If you have data you are planning to analyze, below are some questions to help guide your appraisal of the potential impact of the study and the possible sources of bias.
- What is the question to be answered? What is the current state of knowledge on this question? What is the public health, biological, or clinical significance of this question?
- What type of epidemiologic strategies are you planning to use to test your hypothesis?
- What is the study population? How was it selected? How will this impact generalizability?
- Based on your study design, sample selection, and follow-up, is there any possibility of selection bias? What factors may be contributing to this bias?
- What tissue are you planning to interrogate? Is this the tissue of interest or a surrogate tissue? How is this tissue relevant to your research question? If you are studying a heterogeneous tissue, how do you anticipate cellular heterogeneity to impact your results?
- Considering how the samples were stored and processed, what are some potential batch effects? Were these batch effects measured? (e.g. do you have bisulfite conversion plate, microarray chip, etc. for each sample)
- How are you defining the exposure? How do you get the information? Do you anticipate measurement error and any bias resulting from that error?
- How are you defining the outcome? How do you get that information? Do you anticipate measurement error and any bias resulting from that error?
- Based on subject matter knowledge, what are some potential confounders of your association? How were these variables measured? Are there any potential unmeasured confounders? How do anticipate any unmeasured confounders will impact the estimated association?
- Are there any important potential effect modifiers of the association of interest?
- Do you plan to validate and/or verify your findings? If you have a validation cohort, how does the study population differ from your cohort? How might these differences impact your ability to validate your findings?
- What types of conclusions would you draw from the study? Would the message most likely be for researchers, health care providers, or policy makers?