A b o u t   t h e   P r o g r a m

Installed Sites

National Jewish Health.  Denver, Colorado

Colorado Clinical and Translational Science Institute.  Aurora, Colorado

Harbor-UCLA Medical Center, General Clinical Research Center.  Torrance, California

History of the Project

[10-Feb-2009]  The original version of SDM, which was then called "Study Designer/Manager" was written at the Data Coordinating Center of National Jewish Health in 2002-3 for a large multi-center descriptive study of severe asthma.  That version, 1.0, has been in continuous use for that study since then.

In 2004 we created version 2.0, which lets you use presentation metadata to create forms with sophisticated custom layouts and data checking. It also allows you to import structural metadata from an outside database, which means that you can retrofit SDM for managing that database. Improvements were incremental for several years.  The biggest changes included online discussion for data issues ("queries" in the jargon of clinical trials), free-form notes allowed on any field, more data types, and simplified management of access control for studies involving multiple clinical centers.

In 2006 we licensed a copy of SDM to the General Clinical Research Center (GCRC) at The Children's Hospital of Denver, where it became the standard for data collection of GCRC-approved studies.

In 2007 National Jewish Health developed a way to load data from any SDM study into its institution-wide Research Database.  This database integrates medical record, biobank, and SDM study data into a single resource for sponsored research at National Jewish.

In May of 2008 SDM became the standard data collection and management tool for the CTSA-funded Colorado Clinical and Translational Sciences Institute (CCTSI), which serves researchers at three University of Colorado campuses, The Children's Hospital of Denver, and National Jewish Health.  With this change we have been able to more actively develop and improve SDM, making over 25 enhancements in the first year of the Institute.  There are 7 part-time developers involved, with twice-monthly coordinating meetings, and shared task management, code control and discussion systems to support them.

  • A web server (both IE and Apache have been used).
  • Adobe's ColdFusion, version 7 (8 will be tested soon).  This costs between $1K and $2K, and can be installed on the web server box.
  • SQL Server 2000 or 2005 (2008 will be tested soon)
Functional and Structural Overview

[02-Apr-2009]  Study Design by Metadata (SDM) is a web-enabled system for managing the forms-based acquisition of clinical research data.  The software is driven by metadata describing the data and circumstances of a study.  You use the Study Designer (SD) program to put metadata into a data dictionary that is shared by all studies.  The Study Manager (SM) program then uses study metadata to let you enter, edit, and report on clinical research data.

Data and metadata are maintained in SQL Server databases.  The SD and SM web programs should be used for all access to those databases, other than normal, system-level database administration.  Access to SD and SM (and thus to the data dictionary and study databases) is via password-protected logins.  These logins are controlled by the application itself, and they depend on a separate database that is maintained by a web program called Realm Manager.  SDM works with standard web browsers, using secure data transmission over the internet.  SDM keeps an audit trail recording when changes were made to data and by whom.

The goal of SDM is to allow researchers to manage the data for a clinical research study without the need for custom code development.  SDM is written in the ColdFusion Markup Language (CFML).  In the current version of SDM (version 2.2.135), setting up and operating a study requires using SD to describe the study and its data, creating an empty SQL Server database, and editing a few lines in a CFML configuration file.  Some basic reports are available, but others have to be written in a combination of CFML and SQL, or by using some other reporting tool.  The availability of the data dictionary facilitates writing reports in a more general, re-usable way.

Most of the labor of setting up study data collection is spent entering the metadata using SD.  To make this easier, and to make studies more scientifically comparable, SD allows copying of metadata from one study to another.

SD provides a web form development tool for creating custom data entry screens for various aspects of any study.  You can also use SDM to manage previously existing relational data tables from any other source, providing that they conform to SDM’s CRF data model, explained next.

SDM uses a simple data model in order to make its operation compatible with standard practice in clinical research studies.  The model assumes that each chunk of data will be collected on one of a number of Case Report Forms (CRF’s).  Each different type of CRF is used for a particular part of a study -- almost always documenting the information obtained in a specific encounter with a patient or with a particular laboratory sample from the patient.

The data model, in essence, is this: the data from a single instance of a filled-out CRF becomes a single record in the relational table that corresponds to that type of CRF.   The primary keys of the table are structured so that no duplicate records can occur.  There is no attempt to perform relational normalization of the data across CRF’s.   Some patient encounters can generate a wide variety of data, so for convenience an encounter might be documented with more than one type of CRF.  Multiple instances of a certain type of CRF might be used on a given patient.  For this reason the primary keys of the corresponding data table can include identifiers such as dates or sample numbers, allowing unique identification of possibly multiple records of a study subject.

Developed by: Nationa Jewish Health; Presented by Colorado Clinical and Translational Sciences Institute