Final Exam For Eab 705

Front 1

Study designs are used to explore research questions. Information gained must be published/shared to disseminate knowledge.

Back 1

-Specific types of designs answer certain types of research questions better than others.
-In epidemiology, most study designs are observational because we don't intervene.

Hint 1

week 10 notes

Front 2

Discuss the characteristics of a case-series study design.
What are some weaknesses?

Back 2

-A series of cases
-Carefully describe what is found.
-There is no control/comparison group.
-Used mostly as a basis for further and future research, and to generate hypotheses.

Hint 2

(so you cannot calculate relative risk, attributable risk, or odds ratios).
-Cannot make statements about preventing disease or causation (inferences cannot be drawn).
-Findings must be delivered cautiously as this is a weak(er) study design.
week 10 notes

Front 3

Discuss the characteristics of an ecological study design.
What are some weaknesses?

Back 3

-Compare groups, but do not get data on individuals.
-A crude method to explain associations, usually between occupation or environment and a disease.
-Used mostly to generate hypotheses.

Hint 3

-Susceptible to "ecological fallacy," assuming things are related when there are other explanations (ex: suicides are higher on the "Catholic" side of town.....doesn't mean more Catholics commit suicide, just means the people that live in that area are more apt to).
week 10 notes

Front 4

Discuss the characteristics of a cross-sectional study design.
What are some weaknesses?

Back 4

-Observational, no intervention.
-Effective at providing a picture of disease, health, or psychological phenomenon.
-Effective at assessing several variables at once and associations between those variables.
-Can also be called: survey, slice of time, snapshot, prevalence study.
-Easy & quick; disease & exposure data are collected simultaneously.
-Can determine distribution of disease and exposure.
-Often used to gain initial data.

Hint 4

-Cannot assign cause.
-Temporal sequence is not always known.
-No statement of risk.
week 10 notes

Front 5

Discuss the characteristics of an experimental study design.
What are some weaknesses?

Back 5

-Most "believable" of study designs
-There is an intervention.
-Participants are randomly assigned to a group (experimental [i.e. get a treatment] or control), which takes care of internal validity threats.
-The hope is that both groups are the same in the beginning, and the group receiving the intervention does better.
-Usually use biostatistical methods rather than epidemiological.

Hint 5

-Unethical in many disease cases (ex: you can't assign people to smoking or not).
-Sometimes takes place in an unnatural environment.
week 10, 11 notes

Front 6

Discuss the characteristics of a quasi-experimental study design.

Back 6

-Just like an experimental design, except there is no randomization (there is still manipulation though)

Hint 6

week 10 notes

Front 7

Discuss the characteristics of a case-control study design.
What are some weaknesses?

Back 7

-Also called a retrospective study.
-Is observational (cases have the disease, controls are similar people who do not). Note: controls must be similar UNLESS you want to study age, race, or sex.
-Prevalence is always 100% in case group and 0% in control group.
-Can get odds ratio or relative odds.
-Compare the past exposures of cases and controls: look for past exposures that are common in the ill and rare in the well.
-Data is collected after disease onset in the cases.
-Shows relationships and can describe a problem.
-Quick and easy.
-Works well with rare diseases because you can find participants easily.
-Can get odds ratio.
-Effective for outbreaks.
-Often done with a low level of knowledge.

Hint 7

-Selective survival is a concern (ex: those who died having a heart attack could not be in a study on heart attacks).
-Can be difficult to find controls.
-Cannot get relative risk because you don't know the incidence.
-Temporal sequence is unknown sometimes.
-Recall bias is a concern, also selective memory (those with the disease will remember better than those without it).
week 10, 11 notes

Front 8

Discuss the characteristics of a cohort study design.
What are some weaknesses?

Back 8

-Also called prospective, panel, longitudinal, or follow-up studies.
-All subjects are "well" at onset (usually determined by dr).
-Observe people in "natural settings" and divide based on some exposure, look for outcome in the future.
-Data collected before the disease so the temporal sequence is known.
-Gets disease rates, provide incidence data, and relative risk.
-Is a strong study disease and is the "super-star" in epidemiology.
-Can make statements of relative risk and strength of association.
-Can calculate relative and attributable risk.
-Is good for rare exposures.
-Note: there is also a retrospective cohort study design (also called non-concurrent or historical) where exposure is determined from past records. You can also do a combination of both where you get exposure from records then wait for the outcome.

Hint 8

-Exposure status can change (ex: quit smoking).
-Can be long, you have to wait for cases to occur, then may have to wait after that (ex: most wait 2 years to make sure things have been diagnosed correctly).
-Attrition, which can be detrimental to the study because those that stay involved are likely the successful/health/etc. ones.
-A lot of work: you may have to bring in people every year for medical tests.
-$$
-Not good for rare diseases because you may wait and wait and no cases develop.
-To avoid investigator bias, the person who determines if a disease has occurred should not know if the person was exposed.
-In cases of rare exposure, you must go where the exposure happens.
week 10, 11 notes

Front 9

Discuss the Framingham study.

Back 9

-Most famous cohort study.
-Most of our data on heart disease comes from this study.
-Started in 1947, in a stable population in Framingham, Mass, outside of Boston.
-Looking for risk factors of coronary heart disease.
-Grandchildren are now involved to look at heredity.
-Other diseases have now been added.
-Participants are retested every 2 years.

Hint 9

week 11 notes

Front 10

How does a case-control study differ from a cohort study?

Back 10

-They are opposite!
-In case-control studies, you often start with a low level of knowledge.
-In case-control studies, the end point is the exposure, in cohort studies the end point is the disease.
-How the subjects were originally assigned.

Hint 10

-If they were assigned based on outcome status (if they have disease or not), it's a case-control study.
-If they were assigned based on exposure status (ex: if they smoke or not), it's a cohort study.
week 10, 11 notes

Front 11

What is relative risk?

Back 11

A statistic gained from cohort studies that shows the strength of association of exposure and disease.

Hint 11

week 11 notes

Front 12

How is RR calculated?
What do the numbers mean?

Back 12

(a / (a+b)) / (c / (c+d))
-Basically it's incidence in the high-risk (exposed) group divided by incidence in the low-risk (unexposed) group.

Hint 12

-If RR > 1.0, the factor is said to be causative.
-If RR = 1.0, the factor is unrelated to the disease.
-If RR < 1.0, the factor is said to be protective against the disease.
-Ex: if RR is 1.4, you are 40% more likely to get the disease if you've had the exposure, or you're 1.4 times as apt to get the disease.
week 11 notes

Front 13

What is attributable risk?

Back 13

-A statistic gained from cohort studies that is a measure of how much disease could be eliminated if the exposure was not present.
-Used by health educators because it tells how much disease we could prevent, if an exposure were eliminated.

Hint 13

Note: AR is the risk in the high-risk group.
week 11, 12 notes

Front 14

How is AR calculated?
What do the numbers mean?

Back 14

(a/a+b) - (c/c+d)
-Basically it's incidence in the high-risk group - incidence in the low-risk group.....you are subtracting out the people who got the disease but did not get the exposure (because you're accounting for the fact that some of the people who got the disease would have gotten the disease even if they hadn't have had the exposure).

Hint 14

If you're AR = 0.00358, you can say that 3.58 cases per 1000 could have been eliminated if they did not have the exposure.
week 11 notes

Front 15

What is attributable risk percent?

Back 15

-A statistic gained from cohort studies that (shows what % of a disease could be eliminated if people didn't have an exposure).
-Estimates the proportion of disease among the exposed that is attributable to the exposure or the proportion of the disease that could be eliminated if people did not smoke/drink/have high cholesterol, etc.
-Is also called "etiological fraction."

Hint 15

week 11 notes

Front 16

How is AR% calculated?
What do the numbers mean?

Back 16

((a/a+b) - (c/c+d)) / a/(a+b)

Hint 16

If AR% = 0.397, you can say 39.7% of disease among those exposed is attributable to the exposure, and would be eliminated if people were not exposed.
week 11 notes

Front 17

What is population attributable risk?

Back 17

Indicates the proportion of all cases in the total defined population that can be ascribed to a factor.

Hint 17

Note: population AR is the risk in the whole population.
week 11, 12 notes

Front 18

How is population AR calculated?
What do the numbers mean?

Back 18

((a + c) / total pop) - c / (c + d)
-Basically it's incidence in the total population - incidence in the unexposed.

Hint 18

If your population AR is 0.00078, you can say that 7.8 cases per 10,000 people in the population are attributed to the exposure and would be eliminated if not exposed.
week 11 notes

Front 19

What is population attributable risk percent?

Back 19

Shows the % of disease among the entire population that can be attributed to the exposure.

Hint 19

Note: is easy for lay people to understand
week 11 notes

Front 20

How is population AR % calculated?
What do the numbers mean?

Back 20

((a+c) / (a+b+c+d) - c/ (c+d) / ((a+c) / a+b+c+d)

Hint 20

If population AR% is 0.1264, then 12.64% of all cases can be attributed to the exposure.
week 11 notes

Front 21

What is one weakness of all the kinds of risk that can be calculated?

Back 21

They over-simplify the relationship between the outcome and the exposure, because lots of other things could be going on.

Hint 21

week 11 notes

Front 22

What is odds ratio?

Back 22

-A statistic gained from case-control studies that's an estimate of the relative risk between groups.
-Is a measure of strength of association, just as RR is an expression of strength of association.

Hint 22

week 11 notes

Front 23

How is odds ratio calculated? What do the numbers mean?

Back 23

(a x d) / (b x c)

Hint 23

If your odds ratio = 14, the cases were 14 times as apt to be (smokers, users, chili pepper eaters, etc) as controls.
Note: the odds ratio will be similar to the RR unless incidence in the high-risk group is high (a/a+b)!!
Note: remember you are talking about the disease first here.
week 11, 12 notes