A Guide to CDISC ADaM Standards in 2024

Table of Contents

  1. Introduction
  2. What is CDISC ADaM?
  3. ADaM Datasets and Domains
  4. Conclusion
  5. Frequently Asked Questions

Introduction

In the ever-evolving field of clinical trials, data standardization plays a pivotal role in streamlining data analysis and regulatory submissions. The Clinical Data Interchange Standards Consortium (CDISC) specializes in defining and managing industry standards crucial for data processing in clinical research. Among these standards, the Analysis Data Model (ADaM) stands out for its role in making clinical data analysis-ready. This guide provides a comprehensive overview of CDISC ADaM standards as they stand in 2024, outlining their specifications, types, and applications, particularly in the context of oncology trials.

By the end of this guide, you will have a thorough understanding of ADaM standards, their core principles, and practical applications, enabling you to optimize clinical data analysis processes and ensure compliance with regulatory requirements.

What is CDISC ADaM?

The Analysis Data Model (ADaM) defined by CDISC, is designed specifically for creating datasets and corresponding metadata used in the analysis of clinical trials. ADaM datasets streamline the statistical programming necessary for generating tables, figures, and listings (TLFs) with enhanced efficiency and traceability. This advantage significantly reduces the time required for the approval of regulatory submissions.

ADaM Specifications

ADaM specifications encompass standards for creating both datasets and their metadata. These specifications cover various facets such as variable names, labels, data types, lengths, display formats, controlled terminology, and any necessary derivations or programming notes. Adhering to these guidelines ensures that datasets are ready for statistical analysis, facilitating the evaluation process for regulatory reviewers.

Core Principles of ADaM

ADaM specifications encompass core principles to ensure uniformity and clarity in dataset creation. These principles demand:

  1. Clear alignment with the Statistical Analysis Plan (SAP), TFL shells, and study protocol.
  2. Inclusion of variables required for endpoint analysis based on study needs.
  3. An evolving nature of the document until finalization of TFLs.
  4. Rigorous standards for variable naming, data types, and controlled terminology.
  5. Comprehensive documentation for data processing via Define-XML standards.

ADaM Datasets and Domains

ADaM datasets are derived from Study Data Tabulation Model (SDTM) data, potentially incorporating multiple SDTM datasets into a single ADaM dataset. For example, the ADTTE (time-to-event) dataset might compile data from various SDTM datasets. This integration aids clear documentation of data processing via Define-XML.

Types of ADaM Datasets

ADaM datasets are categorized based on their analysis approach, with standardized structures catering to different analyses such as continuous data values, categorical analyses, and subject-level analyses. Let’s explore these types in more detail:

CDISC ADSL

The ADaM subject-level analysis dataset, ADSL, includes one record per subject with variables pertinent to subject disposition, demographic, baseline characteristics, planned or actual treatment group information, key dates, and randomization information. ADSL serves as a foundational dataset, from which variables may be added to other ADaM domains for analysis output creation or review.

Basic Data Structure (BDS)

The Basic Data Structure (BDS) format is designed for datasets with multiple records per subject and analysis parameter or timepoint. It supports derived analysis parameters when necessary and accommodates analyses such as Last Observation Carried Forward (LOCF) and Worst Observation Carried Forward (WOCF). BDS is particularly useful for continuous value analyses, providing variables for study identifiers, analysis parameter names, codes, and values.

A variant of BDS caters to Time to Event (TTE) analyses, commonly used in fields like oncology, including variables for original risk dates and censoring details.

Occurrence Data Structure (OccDS)

Introduced by CDISC in 2016, the Occurrence Data Structure (OccDS) supports categorical analyses by summarizing frequencies and percentages of occurrences. OccDS uses dictionary coding categories for standardization and accommodates data from event or intervention classes, including exposure data.

ADaM Datasets for Oncology Trials

Oncology ADaM datasets are heavily reliant on RECIST criteria, which standardize how cancer patient responses to treatment are measured. These datasets involve complex data collection and derivations, resulting in the creation of five key ADaM datasets: ADTR, ADRS, ADINTEV, ADEFFSUM, and ADTTE.

  1. ADTR: Tumour Assessment Analysis Data, encompassing valid baseline and post-baseline results from the TR domain.
  2. ADRS: Tumour Timepoint Response Analysis Data, including valid postbaseline results from the RS domain.
  3. ADINTEV: Intermediate Event Analysis Data, capturing intermediate PFS events and relying on ADRS.
  4. ADEFFSUM: Efficacy Summary Analysis Data, deriving categorical endpoints and dependent on ADRS and potentially ADINTEV.
  5. ADTTE: Time-to-event Analysis Data, deriving endpoints like DOR and PFS.

Conclusion

CDISC ADaM standards play a critical role in ensuring clinical trial data is ready for thorough analysis and regulatory submission. By adopting these standards, organizations can achieve higher accuracy, traceability, and efficiency in their data analyses. For those looking to optimize their clinical trial processes, integrating ADaM with statistical programming services is a key step toward improved data quality and faster market access.

Contact Quanticate today for expert guidance and services in automating SDTM dataset production and creating ADaM datasets, ensuring compliance with CDISC standards while saving time and resources.

Frequently Asked Questions

What is the difference between SDTM and ADaM?

SDTM (Study Data Tabulation Model): Focuses on collecting and formatting raw clinical trial data into specific domains, ensuring consistency and facilitating regulatory submissions.

ADaM (Analysis Data Model): Transforms collected data for statistical analyses, supporting various statistical methods while ensuring data clarity and comprehensibility.

What types of analyses can be performed using ADaM datasets?

ADaM datasets support descriptive statistics, regression analysis, subgroup analysis, survival analysis, and more, offering flexibility for diverse analytical needs.

Is it mandatory to use ADaM for regulatory submissions?

While not always mandatory, using ADaM is highly recommended and often expected by regulatory authorities, improving the efficiency of review and approval processes.

What is the ADaM Implementation Guide (IG)?

The ADaM IG provides detailed instructions for creating ADaM datasets, including guidelines on dataset structure, variable naming, and data derivation processes.

Can ADaM datasets be customized?

Yes, ADaM allows for customization to meet specific study needs, provided these adjustments adhere to ADaM core principles and maintain standardization.

For more information on how ADaM standards can enhance your clinical trial data management and regulatory submissions, contact Quanticate today.