Role of a Statistical Programmer
A statistical programmer is someone who works in the statistics department of a company that does research and development for medicines or for a contract research organization (CRO).
A statistical programmer's job is to analyse the data from clinical trials.
They organise data and create reports, make tables, figures, and listings to show the results of the clinical trials. Then these reports are sent to the sponsors and FDA for further analysis.
Drug development process
The drug development process is a long and expensive process. It costs hundreds of millions of dollars and takes up to a decade or more to complete. The drug development process involves several levels of studies that are conducted to ensure the safety and efficacy of the drug.
This process begins with pre-clinical studies conducted in the laboratory and with animals. If these studies are successful then the drug maker usually pursues an Investigational New Drug (IND) application. This allows the drug maker to conduct clinical trials of the new compound on humans.
Phase 1 trials are the first studies of a new drug in humans, usually carried out on small samples of subjects. These studies aim to determine the safety of the drug in a small and usually healthy volunteer study population.
Phase 2 trials explore the efficacy of a drug, with larger study populations aimed at narrowing the dose range for the new medication. Safety is also monitored at this stage, and phase 2 trials are generally conducted in the target study population.
Phase 3 trials are large-scale clinical trials on populations numbering in the hundreds to thousands of patients. These are the critical trials that the drug maker runs to show that its new drug is both safe and efficacious in the target study population. If the phase 3 trials are successful, they will form the keystone elements of a New Drug Application (NDA).
Finally, phase 4 trials, or post-marketing trials, are usually conducted to monitor the long-term safety of a new drug after the drug is already available to consumers.
Trial Study Design
There are different types of clinical trials, but some important things to understand are randomization and blinding.
Randomization means that patients are assigned to different treatments randomly so that there is no bias.
Blinding means that the patients don't know which treatment they are getting, and sometimes the doctors and researchers don't know either. This is to make sure that the study is fair and unbiased.
Pharma regulation and standards
The pharmaceutical industry is heavily regulated by various governing bodies and authorities, which affects the work of statistical programmers. Understanding the following regulations, guidance, and standards of organizations is essential:
- International Conference on Harmonization (ICH): The ICH is a non-profit organization that collaborates with regulatory authorities in the United States, Europe, and Japan to develop common regulatory guidance. The goal is to define a universal set of regulations so that a pharmaceutical regulatory application in one country can be used in another. The FDA usually adopts the guidelines developed by the ICH over time, so keeping an eye on their development can give an indication of what FDA requirements may be forthcoming.
- Clinical Data Interchange Standards Consortium (CDISC): The CDISC is a non-profit organization that defines clinical data standards for the pharmaceutical industry. CDISC has created several data models that statistical programmers should familiarize themselves with. Four of the most important models include:
- Study Data Tabulation Model (SDTM): The SDTM defines the data tabulation data sets that are to be submitted to the FDA as part of a regulatory submission. The FDA has endorsed the SDTM in its Electronic Common Technical Document (eCTD) guidance. The SDTM was originally designed to simplify the production of case report tabulations (CRTs) and is listing-friendly, but not necessarily friendly for creating statistical summaries and analyses.
- Analysis Dataset Models (ADaM): The CDISC ADaM team defines data set definition guidance for the analysis data structures, which are designed for creating statistical summaries and analysis.
- Operational Data Model (ODM): The ODM is an XML-based data model that enables the XML-based transmission of any data involved in the conduct of clinical trials. SAS has provided support for importing and exporting ODM files via the CDISC procedure and the XML LIBNAME engine.
- Case Report Tabulation Data Definition Specification (Define.xml): Define.xml is the upcoming replacement for the data definition file (define.pdf) sent to the FDA with electronic submissions. Define.xml is based on the CDISC ODM model and aims to provide a machine-readable version of define.pdf. The metadata about the submission data sets can be easily read by computer applications, allowing the FDA to work more easily with the data submitted to it.
- Food and Drug Administration (FDA) Regulation and Guidance: The FDA is responsible for ensuring the safety and effectiveness of drugs, biologics, and devices marketed in the United States. Any work that contributes to submission to the FDA is covered by federal regulations. Specific regulations and guidance that statistical programmers must know include:
- 21 CFR – Part 11 Electronic Records; Electronic Signatures: This federal law regulates the submission of electronic records and electronic signatures to the FDA. Of particular interest to statistical programmers are the requirements for validating systems to ensure accuracy and reliability, determining that persons have the necessary education, training, and experience to perform their assigned tasks, implementing adequate controls over the distribution of, access to, and use of documentation, and implementing revision and change control procedures for SAS programming.
- E3 Structure and Content of Clinical Study Reports: This guidance describes what reporting goes into a clinical study report for an FDA submission. Statistical programmers are often required to generate tables, figures, case report tabulations, and clinical narrative support for the clinical study report.
- E9 Statistical Principles for Clinical Trials: This guidance discusses the statistical issues in the design and conduct of a clinical trial, including trial design, conduct, data analysis, and reporting. Although most useful to statisticians, this guide gives an excellent overview of how a clinical trial should be conducted.
- E6 Good Clinical Practice: Consolidated Guidance: The GCPs discuss the overall standards for implementing a clinical trial
Teams that closely work with a Statistical Programmer
There are many groups that work together with the statistics department, below are the main groups/teams you will get a chance to interact with:
- Site Management: This group recruits doctors for clinical trials, trains them in trial conduct, and monitors their compliance with the trial's protocol.
- Data Management: This group is responsible for designing case report forms, setting up the database, entering and cleaning the data, and providing the data to the statistics group for analysis. A well-cleaned and well-coded clinical database saves time and makes the statistical programmer's job easier.
- Information Technology: The IT group is responsible for maintaining computer systems infrastructure, software development, and support.
- Project Management: The project manager provides operational oversight in a clinical trial. They work with the various functional departments, including the statistical programmer, to meet the trial needs.
- Quality Assurance: The QA group helps ensure that operations meet regulatory standards. They can assist in interpreting regulations and help prepare for audits. They also maintain all company standard operating procedures.
- Medical Writing: This group assists in creating various documents, including clinical study reports, NDA submissions, and safety reports. They can help to ensure the consistency and accuracy of statistical analysis before it is sent to authorities.
