Request Collaboration

Investigators seeking statistical support or other collaborations are encouraged to contact the SDBC as early as possible in the design of new research activities. Before making a request, please make sure you have done the following:

1. Defined a research question (see Study Design Resources)
2. Performed a literature review
3. Junior investigators – please make sure you have discussed the project with your research mentor
4. Please make sure your research mentor is able to attend the initial meeting with the SDBC collaborators

Timelines for Productive Collaboration

The following time requirements are essential for SDBC members to work effectively with you. Please note that more complex projects may require more time.

Why Are the Above Timelines Important?

The timelines are important because engaging SDBC members with the initial study design ensures that study hypotheses are clear and testable. Furthermore, effective data analysis often requires the following steps:

1. Development of pre-specified analysis plans
2. Structuring data for analysis
3. Investigation of data quality, missing data, distributions of variables, outliers and other data issues
4. Conduct of pre-specified analyses
5. Documentation of results and interpretation
6. Follow-up analyses to address issues raised by the initial analyses

The quality of the data greatly impacts our timeline. However, even a well-formatted data set can require 6-8 weeks to complete the analysis, particularly if many follow-up analyses are requested. Examples of potentially time-consuming projects include those that involve: a) multiple data sets with many variables, b) repeat measures over time (longitudinal) or within subjects (clustered), and c) specialized statistical analyses. Please keep in mind that SDBC members are always juggling multiple projects, and thus we may not be able to accommodate tight deadlines.

Strategies for expediting your analysis include well-formatted data sets and investing more time up front in the development of a comprehensive statistical analysis plan.

Request for Collaboration

Please complete this form to the best of your capabilities as it helps us determine which SDBC team member(s) will be best for your project, track the progress of your request/project and will make your first meeting more productive. A consultant will contact you within 1 to 5 business days following submission of the request form. When submitting your form, it is recommended that you:

Describe your research question or goal in detail. Consider including the following information:

• Specific aims (potentially in draft form)
• Subjects you intend to study
• Primary outcome variable and how it will be measured
• Predictor variable(s) and how they will be measured
• Any confounding variables or covariates that you have thought about

For those who are considering development of Health Measurement Instruments, it is helpful to include:

• A conceptual description of the variable(s) you want to measure
• Any ideas you have about how they might be measured (include either qualitative, quantitative or both)
• The type of measurement study you intend: either traditional psychometric or computerized adaptive testing

For those who are considering surveys or questionnaires, it is helpful to include:

• The population being surveyed
• The purpose of the survey
• Any instruments or surveys identified
• If you have drafted a survey, please include it

First Meeting With a Consultant

After submitting a request for collaboration an SDBC team member will contact you to schedule a meeting with your assigned SDBC collaborator.  This meeting typically lasts an hour and is used to get details and discuss the issues of your specific research. Depending on your specific type of study or project issues, at your first meeting with an SDBC consultant, you may be asked to provide the following information either at or shortly after the first meeting:

1. Data

We highly encourage initiating collaboration with the SDBC prior to any data collection to ensure the best quality and efficiency of research possible.  We can help develop a data dictionary and define variables that need to be collected in order to most effectively accomplish your research aims.  Please provide any work that has already been done in defining variables or data collection methodology.
If data has already been collected, we will need the data in a clean usable format. Please refer to the data collection & analysis section of this website for details.

2. Sample size calculation

If a sample size estimation is needed, we will need the following information: (It’s okay if the only information available is an “educated guess”)

1). Outcome Variable Type:

• For a continuous variable (such as height or weight), we need: (bring in relevant literature)
  
  • The average value and the standard deviation of the control group;
• For a dichotomous variable (such as: yes/no response), we need:
  
  • The average proportion of the control group or the population

2). The estimated difference between the control and the treatment you would like to detect (effect size).

Both of these are often derived from similar studies that are in the literature AND/OR through pilot studies.

3. Existing Questionnaire/Questionnaire Development

We highly recommend that any “homegrown” questionnaires are reviewed by one of our SDBC consultants BEFORE the start of the study. Though seemingly trivial, proper questionnaire development is critical and often complicated. Time and labor can be saved if you seek our input prior to administering your questionnaire.

4. Health Measurement/Instrument Development

• Bring in any similar studies
• Identify any relevant theoretical perspectives
• If possible, describe the design of the study

5. Health Care Utilization/Cost Data

Interested in cost, charge, or cost-effectiveness?

If relevant, define:

• Total health care cost and utilization
• Multiple years of data or single year data
• Relevant previous studies
• Is the data at the patient or encounter level?
• Sources of cost and effectiveness information (for cost-effectiveness analysis only)