Biometrics partner for sponsors and CROs

Statistical Programming

SDTM, ADaM, TLFs and e-submission packages in SAS and R, independently validated and traceable from raw data to every number.

Statistical programming is where a study's data becomes evidence. Data arrives from EDC systems, central labs, safety databases and external vendors, rarely in the same shape. Our programmers bring it into CDISC SDTM, build ADaM analysis datasets with derivations you can follow, and program the tables, listings and figures set out in the analysis plan.

Every output is independently validated, and every number can be traced back to the record it came from. That is what reviewers check, and it is what we build for from the first dataset.

  • Independent validationEvery dataset and output is checked by a second programmer.
  • Traceable to sourceEvery number can be followed back to the record it came from.
  • SAS, R and PythonProgrammed in your environment or ours, to your standards or ours.
  • Beyond submissionThe same validated data answers the questions that follow.

Planning work on Statistical Programming? Tell us what the study needs and we will come back with scope, timeline and cost.

Beyond the first submission

Most of the work does not stop at database lock. Clinical teams ask follow-up questions, data monitoring committees need unblinded outputs, and regulators send health authority questions (HAQs) during review, often with short deadlines. We program those answers quickly and document them to the same standard as the original outputs.

For integrated summaries of safety and efficacy (ISS and ISE), we pool studies with different designs and versions of the standards into one consistent analysis database. After submission, we provide the programming for DSURs and PSURs/PBRERs from the same validated data.

SAS and R

We work in SAS and R, in your environment or ours, and to your programming standards or ours. Reusable macros and functions take repeat work out of every study.

How the work flows

7 steps, each documented and checked. Choose a step to see what happens there.

Step 1 of 7

Raw data

Data from EDC systems, central labs, safety databases and vendors arrives in different shapes.

Everything we do in Statistical Programming

13 activities, each programmed, checked and documented to the same standard.

1

SDTM dataset programming

Raw study data, including external vendor data, is mapped and programmed into SDTM domains from approved specifications.

2

ADaM dataset programming

Analysis datasets are derived from SDTM to follow the statistical analysis plan, with traceability back to source.

3

Tables, listings and figures

The tables, listings and figures for the clinical study report are programmed to the SAP and shells, blinded and unblinded.

4

Independent QC and double programming

A second programmer independently reproduces each dataset and output, and every difference is resolved before delivery.

5

Interim analysis outputs

Datasets and outputs are prepared for planned interim looks while the study blind is kept.

6

DMC and DSMB outputs

Safety and efficacy outputs are programmed for data monitoring committees, with open and closed sessions handled separately.

7

Integrated summaries (ISS and ISE)

Data from several studies is pooled into integrated datasets, and the outputs for the integrated summaries are produced.

8

Submission data package

XPT datasets, define.xml, annotated CRF and reviewer guides are assembled for eCTD Module 5.

9

Legacy data conversion

Data from older or non-standard studies is converted to SDTM and ADaM so it can be pooled or submitted.

10

PK/PD datasets and outputs

Concentration and parameter data is programmed into PK/PD SDTM and ADaM datasets with matching outputs.

11

Ad-hoc, post-hoc and HAQ analyses

Extra analyses are programmed on request, including answers to health authority questions within the deadline.

12

DSUR and PSUR/PBRER outputs

Exposure summaries, tabulations and line listings are programmed for periodic safety reports.

13

Macro and standards libraries

Validated SAS and R macros and standard output templates are built and maintained for reuse across studies.

Where it fits in the study

The stages of every study, with the part Statistical Programming plays in each.

How we work
  1. 1

    Plan

    Protocol and CRF review from a programming view; specifications and shells agreed.

  2. 2

    Build

    SDTM and ADaM datasets, with dry runs on blinded data.

  3. 3

    Deliver

    Final TLFs, ISS and ISE, and the e-submission package.

  4. 4

    After submission

    Programming for HAQs, DSURs, PSURs/PBRERs and ISS/ISE updates.

Standards and tools

  • SDTM
  • ADaM
  • define.xml
  • cSDRG
  • ADRG
  • Controlled terminology
  • SAS
  • R
  • Python
  • Double programming

Talk to us about Statistical Programming

We use your details only to reply. See our privacy notice.

Have a study that needs statistical support?

Tell us about the study and what you need. A statistician will get back to you.