Consulting & Collaboration

stat keyboard

What is Statistical Collaboration and Consultation?

Because a statistical easy button doesn't exist, when you decide to collaborate or consult with a biostatistician, you are seeking assistance in selecting and using the most appropriate methods for obtaining and analyzing data as well as disseminating the results related to your data.

Do you need a statistical collaborator or consultant?

Collaboration with a biostatistician provides you someone who has expertise and experience in areas such as study design, outcome measurements, data management, statistical analysis and multiplicity. Biostatisticians can assist you in problem solving. The expertise and knowledge of a biostatistician can supplement your area of expertise and enhance the quality, integrity and validity of your study or project.

Roles of a statistical Collaborator or Consultant

The role a biostatistician plays in a research study depends on the needs and resources related to the project. A biostatistician can be a full collaborator or partner in research. The particular work and responsibilities of the biostatistician are agreed upon on a per project basis. These may vary from involvement in all of the project related activities development as little as serving as a technical advisor available to field questions that arise throughout the course of a project. Regardless of the scope of work agreed upon for the biostatistician, there will be certain activities and responsibilities that the biostatistician takes on as standard practice. The biostatistician has the expertise to effectively evaluate the project and requests made and may also identify issues that are not obvious to you the researcher.

How to Involve a Statistical Collaborator or Consultant Prior to Data Collection?

It is important and is typically most effective to involve the biostatistician from the beginning of a project. The biostatistician can verify that the study design, sample size, etc. are adequate to meet the specific aims of the project. The biostatistician can also provide help related to randomization, blinding procedures, developing intervention strategies, timing of procedures and visits as well as other study design issues related to data collection, sampling, repeated measurements, etc. The biostatistician can also look for issues that may compromise the validity of the project. Involving the biostatistician from the beginning of a project can save time, effort and headaches at later stages of the project.

After The Data Has Been Collected

Although it is preferred to have the biostatistician involved early in a project, a biostatistician can be consulted on a project after the data has been collected. A biostatistician can assist in the selection and implementation of appropriate data analysis methods that are most effective for the type of data produced by your study. However, it may be concluded that the statistical design and data collection were flawed and the biostatistician may not be able to salvage the study. In this case, the biostatistician may able to suggest what information can be extracted from your data.

A biostatistician who is first brought into a project at this stage may re-analyze the data using methods they consider more appropriate than the methods initially used. They will examine your data for threats to validity, ranging from missing data to questionable outliers to confounders.

The First Consultation Visit

In the initial meeting with the biostatistician the scope of the project and needs are established. Often there is no charge for this initial, exploratory meeting, but do not assume this. This initial meeting provides an opportunity for the biostatistician to learn about your needs and this also provides you and opportunity to learn about the biostatistician. To set up an initial consultation regarding your project, complete a project registration form, contact Pre-Award Project Manager, Ms. Megan Parks, (913) 588-4795 or email at mparks2@kumc.edu.

Introducing the Problem

Every issue may not be resolved in the initial meeting, but the researcher should strive to briefly address and discuss all the issues. The problem or needs should be described in the early stages of this meeting. It may not be possible to cover all the details in this initial meeting; however enough information should be provided to define clearly the needs and the basis for estimating an appropriate level of statistical support. The biostatistician will most likely ask a variety of questions to ensure that they fully understand your research and statistical needs. Some of these questions may not seem directly relevant to the biostatistician’s role on the project; however, they are often important in uncovering issues that have statistical implications for the project.

What is Needed for the Initial Consultation?

In addition to registering the project online, most biostatisticians’ appreciate some written form of summary material in advance of the initial meeting. The following are suggestions of what to provide:

  • briefly stated background information about the problem, description of the project, procedures to be evaluated, etc.;
  • a protocol, proposal or scope of the work for reference at the initial consultation; and
  • information about any existing database or dataset to be used.
What do you Expect from the Initial Consultation?

Typically researchers either have a clear idea of what they need from the biostatistician and their role in the project or they are seeking help and suggestions on how to proceed. Even if you believe you know what you need from the biostatistician, keep an open mind to alternatives. Based on the biostatistician’s expertise, they may present you with other recommendations related to the best statistical approach to take. They will most likely also present you the possible impact of not taking their expert advice. Sometimes mutual expectations can be completely addressed at the initial meeting; other times it takes a series of discussions to completely clarify the needs. During the discussion of expectations, the deliverables and constraints for the project should be specified such as: a written report of the analysis and results; creation of databases or datasets; programming; timelines; funding; etc.

Responsibilities of a Statistical Consultant

A biostatistician’s level of involvement in a project should be discussed and agreed upon at the start of the collaboration or consulting relationship. The statistician should explain, in a way that is understandable to the investigator, the statistical concepts and methods and their implication for the project. The biostatistician should also provide practical guidance on implementing these concepts and methods. If the biostatistician is responsible for the analyses, they should be willing to provide details of analysis or database definition, such as data dictionaries for defining variables and which subjects are included in an analysis. For reporting study results, the biostatistician should assist and guide you on what can be claimed and concluded based on the available data. The biostatistician should also explain assumptions related to the methods used and limitations of the findings. Please remember that the biostatistician will work hard on the project, but there are limits to what can be learned from any given study.

Responsibilites of a Client

The biostatistician can serve you better if there is good communication between you and the biostatistician. It is the researcher’s responsibility to ensure that the biostatistician has a clear understanding of the objectives of the project by providing relevant information. Feel free to ask for feedback to see whether the biostatistician understands the scope of the project, etc. Remember that the biostatistician may have a learning curve when it comes to the terminology and key components related to your area of research. The researcher is also responsible for providing a complete and accurate description of how the data were acquired and any problems in data collection, protocol deviations, etc. Missing data and protocol deviations should be disclosed as they may impact the conclusions of the study. Many times the biostatistician can suggest a valid approach for proceeding with analysis despite these issues. Finally, any publications or reports produced as a result of the project should acknowledge the participation of the biostatistician, consistent with the value of that involvement. The biostatistician will expect to be a coauthor on articles on which they put forth significant effort. However, you should not associate the biostatisticians name with the project without their consent.

Ethics in Statistical Consulting

The American Statistical Association and International Statistical Institute have published ethical guidelines for professional statisticians.

Last modified: Jul 01, 2013
Important Info

World of Statistics

ID=x7330