CHPS Study Design and Biostatistics Core Services

CHPS statistical support is limited to the studies that are planning to utilize or are currently utilizing at least one additional CHPS Clinical Research Cores at CHOP. As an alternative resource for the studies requesting only statistical support, please contact the Biostatistics and Data Management Core (BDMC).

Statistical consultation in the study-planning phase can help address development of testable hypotheses, sample size estimation and justification, and the statistical analysis plan. Data management plays crucial role in streamlining future statistical analyses. If appropriate, consider requesting database development through the Bioinformatics Core to facilitate management of data at the start of the study.

Once approved, investigators can schedule a meeting with the CHPS statistician to review the analysis plan in detail and set up a timeline of the analysis and manuscript submission. The plan should include authorship placement for the CHPS statistician, and authorship should be discussed prior to analyses for a qualified study.

Data analysis will be conducted when the following conditions are met:

• Recruitment completed
• Final data locked for analysis
• A clear and finalized analysis plan
• Authorship agreement
• Appropriate timeline until manuscript submission

Statistical services are considered general CHPS support, and PIs should acknowledge CHPS and the Clinical and Translational Science Award.

If you have questions or need more information about these services, please contact: John Krall, CHPS Administrative Director, 267-425-1790 or krallj@chop.edu.

Investigators requesting statistical consultation should complete and submit an Application for CHPS Statistical Consultation. The application should include a brief background, aims/hypotheses, description of study participants, data collection methods, the proposed primary outcomes, and a preliminary analysis plan. The description must also include a paragraph on the planned use of CHPS resources.

As this initial biostatistical service is designed to provide consultation in the development of a research protocol to collect pilot or preliminary data, it is understood that the information provided will be preliminary in nature.

Pilot studies are the first step in the research process to provide information needed in planning and designing larger scale, hypotheses-driven studies. For example, a pilot study can be used to assess feasibility of subjects' recruitment, ability to measure and capture outcomes, to provide statistical estimation of parameters, and/or to demonstrate the ability to randomize subjects to interventions. Pilot studies usually involve a small sample size (≤ 30 participants). Sample size should be justified but not estimated to meet a specific statistical power.

Pilot studies do not:

• Test specific hypotheses but can generate data for hypotheses development.
• Evaluate treatment safety, efficacy, and/or effectiveness.
• Generate clinically important and meaningful effect sizes due to the small sample size used.

The statistical sections of the protocol should address the following:

Design
Summarize the study design and the appropriate terms such retrospective, prospective, parallel, crossover, randomized, double blind, controlled, comparative, and observational.

Hypothesis
Describe the corresponding outcome measures and restate the primary hypothesis from the Specific Aims section in terms of a testable statistical hypothesis (i.e., explain how the outcome measure is expected to be affected by the components of the study design). Similar descriptions regarding secondary hypotheses are welcome.

Statistical Methods
Specify the planned statistical methods that will be used to analyze the primary and secondary outcome measures. Analysis plans must relate to stated specific aims. Include statistical references if a nonstandard method will be used to analyze the primary hypothesis.

Randomization
Specify how the randomization will be done, especially if it involves blocking or stratification to control for possible confounding factors.

Sample Size
State the proposed sample size and estimate the statistical power related to testing the primary hypothesis described above. Outline how you arrived at this estimate in terms of the primary outcome measure. The number of subjects proposed must be adequate to achieve the aims of the proposed study. If for any reason, a power calculation cannot be performed, a detailed explanation is required.