• Statistical Guidelines for Protocols

    The following outline includes statistical recommendations for the preparation of protocols. These guidelines assume investigators plan to conduct complete epidemiological and observational studies or clinical trials. Some parts of the guidelines may not be applicable to specific studies, such as a feasibility study or a pilot study.

    1. Study design
    2. Describe subject characteristics:
    3. Include entry and exclusion criteria such as age, sex, race and significant medical conditions.
    4. Indicate how many subjects will be screened and how many will be expected to meet the entry criteria.
    5. Include the variables to be measured and the timing of these measurements.
    6. Sampling design
    7. Indicate whether the study is observational or experimental (interventional) and retrospective, cross-sectional or longitudinal.
    8. For an interventional study, indicate the number of treatment groups and the number of placebo groups, the method of randomization and the duration of the study.
    9. Identify outcome measures (i.e., primary and secondary endpoints) and factors that potentially affect the outcome.
    10. Show the anticipated level of compliance and loss to follow-up, and the strategies to retain subjects, and indicate how compliance will be monitored and optimized.
    11. Data quality control and database management
    12. Describe reliability and validity of the measurements, such as the inter- and intra-laboratory reliability or precision for the measurements; the inter- and intra-interviewer reliability; and validity of the questionnaires. These factors are a necessary aspect of calculating the power or necessary sample size for the study.
    13. Describe the flow of data entry and database management.
    14. Describe methods for data entry and management.
    15. Describe the mechanism of checking and editing the data.
    16. Describe computer data security and subject confidentiality.
    17. Ensure the database program is compatible with the statistical packages proposed for data analysis.
    18. Hypothesis testing
    19. List the specific hypotheses to be tested by the research.
    20. Explain how testing each hypothesis will help achieve the specific aims of the study.
    21. Identify variables relevant to each hypothesis.
    22. Data analysis
    23. Describe subject characteristics, such as mean, standard deviation, maximum and minimum of the outcome variables and covariates.
    24. Describe how data will be handled for non-normally distributed outcome.
    25. Explain how the proposed statistical analysis will test the hypothesis.
    26. Include the timeline for interim and final data analyses and safety monitoring.
    27. Sample size and power calculation
    28. Describe how sample size and power are derived
    29. Describe sample size specific to a hypothesis.
    30. Describe the formula for the calculation and give the source of information for the sample size estimate (i.e., the variability estimates from published reports, previous research by the applicant, a guess).
    31. Discuss how the anticipated (hoped for) difference was determined.
    32. Base the sample size on a two-sided test (if one-sided test is used, explain why).
    33. Applicability of the sample size to all hypotheses
    34. If there are several hypotheses, indicate the one used for sample size calculation.
    35. Explain why this specific hypothesis was chosen.
    36. Describe how the proposed sample size is appropriate to test all the hypotheses.

    How to Reach Us
    For more information, contact Amy Hartkemeyer, business director, 513-636-4273, amy.hartkemeyer@cchmc.org.