Epid 710 - Lecture 8: Cross-Sectional Studies

Front 1

Concerns in cross-sectional studies

Back 1

1. Mixed effects of incidence and duration
2. temporal sequence and causation
3. Recall of exposure (not just in cross-sectional)

Front 2

Cohort Effect

Back 2

characteristics of the cohort that happen to be passing through that age that year

Front 3

Ecological Analysis

Back 3

uses data on populations rather than individuals

Front 4

Compositional Variables

Back 4

aggregate of individuals (e.g. rate)

Front 5

Contextual/integral variables

Back 5

affects everyone (e.g. water quality)

Front 6

Descriptive Epidemiology

Back 6

makes use of data that is already available to examine how rates vary according to demographic variables

Front 7

Analytical Epidemiology

Back 7

designing studies that allow for assessment of hypotheses

Front 8

Purpose of cross-sectional data

Back 8

informs health planners of current disease burden

Front 9

Example of cohort effect

Back 9

prevalence rates increasing for each age group from older to more recent cohorts

Front 10

Period Effect

Back 10

global shift or change in trends that affects the rates across birth cohorts AND age groups (e.g. war, new treatment, massive migration)

Front 11

Cohort Effect

Back 11

the effects resulting from lifetime experiences (like your literal birth) --> cohort effects relating to diet or or smoking habits (large, slow changes)

Front 12

Units of observation for ecologic studies

Back 12

usually geographically defined populations (countries/regions within country)

Front 13

3 different types of variables used in ecological studies

Back 13

1. aggregate measures: mean of characteristics within a group
2. environmental measures: physical characteristics of location of group of interest
3. global measures: characteristics that are not reducible to individual characteristics

Front 14

Ecological Fallacy (i.e. aggregation bias)

Back 14

Association observed at an aggregate level does not necessarily present an association on an individual level (Prussian Protestant example)

Front 15

Another word for ecological fallacy (hint: cross-____ _____)

Back 15

Cross-level inference: literally, when you infer from aggregate results to individual peoples

Front 16

point-source epidemic

Back 16

single source of exposure to a causal agent (e.g. food poisoning, cancer among survivors of atomic blasts)

Front 17

propagated epidemic

Back 17

causal agent itself is transmitted through a population (e.g.influenza)

Front 18

Prevalence Proportion
(sometimes just "prevalence")

Back 18

-measure of disease status
-proportion of people in population with disease
-P/N
-only valid for that one point in time (not good for capturing diseases with short durations)
-not useful for studying causes of disease, good for disease burden

Front 19

Prevalence Odds

Back 19

P/1-P

Front 20

In steady state (incidence rate and duration are same over time), prevalence is related to incidence rate, what's the formula?

Back 20

P/1-P = I * D
where
I--> Incidence rate
D --> Disease Duration

Front 21

Longitudinal Study (as opposed to cross-sectional)

Back 21

if information obtained pertains to more than one point in time

(all cohort studies and MOST case-control studies)

Front 22

Cross-sectional study

Back 22

"snapshots" of population status with respect to disease and/or exposure variables at a specific point in time
-all about disease prevalence

Front 23

Can cross-sectional studies measure disease incidence?

Back 23

No! cross-sectional studies cannot measure risk or rate calculations because both of those require more than one time point

Front 24

Case-Control studies provide

Back 24

Ratios measures of effect only!

Front 25

Cohort studies provide

Back 25

estimates of disease rates and risks for each cohort

Front 26

Cohort study vs. case-control study

Back 26

Cohort: study many different diseases related to same exposure
Case-control: study many different exposures and one disease

Front 27

Cohort Study characteristics (5)

Back 27

1. complete source population denominator experience tallied
2. can calculate incidence rates or risks and their differences and ratios
3. usually very expensive
4. convenient for studying many diseases
5. can be prospective or restrospective

Front 28

Case-Control study characteristics (5)

Back 28

1. sampling from source population
2. can usually calculate only ratio of incidence rates or risks
3. less expensive
4. convenient for studying many exposures
5. can be prospective or retrospective