HUN4445 Exam3 Research

Card Set Information

HUN4445 Exam3 Research
2012-11-19 13:41:28
HUN4445 Exam3 Research

HUN4445 Exam3 Research
Show Answers:

  1. 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
  2. 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
  3. 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
  4. 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
  5. 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
  6. 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
  7. 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?
  8. What is external validity of a scientific experiment
    ability to generalize results of a study beyond the subjects that were in the study