studi kasus kontrol - prof.dr.dr. siti setiati, sppd-kger, m.epid, finasim

8/13/2019 Studi Kasus Kontrol - Prof.dr.Dr. Siti Setiati, SpPD-KGer, M.epid, FINASIM

http://slidepdf.com/reader/full/studi-kasus-kontrol-profdrdr-siti-setiati-sppd-kger-mepid-finasim 1/87

CASE-CONTROL STUDY

Siti Setiati



Major Types ofClinical Epidemiologic Research

Type of ResearchQuestion

Descriptive/Causal Aim

Diagnostic research DescriptivePredict the probability of presence of targetdisease from clinical and non-clinical profile

Prognostic research DescriptivePredict the course of disease from clinical and non-clinical profile

Etiologic research CausalCausally explain occurrence of target disease

from determinant

Intervention research Causal & Descriptive

(1) Causally explain the course of disease asinfluenced by treatment

(2) Predict the course of disease giventreatment (options) and clinical and non-clinical profile



Etiologic research

The research question:

• Is there a relation between a determinant(risk factor) and a disease-outcome?

Research question for causal relation !



HILL’S CRITERIA

• Is it clear that the exposure precede the onsetof the outcome ?( Temporality )

• Is there a dose-response gradient? ( BiologicalGradient )

• Is the association consistent from study tostudy ( Consistency )

• Does the association make biological sense?(Plausibility )



Bradford Hill criteria The Bradford Hill criteria , otherwise known as Hill's criteria for causation , are a groupof minimal conditions necessary to provide adequate evidence of a causalrelationship between an incidence and a consequence, established by the English epidemiologist Sir Austin Bradford Hill (1897 –1991) in 1965.

1. Temporality : The effect has to occur after the cause (and if there is an expected

delay between the cause and expected effect, then the effect must occur afterthat delay).2. Biological gradient : Greater exposure should generally lead to greater incidence of

the effect. In other cases, an inverse proportion is observed: greater exposure leadsto lower incidence

3. Consistency : Consistent findings observed by different persons in different placeswith different samples strengthens the likelihood of an effect.

4. Coherence : Coherence between epidemiological and laboratory findingsincreases the likelihood of an effect.

http://en.wikipedia.org/wiki/England

http://en.wikipedia.org/wiki/Epidemiologist

http://en.wikipedia.org/wiki/Austin_Bradford_Hill

http://en.wikipedia.org/wiki/Austin_Bradford_Hill

http://en.wikipedia.org/wiki/Epidemiologist

http://en.wikipedia.org/wiki/England



Bradford Hill criteria 5. Specificity : Causation is likely if a very specific population at a specific

site and disease with no other likely explanation. The more specific anassociation between a factor and an effect is, the bigger the probabilityof a causal relationship.

6. Strength : A small association does not mean that there is not a causaleffect, though the larger the association, the more likely that it is causal.

7. Plausibility : A plausible mechanism between cause and effect is helpful(but Hill noted that knowledge of the mechanism is limited by current

knowledge).



Etiologic researchWhat study design?

Design of two observational studies to

distinguish between cause and effect:1. Cohort study2. Case-control study



Case-control study

• Also called patient-control study

• Definition – Study in which patients with the disease-outcome and a

control group without the disease-outcome are selectedand in which it is determined how many people in bothgroups have been exposed to the determinant



Case-control study

time

start study

disease +(patients)

disease – (controls)

determinant +

determinant +

determinant -

determinant -



Creutzfeldt- Jakob’s Disease




• Fast, progressive form ofdementia

• In the 90s a new variant ofCreutzfeldt-Jakob wasdiscovered in Europe after anepidemic of mad-cow disease

• Caused by eating beef?

What research question?Why case control?




time

start study

patientswith CJD

controls fromhospital

beef +

beef +

beef -

beef -



Case-control study

determinant-outcome relation

CJD + CJD -

beef +

beef -

a

c

b

d

a/c = oddsbeef+ incases

= a x d / b x cb/d = oddsbeef+ incontrols

Odds Ratio



Case-control study

How do you find patients?• GP; hospital; cancer registration

How to select a control group?• GP; hospital; general population

Patients and controls have to comefrom the same ‘source’ population .



Case-control study

How do you assess exposure todeterminant?

• Interview with participant• Interview with proxy• Medical file



Case Control

Case

Control

+

Population

-+

-

Exposure Outcome Start here



Case-control studysummary

determinant disease-outcome



Measures of association:

Case Control approach• Research question: Does smoking increase

the risk of lung cancer ?

• Patient control study – select cases and controls – Estimate the frequency of smoking among cases

and controls

– prior: % smokers among cases > % smokersamong controls




Case Control approachDisease

Yes No

Yes a bDeterminant

No c d• RR?• Odds ratio = (a/c) / (b/d) = ad / bc



OR and RR

• In a randomized trial or Cohort study: – Relative Risk (RR ) = [a/(a+b) / c/(c+d)]

• In a case-control study:

–Odds Ratio (OR) = (a/b) / (c/d) = ad/bc

• Odds ratios (OR) and relative risk (RR) greaterthan 1 indicate that there is an increased risk ofthe adverse outcome associated with theexposure.

• If OR = RR = 1 the adverse outcome is no morelikely to occur with than without exposure to thesuspected agent.



Relative measures of effect• The relative risk

The relative risk can be calculated as ratio between two incidence proportions (riskratio, see Example 1) or two incidence rates (incidence rate ratio, see Example 2).

- Proportion of patients with CV events in the Ramipril group :651/ 4645=0.14 (14%).

- Proportion of patients with CV events in the placebo group :826/ 4652=0.18 (18%).

The Risk Ratio (RR) is: 0.14/0.18= 0.78

Examp le 1 (Risk Ratio)In the randomized prospective Heart Outcomes Prevention Evaluation (HOPE) study (1)the effect of Ramipril on the risk of cardiovascular (CV) events was investigated by

calculating the ratio between the incidence proportions of CV events in Ramipril treatedand in placebo treated patients.

With CV events Without CV eventsRamipril group (n=4645) 651 3994Placebo group (n=4652) 826 3826

The Heart Outcomes Prevention Evaluation Study Investigators. Effects of anangiotensin-converting-enzyme inhibitor, ramipril, on cardiovascular events inhigh-risk patients. N Engl J Med 2000; 342: 145-153



Relative measures of effect

• The odds ratio

The odds are a way of representing probability , familiar to gamblers (for example, theodds that a single throw of a die produces a six are 1 to 5). In a case-control study theodds of exposure in cases and controls are calculated as the number of exposed individualsdivided by the number of unexposed individuals in each group. If we know the odds ofexposure in cases and controls we can calculate the odds ratio (OR), i.e. the ratio betweenthe odds of exposure in diseased and in non-diseased individuals.

Knoll GA, Wells PS, Young D, et al. Thrombophilia and the risk forhemodialysis vascular access thrombosis. J Am Soc Nephrol 2005;16:1108-1114.

Examp le 2 ( the odds r at io)Knoll et al. (3) investigated the association between vascular access thrombosis andthrombophilia. They considered 107 patients with access thrombosis (cases) and 312 patientswithout fistula thrombosis (controls). Overall, among the 107 patients with access thrombosis, 59had evidence of thrombophilia and 48 did not while among the 312 without access thrombosis 122had thrombophilia and 190 did not.

- Odds of thrombophilia in patients with vascular access thrombosis :59/48=1.229- Odds of thrombophilia in patients without vascular access thrombosis :122/190=0.642

The odds ratio (OR) is: 1.229/0.642= 1.91




Case Control approach• Smoking and lung cancer

(controls = 10% random sampling from cohort)Disease

Yes NoYes 440 300 740

DeterminantNo 212 350 562

• Odds ratio (440/212) / (300/350) = 2.42• RR = (440/740) / (212/562) = 1.57 (shouldn ’ t be

calculated)



SOME IMPORTANT DISCOVERIES MADEIN CASE CONTROL STUDIES

1950's• Cigarette smoking and lung cancer

1970's• Diethyl stilbestrol and vaginal

adenocarcinoma• Post-menopausal estrogens and

endometrial cancer



1980's• Aspirin and Reyes syndrome• Tampon use and toxic shock syndrome• L-tryptophan and eosinophilia-myalgia

syndrome• AIDS and sexual practices

1990's

• Vaccine effectiveness• Diet and cancer



TWO CHARACTERISTICS OF CASES

1. REPRESENTATIVENESS:Ideally, cases are a random sample of allcases of interest in the source population(e.g. from vital data, registry data). Morecommonly they are a selection of available

cases from a medical care facility.(e.g. from hospitals, clinics)



2. METHOD OF SELECTION

Selection may be from incident or prevalentcases:

• Incident cases are those derived fromongoing ascertainment of cases over time.

• Prevalent cases are derived from a cross-sectional survey.



CHARACTERISTICS OF CONTROLS

• Who is the best control?• Where should controls come from?• If cases are a random sample of all cases in the

population, then controls should be a random sampleof all non-cases in the population sampled at thesame time (i.e. from the same study base)



THREE QUALITIES NEEDED IN

CONTROLS• Key concept: Comparability is more important

than representativeness in the selection of

controls • The control must be at risk of getting the

disease.

• The control should resemble the case in allrespects except for the presence of disease



STUDY BASE

Imagining the study base is a usefulexercise before deciding on control

selection.The study base is composed of apopulation at risk of exposure over a

period of risk of exposure .



Cases emerge within a study base.Controls should emerge from thesame study base, except that theyare not cases.

For example, if cases are selectedexclusively from hospitalized

patients, controls must also beselected from hospitalized patients.



It follows from the above that a poolof potential controls must be defined.

This pool must mirror the study base of the cases.



• If cases must have gone through a certain ascertainment process (e.g. screening), controls must have also. (e.g. mammogram-detected breast cancer)

• If cases must have reached a certain age before they can becomecases, so must controls. (thus we always match on age)

• If the exposure of interest is cumulative over time, the controlsand cases must each have the same opportunity to be exposedto that exposure. (if the case has to work in a factory to beexposed to benzene, the control must also have worked wherehe/she could be exposed to benzene)



Case-control study

How do you find cases/patients?

How to selecet a control group?



Case-control study

How do you find patients?• GP; hospital; cancer registration

How to select a control group?• GP; hospital; general population

Patients and controls have to comefrom the same‘

source’

population.



Case-control study

How do you assess exposure anddeterminant?



Case-control study

How do you assess exposure anddeterminant?

• Interview with participant• Interview with proxy• Medical file



Case-control studysummary

determinant disease-outcome



Case Control Start

here

Case

Control

+

Population

-+

-

Exposure Outcome



Prospective Cohort Start here

**

*

+

-

+-

t o t 1

Free ofoutcome

Exposure Outcome



Historical Cohort Start here

**

*

+

-

+-

t o t 1

Free ofoutcome

Exposure Outcome



Study Design

Direction of inquiry

CohortCase-control

Historical cohort

Survey / Cross Sectional

TODAY



Measures of association:Case Control approach

• Research question: Does smoking increasethe risk of lung cancer ?

• Patient control study – select cases and controls – Estimate the frequency of smoking among cases

and controls

– prior: % smokers among cases > % smokersamong controls




DiseaseYes No

Yes a bDeterminantNo c d

• RR?• Odds ratio = (a/c) / (b/d) = ad / bc




• Smoking and lung cancer(controls = 10% random sampling from cohort)

DiseaseYes No

Yes 440 300 740Determinant

No 212 350 562

• Odds ratio (440/212) / (300/350) = 2.42• RR = (440/740) / (212/562) = 1.57 (shouldn ’ t becalculated)



Point Estimates : Odds Ratios

• Example: is gender associated with use of standard adjuvanttherapy (SAT) for patients with newly diagnosed stage III colonor stage II/III rectal cancer? – 53% of men received SAT*

– 62% of women received SAT*• Ages, Sex and Racial Differences in the Use of Standard

Adjuvant Therapy for Colorectal Cancer”,



Odds and Odds Ratios• Odds = p/(1-p)

• The odds of a man receiveng SAT is 0.53/(1-0.53) = 1.13

• The odds of women receiving SAT is 0.62/(1-0.62) = 1,63

• Odds Ratio = 1.63/1.13 = 1.44

• Interpretation : “a woman is 1.44 times morelikely to receive SAT than a man”



Odds Ratio• Odds Ratio for comparing two proportions

OR = p1 / (1-p1) p2 / (1-p2)

= p1(1-p2)

p2(1-p1)

OR > 1 : increased risk of group 1 compared to 2OR = 1 : no difference in risk of group 1 compared to 2OR < 1 : lower risk (“protective”) in risk of group 1 compared to 2

• In our example,

– P1 = proportion of women receiving SAT

– P2 = proportion of men receiving SAT



Odds Ratio from a 2x2 table

SAT No SATWomen a= 298 b= 252 550

Men a= 202 d= 248 450500 500 1000

OR = p1 / (1 - p1) ad p2 / (1 - p2) bc



Why do we so often see OR and not other?

1) Logistic regression

– Allows us to look at association between two variables,adjusted for other variables

– “Output” is a log odds ratio – Ecamples: in the gender ~ SAT example, the odds ratios

were evaluated using logistic regression. In reality, thegender ~ SAT odds ratio is adjusted for age, race, year of

dx, region, marital status,..2) Can be more globally applied. Design of study does not

restrict usage



Case-control study

Advantages and disadvantages

• What are the advantages of a case-control

study?

• What are the disadvantages of a case-controlstudy?



ADVANTAGES AND DISADVANTAGES OF

CASE CONTROL STUDIES Advantages:

1. only realistic study design foruncovering etiology in rare diseases2. important in understanding new diseases3. commonly used in outbreak

investigation4. useful if induction period is long5. relatively inexpensive



Disadvantages:

1. Susceptible to bias if not carefully designed (and matched)2. Especially susceptible to exposuremisclassification3. Especially susceptible to recall bias 4. Restricted to single outcome 5. Incidence rates not usually calculable

6. Cannot assess effects of matching variables



Validity and bias • Validity:

– absence of systematic errors in design, conduct ordata-analysis of the research

• Bias: – degree of disruption of the determinant – outcomerelation caused by systematic errors – leads toreduced validity

• 3 types of bias in etiologic research: – selection bias, information bias, confounding



Any trend in the collection, analysis, interpretation,publication or review of data that can lead to

conclusions that are systematically different fromthe truth (Last, 2001)

A process at any state of inference tending toproduce results that depart systematically fromthe true values (Fletcher et al, 1988)

Systematic error in design or conduct of a study(Szklo et al, 2000)

What is Bias?



Selection bias

definition• Distortion of the determinant-outcome relation

caused by systematic errors in the selection of

study participants (cases and/or controls)• Determinant-outcome relation is different for

those that do and do not participate



Information bias

• Distortion of the determinant-outcome relationcaused by systematic errors in the measurement ofthe determinant and/or outcome .

• Who knows an example?



Information / Measurement / Misclassification Bias

Sources of information bias:

Subject variation

Observer variation

Deficiency of tools

Technical errors in measurement



Information bias

• Misclassification of determinant – Self reporting more accurate for cases than

controls (or the other way around)

• Misclassification of outcome – Disease better diagnosed in people with

determinant



Information / Measurement / Misclassification

BiasReporting bias:Individuals with severe disease tends to have completerecords, therefore more complete information aboutexposures and greater association found

Individuals who are aware of being participants of astudy behave differently (Hawthorne effect)



How to prevent bias?

• Confounding – cannot be prevented – Measure and adjust in data analysis

• Information bias - prevent during design

– Disease status blind for determinant status – Medical files instead of self-reporting – Same way of reporting for cases and controls

• Selection bias - prevent during design – Control selection independent of determinant

status – Good definition of source population



Controlling for Information Bias

- Blindingprevents investigators and interviewers from knowing case/control orexposed/non-exposed status of a given participant

- Form of surveymail may impose less “white coat tension ” than a phone or face-to-face

interview

- Questionnaireuse multiple questions that ask same informationacts as a built in double-check

- Accuracymultiple checks in medical recordsgathering diagnosis data from multiple source s



Confounding

Randomisation, matching and restriction can be tried at the time ofdesigning a study to reduce the risk of confounding.

At the time of analysis:

Stratification and multivariable (adjusted) analysis can achieve the same.

It is preferable to try something at the time of designing the study.



Case-control study

• Advantages – Efficient and relatively cheap – Appropriate for rare outcome – Can study several determinants

• Disadvantages – Cause is measured after effect

– Very sensitive to selection- and infobias – Not appropriate to study several outcomes



ISSUES IN MATCHING CONTROLS IN CASE-CONTROL STUDIES

1. Identify the pool from which controls may come.This pool is likely to reflect the way controls were

ascertained (hospital, screening test, telephonesurvey).

2. Control selection is usually through matching.Matching variables (e.g. age), and matching criteria (e.g. control must be within the same 5 year agegroup) must be set up in advance.



3. Controls can be individually matched orfrequency matched

INDIVIDUAL MATCHING : search for one (or more)controls who have the required MATCHING CRITERIA.PAIRED or TRIPLET MATCHING is when there is one or

two controls individually matched to each case.

FREQUENCY MATCHING : select a population of controlssuch that the overall characteristics of the group matchthe overall characteristics of the cases. e.g. if 15% ofcases are under age 20, 15% of the controls are also.



4. AVOID OVER-MATCHING . match only on factors

known to be causes of the disease.

5. Obtain POWER by matching MORE THAN ONECONTROL PER CASE. In general, N of controls should

be < 4, because there is no further gain of power abovefour controls per case.

6. Obtain GENERALIZABILITY by matching more than

ONE TYPE OF CONTROL



Validity

Absence of:• Systematic errors• Bias (distortion)



4. AVOID OVER-MATCHING . match only on factors

known to be causes of the disease.

5. Obtain POWER by matching MORE THAN ONECONTROL PER CASE. In general, N of controls should

be < 4, because there is no further gain of power abovefour controls per case.

6. Obtain GENERALIZABILITY by matching more than

ONE TYPE OF CONTROL



Precision (or accuracy)

The absence of• Random error

Depends on• Standardisation of measurements• Numbers

– Number of persons – Number of (repeated) observations /

measurements



Selection bias



Selection biasexample 1

• Medical circuit: 'oral anticonception could lead toDVT ’

• Women with DVT complaints who use oralanticonception will be more often referred than thosethat do not use oral anticonception

• Because of this selective r eferral all oralanticonception users will have a higher probability tocome into the study as a case and the effect of oralanticonception on DVT will be overestimated



Point Estimates : Odds Ratios

• Ages, Sex and Racial Differences in the Use of StandardAdjuvant Therapy for Colorectal Cancer”, Potosku, Harlan,Kaplan, Johnson, Lynch, JCO, vol. 20 (5), March 2002, p.1192

• Example: is gender associated with use of standard adjuvant

therapy (SAT) for patients with newly diagnosed stage III colonor stage II/III rectal cancer? – 53% of men received SAT*

– 62% of women received SAT*

• Ages, Sex and Racial Differences in the Use of StandardAdjuvant Therapy for Colorectal Cancer”, Potosku, Harlan,Kaplan, Johnson, Lynch, JCO, vol. 20 (5), March 2002, p.1192



More on the Odds Ratio

• Ranges from 0 to infinity

• Tends to be skewed (1.e not symmetric)

– “protective” odds ratios range from 0 to 1 – “increased risk” odds ratios range from 1 to

• Example:

• “Women are at 1.44 times the risk/ chance of men” • “Men are at 0.69 times the risk/chance of women”




• Sometimes, we see the log odds ratio insteadof the odds ratio

• Example:• “Women are at 1.44 times the risk/ chance of men”

• “Men are at 0.69 times the risk/chance of women”



Related Measures of Risk• Relative Risk : RR = p1/p2

– RR = 0.62/0.53 = 1.17

– Different way of describing a similar idea of risk

– Generarlly, intepretation “in words” is the similar” “Women are at 1.17 times as likely as men to

receive SAT” – RR is appropriate in trials often

– But RR is not appropriate in many settings (e.g. case-controls studies)

– Need to be clear about RR versus OR:

• p1

= 0.50, p2 = 0.25• RR = 0.5/0.25 = 2

• OR = (0.5/0.5)/ (0.25/0.75) = 3

• Same results, but OR and RR give quite different magnitude



Why do we so often see OR and not other?

1) Logistic regression

– Allows us to look at association between two variables,adjusted for other variables

– “Output” is a log odds ratio – Ecamples: in the gender ~ SAT example, the odds ratios

were evaluated using logistic regression. In reality, thegender ~ SAT odds ratio is adjusted for age, race, year of

dx, region, marital status,..2) Can be more globally applied. Design of study does not

restrict usage



Point Estimates : Hazard RatiosRandomized Controlled Trial of Single – Agent Paclitaxel Versus Cyclophosphamide,Doxorubicin and Cisplatin in Patients with Recurrent Ovarian Cancer Who Responded ti First-line Platinum-Based Regimens”, Cantu, Parma, Rossi, Floriani, Bonazzi, Dell’Anna . Torri,Colombo, JCO, vol. 20(5) March 2002, p.1232

• What is the effect of CAP on overall

survival as compared to paclitaxel?” • Median survival in CAP group was 34.7months

• Median survival in paclitaxel group was25.8 months

• But, median survival doesn’t tell thewhole story…



Hazard Ratio

• Compares risk of event intwo populations or samples

• Ratio of risk in group 1 torisk in group 2

• First things first• Kaplan – Meler Curves

(product-limit estimate)• Makes a “picture” of

survival



Hazard Ratio

• Assumption : “Proportional hazards” • The risk doest not depend on time

• That is, “risk is constant over time”

• But that is still vague….

• Example: Assume hazard ratio is 0.7 – Patients in temsirolimus group are at 0.7 times the

risk of death as those in the interferin-alpha arm,at any given point in time



Survival (S(t)) vs Hazard (h(t))

• Not the same thing• Hard to “envision” the difference

• Technically, it is the negative of the slope of the log of survivalcurve

• Yikes… don’t worry though

• The statisticians deal with this stuff

• Just remember: – The hazard ratio is not the ratio of the survival curves

– It is a ratio of some function of the survival curves



Hazard Ratios




• Ranges from 0 to infinity

• Tends to be skewed (1.e not symmetric)

– “protective” odds ratios range from 0 to 1 – “increased risk” odds ratios range from 1 to

• Example:

• “Women are at 1.44 times the risk/ chance of men” • “Men are at 0.69 times the risk/chance of women”



Hazard Ratio

• Compares risk of event intwo populations or samples

• Ratio of risk in group 1 torisk in group 2

• First things first• Kaplan – Meler Curves

(product-limit estimate)• Makes a “picture” of

survival

H d R i



Hazard Ratio

• Assumption : “Proportional hazards” • The risk doest not depend on time

• That is, “risk is constant over time”

• But that is still vague….

• Example: Assume hazard ratio is 0.7 – Patients in temsirolimus group are at 0.7 times the

risk of death as those in the interferin-alpha arm,at any given point in time

S i l (S( )) H d (h( ))



Survival (S(t)) vs Hazard (h(t))

• Not the same thing• Hard to “envision” the difference

• Technically, it is the negative of the slope of the log of survivalcurve

• Yikes… don’t worry though

• The statisticians deal with this stuff

•Just remember:

– The hazard ratio is not the ratio of the survival curves

– It is a ratio of some function of the survival curves

studi kasus kontrol - prof.dr.dr. siti setiati, sppd-kger, m.epid, finasim

Documents