Skip to main content
Clinical and Translational Science Center

Clinical and Translational Science Center

BIOSTATISTICS

Biostatistics is an integral and important component of research projects and clinical trials. The Biostatistics program provides biostatistical services to UC Davis investigators undertaking clinical and translational research.

Request Services

Primary Contact:
Sandra Taylor, Ph.D.
Senior Statistician
sltaylor@ucdavis.edu
916-703-9171

Services We Offer

The Biostatistics Group assists researchers with all sizes and types of projects, from simple data analyses to large, multi-center clinical trials. Specific services we can provide include:

  • Grant proposal preparation
  • Study design/sample size calculation
  • Statistical analysis plan
  • Data analysis and interpretation
  • Manuscript review and preparation

How to Get Help

Weekly Biostatistics Office Hours

Receive input from a statistician about the design, analysis or presentation of your medical research in a friendly, informal setting.

Request Services

Request Services

Investigators can request more in-depth biostatistics assistance through the CTSC, and one of our statisticians will be assigned to work with you individually. Investigators needing support with proposal preparation or data analysis should request services at least 4-6 weeks in advance of the due date of the proposal or abstract/paper submittal.

Financial Considerations

A detailed “Guidelines for Estimating Biostatistician Effort and Resources on Grants” is available as a PDF.

Costs and Charges

The CTSC Biostatistics and Study Design program is supported in part by funding from the CTSA NIH grant and the UC Davis School of Medicine. This funding allows us to provide some services without charge. Grant proposal preparation and analysis of preliminary data in support of grant applications receive highest priority for unfunded efforts. Grant proposals that require statistical assistance for preparation usually will require statistical services to accomplish the project goals; we will assist with preparing the budget to provide an appropriate level of statistical support. Additional information on estimating the level of effort for biostatistics required on grants is provided below.

For other biostatistical assistance (e.g., data analysis), unfunded support is limited to 10 hours. For this level of effort, we can provide advice on appropriate statistical methods for data analysis, assist with interpreting the results, review draft manuscripts, assist with response to reviewer comments and consult on general statistical questions. Usually, this level of effort is not sufficient for us to be able to actually conduct statistical analyses except for simple, straight-forward projects. A “clean”, well-documented and structured data set is required if we conduct the statistical analyses. Detailed information on preparing your data for analysis is provided under the “Statistical Resources” heading. Beyond 10 hours, our standard recharge rate of $89/hour will be charged. All projects supported by extramural funding must provide salary cost recovery for professional statistical services at a level commensurate with the available funding and level of effort.

Why Include Biostatistics on Grants?

Successful investigators recognize the benefits of including statisticians as collaborators on their research projects. Statisticians help ensure that the study’s design is powerfully and cost-efficiently matched to study objectives, that necessary data to address those study objectives are collected and that appropriate statistical analyses are conducted and reported. These conditions are vital for successful research. In addition, biostatisticians have expertise in translating scientific hypotheses into actionable analyses and they also function as intellectual brokers who are exposed to, and contribute, innovative methods from a wide range of disciplines. Hence, they are well-poised to make substantial improvements in a proposal’s scientific aims and methods, both by identifying and correcting weaknesses and by spotting opportunities to apply and develop promising innovations from other fields. Early statistical collaboration thus leads to much better chances for a research proposal to be funded and for that research to lead to sustained and significant impacts. Within the headings that follow is guidance regarding the percent effort and the level of funding to allocate for biostatisticians on research projects. It is based on the collective experience of faculty and staff biostatisticians in the Division of Biostatistics. These guidelines should serve as a starting point for budget discussions. We strongly recommend that biostatisticians be actively involved throughout the grant proposal development and submission process, including the specification of research objectives and approach and proposal writing, as well as in budgetary decisions about biostatistician and programmer FTE, computer and software purchases, and scientific travel.

Effort Allocation Guidelines

In general, funding for faculty and staff should not fall below 10% of total effort per statistician per time period on a single project. Although there occasionally are valid reasons for a lower level of effort on particular projects, intervals with funded effort falling below 10% require approval by the division chief for faculty and by the CTSC biostatistics director for CTSC staff. Funding should be matched to the size, scope and complexity of the data analysis and study design. Key determinants include the number of primary and derived study variables that will be collected and analyzed, the quality and completeness of the data to be supplied for analysis, and the complexity of the programming necessary to assemble input data and implement descriptive and analytical statistical methods.
  1. Large or complex projectsTotal biostatistics FTE 50–100+%, such as 20% or more of Ph.D. biostatistician plus 30–100% of an MS biostatistician

    High level of involvement in the development and implementation of the research project and communication of study results, which may take many forms, including:
    1. Development and/or implementation of complex study designs
    2. Assembly of datasets from large, complex or poorly documented administrative or survey databases
    3. Development and/or implementation of interim data analyses during data collection phase of prospective studies
    4. Coordination of analyses for multi-site projects.
    5. Development of and/or use and interpretation of novel or complex statistical methods
    6. Active participation in publications, with opportunity for first authored papers
  2. Regular Projects: Total biostatistics FTE 30–65%, such as 10–15% Ph.D. biostatistician plus 20–50% of an MS biostatistician.

    This effort profile is suitable for straightforward projects with uncomplicated analyses.
    1. Active participation in publications, with opportunity for first authored papers
    2. Routine study design and analysis, e.g., analyses carried out using off-the-shelf procedures available in statistical software packages
    3. Involvement in study design, implementation and data collection
  3.  Limited Scope ProjectsTotal biostatistics FTE 20–35%, such as 10–15% of PhD Biostatistician plus 10-20% of an MS biostatistician.
    1. Ongoing occasional consultations with PI about choice of statistical methods to use. This FTE level is typically too low for a Ph.D.-level biostatistician to carry out analyses
    2. This FTE level may be too low to support attendance at weekly or biweekly project meetings by the Ph.D. biostatistician.
    3. This level of effort commitment and support for the Ph.D. biostatistician is generally not compatible with smooth workflows and readily available consultation support, unless an experienced and capable M.S. biostatistician is supported on the project as well.
For multi-year projects, effort commitments may vary throughout the study timeline, according to the needs in various phases, including randomization schemes for sampling and experimental assignment (early), the development and implementation of data and safety monitoring plans (during the middle phases of prospective studies) and the implementation of statistical analyses and communication of study results (later).

Other Budget Considerations

  • In general, biostatisticians help to develop proposals without compensation, including such aspects as calculation of samples sizes, analysis of preliminary data, and writing of statistical sections of grants, since it is assumed that the major biostatistical effort on the project will be via allocated funded effort post-award.
  • There are some grant mechanisms that do not support funded effort by biostatisticians; this may include some K awards. In this case, the PI should discuss the proposal with the division chief or CTSC biostatistics director.
  • Any changes in percent support made during proposal writing or after research has been funded must be made jointly between the PI and biostatisticians. If biostatistics percent efforts are reduced due to budget cuts, then the work of the biostatistician must also be reduced, and this requires consultation between the PI and the biostatisticians. Simply reducing sample sizes does not reduce the amount of work to do the analysis.
  • Sometimes additional resources are required, such as computer hardware, specifically required software and travel. Please discuss this in advance.
  • Letters of support are usually not needed if the letter writer is included as a funded investigator in the proposal, though we are happy to write one if it would be helpful. If the statistical support for the project will come from one of the CTSC faculty or statistical staff, and a letter is desired from CTSC Biostatistics program director, then we will be happy to provide one. With some exceptions, we usually will not write a letter of support if we do not have a defined role in the project. In general, biostatistical letters of support without accompanying biostatistical investigator funding is not a convincing component of a grant proposal. If the project needs statistics, a method of providing it should be in the proposal. If it does not, then a letter of support is usually not needed.
  • Please notify us about the funding decision on the proposal, whether positive or negative. If the project is funded at a reduced amount, it may be appropriate to reduce the statistical support commensurately; it is rarely justifiable to reduce support by a much larger fraction than that. On rare occasions, investigators will re-budget upon the grant being awarded to eliminate statistical support. As one can imagine, this will considerably reduce our enthusiasm for future collaboration.

Manuscript Information

Authorship

Authorship of a manuscript for biostatisticians (and others) is appropriate when the biostatistician has made a contribution to the work described in the manuscript and when the biostatistician has contributed to the manuscript, including writing a statistical methods section, providing statistical results, or reviewing it for correctness. The appropriateness of authorship is not related to whether the biostatistician has been compensated, just as it is not related to authorship for principal investigators or other co-investigators. See the following link for further information. Please note that omission of an author who meets the requirements of this policy is considered to be a violation of standard research ethics.

 International Committee of Medical Journal Editors
“Recommendations for the Conduct, Reporting, Editing, andPublication of Scholarly Work in Medical Journals” 

 

Citing the Grant

Investigators are required to cite the CTSC grant in publications that result from studies that received support from any CTSC resource.  Publications that have cited the CTSC grant must also be made publicly available on
PubMed Central and have a valid PubMed Central ID (PMCID).

Further details are found on the following page: 

http://www.ucdmc.ucdavis.edu/ctsc/audience/investigators/ctsc_publication_citing.html

Other Resources

Data Preparation

A clean, suitably-structured, and well-documented data set is critical for efficient and accurate statistical analysis. Most commonly, data is imported into statistical analysis programs as a comma delimited text file. For easy and accurate importation of data into statistical software, it is essential that the data adhere to a regular structure with consistent entries.

While it is not required, using REDCap (Research Electronic Data Capture) can greatly simplify data collection and minimize costly and time-consuming data clean-up activities. REDCap is a secure web-based application for building and managing online databases for research and is supported by the CTSC Biomedical Informatics team. Regardless of the software used to record data, adhering to the following guidelines will facilitate importation of the data into statistical software. In addition, every data set must include a data dictionary that describes each variable and identifies acceptable values. Additional information on data dictionaries is available on the UC Davis REDCap website.

Additional tips for data management are available in the PDF document, “Guidance for Database Developers for Efficient Import to Statistical Software.” 

Statistical Analysis Software

Interactive Statistical Calculation Pages – Comprehensive list of sites for many statistical analyses, including power and sample size calculations. The website has a page listing websites for interactive analyses (“Interactive Stats”) and for free software (“Free Software”) packages that can be downloaded and run on your local computer. This website also has links to many technical resources on statistics, including general introductory material.

The R Project for Statistical Computing. R is a free statistical programming language that can be used for any and all statistical analyses. It is commonly used by CTSC statisticians. The down side is that it is a programming language and hence has a bit of a learning curve. However, the R Project site contains many documents to help users learn how to use R and many other resources are available online detailing how to conduct specific types of analyses.

MINITAB. Minitab is a commercial, easy to use statistical package with a drop-down menu interface. You can download a 30-day trial version for free.

SPSS, SAS, and JMP can be obtained at a reasonable cost through UC Davis Information and Educational Technology.

Power/Sample Size Calculations

Southwest Oncology Group Statistical Center Power and Sample Size Calculators. On-line sample size/power calculators for one and two sample tests of means and proportions as well as for simple survival analyses

G*Power: Statistical Power Analyses for Windows and Mac. Freely downloadable software that is easy to use with a detailed and helpful user manual. Wide range of statistical procedures are supported including common mean and proportion tests as well as multiple linear regression, logistic regression and poisson regression.

Russ Length’s Power and Sample Size Calculators. Comprehensive site for conducting power/sample size calculations on-line. Functions available for wide diversity of study designs.

Centre for Clinical Trials Power and sample size tools for one and two sample tests of means, proportions and survival data. Sample size calculators for tests of equality, non-inferiority/superiority and non-equivalence. Also has cross-over designs and Phase II clinical trial calculators. Examples provided for each situation. Companion text "Sample Size Calculations In Clinical Research" is freely downloadable from UC Davis library website. The site also has calculators for confidence intervals for proportions, correlation, relative risk, odds ratios, and diagnostic tests, and will perform McNemar’s test for paired binary data

Educational Resources

Introduction to Clinical Research for Residents - This online course consists of readings compiled by the UC Davis and CTSC Biostatisticians.

The Little Handbook of Statistical Practice - Nice, relevant overviews of common statistical analyses are presented. Gives applied examples and interesting discussion of various topics relevant to applied data analysis.

UCLA’s Institute for Digital Research and Education – A wealth of information on conducting statistical analyses using SAS, R, SPSS, Stata, and Mplus is available from this site. The content includes examples of different types of analyses by explaining a motivating data set, providing code to analyze the data in one of the statistical packages, and reviewing and interpreting the output.

Clinical Research Case Studies - CTSPedia entries on selected clinical research topics, including step-by-step tutorials on common sample size calculations; handling outliers; dealing with selection bias in observational studies; and others.

Biostatistics for Non-statisticians – on-line video series from the University of Colorado CTSI.

Ohio State University CCTS – Papers on topics in study design and planning, power and sample size, statistical analysis.

Columbia University Irving Institute – List of references for study design and biostatistics for clinical trials.

University of Utah CCTS Seminar Series – Powerpoint presentations on diverse biostatistical topics including exact statistical inference, multiple imputation, mixed effect models, generalized linear models, survival analysis, epidemiology, and Bayesian methods.

Medical College of Wisconsin – YouTube seminar series with seminars on longitudinal analysis, survival analysis, propensity scores, Bayesian statistics, linear regression, sample size calculations, ANOVA, multiple comparisons, logistic regression among others.

Past Seminars

Did you know...

CTSC biostatisticians

  • collectively provide more than 100 years of statistical expertise.
  • have helped secure $170 million in grant funding since 2011.
  • coauthored nearly 50 peer-reviewed publications in 2013.
  • are available on both the Sacramento and Davis campuses.