Drug Lit Exam 1- Cohort Study

Front 1

epidemiology definition- study of...(2)

useful for?

Back 1

Study of frequency of health‐related outcomes

Study of the distribution and determinants of disease in human
population.

Useful for prognostic information

Front 2

epidemiology generates what 2 things

Back 2

generates a numerator AND a denominator

Front 3

"numerator" and "denominator" : describe what they are in epidemiology

Back 3

• In clinical medicine the numerators or individual patient
outcomes are of primary interest
• In epidemiology, the denominator (all individuals at risk or “exposed”) is defined first, then numerators are counted.

Front 4

Epidemiology uses what type of study design?

Back 4

Observational Study design and is NON EXPERIMENTAL

Front 5

definition of an observational study

Back 5

Records specific events happening in the world among a defined population without any intervention by the
researcher

Front 6

3 types of observational designs

Back 6

– Cohort Study
– Case Control Study
– Cross‐Sectional Study

Front 7

data collection in an observational study can be what 3 methods

Back 7

– prospective
– retrospective (historical)
– combination of both prospective and retrospective

Front 8

RCT: cause and effect- in an RCT how are things handled so that you can determine cause/effect? (3)

Back 8

• In an RCT, the investigator assigns the “exposure” or treatment to patients in a randomized manner.
• thus You KNOW the exposure occurred before the outcome AND
• Randomization has created “equivalent” groups in
terms of the expected value on all variables.
• Then you compare the extent of outcomes,

Front 9

do cohort studies show cause and effect? why or why not? (2)

Back 9

You have NOT randomized patients to receive the treatment or “exposure”. You’re following a group of people and some have it and others do not

We are concerned whether the exposure truly caused the outcome,

Front 10

things to ask yourself when determining whether there is cause and effect (3)

Back 10

• Is there an association?
• Is this a temporal relationship,?
• Can we exclude competing causes?

Front 11

advantages of cohort studies (4)

Back 11

• Ascertain incidence and natural history of a disorder
• Can investigate multiple outcomes
arising from a single exposure
• Lack of ethical concerns since
investigator does not “assign” exposure
• Allow calc of incidence rates, RR and CIs

Front 12

4 disadvantages of cohort study

Back 12

• Selection bias built in
• Not good for extremely rare diseases of those which take a long time to develop. However we have large long‐term studies (Framingham, Nurses Health Study) which have provided data
• Lost to follow‐up
• Change in exposure status

Front 13

prospective cohort design- indicate which steps are in the present and which are future (3)

Back 13

take a study population and exclude those who already have the disease or outcome you are looking at

measure/classify into risk/factor+ and risk or factor -

(these are in present)

measure outcomes in each "arm" and compare (future)

Front 14

retrospective cohort study- what makes it retrospective

Back 14

dates on the data obtained about exposure/risk factor & outcome occurred in the PAST, but data is collected now

Front 15

retrospective cohort design (3) indicate which steps are present/future

Back 15

take population of clinic patient data- exclude those who have the disease/outcome you are measuring already

classify into your 2 groups (exposure to risk factor vs. no exposure)

measure outcomes in each "arm" and compare

Front 16

biostatistics seen in cohort studies (like...RR, etc) (7)

Back 16

• Incidence
• Prevalence
• Relative risk
• Relative risk reduction
• Absolute risk reduction
• Number needed to treat
• Number needed to harm

Front 17

prevalence definition

Back 17

The probability of having a disease (or
characteristic) at one point in time

Front 18

incidence definition
be sure to notice what

Back 18

the number of new cases (or events) that
occur per unit of time (often calculated per
year but be sure you notice the specific
interval)

Front 19

relative risk definition

Back 19

Incidence rate in exposed/Incidence rate in
unexposed

Front 20

absolute risk reduction definition

Back 20

Absolute difference in incidence in exposed and incidence in unexposed

Front 21

Relative Risk Reduction: compare to ARR
equation

Back 21

– ARR /Control group rate
– Will likely appear more impressive than ARR

Front 22

NNT or NNH -what is it, how to calculate

Back 22

How many people treated over time frame to
see for one event to see one person have the outcome

1/ARR

Front 23

in the prospective cohort study regarding coffee and type 2 diabetes incidence...how was the "exposure" measured?

Back 23

Coffee consumption assessed every 2‐4 years through validated questionnaires

Front 24

reported relative risk in coffee study...
how was RR adjusted?
how was data stratified??
how was it reported?

Back 24

– Multivariate adjusted RR
– Age and BMI Adjusted RR
– Reported per range of “cups per day” (like had RR depending on how many cups per day)
– Also broken down by gender as well as beverage of choice (tea, coffee, decaf)

Front 25

5 types of bias found in cohort studies

Back 25

Selection Bias
non-respondent sampling bias
surveillane bias
recall bias
changes in habits or exposures

Front 26

effect of selection bias on internal and external validity

Back 26

Internal Validity…Affect your IR (incidence rate) and RRs
External Validity…Affect your ability to generalize results

Front 27

Non‐respondent sampling bias- what does this mean? what is the opposite bias called?

Back 27

Loss to Follow‐up
• Persons who fail to respond to a request for information or to a survey may be systematically different from people willing to respond, the opposite bias is called volunteer bias

Front 28

surveillance bias- what is it, when does it occur

Back 28

Need to follow‐up the same amount and with similar intensity with all subjects

surveillance bias, what some texts call detection bias, occurs when one group is followed more closely than the other group.This could lead to an outcome being diagnosed more often in the more closely followed group, but not because it truly occurred more often in that group

Front 29

recall bias definition

when does this happen

Back 29

Problem with remembering past activities; if collected at appropriate intervals can be OK

unique to case control studies that rely on information provided by the subjects resulting in non-random misclassification bias. The notion is that because subjects are aware of their health status as cases or controls, such knowledge might lead to a differential recall of an exposure status

Front 30

changes in habits/exposure- what to do

Back 30

cannot control but must measure!

Front 31

confounders- what is they, distort what

may be responsible for what?

Back 31

Confounders are factors/variables/ characteristics that distort the RR between an exposure and outcome.

• May be responsible for the difference noted between groups

Front 32

how are confounders detected? (2)

Back 32

NO statistical test for confounding!
• Found via stratification and multivariate analysis to --> “adjusted” RR.

Front 33

confounders are associated with...(2) how are they associated

Back 33

• both exposure and outcomes.
The confounder is associated with the exposure and is a cause or contributes to the development of the outcome

– Classic example…cigarette smoking is a confounder of the relationship between coffee drinking and MI. Cigarette smoking is associated
with coffee drinking and is a cause of MI

Front 34

draw/describe confounding triangle

Back 34

confounding triangle

confounder with arrows pointing to characteristic/factor

also points to disease/outcome effect (relationship)

characteristic/factor has an "apparent relationship" arrow drawn to disease/outcome

Front 35

look at how to calculate adjusted RR and shit (stratification)

Back 35

---

Front 36

approaches to control confounding in the design of the study

Back 36

Exclude subjects with known confounding factors

Front 37

approaches to control confounding in the analysis of the study (2)

Back 37

– Stratify on the confounding factors e.g. by age, weight, sex,
disease severity
– Use multivariate analysis to adjust results

Front 38

adjusted RR definition

Back 38

relative risk or rate ratio that
is “corrected” for confounding

Front 39

why do we adjust for confounding?
how do we approach confounders when analyzing a study?

Back 39

To minimize the potential for differences in
the groups to be affected by some factor other than that which was measured

Important when we consider limitations of study
– Did the investigators consider all possible confounders?

Front 40

purpose of cohort studies

Back 40

Exposure and outcome are obtained on all
population members in order to estimate effects of the exposure on incidence of outcomes.

Front 41

what are cohorts?
example

what can be used as a control?

Back 41

Study Population are called cohorts

e.g. e.g., nurses 25‐42 in 11 most populous states in 1976; Control group or cohort is similar to the study cohort, except for the exposure

Data obtained in the general population may be used as control

Front 42

2 things to measure in the study...

Back 42

– Incidence data
– Risk factors (exposure status)

Front 43

Relative risk, Risk ratio, Incidence rate ratio...

Back 43

are all the same thing