Development For Standardizing Clinical Data Health And Social Care Essay

ABSTRACT

Modern clinical research requires that biopharmaceutical companies, medical device companies, and their partners easily be able to exchange and review high-quality clinical trial data. We require clinical research standards across the industry to optimize the drug development lifecycle and improve the regulatory review process. Data standards are a critical component in the quest to improve global public health. Inefficiencies in the collection, processing and analysis of patient and health-related information drive up the cost of drug development for life sciences companies and negatively affect the cost and quality of health care delivery for patients and consumers.The solution is to make use of data standards, such as those from CDISC (Clinical Data Interchange Standards Consortium), to provide more efficient and effective use of medical information by all members of the health care and life sciences ecosystem.This paper focuses on the need to standardize clinical research data and the steps taken to convert raw clinical data (SDTM-) to SDTM+ and further converted to ADaM dataset for analysis and generate final report.

KEYWORDS: CDISC, SDTM, ADaM, CDASH, CRF, Clinical Trial, Macro, Domain, Observations, Variables, SAS, Randomization, Adverse Event, Treatment Arms, Inclusion Criteria, Exclusion Criteria, ADSL, BDS, Traceability

INTRODUCTION

The objective of this project is to develop and support global, platform-independent data standards that enable information system interoperability to improve medical research and related areas of healthcare.The purpose is to take raw clinical data (i.e., SDTM-) and convert it to SDTM+ or SDTM (Study Data Translation Model) using some macros and then convert it to ADaM (Analysis Data Model) again using some macros based on the SDTM and ADaM requirements given by the customer. Finally a report is generated in the form of TLF (Tables/Lists/Figures). The flow is shown below:

TLF

First we need to gather the SDTM and ADaM requirements on the Case report Form (CRF) write some generic code to convert SDTM- to SDTM+ based on the customer requirements. These generic codes are known as macros. Macros allow writing a piece of code once and using it over and over. These macros are written in SAS (Statistical Analysis System).

SDTM is a data standard developed by CDISC, describing the contents and structure of data collected during a clinical trial. Purpose is to submit Case Report Tabulation (CRT) data to a regulatory agency such as FDA (Food and Drug Administration), in a standardize format.

Next we need to write macros to convert the SDTM to ADaM based on ADaM requirements provided by the customer. ADaM includes all the variables that are enough to generate any output and hence is used for further analysis of the standard clinical data generated in SDTM. We may require integrating various domains in SDTM+ to generate ADaM output, therefore SDTM act as input for ADaM.

CDISC has recommended using SDTM data structure to store source data in a standard format and the ADaM data structure to store data for statistical analysis and reporting.

CDISC (Clinical Data Interchange Standard Consortium) is a global, open, multidisciplinary, non-profit organization that has established standards to support the acquisition, exchange, submission and archive of clinical research data and metadata. CDISC standards, implementations and innovations can improve the time/cost/quality ratio of medical research, to speed the development of safer and more effective medical products and enable a learning healthcare system.

CDISC uses various standards:

CDASH (Clinical Data Acquisition Standards Harmonization) used for data collection. It is a content standard for the data elements that should be captured on a Case Report Form (CRF).

SDTM (Study Data Tabulation Model) used for data tabulation. The SDTM is a CDISC content standard that describes the core variables and domains (like DM, AE, CM, CO etc) to be used as a standardized submission dataset format for the Food and Drug Administration (FDA). Define.xml holds the metadata for these datasets.

ADaM (Analysis Data Model) used for statistical analysis. The Analysis Dataset Model specifies data structures and associated metadata for analysis datasets.

CLINICAL TRIAL

Clinical Trials are sets of tests in medical research and drug development that generate safety and efficacy data (or more specifically, information about adverse drug reactions and adverse effects of other treatments) for health interventions (e.g. - drugs, diagnostics, devices, therapy protocols).

It is a rigorously controlled test of a new drug or a new invasive medical device on human subjects; in US it is conducted under the direction of the FDA before being made available for general clinical use.

In clinical trial the patients are assigned to groups that receive different treatments. The process of assigning patients to these groups by chance is called randomization. At the end of clinical trial researchers compare the groups to see which treatment is more effective or has fewer side effects. Randomization, in which people are assigned to groups by chance alone, helps prevent bias.

Each choice a patient has for treatment is called an Arm. Treatment arm includes group of patients receiving a certain type of therapy or treatment.

We need some standards to be adopted for exchange of these clinical trial data since these trials can be multicenter or single centered and in both cases we need to exchange this data in some standard format.

Multicenter Research Trial is a clinical trial conducted at more than one medical center or clinic whereas Single Center Clinical Trial is initiated by one researcher that is only available at one center.

The subjects for clinical trial are chosen on the basis of some criteria. For e.g. - age, sex, type, stage of a disease, treatment history and other medical conditions.

Inclusion Criteria are a set of conditions that must be met in order to participate in a clinical trial.

Exclusion Criteria are the standards used to determine whether a person may or may not be allowed to participate in a clinical trial.

Phases of Clinical Trial:

There are 5 phases in clinical trial:

Phase 0: Pharmacodynamics and Pharmacokinetics

Phase 1: Screening for Safety

Phase 2: Establishing the testing protocol

Phase 3: Final Testing

Phase 4: Post approval studies

Phase 0 trials are the first in-human trials, single sub therapeutic doses of the study drug are given to a small number of subjects (10-15) to gather preliminary data on the agent's pharmacodynamics (what the drug does to the body) and pharmacokinetics (what the body does to the drugs).

In Phase 1 trials, researchers test an experimental drug or treatment in a small group of people (20-80) for the first time to evaluate its safety dosage range, and identify side effects.

In Phase 2 trials, the experimental treatment is given to a larger group of people (100-300) to see if it is effective and to further evaluate its safety.

In Phase 3 trials, the treatment is given to a larger groups of people (1000-3000) to confirm its effectiveness, monitor side effects, compare it to commonly used treatments, and collect information that will allow it to be used safely.

In Phase 4 trials, post marketing studies delineate additional information, including the treatment's risk, benefits and optimal use.

These trials may result into any untoward medical occurrence in a patient or clinical investigation subject administered a pharmaceutical product and which does not necessarily have a casual relationship with this treatment.

Such unfavorable and unintended sign, symptom or disease temporally associated with the use of medicinal product, whether or not related to the medical product is called Adverse Event (AE).

To avoid such events on actual subjects (patients) pre-clinical trial is performed before starting these phases. These pre-clinical trials are done on animals and after approving this pre-clinical trial, the actual trail starts on a human being.

FUNDAMENTALS OF SDTM

SDTM is a data standard developed by CDISC, describing the contents and structure of data collected during a clinical trial.

The purpose is to submit Case Report Tabulation (CRT) data to a regulatory agency such as the FDA, in a standardized format that:

- provides clear description of the structure, attributes and contents of each dataset and variables submitted

- Compatible with available software tools that allow efficient access and correct interpretation of the data submitted.

SDTM does not specify what data to collect the sponsor decides this based on science and regulation

SDTM Model - Building Blocks

The SDTM provides a general framework for describing the organization of information collected during Clinical studies.

• OBSERVATIONS:

Observations normally correspond to rows in a dataset.

• VARIABLES:

Variables, normally correspond to a column in a dataset; can be classified according to its Role.

• DOMAINS:

Observations are reported in a series of Domains, usually corresponding to data that were collected together as a dataset.

• CLASSES:

Data domains are grouped into classes based on the nature of observations collected in each domain.

- Interventions Class: Investigational treatments, therapeutic treatments, and procedures administered to or taken by the subject. One record per constant dosing/treatment interval

- Events Class: Occurrences or incidents independent of planned study evaluations occurring during the trial or prior to the trial. One record per event

- Findings Class: General subject observations such as questions and tests. Observations resulting from planned evaluations. One record per finding result or measurement

Class

Domain

Variables

Observation

ADaM FUNDAMENTALS

Observations consist of discrete pieces of information collected during a study. Observations normally correspond to rows in a dataset. A collection of observations on a particular topic is considered a Domain (e.g.- Demographic (DM), Adverse Event (AE), Concomitant (CM), Comment (CO), TA (Trial Arms), Medical History (MH), ECG Tests (EG), Laboratory Tests (LB) Vital Signs (VS), etc).

An observation example:

"Subject 123 had a severe headache starting on study day 2"

CDISC categorizes variables into five roles:

- Identifier: identify the study, subject of the observation, the sequence number

- Topic: specify the focus of the observation (such as the name of the lab test)

- Timing: describe the timing of the observation (Visit, Start/End date, Days, Time Points, Duration)

- Qualifier: additional text or numeric values

- Rule: express an algorithm or method to define start, end or looping conditions in the Trial Design model

In the above example, 123 is a Unique Identifier of the subject which is a required variable, Severe is a Record Qualifier which is a permissible variable, Headache is a Topic Variable (required) and Day 2 is a Timing Variable (permissible).

FUNDAMENTALS OF ADaM

Analysis Data Model (ADaM) consists of two data structure:

Subject-level analysis dataset (ADSL): It consists of one record per subject and is used to provide the variables that describe attributes of a subject. It is similar to DM domain in SDTM.

Basic Data Structure (BDS): it consists of one or more record per subject, per analysis parameter, per analysis time point. It supports parametric and non-parametric analysis. We may require integrating various SDTM domains to generate this structure.

General ADaM Definitions:

Analysis-enabling: Required for analysis. A row or column is analysis-enabling if it is required to perform the analysis.

Traceability: Property that enables the understanding of the data's lineage and/or the relationship between an element and its predecessors facilitates transparency. This permits the understanding of relationship between the analysis results, the analysis datasets and the SDTM domains.

Supportive: It enables traceability. A column or row is supportive if it is not required in order to perform an analysis but is included in order to perform analysis and facilitate traceability.

Record : Row in a dataset

Variable : Column in a dataset

A CDISC compliant submission includes both SDTM and ADaM datasets; therefore, it follows that the relationship between SDTM and ADaM must be clear. This highlights the importance of traceability between the input data (SDTM) and the analyzed data (ADaM).

SDTM vs. ADaM

In SDTM structure, variables and variable names are pre-specified but ADaM can include sponsor-defined derived variables and observations. It can include SDTM variables and observations. It may also include replication of core variables and other variables that are needed to replicate the analysis with minimal programming.

SDTM does not allow any imputations and mostly include textual data to facilitate clinical review whereas ADaM allows imputations and can include numeric values when needed for statistical programming.

METHODOLOGY

The entire work will be performed based on the Software Development Life Cycle Methodology.

The requirements for SDTM and ADaM are gathered from the customer in CRF which will follow the CDASH standards. This may include any row data in the form of SDTM-. We are required to convert this data to a standard storage structure i.e., SDTM+. Read the SDTM+ specification and then determine each variable's standard name, label, length and type. Then analyze the data and study the requirements and perform any pre-processing if required. Now based on user requirement we need to develop generic macros in SAS that will convert each raw data to specified standards. Each SDTM domain is separately defined and standardized.

Final testing is performed on each domain and design document for each domain is created separately as deliverables.

Now we consider SDTM+ as our input data to generate ADaM datasets as required by the customer. ADaM is used for analyzing the data derived to generate various clinical results. To standardize this data again we write generic codes (macros) in SAS that will convert SDTM standard data to ADaM standards that will be used for further analysis. For example, suppose SDTM standard format for date is DD/MM/YYYY and ADaM requirement is 7-January-2013, then we are required to generate macros that will perform this conversion.

ADaM must include all the variables that are enough to generate any output. Therefore we may require integrating various SDTM domains together to derive ADaM domain variables, which was not possible in case of SDTM derivation. SDTM only contains variables defined in SDTM-.

Finally, ADaM standards are tested and final report is generated in the form of TLF.

TESTING TECHNOLOGIES

Black Box Testing or Functional testing of each SDTM domain requirements which focuses solely on the outputs generated in response to selected inputs and execution conditions.

Unit testing of macros: Functionality of each macro is tested corresponding to the standards specified in the SDTM and ADaM requirements.

System Testing to make sure that the software satisfies its specification and to detect implementation faults.

Compliance Testing to check whether system is developed in accordance with standards and procedures.

LIMITATIONS

Limitation of establishing traceability: Establishing traceability in ADaM datasets is not an easy task. It requires lot of effort and overhead like extra SAS code, creation of intermediate datasets and large analysis datasets with additional variables and records. Even though this increases size of analysis datasets and complexity of programs it is strongly recommended by ADaM implementation guide to include as much supporting data as necessary to build traceability.

Limitations of SAS:

Short numeric and mixed type variables: In order for SAS data sets to be accessed across architectures, they should not include two-byte numeric variables. This length is allowed on IBM mainframe machines, but other hosts on which SAS software runs have a minimum numeric variable length of three. As a result, a data set that contains a two-byte numeric cannot be accessed across architectures from other types of hosts.

Catalog Access: When the user and server run cross-architecture, access to several types of data that is normally available through SAS/SHARE software is not supported, including catalog access.

CONCLUSION

The entire scope of work will be performed by qualifier personnel and based on the Software Development Life Cycle Methodology. Contractor shall perform consulting services for the customer by performing statistical programming activities to create re-usable SAS template programs for generating SDTM+ datasets (SDTM+ template programs) and safety ADaM datasets (ADS template programs).The generic template program would replace the manual creation of the SDTM+ and ADaM datasets and shall be applicable to several clinical trial studies falling in the general definition specified by the customer. The Program will be written using SAS Software and based of ADaM and SDTM+ template specification supplied and maintained by the customer.