Hypothesis development and definition of study objectives
The transformation of research questions into statistically testable hypotheses is an important first step, providing guidance for the design of the study. Our group is experienced in working with investigators to translate their scientific questions into clearly stated aims with hypotheses that can be measured and tested.
Selection of study outcome measures
The outcome measure is an event or change measure that is the target for treatment and addresses the study objective. Outcome measures should be clinically relevant, easy to observe and reliable. The selection of an outcome often requires a balance between cost and precision. Our researchers are skilled in summarizing and weighing the advantages and disadvantages of various outcomes.
Determination of the assessment schedule
While selection of the primary, secondary and exploratory outcomes is an essential step, when and how often these outcomes are assessed is just as critical. Our experts collaborate with investigators to determine the most efficient and effective assessment schedule (i.e. optimal number and timing of outcome assessments) weighing the statistical efficiency against practical (e.g. participant burden) and economic limits.
Determination of the statistical design
Trained and experienced YDCC researchers collaborate with investigators to determine the optimal statistical design, including standard designs (e.g., parallel group, factorial, cluster, crossover, noninferiority, two-stage) as well as newer types of designs such as SMART and MOST. Our researchers provide sample size/power estimation for a variety of designs, outcomes and hypotheses; develop randomization plans and treatment allocation procedures; outline procedures for bias/variance control; and create the statistical analysis plans (SAP), including methods for interim monitoring of the data and reporting to Data and Safety Monitoring Boards (DSMBs). Our faculty and staff are experienced with several free (e.g. R) and commercially available (e.g. PASS, SAS) software packages for the estimation of sample size and are also equipped to perform simulations. We have conducted sample size/power estimation for:
- Continuous/Discrete/Categorical/Censored Outcomes
- Superiority/Non-Inferiority/Equivalence Hypotheses
- Parallel Group/Crossover/Factorial/Cluster Randomized Trials
- Case-Control/Matched Case Control/Cohort studies
- Hierarchichal and Repeated Measures Studies
- Fixed/Sequential Designs
Development of case report forms (CRF)
Well-designed case report forms are critical to the conduct of a study, with typical studies utilizing dozens of forms. As such they require the development by personnel with experience in form construction and familiarity with methods for data collection and processing. Our data management experts understand that appropriate form layout and item construction can help to avoid confusion from study personnel and lead to more accurate and complete data collection.
Design, implementation and maintenance of clinical data management systems
The YDCC offers several solutions for trial management, data collection and storage including local and web-based systems for centralized or remote entry (e.g. entry by study participants). Our programmers also offer custom solutions to transfer data from existing ancillary sources. We provide training of study personnel and tech support throughout the trial for the data management system.
Our data entry specialists are trained in the data collection process specific to each study and follow detailed protocols to avoid entry errors and identify data errors.
Data quality control
Data intake systems are designed to encourage timely and continuous data flow for quality control. Our data management experts work with investigators to develop a tailored approach to quality control that may include automated system edit checks at data entry, standard and custom reports for data edit checks and a process for edit queries, resolution and tracking through audit trails.
The YDCC offers investigators a secure environment to store their data with general precautions and safeguards to protect against loss and unauthorized use. Systems meet requirements of the Code of Federal Regulations (CFR), 21 Part 11.