RCDRC Guidelines for Clinical Trial Design

Applications to the RCDRC should address the following principles of study design. The guidelines are written primarily for clinical trials but most of the issues pertain also to observational studies including case-control and cohort studies. Biometry Core biostatisticians are available to consult with investigators on study design (contact Dr. Brian Leroux at 543-7304 or leroux@biostat.washington.edu). Seeking advice in the early planning stages is recommended. A short list of references is also provided below for additional reading on clinical trial design.

A. Study Objective and Hypotheses

Give a clear and concise statement of the research question(s) (eg, does treatment X have an effect on variable Y in population Z compared with placebo?). State the type of comparison to be made, ie, whether the goal is to establish superiority of one treatment over another or non-inferiority of one treatment with respect to another. State one or more specific null hypotheses based on the research questions (eg, mean levels of variable Y in treatment and control groups will not be different xx months after treatment) and the corresponding alternative hypotheses. The context of the study within an overall research plan should be given, eg, is this an initial exploratory study or a confirmatory trial to test a predefined hypothesis?

B. Outcome Variables

Give one or more primary outcome variables. These should be the variables most relevant to the study objective. Sample size calculations should be based on the primary outcome variables. If more than one primary outcome variable is used, explain why no single primary outcome variable could be selected. If appropriate, define the secondary outcome variables which are measurements made to provide supporting evidence for the research question or to provide information on treatment mechanisms.

C. Study Population

Define the specific population from which study subjects will be selected and give the mechanism for selection (eg, random sampling, clinic population, recruitment through advertisements). State clearly defined inclusion and exclusion criteria.

D. Treatment Assignment

State the method of assigning treatments to subjects (or experimental units such as teeth within subjects). This method should be selected to minimize bias due to selective assignment of certain treatments to particular subsets of subjects based on demographic, prognostic, or other information. Ideally, randomization will be used for assigning treatments. Randomization involves the deliberate use of an element of chance such as flipping a coin, computer generated random numbers, or tables of random numbers. The procedures for generating random treatment assignments should be explicitly defined, including blocking or stratification as appropriate.

Randomization tends to guard against bias due to selective assignment of treatments and/or imbalances in prognostic factors, and also serves as the fundamental basis for the statistical analysis of the results. If randomization is not proposed, carefully justify why it is not feasible and what procedures will be used to guard against bias.

E. Blinding

State the type of blinding that will be used for subjects, clinicians, and evaluators. Ideally, double blinding will be used, in which the clinicians and evaluators, as well as the subjects, do not know the identity of the treatment used. Blinding is used to minimize bias of clinicians or evaluators and the placebo effect. If blinding is not to be used state why it is not feasible and describe the procedures that will be used to guard against bias.

F. Study Design

State the specific type of design to be used. This should be either a parallel group design or a within-subject design. In a parallel group design, the group to which a subject is assigned determines the interventions received. In a within-subject design each subject receives more than one of the treatments and comparisons of treatments can be made within subjects. Examples of within-subject designs are split-mouth designs in which different treatments are assigned to different parts of the mouth, cross-over designs in which each subject receives two or more treatments in succession.

G. Sample Size

Give a justification of the sample size to be used. Ideally, the sample size is selected based on calculation of the power of the statistical tests of the primary hypotheses. The sample size will depend on the the magnitude of the treatment effect, the level of the type I and type II errors. Depending on the type of statistical analysis to be used (comparison of means, proportions, event rates, etc), the sample size may also depend on the variability of the primary outcome variables or the baseline disease prevalence or incidence rate in the population. If the sample size is not based on calculation of power, explain why this is not feasible, eg, because the study is exploratory in nature or because sufficient pilot data is not available for power calculations.

H. Data Collection

Describe the procedures for collection, including timing of all subject visits, personnel for performing data collection, and instruments/questionnaires to be used in data collection. Describe also procedures for managing the data, including computer entry of the data.

I. Statistical Analysis

Describe the statistical analysis methods to be used, paying particular attention to methods for testing the primary hypotheses. Give plans for evaluating baseline comparability of treatment groups and any statistical assumptions necessary. Analyses of secondary outcome variables should also be described.

References:

1. International Conference on Harmonisation; Draft Guideline on Statistical Principles for Clinical Trials. U.S. Food and Drug Administration, 1997.
http://www.fda.gov/cder/guidance/index.htm

2. Statistical guidance for clinical trials of non-diagnostic medical devices. Division of Biostatistics, Center for Devices and Radiological Health, U.S. Food and Drug Administration, 1996. http://www.fda.gov/cdrh/manual/statgde.html

3. Page RC, Armitage GC, DeRouen TA, Genco RJ, Hujoel P, Jeffcoat MK, Kornman KS, Williams RC: Design and Conduct of Clinical Trials of Products Designed for the Prevention, Diagnosis and Therapy of Periodontitis. The American Academy of Periodontology, 1995.

4. Friedman LM, Furberg CD, DeMets DL: Fundamentals of clinical trials (3rd ed.). St. Louis: Mosby-Year Book, 1996.

5. Pocock SJ: Clinical trials: a practical approach. Chichester: Wiley, 1983.