burdenIData collection formsOutlineA Clinical Scenario: Are Kidney Stones and HypertensionConnected?

8 The Cross-Sectional StudyThe LogicThe StructurePros and ConsConducting a Cross-Sectional StudySteps in building a Cross-Sectional studyPractical IssuesDiscussion of clinical exampleThe data setThe questionsConfounding and interactionData analysis issuesSuggested ReadingKidney Stones and BPGillen,et al,used NHANES III data (publicly available) to addresstheir research 1988 1994 by NCHSIN ational population-based sampleINoninsitutionalized persons aged>2 monthsIn= 33 994 IStone history available onn= 20 key questionsHave you ever had a kidney stone?919 answer Yes ( )This is a simpleprevalencecalculationBP you ever been told you have high blood pressure? , DBP, Pulse pressureRelationship of sex to stone formation:Kidney stone Hx : 54% womenKidney stone Hx +: 40% womenp< of stone history higher in men than womenRelation of stone formation to hypertension: SFs were more likely to report a previous diagnosis ofhypertension compared with non-SFs ( vs ;P= ).

9 [Univariate relationship]Answering key questionsHave you ever had a kidney stone?919 answer Yes ( )This is a simpleprevalencecalculationBP you ever been told you have high blood pressure? , DBP, Pulse pressureRelationship of sex to stone formation:Kidney stone Hx : 54% womenKidney stone Hx +: 40% womenp< of stone history higher in men than womenRelation of stone formation to hypertension: SFs were more likely to report a previous diagnosis ofhypertension compared with non-SFs ( vs ;P= ). [Univariate relationship]Answering key questionsHave you ever had a kidney stone?919 answer Yes ( )This is a simpleprevalencecalculationBP you ever been told you have high blood pressure? , DBP, Pulse pressureRelationship of sex to stone formation:Kidney stone Hx : 54% womenKidney stone Hx +: 40% womenp< of stone history higher in men than womenRelation of stone formation to hypertension: SFs were more likely to report a previous diagnosis ofhypertension compared with non-SFs ( vs ;P= ).

10 [Univariate relationship]Answering key questionsHave you ever had a kidney stone?919 answer Yes ( )This is a simpleprevalencecalculationBP you ever been told you have high blood pressure? , DBP, Pulse pressureRelationship of sex to stone formation:Kidney stone Hx : 54% womenKidney stone Hx +: 40% womenp< of stone history higher in men than womenRelation of stone formation to hypertension: SFs were more likely to report a previous diagnosis ofhypertension compared with non-SFs ( vs ;P= ). [Univariate relationship]Answering key questionsHave you ever had a kidney stone?919 answer Yes ( )This is a simpleprevalencecalculationBP you ever been told you have high blood pressure? , DBP, Pulse pressureRelationship of sex to stone formation:Kidney stone Hx : 54% womenKidney stone Hx +: 40% womenp< of stone history higher in men than womenRelation of stone formation to hypertension: SFs were more likely to report a previous diagnosis ofhypertension compared with non-SFs ( vs ;P= ).