Collaborating With BEAR
Seattle Children’s Biostatistics, Epidemiology and Analytics in Research (BEAR) Core is committed to providing statistical collaboration with investigators at Seattle Children's and beyond to support pediatric research.
Funding BEAR Support and Collaboration
There are three main ways to fund BEAR support:
- Grant FTE
- Program or department FTE
- Hourly rates are billed monthly for smaller projects and may be used for projects without grant funding or internal grants.
Please email us if you have questions, or to inquire about current hourly rates.
Services Subsidized by the Research Institute
- Initial consult with the BEAR team (up to 2 hours)
- Grant preparation, for grants where BEAR services are written into the grant (depending on grant complexity, 10-20 hours)
- Access to and building cohorts in CDR (Clinical Data Repository)
Request Lead Time
BEAR requires different amounts of lead time depending on the type of deliverable requested, e.g. grant application, abstract, manuscript or meeting presentation. This lead time is necessary both to arrange for team member availability and to allow sufficient work time.
The times shown below are minimum amounts of advance notice. It is advisable to allow more time.
Please be aware that BEAR may not be able to accommodate last minute requests.
Note: "Budget" refers to a signed budget estimate with LAN and ready-to-analyze data
- Internally-funded grants
Intake Form: 6 weeks
- Externally-funded grants
Intake form: 10 weeks
- Abstracts Intake form:
Budget: 8-12 weeks
- Presentation slides/poster
Intake form: 14-16 weeks
Budget: 12-14 weeks
Intake form: 4-6 months
Budget: 3-5 months
Levels of Support
When requesting BEAR services, these are some general guidelines to keep in mind regarding levels of support and scope of work.
Very short-term, ad-hoc support
- Advising PI about study design or choice of statistical methods to use
- Brief consultation on database design
- Helping PI respond to reviewer comments on manuscript that biostatistician has not helped prepare
- Brief follow-up to previous collaborative effort, e.g., re-run analyses with additional data, help with abstract or poster presentation
Ongoing but very limited support
- Statistical oversight: Supervise a colleague who is FTE-funded to carry out analysis
- Statistical oversight of junior investigator (e.g. for NIH K- awards, post-doctoral and training grants)
- Limited database support
- DSMB membership
- May include regular billable meeting
Ongoing but limited support
- Assist experienced PI with very small-scale projects such as pilot studies, phase 1 trials, modest proof-of-concept animal studies; help with study design, provide some limited-scope statistical analysis and limited manuscript support
- Regular, frequent meetings and discussions with an investigator or team for an ongoing research program; mentoring/oversight that does not include conducting analyses. Mentoring may take additional time (e.g., to reproduce analyses or meet with investigators)
- Note: If it becomes clear that biostastician will have do most of the data cleaning and manipulation including performing complex programming to create analytic dataset, the scope will change and level of effort would have to be revised accordingly and escalated to a higher level.
Ongoing support for simple research projects
- Involves support for projects that require straightforward analyses and do not involve extensive data preparation or mentoring. PI provides clean and well-documented dataset, in a format that is easily imported into a statistical package for analysis.
Comprehensive support for average-size research projects
This level of support is suitable for straightforward projects that require uncomplicated data management/manipulation or statistical support, such as:
- Routine study design and analysis using off-the-shelf procedures available in standard statistical software
- Projects in which investigator provides clean and well-documented primary datasets for statistical analysis
- Comprehensive support for publications and other deliverables including 1–2 manuscripts, abstracts, work-in-progress reports, posters, etc.
- Database support and data monitoring
Comprehensive support for large or complex projects
This level of support involves high-level involvement of the statistician in the development and implementation of the research project and communication of study results. Responsibilities may include:
- Development and/or implementation of complex study designs
- Assembly of datasets from large, complex or poorly documented sources (e.g., administrative or survey databases)
- Planning and/or implementation of interim data analyses
- Analysis and coordination of analyses for multi-site projects
- Multiple imputation of missing data
- Developing algorithms to identify units of analysis and define analysis variables
- Analysis of data from trials with unique clinical trial design that require a high level of sophistication in developing biostatistical solutions
- Application of specialized statistical methods that may include:
- Application of statistical methodology new to biostatistician, which requires extensive time to research literature and obtain corresponding software if available, or develop software to implement the method
- Application of methodology new to statistical community (i.e., recently published), which may not have corresponding software to implement it, thus necessitating considerable effort to develop it
- Development of new analytic methodology because an appropriate one doesn’t currently exist in the literature
- Active participation in publications, with opportunity for first authored papers
- Project-related travel and external presentations
- Database design, testing and data quality control
Preparing for Your Consultation Appointment
For study design consultations, it is helpful to bring background materials such as a draft of the grant application, Institutional Review Board (IRB) application or study protocol. Copies of key publications related to your topic can also be useful.
If you will be requesting a statistical power or sample size calculation for a randomized trial, your consultant will need information about the expected outcomes in the control group and the magnitude of effect that you would like the study to be able to detect. Most commonly, estimates of mean and standard deviation of the primary outcome measurement for each group are adequate. Studies for which the primary outcome measurement is reported as a proportion do not require a standard deviation estimate. Maximum recruitment potential is also a key factor in evaluating sample size options, so information on the availability of eligible subjects at study sites, budgetary constraints or logistical limitations is also useful.
For projects needing statistical support for data analysis, please bring copies of case report form templates, data coding keys, electronic file templates and/or the actual data file(s).