PUBH 581 EXTRA CREDIT - Study Designs 30 points You are going to design a study. Pick an exposure and disease that you are interested in. It could be related to something you are already working on, something that you will be working on in the future, or something you have never worked on that you find interesting. Next, pick a study design that you feel would be appropriate for examining this particular exposure-disease relationship (any design except an ecological study design is appropriate). Then, answer the following questions. When asked why (or to justify your answer) bullet points or brief statements will suffice. 1. What are your exposure and disease/outcome of interest? a. What is your research question? b. What is your hypothesis (state as a null hypothesis)? 2. What study design will you use to study this relationship? a. Why is this design appropriate for studying this relationship? That is, is it the best design to use? Why or why not? 3. What are some possible confounding variables that you will consider in your study? Will you assess these variables in the design or analysis phase? Why? 4. Will you be able to estimate risk in this study design? Why or why not? 5. Can this study design provide evidence of a causal relationship? Why or why not? (Think about temporality and other criteria.) 6. Based on your chosen study design, answer the following questions. a. If you picked a cross-sectional study? i. From what population will you sample, and why? ii. How will you assess/measure the exposure? iii. How will you assess/measure the disease/outcome? iv. What are some of the problems you might experience in this study design? In relationship to finding your sample and in drawing conclusions. b. If you picked a case-control study: i. How will you select your cases? ii. How will you select your controls? iii. How will you measure/assess your exposure and possible confounders? iv. Will you use matching? Why or why not? If so, what variables will you match on? c. If you picked a cohort study: i. How will you pick your cohort? ii. How will you identify/measure/classify the exposed and unexposed? iii. How will you identify disease? What disease-related information would be important for you to collect? iv. How might you handle or reduce people who are lost to follow-up? d. If you picked an intervention study? i. Will you randomize? Why or why not? ii. Will you use any blinding? Why or why not? If so, what level of blinding will be used (i.e. who will be blinded)? iii. What kind of outcome data might you want to collect? iv. How might you handle or reduce non-compliance in your design? 7. What are your inclusion and exclusion criteria? 3 8. What measures of association and/or impact will you calculate?