Study Design

Question 1

Characteristics Of Descriptive study

Answer 1

• case report, case series , survey
• describe but do not compare
• hypothesis generating
• No inferences

Question 2

Characteristics of Analytic studies

Answer 2

• experimental and Observational
• Make comparisons btwn groups
• inferences abt E and O associations
• hypothesis testing

Question 3

Characteristics of Cross - sectional studies(advantages/disadvantages)

Answer 3

lower causal inference because they measure prevalence not incidence and b/c of the inability to refute revers-causation (ie which camefirst- E or O )

• best for permanent exposures
• snapshot- one point in time
• fast
• Inexpensive
• can use all MoA if random sampling
• less bias than C-Cstudies
• recall bias

Question 4

Steps in Cross Sectional study.

Answer 4

I. Sample without regard to E or O

2. Randomly select sample from source pop
3. take measurements of each person in study group to determine E and O status
4. Classify based on E and O
5. Compare prevalence or odds of outcome in E positive and E negative

Question 5

Steps in a cohort study

Answer 5

I. Clear/ consice objectives

2. Define target and source pop
3. Define unit of concern (ie indindual or groups)
4. Define E and O
5. set followup period

Question 6

Longitudinal Cohort.

Answer 6

You do not know E status of ppl so you select a single group of participants, heterogenous with respect to E of Interest

Question 7

How do you sample in Cohort?

Answer 7

sample based on E Status or large hetero group with respect to E

Question 8

How do you ensure validity in Cohorts?

Answer 8

1. Ensure representation of target pop (external)
2. Ensure comparability with respect to other factors that are not associated with E of interest. (internal) , by:
   a) exclusion /restricted sampling : Identify CFV’s and include only 1 level
   b) Matching based on CFV
   ) Analytic Control : through Mantel Haenazel Or regression
   3 . Follow up equally
3. clear diagnostic criteria

Question 9

Risk-based cohort

Answer 9

• closed pop/ group: all subjects Observed for full study period
  -short risk period
• measures risk of O in E positive and E negative groups
  -Egroups defined at beginning of study period.
• Analysis:
  I . bi variable risk-based (MOE and M0A)
  2. Stratified Analysis
  3. Multivariable regression cotten logistic)

Question 10

Rate Based cohort

Answer 10

• open pop
  -rates used to measure disease Freq.
• Incidence of disease is expressed relative to the time at risk in each level of exposure :
  i. e: sum of amount of time at risk contributed by each memher in each grap (denominator)
  → contribute to time at risk until they develop O change E Status, are lost to follow up , W D or study period ends
  Analysis:
  1. Rates
  2. Surviva l Analysis

Question 11

Sampling in case control

Answer 11

sample purposively based on O status

Question 12

Simple Overview : Case control

Answer 12

1. select a ft of individuals that have newly the O of interest(cases)
2. compare to a group who do not have O of interest(controls)
3. frequency of exposure factors are compared between cases + controls

Question 13

controls need to be:

Answer 13

FROM THE SAME STUDY BASEAS CASES
Not just “healthy ppl” they are Simply ppl without O of interest
they have the same exposure to exposure factors as cases

Question 14

what is the goal of rate based sampling .

Answer 14

controls Mirror exposure-time at risk of non-cases in source pop

Question 15

What is the goal of risk based sampling ?

Answer 15

controls represent exposure in source Pop

Question 16

How to select controls for primary base rate-based sampling

Answer 16

1. select from source pop at the end of study period proportional to time at risk. → requires stable pop (Ie exposure does not vary) TAR data known
2. selected at fixed time points during study from risk set→ assumes Stable pop TAR data NOT known
3. selected from risk set, matched timewise to cases → AKA incidence density sampling select control when case occurs , matched analysis necessary if exposure varies with time

Question 17

How to select controls from a rate-based, secondary base study:

Answer 17

1. select Controls from all non-case ADMISSIONS at end of Study → select controls from norn- case diagnoses that are not associated with exposure
2. selected from registry at fixed time points
3. selected from regsistry Matched - time wise to case admissions

Question 18

How to select controls for a risk-based primary base Study

Answer 18

1. Random selection of non-cases in source pop → generally occurs atthe end of the risk/study period(However can be selected matched timewise to cases )

Question 19

How to select controls for risk-based secondary base study

Answer 19

1. Random selection of non- cases in the same secondary base registry → Generally occurs atthe end of the risk /study period - but can be selected time wise to cases.

Question 20

primary Base

Answer 20

Well- defined source pop ( literally or conceptually have a list of all ppl in source pop )
ex federal disease registry.

Question 21

secondary Base

Answer 21

one or more steps removed from the actual source pop

ex refferals to a clinic

Question 22

Risk set

Answer 22

Those non-cases eligible to be cases at that point in time

Question 23

incidence Density sampling

Answer 23

# of controls randomly selected at the time a case arises from the risk set
→ well suited for when exposure changes with calendar time (Matched analysis needed) 
→ in rate based design  people initially identified as controls can become cases 
→ mostcommon, do not need stable pop Or TAR data