Slide 1Cause and EffectSlide 3Descriptive studiesClarence DeMarCorrelational studySlide 7Cross sectional studyExample cross-sectional studyEXAMPLE – CROSS SECTIONAL STUDYCross sectional studyCase-control studiesExample case-control studySlide 14Slide 15Slide 16Cohort studiesSlide 18Slide 19Randomized controlled trial studiesSlide 21Randomized controlled trial studiesCross-sectional studiesCase control studiesProspective cohort studiesRandomized controlled trialsTypes of EvidenceSummaryStudy DesignEvaluating EvidenceEpidemiology Concepts Study designCause and EffectGoal: to determine whether an exposure “causes” an outcome–Example:BUT… an association does not equal causality–Increases in ice-cream sales are associated with increased murder rates in urban environments–Does eating ice cream cause murder??•Confounding variable – an extraneous variable in a design that correlates with both the dependent and independent variablesTypes of epidemiological studiesEpidemiology is the study of the patterns, causes, and effects of health and disease conditions in defined populations.Descriptive studies•Case-study/case report–usually on one person (one “case”) or small group of people•Not widely used in physical activity research•Example: Investigating the diameter of arteries of the best runners in the world and noting the artery diameters are larger than similarly aged people who are not runnersClarence DeMar•Mr. Marathoner – Boston 7x!•Died in 1958•Autopisied by Dr. Paul Dudley White•Reported ‘….coronary artery diameters were 2 -3 times normal size’ (New England Journal of Medicine, 1961)1930 Boston MarathonClarence DeMar running to work passing through Everett, MACorrelational study•Evaluating population level associations between an exposure and outcome•Unit of exposure is a ________•CORRELATION DOES NOT EQUAL CAUSALITY•Shows level of association between exposure (e.g. physical activity) and outcome = disease (e.g. diabetes) or disease mortality (e.g. deaths due to cardiovascular disease)Correlational studyCross sectional study•Measure exposure and outcome at the same time•Provides a “snapshot” of the relationship between exposure and outcome•Commonly used in population surveillance analyses (i.e. NHANES)Example cross-sectional study The class is given a questionnaire that asks about their physical activity levels and on the same day measure LDL levels in their bloodEXAMPLE – CROSS SECTIONAL STUDYLow Medium High020406080100120140160PA LevelLDL LevelsLevine et al, Science, 2005Cross sectional studyCase-control studies•First identified based on outcome•Examine exposure variables in cases and controls to see if there are differences •Retrospective – looks back (vs. prospective which evaluates future events)•Example: Cases:Controls:Example case-control study Ask them about their physical activity from their childhood up until nowAsk them about their physical activity from their childhood up until nowEvaluate differences between cases and controls for the outcomeEvaluate differences between cases and controls for the outcomeFriedenreich et al., Am J Epi, 2001OR=The odds ratio is the ratio of the odds of an event occurring in one group to the odds of it occurring in another group.Case-control studyNo Caption Available.Lifetime PAquestionnairerearrange: a*d/c*bODDSRATIOCohort studies•At the beginning of the study•Prospective design•Follow subjects over time to see who develops outcome of interest–Multiple outcomesvs. ______________Lancet, 2011Lancet, 2011Randomized controlled trial studies•Identify factor believed to impact health (physical activity) •Take group of people and •Do treatment for period of time•Compare health outcome•Also called experimental trialsSlentz et al, JAP, 2005STRRIDE= Studies of Targeted Risk Reduction Interventions through Defined ExerciseRandomized controlled trial studiesDiabetes Prevention Program (DPP) Knowler, et al 2002 NEJMCross-sectional studiesAdvantages•Inexpensive•Prevalence•Study multiple outcomes•TimeDisadvantages•Sampling methods?•Cannot conclude causality•“Chicken or the egg” Cross sectional: Measure Exposure and Outcome at the same timeCase control studiesAdvantages•Good for rare conditions•Efficient•Low cost Disadvantages•Bias–Selection–Recall–How do you select controls?•Cannot determine causality Case Control: Measure Outcome in people who have (cases) and do not have (control) the exposure - RetrospectiveProspective cohort studiesAdvantages•Assess multiple outcomes•Cause/effect – Maybe yes•Longitudinal natural study–Not changing anything Disadvantages•Expensive•Takes time! •Cause/effect –maybe no•Social desirability: change in behavior because you know you are being measured Cohort: Prospective study, tracks development of outcome over timeRandomized controlled trialsStrengths•Intervene and change the variable (exposure) hypothesized to be affecting outcome•Best for cause /effectWeaknesses•Expensive and labor intensive if done right•Ethical? –When would it not be ethicalRCT: Experimental study, random assignment to treatment and control groups and tracks development of outcome over timeTypes of EvidenceType 1: randomized controlled trails without major limiations Type 2 Type 3 Type 4 http://www.health.gov/paguidelines/Report/E_integration.aspx2a – RCTs (or meta-analyses) with important limitations2b – non-randomized clinical trials 3a – Well-designed prospective cohort studies and case-control studies3b – other observational studies, weak prospective cohort studies or case-control studies, cross-sectional studies or case series Inadequate, very limited, or no data in population of interest. Anecdotal evidence or no/little clinical experienceSummary •Cause and effect–Best study to determine this?•Types of study designs•Advantages/disadvantages of each type of study