study design for chemoprevention · 2008-05-20 · 2.2. purpose of randomization:purpose of...

Study Design for Study Design for ChemopreventionChemoprevention

Cancer Epidemiology, Prevention and Cancer Epidemiology, Prevention and Control WorkshopControl Workshop

Shanghai, March 12, 2008Shanghai, March 12, 2008

I. INTRODUCTIONI. INTRODUCTION

Experimental studies are conducted to assess the Experimental studies are conducted to assess the effect of a treatment using a drug, or effect of a treatment using a drug, or intervention using a preventive agent, etc. intervention using a preventive agent, etc.

•• TreatmentTreatment•• PreventionPrevention•• Early detection/screeningEarly detection/screening•• DiagnosticDiagnostic•• Quality of life/supportive careQuality of life/supportive care

A. Defined:A. Defined:

•• A study design in which the investigator actively A study design in which the investigator actively controls who is exposed and who is not. controls who is exposed and who is not. Subjects are randomly assigned to various Subjects are randomly assigned to various treatment groups and followed to observe treatment groups and followed to observe outcomes.outcomes.

Comparison of Two Study DesignsComparison of Two Study Designs

Stratified RandomizationStratified Randomization

B. Experimental studies compared to B. Experimental studies compared to cohort studiescohort studies

1.1. Similarities between cohort and experimental studies. Similarities between cohort and experimental studies. Both:Both:

a.a. Subjects must be free of the outcome at the start of the Subjects must be free of the outcome at the start of the study.study.

b.b. People are grouped into “exposed”/ “not exposed” People are grouped into “exposed”/ “not exposed” categories.categories.

c.c. Groups are followed for a period of time to determine Groups are followed for a period of time to determine outcome.outcome.

d.d. Yield incidence data so allow the calculation of risk and Yield incidence data so allow the calculation of risk and related measures.related measures.

e.e. Susceptible to lostSusceptible to lost--toto--followfollow--up bias.up bias.

B. Experimental studies compared to B. Experimental studies compared to cohort studiescohort studies

2.2. Differences between cohort and experimental studies:Differences between cohort and experimental studies:a.a. Experimental studies involve active manipulation of Experimental studies involve active manipulation of

exposure (treatment/alternative treatment), whereas in exposure (treatment/alternative treatment), whereas in cohort studies, the investigator must merely observe the cohort studies, the investigator must merely observe the effect of exposure.effect of exposure.

b.b. Random allocation (or randomization) is an essential part Random allocation (or randomization) is an essential part of a good experimental study. Not possible in a cohort.of a good experimental study. Not possible in a cohort.

c.c. Ethical issues often a major issue in experimental Ethical issues often a major issue in experimental epidemiological studies.epidemiological studies.

d.d. Compliance with study protocol is an important concern in Compliance with study protocol is an important concern in experimental studies.experimental studies.

C. Random allocation/C. Random allocation/randomizationrandomization

1.1. Defined:Defined:A procedure for assigning patients to experimental A procedure for assigning patients to experimental treatment and other treatment groups so that treatment and other treatment groups so that chance alone is responsible for the group chance alone is responsible for the group assignment…each subject has an equal chance of assignment…each subject has an equal chance of being in any of the treatment groups.being in any of the treatment groups.

2.2. Purpose of randomization:Purpose of randomization:To (attempt to) assure comparability of the study To (attempt to) assure comparability of the study groups with respect to factors which may be groups with respect to factors which may be related to outcome.related to outcome.


3. 3. IMPORTANT: Randomization is IMPORTANT: Randomization is done after informed consent is done after informed consent is obtained!!!obtained!!!


4. 4. Do NotDo Not Confuse Random Allocation with Random Confuse Random Allocation with Random Selection!!!Selection!!!

Random Random selectionselection of subjects:of subjects:A procedure for A procedure for selectingselecting subjects so that each has the subjects so that each has the same chance of being included in the study. When we can’t same chance of being included in the study. When we can’t afford to use all possible subjects in the source population.afford to use all possible subjects in the source population.

Purpose: Purpose: To assure To assure representativenessrepresentativeness of subjects (of source pop.)of subjects (of source pop.)

When used?When used?In any type of study design where a sample of the In any type of study design where a sample of the population is being selected.population is being selected.


5. Randomization does not guarantee similarity of groups5. Randomization does not guarantee similarity of groups

D. Uses of experimental study D. Uses of experimental study design:design:

1. Evaluate benefits of an intervention:1. Evaluate benefits of an intervention:•• TherapeuticTherapeutic•• PreventivePreventive

2. Confirm etiologic relationship2. Confirm etiologic relationship

II. TWO MAJOR TYPES OF II. TWO MAJOR TYPES OF EXPERIEMENTAL EXPERIEMENTAL

STUDIES: Clinical and STUDIES: Clinical and community trialscommunity trials

A. Randomized clinical trial (RCT):A. Randomized clinical trial (RCT):

1.1. Defined (phase III):Defined (phase III):An experimental study where the effectiveness of the An experimental study where the effectiveness of the intervention is being tested on individuals. intervention is being tested on individuals.

Phase I trialsPhase I trials•• How does the agent affect the human body?How does the agent affect the human body?•• What dosage is safe? What dosage is safe?

Phase II trialsPhase II trials•• Does the agent or intervention have an effect on the disease? Does the agent or intervention have an effect on the disease?

Phase III trialsPhase III trials•• Is the new agent or intervention (or new use of a treatment) Is the new agent or intervention (or new use of a treatment)

better than the standard?better than the standard?•• Participants have an equal chance to be assigned to one of two Participants have an equal chance to be assigned to one of two

or more groupsor more groups

A. Randomized clinical trial (RCT):A. Randomized clinical trial (RCT):

2.2. Clinical trials are conducted for the Clinical trials are conducted for the treatment/prevention of both infectious treatment/prevention of both infectious diseases and chronic diseases:diseases and chronic diseases:

a.a. Infectious diseases: field trials (often refers to Infectious diseases: field trials (often refers to vaccine trials)vaccine trials)

b.b. Chronic diseases: e.g. Women’s Health Initiative Chronic diseases: e.g. Women’s Health Initiative (WHI):(WHI):

B.4. Recruit study Population:B.4. Recruit study Population:

Effectiveness of InterventionEffectiveness of Intervention

From: Writing Group for the Women’s Health Initiative InvestigatFrom: Writing Group for the Women’s Health Initiative Investigators. Risks and benefits ors. Risks and benefits of the estrogen plus progestin in healthy postmenopausal women. of the estrogen plus progestin in healthy postmenopausal women. Principal results Principal results from the Women’s Health Initiative Randomized Controlled from the Women’s Health Initiative Randomized Controlled Trial.JAMA.2002;288:321Trial.JAMA.2002;288:321--3333

B. Community trialB. Community trial

1.1. Defined:Defined:An experimental study where the effectiveness of An experimental study where the effectiveness of an intervention is tested on a community.an intervention is tested on a community.

2.2. Example of a community trial: fluoridation of Example of a community trial: fluoridation of water.water.

Fluoridation of WaterFluoridation of Water

B. Community trialB. Community trial

3.3. Problems of conducting community trials:Problems of conducting community trials:a.a. Obtaining an appropriate control group:Obtaining an appropriate control group:

1)1) Same community before and after interventionSame community before and after intervention2)2) A control community: similar to experimental A control community: similar to experimental

community with respect to possible confounders.community with respect to possible confounders.

b.b. Other problems: Other problems: 1) It is hard to get individual’s informed 1) It is hard to get individual’s informed concentconcent2) Intervention not at individuals level2) Intervention not at individuals level3) Collaboration of communities3) Collaboration of communities

III. STEPS IN III. STEPS IN CONDUCTING CONDUCTING

RANDOMIZED CLINICAL RANDOMIZED CLINICAL TRIALSTRIALS

Diagram of a design flow:Diagram of a design flow:

B.1. StepsB.1. Steps

•• 1. Specify Hypothesis1. Specify Hypothesis

B.2. Steps B.2. Steps

•• 2. Specify target and source of populations:2. Specify target and source of populations:

B.3 Define endpoints/possible side B.3 Define endpoints/possible side effects of intervention:effects of intervention:

B.5. Obtain informed consent from B.5. Obtain informed consent from those willing to participate:those willing to participate:

1.1. Informed consent: Informed consent: An agreement (signed) that the subject understands the An agreement (signed) that the subject understands the benefits and risks of the study.benefits and risks of the study.

2.2. Human Subjects Protection Committees: Human Subjects Protection Committees: act as watchdogs…review research applications the be sure act as watchdogs…review research applications the be sure they comply with issues pertaining to protection of human they comply with issues pertaining to protection of human subjects. To receive research money from NIH (and other subjects. To receive research money from NIH (and other agencies), these committees must approve grand proposals. agencies), these committees must approve grand proposals. Require that all subjects must read and sign “Informed Require that all subjects must read and sign “Informed Consent” form.Consent” form.

B.6 Randomize to treatmentB.6 Randomize to treatment

1.1. Experimental treatment.Experimental treatment.2.2. Alternative treatment (“controls”):Alternative treatment (“controls”):

a.a. The current standard treatment:The current standard treatment:b.b. Placebo:Placebo:

1)1) Defined: An inert substance prepared to look as similar Defined: An inert substance prepared to look as similar as possible to the experimental treatment.as possible to the experimental treatment.

2)2) Placebo effect: Am improvement on health, symptoms, Placebo effect: Am improvement on health, symptoms, due to fact of being treated…and not due to the due to fact of being treated…and not due to the treatment!treatment!

RandomizationRandomization

Three advantages of the Three advantages of the randomized design:randomized design:

•• Randomization removes the potential of bias in the Randomization removes the potential of bias in the allocation of subjects to the intervention group or to the allocation of subjects to the intervention group or to the control group;control group;

•• Randomization tends to produce comparable groups; that Randomization tends to produce comparable groups; that is, the measured or unknown prognostic factors and other is, the measured or unknown prognostic factors and other characteristics of the subjects at the time of randomization characteristics of the subjects at the time of randomization will be, on the average, evenly balanced between the will be, on the average, evenly balanced between the intervention and control groups;intervention and control groups;

•• The internal validity of statistical tests of significance is The internal validity of statistical tests of significance is guaranteed.guaranteed.

B.7. BlindnessB.7. Blindness

•• Fundamental PointFundamental Point: To avoid potential problems of bias : To avoid potential problems of bias during data collection and assessment a clinical trial Ideally during data collection and assessment a clinical trial Ideally we should have a doublewe should have a double--blind design. In studies where blind design. In studies where such a design is impossible, a singlesuch a design is impossible, a single--blind approach and blind approach and other measures to reduce potential bias are favored.other measures to reduce potential bias are favored.

•• UnblindedUnblinded: In an : In an unblindedunblinded or open trial, both the subject or open trial, both the subject and the investigator know to which intervention the subject and the investigator know to which intervention the subject has been assigned. The studies involving most surgical has been assigned. The studies involving most surgical procedures, changes in life style (eating habits, exercise, procedures, changes in life style (eating habits, exercise, smoking) or learning techniques can be conducted only in smoking) or learning techniques can be conducted only in this manner. The this manner. The advantageadvantage is that it is usually simpler to is that it is usually simpler to carry out and the carry out and the disadvantagedisadvantage is the possibility of bias.is the possibility of bias.


•• SingleSingle--BlindBlind: : Patient does not know which treatment group Patient does not know which treatment group he/she is in. he/she is in. Only the investigator is aware of which Only the investigator is aware of which intervention each subject is receiving. The advantage and intervention each subject is receiving. The advantage and disadvantage are similar to those of disadvantage are similar to those of unblindedunblinded trials.trials.

•• DoubleDouble--BlindBlind: Neither the subjects nor the investigators : Neither the subjects nor the investigators responsible for following the subjects know which responsible for following the subjects know which intervention the subject has been assigned. Such designs are intervention the subject has been assigned. Such designs are usually restricted to trials of drug efficacy. The usually restricted to trials of drug efficacy. The advantageadvantageis that the risk of bias is reduced. The is that the risk of bias is reduced. The disadvantagedisadvantage is that is that certain responsibilities, which in open or singlecertain responsibilities, which in open or single--blind blind studies could be accomplished by the investigators, must be studies could be accomplished by the investigators, must be taken over by others in order to maintain the blindness.taken over by others in order to maintain the blindness.


•• TripleTriple--BlindBlind: Neither the subjects, nor the : Neither the subjects, nor the investigators, nor the committee monitoring investigators, nor the committee monitoring response variables is told the identity of the response variables is told the identity of the groups. The theoretical groups. The theoretical advantageadvantage is to allow the is to allow the monitoring committee to evaluate the response monitoring committee to evaluate the response variable results more objectively. The variable results more objectively. The disadvantagedisadvantage is that in a trial where the is that in a trial where the monitoring committee has an ethical responsibility monitoring committee has an ethical responsibility to ensure subject safety, such a design may be to ensure subject safety, such a design may be counterproductive. counterproductive.

BlindnessBlindness

B.7. Blinding (cont.)B.7. Blinding (cont.)

b. Purpose of blinding:b. Purpose of blinding:To prevent biases in assessing outcome, which may be To prevent biases in assessing outcome, which may be influenced by knowledge of treatment group.influenced by knowledge of treatment group.

c. Blinding is most important whenc. Blinding is most important whend. Blinding is less important whend. Blinding is less important whene. Not always possible to blind subjects and/or e. Not always possible to blind subjects and/or

investigatorsinvestigators

B.8 FollowB.8 Follow--up study groups for up study groups for outcomesoutcomes

B.9. Analyze results: B.9. Analyze results: Use “Intention to treat” analysesUse “Intention to treat” analyses

1.1. Defined:Defined:Data are analyzed so that subjects remain in Data are analyzed so that subjects remain in groups as originally assigned…even if subjects do groups as originally assigned…even if subjects do not comply or change treatments on their own.not comply or change treatments on their own.This means that, even if the day after a subject was This means that, even if the day after a subject was assigned to the control group, he starts the assigned to the control group, he starts the experimental treatment on his own…the outcome experimental treatment on his own…the outcome for that patient is analyzed as if he were still in the for that patient is analyzed as if he were still in the control group! …..In short: control group! …..In short: “Once randomized, “Once randomized, always analyzed.”always analyzed.”

B.9. Analyze results: B.9. Analyze results: Use “Intention to treat” analysesUse “Intention to treat” analyses

2.2. Why we must use intention to treat analyses:Why we must use intention to treat analyses:a.a. The results reflect what happens in the real world when the The results reflect what happens in the real world when the

treatment is offered.treatment is offered.b.b. The trial concerns whether the offering of a new treatment The trial concerns whether the offering of a new treatment

is more effective.is more effective.c.c. NonNon--compliers are often different with respect to outcome, compliers are often different with respect to outcome,

so likely to be less bias in keeping them in original groups so likely to be less bias in keeping them in original groups than analyzing data in other ways.than analyzing data in other ways.

d.d. Though final results must reflect intentionThough final results must reflect intention--toto--treat analysis, treat analysis, investigators should look at compliers and noninvestigators should look at compliers and non--compliers compliers and their possible effect on the results.and their possible effect on the results.

B.10. Establish procedures for B.10. Establish procedures for terminating the trial and informing terminating the trial and informing

subjects of results.subjects of results.

To Protect the welfare of subjects, data must be To Protect the welfare of subjects, data must be monitored regularly and analyzed for any monitored regularly and analyzed for any obvious benefits or clear risks to subjects. If obvious benefits or clear risks to subjects. If either occurs, the trial must be stopped and either occurs, the trial must be stopped and appropriate action taken (if there is a benefit, appropriate action taken (if there is a benefit, then controls must be offered new treatment. If then controls must be offered new treatment. If a risk, a risk, experimentalsexperimentals must be taken of must be taken of experexper. . treatment.treatment.

B.10. Establish procedures for B.10. Establish procedures for terminating the trial and informing terminating the trial and informing

subjects of results.subjects of results.

Caveat: Small, transient differences between the Caveat: Small, transient differences between the groups may be observed early in the trial. So groups may be observed early in the trial. So there must be a balance between a longthere must be a balance between a long--enough enough followfollow--up to see whether true benefits occur vs. up to see whether true benefits occur vs. depriving the controls of a better treatment.depriving the controls of a better treatment.

IV. SOME MEASURES OF IV. SOME MEASURES OF EFFECT IN EFFECT IN

EXPERIMENTAL EXPERIMENTAL STUDIESSTUDIES

A. Risk ratio:A. Risk ratio:

IIC C / I/ ITT

B. Risk difference:B. Risk difference:

IICC -- IITT

C. Efficacy:C. Efficacy:

IcIc –– ItIt-------------------------- x 100x 100

IcIc

D. Example:D. Example:

849 working adults between the ages of 18849 working adults between the ages of 18--64 took part 64 took part in a doublein a double--blind, placeboblind, placebo--controlled, randomized trial controlled, randomized trial of the effect of influenza vaccination on three of the effect of influenza vaccination on three outcomes: upper respiratory illness (URI), absenteeism outcomes: upper respiratory illness (URI), absenteeism from work due to URI, and visits to doctors for URI. from work due to URI, and visits to doctors for URI. Baseline characteristics were compared to assure Baseline characteristics were compared to assure comparability of the two groups. The subjects were comparability of the two groups. The subjects were followed from Dec. 1, 1994, to March 31, 1995. followed from Dec. 1, 1994, to March 31, 1995. (Nichol, K.L. et al. N (Nichol, K.L. et al. N EnglEngl J Med 1995; 333: 889J Med 1995; 333: 889--93.)93.)

D. ExamplesD. Examples

•• a. The following tables shows the baseline a. The following tables shows the baseline characteristics of the two groups: does the characteristics of the two groups: does the randomization appear to have been successful, randomization appear to have been successful, that is, do the two groups appear to be similar that is, do the two groups appear to be similar for those variables?for those variables?

Baseline AssessmentBaseline Assessment

Fundamental PointFundamental Point: Relevant baseline data should be: Relevant baseline data should bemeasured in all study subjects before the start of the study.measured in all study subjects before the start of the study.

•• Baseline data include risk factors, prognostic factors, Baseline data include risk factors, prognostic factors, demographic and socioeconomic factors, and medical demographic and socioeconomic factors, and medical history. Baseline assessment can be used to assess history. Baseline assessment can be used to assess comparability of the study groups. Although the comparability of the study groups. Although the randomization on the average produces balance between randomization on the average produces balance between comparison groups, comparison groups, it does not guarantee balanceit does not guarantee balance in any in any specific trial. If baseline factors are not balanced, statisticaspecific trial. If baseline factors are not balanced, statistical l adjustment methods such as stratified analysis and adjustment methods such as stratified analysis and multivariate analysis can be used. multivariate analysis can be used.

Baseline AssessmentBaseline Assessment


•• b. One concern in any experimental study, even b. One concern in any experimental study, even if subjects are blinded, is whether subjects can if subjects are blinded, is whether subjects can tell whether they are in the treatment or control tell whether they are in the treatment or control group. One way to determine is to ask the group. One way to determine is to ask the subject which group they think they are in. In subject which group they think they are in. In the above study, 60% of the placebo and 54% of the above study, 60% of the placebo and 54% of the vaccine recipients correctly identified their the vaccine recipients correctly identified their group (slightly better than chance along).group (slightly better than chance along).


•• c. Another way to look at this is to see if the two c. Another way to look at this is to see if the two groups differ in reporting side effects. Table 2 groups differ in reporting side effects. Table 2 shows the side effects reported by each of the shows the side effects reported by each of the groups. Does it look as if any of the side effects groups. Does it look as if any of the side effects night give away a subject’s status?night give away a subject’s status?


d. Table 3 summarizes the data regarding the d. Table 3 summarizes the data regarding the three outcomes, using rates per 100 subjects. three outcomes, using rates per 100 subjects. For “Episodes of UPI”, calculate the three For “Episodes of UPI”, calculate the three measures of effect:measures of effect:

•• Risk ratio:Risk ratio:•• Risk differenceRisk difference•• Efficacy (or “Vaccine Effectiveness”)Efficacy (or “Vaccine Effectiveness”)

From: Writing Group for the Women’s Health Initiative InvestigatFrom: Writing Group for the Women’s Health Initiative Investigators. Risks and ors. Risks and benefits of estrogen plus progestin in healthy postmenopausal wobenefits of estrogen plus progestin in healthy postmenopausal women. Principal results men. Principal results from the Women’s Health Initiative Randomized Controlled from the Women’s Health Initiative Randomized Controlled Trial.JAMA.2002;288:321Trial.JAMA.2002;288:321--33.33.

Major potential biases in Major potential biases in experimental studies:experimental studies:

•• LostLost--toto--followfollow--upup

•• Compliance/adherence to Compliance/adherence to study’s protocol, e.g., study’s protocol, e.g., WHI:WHI:

Reasons for nonReasons for non--compliance:compliance:

1.1. Misunderstanding of instructions.Misunderstanding of instructions.2.2. Inconvenience of participation.Inconvenience of participation.3.3. Side effects of treatment.Side effects of treatment.4.4. Cost of participation.Cost of participation.5.5. Forgetfulness.Forgetfulness.6.6. Disappointment with results.Disappointment with results.7.7. Preference for another treatment.Preference for another treatment.

Ways to improve compliance:Ways to improve compliance:

1.1. Select motivated persons.Select motivated persons.2.2. Pretest ability and willingness of participants to comply.Pretest ability and willingness of participants to comply.3.3. Provide simple and lucid instructions to subjects.Provide simple and lucid instructions to subjects.4.4. Offer incentives to comply (e.g., no charge for therapeutic Offer incentives to comply (e.g., no charge for therapeutic

intervention and associated examinations).intervention and associated examinations).5.5. Provide positive reinforcements to subjects for adherence to Provide positive reinforcements to subjects for adherence to

treatment regimen.treatment regimen.6.6. Maintain frequent contract with participants and remind them Maintain frequent contract with participants and remind them

about importance of adherence to the regimen.about importance of adherence to the regimen.7.7. Measure adherence through pill counts or sampling of biologic Measure adherence through pill counts or sampling of biologic

specimens.specimens.8.8. Limit duration of intervention.Limit duration of intervention.

Sample Size:Sample Size:

Fundamental PointFundamental Point: Clinical trials should have sufficient : Clinical trials should have sufficient statistical power to detect differences between groups statistical power to detect differences between groups considered to be of clinical interest. Therefore, considered to be of clinical interest. Therefore, calculation of sample size with provision for adequate calculation of sample size with provision for adequate levels of significance and power is an essential part of levels of significance and power is an essential part of planning.planning.

Clinical trials tend to be small. Even if there is a real Clinical trials tend to be small. Even if there is a real difference in two treatments, the difference may not be difference in two treatments, the difference may not be statistically significantstatistically significant because of the small numbers. because of the small numbers. That is, the study did not have the That is, the study did not have the powerpower to detect to detect (statistically) a real difference.(statistically) a real difference.

Sample sizeSample size•• False positive (alphaFalse positive (alpha--level, or Type I error). The level, or Type I error). The

alphaalpha-- level used and accepted traditionally are 0.01 or level used and accepted traditionally are 0.01 or 0.05. The smaller the level of alpha, the larger the 0.05. The smaller the level of alpha, the larger the sample size.sample size.

•• False negative (betaFalse negative (beta--level, or Type II error). (1level, or Type II error). (1--beta) is beta) is called the power of the study. Investigator like to have a called the power of the study. Investigator like to have a power of around 0.90 or 0.95 when planning a study, power of around 0.90 or 0.95 when planning a study, which means that there have a 90% or 95% chance of which means that there have a 90% or 95% chance of finding a statistically significant difference between finding a statistically significant difference between study and control groups.study and control groups.

•• The difference between study and control groups The difference between study and control groups (delta). Two factors need to be considered here: one is (delta). Two factors need to be considered here: one is what difference is clinically important, and the another what difference is clinically important, and the another is what is the difference reported by previous studies.is what is the difference reported by previous studies.

V. The Validity of Results V. The Validity of Results of Epidemiologic Studiesof Epidemiologic Studies

A.A. Internal ValidityInternal Validity

1.1. Defined:Defined:

The truth or accuracy of results from a study with respect to The truth or accuracy of results from a study with respect to its defined source populationits defined source population

2. Factors affecting internal validity:2. Factors affecting internal validity:

•• Information bias: Information bias: distortion of true effects due to distortion of true effects due to systematic errors in collection of data on exposure OR systematic errors in collection of data on exposure OR outcomeoutcome

•• Selection bias: Selection bias: distortion of true effects due to error in distortion of true effects due to error in selection of subjectsselection of subjects

•• Confounding: Confounding: distortion of true effects from other risk distortion of true effects from other risk factors or exposures extraneous to the studyfactors or exposures extraneous to the study

3. Algorithm for assessing flaws in an 3. Algorithm for assessing flaws in an experimental study:experimental study:

B. External ValidityB. External Validity

1. Defined:1. Defined:

Applicability of study results to other populations (Applicability of study results to other populations (ieie, to , to specified target or reference population) specified target or reference population)

GeneralizabilityGeneralizability

2. Factors affecting external validity:2. Factors affecting external validity:

a.a. Does the study have internal validity? Does the study have internal validity? external validity is not possible without internal (external validity is not possible without internal (ieie, a , a study must be internally valid first before it can also be study must be internally valid first before it can also be externally valid)externally valid)

a.a. Is the study group representative of the target Is the study group representative of the target population?population?if the study group is nonif the study group is non--representative, then the study is representative, then the study is not externally validnot externally valid

3. Algorithm for assessing external 3. Algorithm for assessing external validity:validity:

Is experimental population representative of the reference population?

YESNo external validity problem

You CAN extrapolate

NOWould reference population

have same response to treatment as experimental population?

YESNo external validity problem

You CAN extrapolate

NOExternal validity problem

You CANNOT extrapolate

study design for chemoprevention · 2008-05-20 · 2.2. purpose of randomization:purpose of...

Documents