EBM Exam 1

  1. Definitions of bias
    • systematic error in measurement or difference between groups
    • process which over or underestimates conclusions and distorts comparisons
  2. potential for bias exists
    in all clinical trials
  3. bias cannot be ____ only ____.
    • eliminated
    • minimized
  4. Dismiss a study as a bad one based solely upon the presence of bias?
    not reasonable
  5. Biase must be identified and
    taken into consideration when evaluating literagure
  6. Internal validity
    • the degree to which the results of an individual trial are likely to be true
    • observations during the trial can be attributed to the intervention
  7. Internal validity is "internal" because
    it applies only to the exact conditions within the trial
  8. Internal validity =/=
  9. External Validity
    degree to which the results obtained from a trial can be extended to make predictions in settings outside the trial
  10. two types of external validity
    • population validity
    • ecological validity
  11. population validity
    the sample population studies sufficiently represents the larger population to make conclusions
  12. ecological validity
    the conditions of the study accurately reflect the "real world" conditions
  13. external validity is also referred to as
  14. What most accurately defines "bias" as it relates to clinical trials?
    a systematic error that results in overestimate or underestimated conclusions
  15. Volunteer bias
    • those who volunteer are not representative of the population
    • concern in trials with incentives
    • can be minimized with appropriate inclusion/exclusion criteria
    • occurs during enrollment
  16. Volunteer bias affects
    external validity
  17. Subject Bias
    • occurs from participants changing their behavior
    • occurs during data collection
  18. types of subject bias
    • response bias
    • social desirability bias
    • hawthorne effect
  19. response bias
    pt give response they think the investigaotro wants
  20. social desirability bias
    • reluctance of pt to give socially unacceptable answers out of fear of being judged
    • smoking
    • drinking
  21. Hawthorne effect subject bias
    pt acts differently bc they are aware of being monitored
  22. subject bias affects
    internal and external validity
  23. selection bias
    • systematic differences between age groups
    • occurs if subjects do not have an equal chance to develope outcome
  24. reduces incidence of selection bias
  25. compliance bias
    • treatments being compared are dissimilar, leading to decreased compliance/adherance for one study group
    • bad tasting meds
    • complicated dosing schedule
  26. compliance bias minimized with
    • blinding
    • double dummy blinding
  27. measurement bias
    differences in measurement techniques
  28. to minimize measurement bias
    • central laboratory
    • training of investigators taking measurements
    • calibration of instruments
    • take average of repeated measurements
  29. surveillance bias
    investigator gives subtle cues that influence subject's response
  30. survellance bias is best minimized by
    double blinding
  31. surveilance bias can also be minimized by
    • standard time spent with each pt
    • standard questionaires
    • training investigators
  32. Attrition
    reduction or decrease in number or size
  33. attrition bias caused by
    • subjects to not respond to questionaires
    • subjects relocate or cannot be located
    • subjects withdrawl consent
    • subjects drop out due to side effects
    • death/ill health, coginitive decline
  34. publication bias
    • tendancy of published literature to be dependent on nature of the study and results
    • "hot" topics are more likely to be published
    • investigators want to be published in well known journals (impact factor)
  35. Language Bias
    • only published in author's native language
    • may not pass peer review due to poor translation
    • non-english investigators more likely to publish in english if their results are positive
  36. time lag
    • speed of publication depends on findings
    • studies with positive results are published twice as quickly
  37. title bias
    title should be free of author's preferences
  38. T/F. When performing literature search, disreguard trials w/ titles that indicate investigator preference.
Card Set
EBM Exam 1
Trial Bias