1. Study Design

2. Study Designs

3. Descriptive Studies Record events, observations or activities,documentaries No comparison group or intervention Describe an event that needs to be brought to the attention of colleagues May lead to additional discoveries Exercise caution with interpretation and application of information

4. Observational Studies Include case-control, cohort (follow- up), and cross-sectional studies Goal is to answer questions, not simply describe a situation Include a comparison Report the natural course of events Researchers are bystanders

5. Subjects With Disease (case) Disease (case) Subjects Without Disease (control) StudyPopulation Prior Exposure to Suspected Risk Factor Prior Exposure to Suspected Risk Factor No Prior Exposure to Suspected Risk Factor No Prior Exposure to Suspected Risk Factor Present-------------Time---------------Past Case Control Studies

6. Case-control Studies: Strengths Efficient –Time –Money Useful for the study of rare diseases Ethical concerns are limited

7. Case-control Studies: Weaknesses Incomplete or biased records Recall of subjects may be selective or biased Selection of controls –From clinic, acute care facility, or community –Who included and excluded –Matched controls Selection of cases Info gathering, differences between groups, etc.

8. Subjects Exposed Subjects Not Exposed StudySample at Risk at Risk Develop Measured Outcome Outcome Do Not Develop Measured Outcome Do Not Develop Measured Outcome Time Cohort Studies

9. Strengths –Data is collected prospectively and in a more controlled fashion –Subjects are not required to “recall” data –More accurate assessment of “which” came first Weaknesses –High cost and much time required –Population selection –Loss to follow-up –Changes in subject characteristics –Surveillance bias

10. Case-control study: takes the outcome as the starting point of the inquiry and looks for precursors or risk factors Cohort Study: starts with a risk factor or exposure and looks at consequences Comparison Case Control and Cohort

11. Cross-sectional Studies Assesses exposure and outcome status simultaneously “Slice-in-time” design or prevalence Assessment made during a single examination

12. Cross-sectional Studies Subjects With Disease Subjects Free Of Disease StudySample Risk Factor Present Present No Risk Factor Present Present PRESENT

13. Cross-sectional Studies: Evaluate a new test Evaluate the predictive capability of clinical features Identify etiological agents or causative factors Determine prevalence

14. Cross-sectional Design Strengths –Efficient use of time –Efficient use of money Weaknesses –Population selection –Subject selection –Antecedent-Consequent relationship

15. Review : Study Designs

16. Experimental Studies Also known as a controlled trial, clinical trial, health-care trial, or intervention trial Goal is to answer questions, not simply describe a situation Include a comparison An intervention is planned An outcome is measured Subjects followed forward in time

17. Definitions Retrospective Trial –Case-control study –A study that analyzes previously collected data –Carefully kept records are needed Prospective Trial: –Cohort or follow-up study –A study in which data are collected forward in time from the date of study initiation

18. Clinical Trials: Study Design and Methods of Control. \

19. Study Designs

20. The True Experiment Clinical trial, controlled trial, health-care trial, intervention trial Tool to evaluate clinical problems Investigator actively intervenes Intervention applied (independent variable) Outcome is observed (dependent variable) Goal is to establish a causal relationship

21. The True Experiment Most effective for controlling extraneous factors affecting interpretation of results Limitations include: –Expense –Ethics –Time required –Complexity

22. The Hypothesis Stated in introduction section Focuses a study Null hypothesis –Hypothesis of no difference or no relationship –Results held to be true until evidence shows otherwise Research hypothesis: expected results of the experiment based on theory or related research

23. The Hypothesis Two Tailed Hypothesis States that an association or difference exists No direction of association specified Statistically stronger One Tailed Hypothesis States that an association or difference exists Direction between the predictor & outcome variables is specified Statistically easier to prove

24. Subject Identification Population selection Subject selection –Convenience sample –Systematic sample –Random sample Inclusion & exclusion criteria Author should clearly explain process of identifying subjects

25. Methods of Control Randomization Placebo, active, or historical control Blinding Stratification Matching Cross-over design

26. Randomization Always implies there is a control group Assignment procedure Improves validity of a study Subjects are assigned to groups using a system independent of personal judgment or subjects themselves Every member of population has a predetermined probability of being chosen for a group

27. Randomization Goal is to decrease/eliminate selection bias Individuals randomized may or may not be representative of a larger population Differs from random sampling (individuals are selected for a study by chance from a larger group or population; Aim is to obtain a sample that reflects characteristics of a larger group)

28. Blinding Eliminates bias that randomization cannot eliminate Errors occur when patient or individual making assessment are aware of treatment administered, especially if outcome is subjective Important to guarantee that the observed outcome is due to the applied therapy Attempts to overcome placebo effect

29. Blinding Single blind: subject unaware of therapy received Double blind: neither the observer nor subject aware of therapy received Triple blind: subject, investigator, and researcher measuring outcome unaware of therapy received Often impractical or unsuccessful

30. Stratification Method for dealing with inequalities remaining even with randomization Subjects are grouped based on risk factors or co-morbid conditions capable of influencing the results prior to randomization

31. Matching Assignment procedure Goal is to decrease/eliminate selection bias Study and control group are selected to ensure that an identified variable is the same in both groups Ability to evaluate influence of “matched characteristic” on outcome is lost

32. Cross-over Design Uses subjects as their own control Good for a small number of patients Subjects assigned to all interventions and measurements obtained for each treatment Control for order effect (randomize) Control for carryover effect (sufficient wash-out period) Blind the cross-over point

33. Cross-over Design Group 1 Group 2 Group 1 Cross-over Point Tx A Tx B Tx A

34. Cross-over Design Group 1 Group 2 Group 1 Wash-out Pd Tx A Tx B Tx A

35. Case Study Sample Treatment Measure Outcome Variables Population X

36. One Group Pre-treatment & Post- treatment Design Treatment Measure Outcome Variables Population X Sample Measure Baseline Variables

37. Static Group Comparison Sample Treatment Measure Outcome Variables Measure Outcome Variables Population X

38. Non-equivalent Pre and Post- treatment Control Group Design Sample Treatment Measure Outcome Variables Measure Outcome Variables Population X Measure Baseline Variables Measure Baseline Variables

39. Post-treatment Only Control Group Design Sample Treatment Measure Outcome Variables Measure Outcome Variables Population R

40. Pre and Post-treatment Control Group Design Sample Treatment Measure Outcome Variables Measure Outcome Variables Population R Measure Baseline Variables Measure Baseline Variables