-
How can one control for confounding (5)
- · Homogeneous sample: by restricting age range
- · Match your experimental and control groups in regards to demographics, such as age (even out the demographics in each group)
- · Randomization of treatments to groups
- · Stratification in randomization or in analysis
- · Covariate analysis
-
What are the characteristics and the strengths and limitations (3) of the ecological study designs? (5)
- Use existing data gathered on populations to examine associations between diseases and other factors
- Data are not usually gathered for this purpose
- Confounding factors cannot be controlled
- No control over the quality of the data
- The weakest study design for making inferences about causality
-
What are the characteristics and the strengths (3) and limitations (5) of the case-control study designs?
- Compares characteristics of individuals with a disease to those without, matching the cases and controls as closely as possible on variables such as age, sex, socioeconomic status
- Retrospective study (ppl who already have disease); lung cancer and smoking (benzene can also cause lung cancer)
- Results are obtained quickly
- Useful with rare diseases
- Concerns:
- Comparability of cases and controls
- Selection bias in choosing controls
- Alternative explanations of causality
- Does the putative causal factor precede disease or did it result from it
- Recall bias may be present b/c pts are recalling back in time what they ate
-
What are the characteristics and the strengths (4) and limitations (6) of the cohort study designs?
- Prospective (have to wait for ppl to get disease)
- Compare outcomes over time of individuals with differing exposures
- Advantages
- · Diet and other exposure data are more reliable b/c prospective
- · Exposure precedes the disease
- · Selection bias minimized
- · Many hypotheses can be addressed
- Concerns:
- Can be very lengthy
- Can be very expensive
- Requires large numbers of subjects unless disease under study is very common
- A change in exposure status may occur during the study (such as half way thru the nurse’s study, folic acid fortification began)
- Loss to follow-up such as if ppl move, drop out, etc
- Alternative explanations of causality
-
What are the characteristics and the strengths and limitations (4) of the experimental study designs?
- -only one that you can for sure conclude that one thing
- causes another
- The gold standard of analytic research
- All factors are held constant except those manipulated by the investigator
- Subjects are randomized to a treatment or control group
- Double-blinding where possible
- The only design that definitively establishes causality
- Concerns:
- Expensive
- Time consuming- compared to case control or ecological study where you can just pull some records
- Only very specific hypotheses can be tested
- Drop out by subjects can compromise results
-
Factors that increase confidence in results of studies (especially in observational studies) (6)
- Dose response relationship
- Consistency of findings with other studies, such as many animal studies showing same result
- Biological plausibility
- Specificity of the association such as cigarettes and lung cancer
- Control of other possible explanations (confounders) in the design or analysis
- Strength of the association
-
Questions to ask when assessing the internal validity of a scientific experiment
- • Is there a control group?
- • Were subjects randomized?
- • Was double-blinding present?
- • Were groups treated equally except for the intervention?
-
What is external validity of a scientific experiment
ability to generalize results of a study beyond the subjects that were in the study
|
|