BERD: HOW TO WORK WITH THE BERD CORE
Requesting BERD Support
BERD support is requested through an online form. Investigator and project information from this form will populate fields of the BERD Time and Accomplishment tracking database, where BERD personnel will record project work and product information. BERD members should regularly update the Access BERD Time and Accomplishment database to reflect their project work time and project accomplishments, including abstract and manuscript submissions and publications, grant submissions and awards, and presentations.
1. Requests are reviewed, and a BERD member is assigned within two working days.
2. Grant proposals must be submitted to the BERD for support at least three weeks before the grant submission deadline. Exceptions to this policy are rare.
3. Data cleaning and processing requires a minimum of two weeks.
4. After the data have been cleaned and queries addressed, please allow a minimum of two weeks for descriptive and univariate analyses and a minimum of four weeks for multivariate analyses or analyses involving weighted, hierarchical sampling, or longitudinal data.
1. Chargebacks will be discussed with investigators at the start of each new project. BERD support for some projects may be subsidized, depending on OSCTR and institutional priorities and whether the project has external funding.
2. For grant applications including BERD personnel, BERD personnel must approve the final grant document prior to submission. The application will include salary support for BERD personnel when allowable under the funding mechanism. The percent effort on a grant will be decided with the BERD Director; the minimum is generally 10% FTE for clinical studies and 5% FTE for basic science studies. Additional support may be required for programmer/analyst or project coordinator time. When faculty salary is not allowed under the funding mechanism, contributed effort must be approved by the BERD Director if the OSCTR BERD budget supports the time. In addition, when faculty time is not allowed under the funding mechanism, the grant budget will request design and data analysis support from staff. For one- to two-year pilot projects, $2000 to $5000 per year is typically requested for BERD staff support.
3. Supplemental funds will be requested if an investigator or group requires more than 10% FTE of dedicated time from a given faculty, staff, or graduate research assistant member of the BERD. A Memorandum of Understanding with the investigator or group will be developed to cover the salary and fringe of the BERD member and a 3% computer/server overhead charge. Typically, such funds will offset the additional salary of the BERD member without requiring a reduction in the BERD funding for that member.
4. Graduate research assistant time for data entry is limited to 20 hours of support funded by the BERD for a given project. Efforts beyond the initial 20 hours will be funded through supplemental arrangements to cover the salary and fringe of the student.
At the initial consultation meeting, BERD members will develop and communicate with the project investigator within two weeks of the initial meeting a project timeline, a summary of the scope of work, and a listing of deliverables related to the project. Project planning and report templates are available for BERD members to standardize communications with investigators.
Please bring the following to the initial consultation:
1. Statement of research questions, including primary and secondary objectives. Primary objectives will eventually determine the research design, data measurement methods and protocol, and statistical strategy, along with sample size, effect size, or statistical power calculations.
2. Details regarding the primary research question.
Outcome of interest
Level of measurement (nominal, ordinal, interval, ratio)
Assessment of measurement error and misclassification
3. Design issues related to the primary question.
Comparison groups, based on intervention, based on interest in stratification by important cofactors, covariates, or confounders
Hierarchies, clusters, or correlations among the outcome measures. Examples include repeated measures over time (i.e., measures taken each week), spatially correlated data (i.e., measures from left and right eye), and natural groupings (i.e., families, clinics, etc.).
Plans to measure continuous or categorical covariates that will permit: Exploration of subgroup associations (statistical interaction or effect modification); Adjustment for confounding of the primary association of interest by factors known or suspected to be associated with the outcome from previous literature; and Assessment of measurement error and misclassification
4. Relevant preliminary data or published papers for calculating sample size, effect size, and power. These should provide estimates of:
The magnitude of the outcome of interest (such as group means or proportions)
The magnitude of the variability you expect in your outcome (usually standard deviation)
5. Details regarding the participants in your study.
Inclusion criteria to help define the population(s) of interest from which group(s) will be selected or sampled
Exclusion criteria that, by restricting the sample, limit the influence of potential confounding variables or known sources of variation
Sampling plans (random, stratified, paired)
Delivery or Sharing of Study Data
1. The study protocol and IRB approval must accompany the data.
2. A data dictionary should accompany data and should include the following:
Units of measurement for each continuous variable
Definitions for levels of categorical variables with underlying continuous measures
Definitions of codes for nominal categorical variables like race, ethnicity, county, etc.
3. Data must be de‐identified so that it has no direct identifiers, such as:
Street name, street address, or post office box (i.e., not including city, state, or ZIP code)
Telephone and fax numbers
Social security number
Vehicle identifiers and serial numbers
URLs and IP addresses
Full‐face photos and other comparable images
Medical record numbers, health plan beneficiary numbers, and other account numbers
Device identifiers and serial numbers
Biometric identifiers, including finger and voice prints
4. If data is not de‐identified, you must first Zip the file(s) and then send them via the link: https://securefiletransfer.ouhsc.edu/. Visit the following link for more information and directions: http://it.ouhsc.edu/services/SecureFileTransfer.asp.