Introduction to observational medical Introduction to observational medical studies and measures of associationstudies and measures of association

HRP 261 HRP 261 January 5, 2005January 5, 2005

Read Chapter 1, Agresti Read Chapter 1, Agresti

To Drink or Not to Drink?Volume 348:163-164 January 9, 2003 Ira J. Goldberg, M.D.

A number of epidemiologic studies have found an association of alcohol intake with a reduced risk of cardiovascular disease. These observations have been purported to explain the so-called French paradox: the lower rate of cardiovascular disease in….

…..With this in mind, is it time for a randomized clinical trial of alcohol?

According to scientists, too much coffee may cause... 1986 --phobias, --panic attacks 1990 --heart attacks, --stress, --osteoporosis 1991 -underweight babies, --hypertension 1992 --higher cholesterol 1993 --miscarriages 1994 --intensified stress 1995 --delayed conception But scientists say coffee also may help prevent... 1988 --asthma 1990 --colon and rectal cancer,... 2004—Type II Diabetes (*6 cups per day!)

Coffee Chronicles BY MELISSA AUGUST, ANN MARIE BONARDI, VAL CASTRONOVO, MATTHEW

JOE'S BLOWS Last week researchers reported that coffee might help prevent Parkinson's disease. So is the caffeine bean good for you or not? Over the years, studies haven't exactly been clear:

June 05, 2000

February 14, 1996

Personal Health: Sorting out contradictory findings about fat and health. By Jane E. Brody

MANY health-conscious Americans are beginning to feel as if they are being tossed around like yo-yos by conflicting research findings. One day beta carotene is hailed as a life-saving antioxidant and the next it is stripped of health-promoting glory and even tainted by a brush of potential harm. Margarine, long hailed as a heart-saving alternative to butter, is suddenly found to contain a type of fat that could damage the heart.

Now, after women have heard countless suggestions that a low-fat diet may reduce their breast cancer risk, Harvard researchers who analyzed data pooled from seven studies in four countries report that this advice may be based more on wishful thinking than fact.

The researchers, whose review was published last week in The New England Journal of Medicine, found no evidence among a number of studies of more than 335,000 women that a diet with less than 20 percent of calories from fat reduced a woman's risk of developing breast cancer. Nor was risk related to the types of fats the women ate, the study reported.

Is Fat Important? …….

Statistics HumorStatistics Humor The Japanese eat very little fat and suffer fewer heart attacks than the British

or the Americans.

On the other hand, the French eat a lot of fat and also suffer fewer heart attacks than the British or the Americans.

The Japanese drink very little red wine and suffer fewer heart attacks than the British or the Americans.

The Italians drink excessive amounts of red wine and also suffer fewer heart attacks than the British or the Americans.

Conclusion: Eat and drink whatever you like. It's speaking English that kills

you.

Assumptions and aims of Assumptions and aims of medical studiesmedical studies

1) Disease does not occur at random but is related to environmental and/or personal characteristics.

2) Causal and preventive factors for disease can be identified.

3) Knowledge of these factors can then be used to improve health of populations.

Medical StudiesMedical Studies

Evaluate whether a risk factor (or preventative factor) increases (or decreases) your risk for an outcome (usually disease, death or intermediary to disease).

The General Idea…

Exposure Disease?

Observational vs. Observational vs. Experimental StudiesExperimental Studies

Observational studies – the population is observed without any interference by the investigator

Experimental studies – the investigator tries to control the environment in which the hypothesis is tested (the randomized, double-blind clinical trial is the gold standard)

Confounding: A major problem Confounding: A major problem for observational studiesfor observational studies

Exposure Disease

Confounder

?

Alcohol Lung cancer

Smoking

Confounding: ExampleConfounding: Example

Why Observational Studies?Why Observational Studies?

CheaperFasterCan examine long-term effectsHypothesis-generatingSometimes, experimental studies are not

ethical (e.g., randomizing subjects to smoke)

What is What is risk risk for a biostatistician?for a biostatistician?Risk = Probability of developing a disease or other adverse

outcome (over a defined time period)In Symbols: P(D)

Conditional Risk = Risk of developing a disease given a particular exposure

In Symbols: P(D/E)

Odds = Probability of developing a disease divided by the probability of not developing it

In Symbols: P(D)/P(~D)

)(1

)( disease of odds

DP

DP

Possible Observational Possible Observational Study DesignsStudy DesignsCross-sectional studies

Cohort studies

Case-control studies

Cross-Sectional (Prevalence) Cross-Sectional (Prevalence) StudiesStudies

Measure disease and exposure on a random sample of the population of interest. Are they associated?

Marginal probabilities of exposure AND disease are valid, but only measures association at a single time point.

Introduction to the 2x2 TableIntroduction to the 2x2 Table

  Exposure (E) No Exposure (~E)

 

Disease (D) a b a+b = P(D)

No Disease (~D) c d c+d = P(~D)

  a+c = P(E) b+d = P(~E)

Marginal probability of disease

Marginal probability of exposure

Agresti Example: Belief in Agresti Example: Belief in AfterlifeAfterlife

582

509

  Yes No or undecided

 

Females 435 147

Males 375 134

  810 281

737.509

375;747.

582

435// malebelievefemalebelieve pp

1091

37.027.

01.

509)737.1)(737(.

582)747.1)(747(.

737.747.

).(.

differencees

differenceZ

Cross-Sectional StudiesCross-Sectional Studies

Advantages: – Cheap and easy– generalizable– good for characteristics that (generally) don’t

change like genes or gender

Disadvantages – difficult to determine cause and effect

2. Cohort studies2. Cohort studies::

1. Sample on exposure status and track disease development (for rare exposures)

Marginal probabilities (and rates) of developing disease for exposure groups are valid.

Example: The Framingham Example: The Framingham Heart StudyHeart Study

The Framingham Heart Study was established in 1948, when 5209 residents of Framingham, Mass, aged 28 to 62 years, were enrolled in a prospective epidemiologic cohort study.

Health and lifestyle factors were measured (blood pressure, weight, exercise, etc.).

Interim cardiovascular events were ascertained from medical histories, physical examinations, ECGs, and review of interim medical record.

Cohort StudiesCohort Studies

Target population

Exposed

Not Exposed

Disease-free cohort

Disease

Disease-free

Disease

Disease-free

TIME

  Exposure (E) No Exposure (~E)

 

Disease (D) a b

No Disease (~D) c d

  a+c b+d

)/()/(

)~/(

)/(

dbbcaa

EDP

EDPRR

risk to the exposed

risk to the unexposed

The Risk Ratio, or Relative Risk (RR)

400 400

1100 2600

0.23000/4001500/400 RR

Hypothetical DataHypothetical Data

  Normal BP

Congestive Heart Failure

No CHF

1500 3000

High Systolic BP

Advantages/Limitations:Advantages/Limitations:Cohort StudiesCohort Studies

Advantages:– Allows you to measure true rates and risks of disease for the

exposed and the unexposed groups.– Temporality is correct (easier to infer cause and effect).– Can be used to study multiple outcomes. – Prevents bias in the ascertainment of exposure that may occur

after a person develops a disease. Disadvantages:

– Can be lengthy and costly! More than 50 years for Framingham.– Loss to follow-up is a problem (especially if non-random).– Selection Bias: Participation may be associated with exposure

status for some exposures

Case-Control StudiesCase-Control Studies

Sample on disease status and ask retrospectively about exposures (for rare diseases) Marginal probabilities of exposure for cases and

controls are valid.

• Doesn’t require knowledge of the absolute risks of disease

• For rare diseases, can approximate relative risk

Target population

Exposed in past

Not exposed

Exposed

Not Exposed

Case-Control StudiesCase-Control Studies

Disease

(Cases)

No Disease

(Controls)

Example: the AIDS epidemic Example: the AIDS epidemic in the early 1980’sin the early 1980’s

Early, case-control studies among AIDS cases and matched controls indicated that AIDS was transmitted by sexual contact or blood products.

In 1982, an early case-control study matched AIDS cases to controls and found a positive association between amyl nitrites (“poppers”) and AIDS; odds ratio of 8.6 (Marmor et al. 1982). This is an example of confounding.

Case-Control Studies in Case-Control Studies in HistoryHistory

In 1843, Guy compared occupations of men with pulmonary consumption to those of men with other diseases (Lilienfeld and Lilienfeld 1979).

Case-control studies identified associations between lip cancer and pipe smoking (Broders 1920), breast cancer and reproductive history (Lane-Claypon 1926) and between oral cancer and pipe smoking (Lombard and Doering 1928). All rare diseases.

Case-control studies identified an association between smoking and lung cancer in the 1950’s.

The proportion of cases and controls are set by the investigator; therefore, they do not represent the risk (probability) of developing disease.

bc

ad

dcba

dcddccbabbaa

ORDEP

DEP

DEPDEP

)/()/()/()/(

)~/(~)~/(

)/(~)/(

  Exposure (E) No Exposure (~E)

 

Disease (D) a b

No Disease (~D) c d

 

The Odds Ratio (OR)

a+b=cases

c+d=controls

Odds of exposure in the cases

Odds of exposure in the controls

bc

ad

dcba

dcddccbabbaa

ORDEP

DEP

DEPDEP

)/()/()/()/(

)~/(~)~/(

)/(~)/(

  Exposure (E) No Exposure (~E)

 

Disease (D) a b

No Disease (~D) c d

 

The Odds Ratio (OR)

a+b=cases

c+d=controls

The Odds RatioThe Odds Ratio

RR

OR

EDPEDP

EDPEDP

EDPEDP

EDPEDP

EDPEDP

DEPDEP

DEPDEP

)~/()/(

)~/(~)~/(

)/(~)/(

)~&(~)&(~

)~&()&(

)~/(~)~/(

)/(~)/(

When disease is rare: P(~D) 1

“The Rare Disease Assumption”

Via Bayes’ Rule

1

1

bc

adOR

db

ca

  Exposure (E) No Exposure (~E)

 

Disease (D) a = P (D& E) b = P(D& ~E)

No Disease (~D) c = P (~D&E) d = P (~D&~E)

 

The Odds Ratio (OR)

Odds of disease in the exposed

Odds of disease in the unexposed

0 0.35 0.7 1.05 1.4 1.75 2.1 2.45 2.8 3.15 3.5 0

1

2

3

4

5

6

P e r c e n t

Simulated Odds Ratio

Properties of the OR (simulation)

Properties of the lnOR

-1.05 -0.75 -0.45 -0.15 0.15 0.45 0.75 1.05 1.35 1.65 1.95 0

2

4

6

8

10

P e r c e n t

lnOR

Standard deviation =

dcba

1111

Standard deviation =

Hypothetical DataHypothetical Data

0.8)10)(6(

)24)(20(OR

25.8) - (2.47(8.0)e ,(8.0)e CI %95 24

1

10

1

6

1

20

196.1

24

1

10

1

6

1

20

196.1

  Amyl Nitrite Use No Amyl Nitrite

 

AIDS 20 10

No AIDS 6 24

 

30

30

Note that the size of the smallest 2x2 cell determines the magnitude of the variance

Odds Ratios in the literatureOdds Ratios in the literature

OR= 1.47 (.99-2.14)

•Things to think about:

•What does an Odds Ratio of 1.47 mean?

•“An increased risk of 47%”—is this misleading?

Highest Quintile of Mercury (in toenails) and Risk of Heart Attacks (NEJM Nov 02)

When can the OR mislead?When can the OR mislead?

When is the OR is a good When is the OR is a good approximation of the RR?approximation of the RR?

General Rule of Thumb:

“OR is a good approximation as long

as the probability of the outcome in the

unexposed is less than 10%”

Volume 340:618-626February 25, 1999

From: “The Effect of Race and Sex on Physicians' Recommendations for Cardiac Catheterization”

Volume 340:618-626February 25, 1999

From: “The Effect of Race and Sex on Physicians' Recommendations for Cardiac Catheterization”

Study overview: Researchers developed a computerized survey

instrument to assess physicians' recommendations for managing chest pain.

Actors portrayed patients with particular characteristics (race and sex) in scripted interviews about their symptoms.

720 Physicians at two national meetings viewed a recorded interview and was given other data about a hypothetical patient. He or she then made recommendations about that patient's care.

Media headlines on Feb 25Media headlines on Feb 25thth, , 1999…1999…

Wall Street Journal: “Study suggests race, sex influence physicians' care.”

New York Times: Doctor bias may affect heart care, study finds.”

Los Angeles Times: “Heart study points to race, sex bias.” Washington Post: “Georgetown University study finds

disparity in heart care; doctors less likely to refer blacks, women for cardiac test.”

USA Today: “Heart care reflects race and sex, not symptoms.” ABC News: “Health care and race”

Their results…Their results…

The Media Reports: “Doctors were only 60 percent as likely

to order cardiac catheterization for women and blacks as for men and

whites.”

A closer look at the data…A closer look at the data…

The authors failed to report the risk ratios:

RR for women: .847/.906=.93

RR for black race: .847/.906=.93

Correct conclusion: Only a 7% decrease in chance of being offered correct treatment.

Lessons learned:Lessons learned:

90% outcome is not rare! OR is a poor approximation of the RR here,

magnifying the observed effect almost 6-fold. Beware! Even the New England Journal doesn’t

always get it right!

SAS automatically calculates both, so check how different the two values are even if the RR is not appropriate. If they are very different, you have to be very cautious in how you interpret the OR.

SAS code and outputSAS code and outputfor generating OR/RR from for generating OR/RR from

2x2 table2x2 table  Cath No Cath  

Female 305 55

Male 326 34

 

360

360

data cath_data;

input IsFemale GotCath Freq;

datalines;

1 1 305

1 0 55

0 1 326

0 0 34

run;

data reversed; *Fix quirky reversal of SAS 2x2 tables;

set cath_data;

IsFemale=1-IsFemale;

GotCath=1-GotCath;

run;

proc freq data=reversed;

tables IsFemale*GotCath /measures;

weight freq; run;

SAS outputSAS output

Statistics for Table of IsFemale by GotCath

Estimates of the Relative Risk (Row1/Row2)

Type of Study Value 95% Confidence Limits ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ Case-Control (Odds Ratio) 0.5784 0.3669 0.9118 Cohort (Col1 Risk) 0.9356 0.8854 0.9886 Cohort (Col2 Risk) 1.6176 1.0823 2.4177

Sample Size = 720

Furthermore…stratification Furthermore…stratification shows…shows…

Advantages and Limitations: Advantages and Limitations: Case-Control StudiesCase-Control Studies

Advantages:– Cheap and fast– Great for rare diseases

Disadvantages:– Exposure estimates are subject to recall bias (those

with the disease are searching for reasons why they got sick and may be more likely to report an exposure) and interviewer bias (interviewer may prompt a positive response in cases).

– Temporality is a problem (did exposure cause disease or disease cause exposure?)

Final Note:Final Note: controlling for controlling for confounders in observational confounders in observational

studiesstudies1. Confounders can be controlled for in the

design phase of a study (restriction or matching).

2. Confounders can be controlled for in the analysis phase of a study (stratification or multivariate regression).