Skip to main content
Sign In

University of Colorado Denver


Study Design

When to contact us

Contact at the earliest possible time, usually more than 6 weeks prior to any deadline.

What to expect

After a scope of work is developed and signed, be prepared to spend much of a first meeting with the collaborating biostatistician discussing your science before moving onto specific design discussion. You will then move to discussing measurements and design choices (pre-post, parallel arms, observational, etc.). The end product is a set of testable hypotheses mapped to data analysis sections.

Elements of a study design
Defining primary, secondary and exploratory hypotheses:

Although all aspects of your study are important, for valid conclusion drawing it is important to select a primary hypothesis and a primary outcome for that hypothesis. That minimizes the chance that you will have false positive findings.

Defining primary, secondary and exploratory outcomes:

There are often many ways to measure an outcome especially when the outcome is difficult to biologically define. However, to minimize the chance that your findings will be false positives or negative, we will work with you to identify a primary outcome that best answers your primary hypothesis.

How many measurements on an individual:

Is it better to collect more people with one measurement or consider multiple measurements on an individual? It turns out it depends on how correlated measures are on the same individual and the variation of your measurement. We will help you consider the strengths and weaknesses of all applicable designs and how the different designs link with your ability to answer your primary hypothesis.

Confounders and precision variables:

In non-randomized trials, it is important to collect additional measures that might mask (or enhance) any associations between your primary variable of interest and the outcome. These additional measures are often called confounders. We will work with you to incorporate potential confounders in your analysis and help you identify the number of confounders that can reasonably be accounted for given the limitations of your study. Even in randomized trials there is value in collecting additional information on other variables that are known to be associated with the outcome (precision variables). Including these variables in your final analysis can increase power.

Data analysis plan:

We will draft a data analysis plan for each hypothesis and each outcome that is appropriate based on the design selected.

Sample size justification:

See the sample size resource for details on what you need to provide for sample size calculations. As part of design we will either perform calculations to help you determine the sample size necessary to detect clinically meaningful differences or if the sample size is fixed (e.g., secondary data analysis) we will perform calculations to estimate what size of difference (or association) you are powered to detect with the given sample size.

Data, Safety, and Monitoring:

For clinical trials we will help you develop a data, safety and monitoring plan and identify "independent" biostatisticians that might serve on a review panel.


We will help you develop a randomization plan and will create the randomization scheme as necessary. During the design phase we will draft a paragraph defining the plan. During implementation we can implement the randomization if your study statistician needs it to be independent from them.

Colorado Biostatistics Consortium

13001 17th Place | Mail Stop B119 | Room W3129, Building 500 | Aurora, CO 80045
303-724-4370 |

Request Biostatistics | Department of Biostatistics and Informatics | BERD | BBSR | ColoradoSPH

© The Regents of the University of Colorado, a body corporate. All rights reserved.

Accredited by the Higher Learning Commission. All trademarks are registered property of the University. Used by permission only.