Lecture 20: Cross Sectional Studies Flashcards
Review: What are the 3 types of observational studies?
-Cross sectional (sample without regard to exposure or outcome status)
-Case-control (sample based on outcome status)
-Cohort (sample based on exposure status)
*main difference is the way participants are selected
Give an overview of cross sectional studies?
-We choose our study participants without regard to exposure or outcome
-Then we measure the exposure and outcome at the same time (these studies are often referred ti as taking a “snapshot”)
-Studies involve a random (not always but best to avoid selection bias) selection of participants without regards to their exposure or outcome then participants are questioned about the exposure and outcome of interest AT THE SAME TIME
What is the study design of cross sectional studies?
Objective: prevalence estimates
-How many of our study subjects have the exposure of interest? How many have the outcome of interest?
Sampling: Without regard to exposure or outcome status, snap shot in time
Analysis: Compare prevalence or odds of outcome in exposed and unexposed groups
What are the steps in cross sectional studies?
- Randomly sample individuals from the source population (to avoid selection bias)
- Take measurements on each individual at one point in time to determine both exposure and outcome status (measurement can be through a questionnaire, medical test, survey etc)
- Classify each based on their OUTCOME (O) (prevalence only study would end here)
- Classify each based on their EXPOSURE (E)
- Compare prevalence or odds of outcome in E+ vs E- groups (or prevalence or odds of exposure in O+ vs O- groups)
What is the only thing we can measure in a cross sectional study?
We can only measure prevalence NOT incidence
Prevalence: new and old cases (have the outcome)
-Point P: at a specific point in time
-Period P: over a defined period/time range
Incidence: new cases (get the outcome)
-This is bc we can’t establish temporality since we are determining the exposure and outcome at the same time (exception is a longitudinal study)
What are the advantages of cross-sectional studies?
-Can determine prevalence of E and O in the population
-Can study multiple exposures and multiple outcomes
-Appropriate for studying permanent factors (not changing: sex, ethnicity, stable environmental factors)
-Relatively fast and inexpensive
-No loss to follow-up selection bias
-Less potential for bias than case-control studies
-Can estimate almost all MOE and MOA except for rate bc can’t measure incidence
What are disadvantages to cross-sectional studies?
-not good for rare exposures, rare outcomes, or diseases with short duration
-Rare outcomes (brain tumour) bc if have rare outcome have a hard time finding participants that have the outcome of interest
-Rare exposure (radon exposure)
-Doesn’t measure incidence
-Cant establish temporal sequence (did E or O come first)
-Need to be carful with non-permanent (transient) factors
-Bias (selection, info and confounding)
How does selection bias relate to cross sectional studies?
Recall selection bias: systemic differences in those participating in the study and those not
types of selection bias
-Non-response/volunteer
-Selective entry/ healthy worker effect
-Detection/surveillance bias
-Loss to follow up bias (not a concern in cross-sectional studies)
ex: conducting research on a health outcome in general horses but we only sapling racehorses (selective entry)
How does info bias relate to cross sectional studies?
Recall info bias: systemic error in the way data is collected about exposures and outcomes
types of info bias
-Measurement error: ie scale isnt calibrated
-Misclassification bias: ir individuals are assigned to the wrong cell in 2x2 table (2 kinds differential and non-differential)
Ex: individuals who were sick (O) after a party are more likely to remember what they ate (E) is recall bias which is part of differential
