Biostats: study designs

Question 1

Define and list examples of descriptive studies

Answer 1

def: describes patterns of disease occurence. used to fomulate hypotheses but can’t test it.
- case reports/case series
- ecological studies
- studies of disease frequency

Question 2

List guidelines for determining if association is causal

Answer 2

• temporal relationship- E before D
• strength of association - measured by RR or OR
• Dose-Response relationship - when strength of assoc increases as function of exposure dose
• consistency
• biological plausibility-reasonable explantion
• consideration of alternative explanations
• cessation of exposure

Question 3

What are the advantages and disadvantages of case report/series studies?

Answer 3

Advantages:
-easy and cheap
-early warning
-generates hypothesis
Disadvantages
-can't test hypothesis
-limited generalizability

Question 4

What are the advantages and disadvantages of ecological studies?

Answer 4

(collects E and disease freq info in populations)
Advantages
-efficient
-population based
-prelim test of hypothesis
Disadvantages
-can't control for confounding factors
-ecological fallacy - no guarentee that persons with disease are same persons who were exposed

Question 5

What is a cross sectional study?

Answer 5

Measures E and D at the same time for all subjects. Calculate RR using prevalence or OR.
E.g. Stool samples taken from patients admitted to nursing home. Height and weight taken from chart. Associated thin patients with positive test for intestinal parasites.

Question 6

What are the advantages and disadvantages of Cross Sectional studies?

Answer 6

Advantages:
-quick and low cost
-generates hypothesis
-generalizability of results
Disadvantages
-Temporal rel - hard to say that D didn't cause E
-prevalence bias- prevalent cases who've had disease for longer time may have dif risk factors
-selection bias
-confouding
-hard to study rare disease

Question 7

What is a cohort study?

Answer 7

Individuals without outcome followed over period of time. Includes both exposed and unexposed individuals. Asses baseline and follow disease status.

Question 8

What are the advantages and disadvantages of cohort studies?

Answer 8

Advantages
-time seq: E before D
-rare exps can be studied
-multiple outcomes
-less prone to selection bias
-information bias less likely. 
-IR can be measured
Disadvantages
-cost
not appropriate for rare disease outcomes
-validity: confounding, loss to follow up, information bias
-nonparticipation

Question 9

What is a case control study?

Answer 9

-finds subjects with and without disease. Obtains data regarding past exposures. Assumptions: cases and controls from same population.
cases representative of all cases from the area, same with controls. Measure using OR.

Question 10

What are the advantages and disadvantages of case control studies?

Answer 10

Advantages: 
-can study rare diseases
-can study multiple exposures
-efficient
-OR good approx for RR
Disadvantages
-information bias: reporting of past exposure. Ppl with disease may remember past exposures differently than controls
-selection bias
-prevalence bias : Cases with severe emphysema more likely to smoke, have higher fatality than cases with less severe emphysema, so the prevalence of emphysema in smokers will be 
underestimated compared to incidence
-confounding
-temporal relationship problems

Question 11

What is a nested case control study?

Answer 11

cases are members of cohort who develop the disease. Controls are from other members of cohort. Exposures based on info previously collected on all members of cohort uniformly.

Question 12

What are the advantages and disadvantages of intervention studies?

Answer 12

Advantages
-can prove causality
-controlled time sequence
-ruled out selection bias
-multiple outcomes
-IRs can be measured
Disadvantages
-costs
-limited exposures and limited outcomes
-placebo effect
-information bias: recal bias, interview bias
-lost to follow up and noncompliance
-confounding

Question 13

What are the main factors that affect Internal Validity?

Answer 13

Selection Bias-arises when selection of study subjects results in change in value of association: which includes diagnostic bias, surveillance bias, non participation, self selection, lost to follow up, noncompliance

Information Bias-on observer end-airses when measurement of E or D is incorrect: non differential and differential

• recall bias-inaccurate recall of past exposures
• interview bias-when investigator know hypothesis, exposure, or outcomes
• surveillance bias - when exposure associated with increased monitoring
• misclassification of E or D: taking only one blood pressure leading to error

Confounding
confounder can be associated with E and D. True confounder when RR changes by 10% after adjustment.