PCM1: Observational and Case-control studies

=Objectives=

Ecological (Correlation Study)

 * A measure of disease freqency in an entire population
 * date is not obtained from individuals
 * evidence of a linear relation between the exposure frequency and an outcome
 * example: graph of tobacco use per capital consumption (vs.) time AND age-adjusted lung cancer (vs.) time. The graph shows a lag in the effect of smoking on a population basis.  Also, when you separate men and women (who started smoking later), the correlation between these smoking and lung cancer become apparent.
 * Advantage:

Surveillance

 * descriptive
 * follow disease incidences over time and stratify by identifiers (eg, race, sex, age, etc). For example, we can look at cancer incidence data throughout time to identify trends. We can see spikes in the onset of male cancer when PSA testing became more prevalent in the clinic.  We can also see how effective we are with public health initiatives.  Example2: Diabetes prevalence in the population from 1990 to 2007 shows the rising epidemic.
 * Advantage: descriptive, comprehensive descriptions in an entire population
 * Disadvantage: it can't tell us why.

Cross-sectional Studies

 * snapshot: information about the exposure measure and the outcome measure are obtained at the same time.
 * Frequency measure is usually prevalence
 * Example: connection between diabetes and inflammation looked at using population based survey. Looks at diabetes status in 10k patients and stratified them.  Then in each of the subsets of patients, he looked at a biochemical readoout (C-reactive protein) to see if it's elevated.  Data readout is usually a prevalence ratio (ie 34% of diabetics have elevated levels but only 13% of normal people have elevated CRP --> 2.3 prevalence ratio.
 * generates "crude prevalence ratio" (univariate) and an "adjusted: or multivariate
 * disadvantage: no followup

Case controlled studies

 * motivation for these is cost and feasibility for low incidence diseases, (eg invasive colon cancer in oregon 50/100k).
 * How are subjects selected?
 * Identify cases newly diagnosed with the condition
 * Identify controls. Most difficult part....people in the source population who have never been diagnosed with the condition but are at risk. controls don't have to be the "general population" but need to get at the background population.
 * How are subjects evaluated? query both groups about past history of exposure. can be done with medical records, questionnaires, etc. case and control status should be blinded in the collection of data/survey/etc.
 * What measure of frequency is used?
 * Odds ratio
 * crude odds ratio (univaraiate) or adjusted odds (multivariate)
 * Note that the odds ratio is not the underlying disease incidence rate; rather, there's a mathematical relationship between underlying disease and the odds ratio.
 * What measure of association is used?
 * What are the major advantages and disadvantages?
 * Number are fixed. no follow up occurs
 * Number are fixed. no follow up occurs

Describe the main features of cross-sectional and case-control studies and be able to answer the following questions for each: a) How are subjects selected? b) How are subjects evaluated? c) What measure of frequency is used? d) What measure of association is used? e) What are the major advantages and disadvantages?

Demonstrate knowledge of an ecological, cross-sectional or case-control study.

Describe what information can be provided by Ecological, Cross-Sectional, and Case-Control studies that other study types can't provide.