The Role and Responsibilities of Biostatisticians in Clinical Trials

The Role and Responsibilities of Biostatisticians in Clinical Trials

John Kittelson, PhD

Biostatisticians have primary responsibility for the scientific integrity of a clinical trial. Clinical trials are the epitome of multidisciplinary research, and in this collaborative endeavor the biostatistician provides leadership in the methods of science and is responsible for the integrity of the clinical trial as a scientific experiment. Some of specific responsibilities of the clinical trialist biostatistician include:

Trial planning:

·  Scientific design including:

·  Specification and prioritization of trial questions/hypotheses

·  Selection and complete definition of trial endpoints

·  Identification and specification of the target population and the study population

·  Selection and definition of study treatments (active and control) including:

§  Specification of treatment duration

§  Specification of ancillary treatments

§  Specification of follow-up (rescue) treatments

§  Excluded therapies

·  Statistical design

·  Refinement of clinical hypotheses into testable scientific and statistical hypotheses:

§  Superiority, non-inferiority, equivalence, and bioequivalence designs

§  Context: phase I/II/III

·  Refinement of trial endpoints into precisely defined statistical variables; surrogate versus clinical endpoints

·  Identification of primary analysis

·  Sample size evaluation including power, hypotheses discriminated, design resolution.

·  Protocol preparation

·  Preparation of the full trial protocol (not the same as an NIH grant proposal)

·  Complete specification of each of the above elements.

·  Specification of trial procedures, measurement methods, follow-up visits, adverse event monitoring,

·  Randomization and blinding procedures; stratification and blocking factors.

·  Outline statistical analysis plan

·  Trial monitoring procedures and interim statistical analyses

·  Quality control procedures.

·  Recruitment and retention plans

·  Planning for missing data

·  Prepare statistical analysis plan (formal written document) including:

·  Data set construction, preliminary evaluation of data

·  Description of study process (screened, enrolled, randomized, retained, analyzed)

·  Description of participants (“table 1”)

·  Primary analysis

·  Secondary analyses

·  Interpretation of multiple analyses (e.g., confirmatory, co-primary, mechanistic)

·  Subgroup analyses

·  Robustness evaluation

·  Approaches to handling missing data

·  Analyses of adverse events, safety endpoints

·  Table and figure templates

·  Prepare interim statistical analysis plan (formal written document) including:

·  Ongoing safety evaluation procedures

·  Interim analysis plans of primary (and secondary?) outcome data

·  Rationale for early termination

·  Specification and evaluation of interim decision rules

·  “Adaptive” elements to the design

·  Steering committee participation

Trial initiation:

·  Structure and content of CRF’s

·  Endpoint adjudication systems

·  Measurement procedures

·  Design of data collection and data management systems

·  Design and structure of trial informatics system; data transfer, data security

·  Trial quality control systems including:

·  Routine reporting

·  Adverse event reporting

·  Troubleshooting scientific integrity

·  Procedures for routine evaluation of data quality control

·  Enrollment and randomization procedures

·  Assuring quality of all trial elements across multiple centers

·  Establish and train trial staff

·  Drug tracking procedures; assuring proper treatment labeling, prescription, and distribution.

·  Procedures for assuring adequacy of blinding

·  Recruitment and retention planning

Trial conduct:

·  Manual of procedures

·  Ongoing quality control:

·  Actual accrual compared with target rate

·  Site quality monitoring

·  Data completeness; range checks; consistency checks

·  Construction and QC of analysis data sets

·  DSMB management; report preparation

·  Troubleshooting:

·  Identify problems before (or immediately after) they develop

·  Inconsistent or unanticipated results that merit DSMB consideration

·  Avoid “blind stupidity”

·  Planning for trial closeout

·  Updating protocol, SAP, ISAP, manual of procedures

·  Steering committee participation

Trial close-out:

·  Pre-planning:

·  Confirm SAP, data QC, construction of analysis data sets

·  “Dry-run” of data analysis using dummy data (generate trial report)

·  Database lock

·  Data cleaning process

·  Final data set construction

Analysis of results:

·  Report generation; implementation of SAP

·  Unanticipated statistical analyses (minimize)

Interpretation of results, manuscript preparation:

·  Interpretation should follow SAP

·  Preparation of manuscript tables and figures

·  Writing and reviewing