gms | German Medical Science

GMDS 2015: 60. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS)

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie

06.09. - 09.09.2015, Krefeld

Development and implementation of the open source electronic randomization system RANDI2 and its connectivity to the electronic data capture system (EDCS) OpenClinica

Meeting Abstract

Suche in Medline nach

  • Daniel Schrimpf - DKFZ, Heidelberg, Deutschland
  • Lothar Pilz - Universität Heidelberg - Medizinischen Fakultät Mannheim, Mannheim, Deutschland

GMDS 2015. 60. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e.V. (GMDS). Krefeld, 06.-09.09.2015. Düsseldorf: German Medical Science GMS Publishing House; 2015. DocAbstr. 223

doi: 10.3205/15gmds057, urn:nbn:de:0183-15gmds0571

Veröffentlicht: 27. August 2015

© 2015 Schrimpf et al.
Dieser Artikel ist ein Open-Access-Artikel und steht unter den Lizenzbedingungen der Creative Commons Attribution 4.0 License (Namensnennung). Lizenz-Angaben siehe http://creativecommons.org/licenses/by/4.0/.


Gliederung

Text

Introduction: In evidence based medicine clinical studies are the gold standard in introducing new treatments to show efficacy and clinical benefit. If possible it should be randomized and an experimental treatment should be tested against a placebo or the established standard. In some medical fields the expected measurable benefit could be small or there are several new treatments such that adaptive randomization is an opportunity to accelerate approval. The aim of the presented development is the implementation of the open source electronic randomization system RANDI2 and its connectivity to the EDCS OpenClinica supporting the (adaptive) randomization, data management, and entry in medical studies.

In this field the main used data interchange and communication standards are based on the extensible markup language (XML) like the standards published by the „Clinical Data Interchange Standards Consortium“ (CDISC). For the easy accessibility of applications web standards as HTML (hypertext markup language – structure of a web-page), CSS (Cascading Style Sheets – for layout definitions), JavaScript (a dynamic scripting language supported by most of the browsers), and for the server connectivity the standard SOAP are often employed.

Methods: For the software development the following programming languages and tools have been used: The object oriented language JAVA, the object and functional language Scala, Liquibase to handle database updates/versioning, Slick for the database access, the Lift framework to generate web application, and JUnit/ScalaTest to create automated software tests. Legal standards/regulations like good clinical practice (GCP-V) and European and FDA requirements are obeyed.

The profile of requirements for the randomization/data capture system should cover amongst others the following features for adaptive randomization methods:

1.
Implementation of adaptive randomization systems in a simple way;
2.
Simulation of selected randomization techniques;
3.
Accelerated connection of randomization and data capture.

Results: Generally, the following realizations for combining randomization and electronic data capture are possible:

1.
Two separate systems for randomization and data capture with communication by the user itself, implicating that costs for installation, management, maintenance, and education can almost be doubling despite the fact that data must be captured twice, which is an avoidable source of error.
2.
The direct integration of the randomization system in the EDCS minimizing the disadvantages of the first realization but showing less flexibility in the randomization part as in pseudo-random number generation, probabilistic distributions, and in-numero experiments.
3.
The combination of both systems via defined interfaces with direct communication between both systems having the advantage that each application can be developed separately, except the interface definition, which results in a higher flexibility. The necessary interfaces for the automated communication between the systems are defined.

Best flexibility offers version three showing the opportunity to optimize tasks in single systems. The choice of RANDI2 for the platform of (adaptive) randomization and OpenClinica for the management of clinical data was mainly for their open source characteristics. The data interchange is performed with CDISC ODM messages over SOAP-web-services:

a) Generating a list of relevant studies;
b) Meta-data of the selected study;
c) Patients’ lists of this study;
d) Issuing the necessary patient information;
e) Documentation of the randomization result.

The process is secured by the authentication of users and in assigning user roles.

The used system RANDI2 has the advantage that it is flexible to react on different requirements of the clinical study. All main classical randomization methods are implemented and new procedures can be easily incorporated by a plug-in mechanism.

The key interfaces of the chosen open source EDCSOpenClinica are detailed analyzed opening the communication in study information, patients’ lists, study events (for event driven adapted randomization), and the documentation of randomization results. New is the development of an interface to export saved patient data; for this purpose the source code of OpenClinica had to be changed in a none interfering way. With this approach it was possible to introduce an EDC-component that makes it possible to communicate with OpenClinica and access all the necessary clinical data. As an implication it is now possible to implement and start a new study in the user interface of RANDI2, where even aspects of single study centers can be treated. The EDC-menu in RANDI2 compromises entries for the management of EDC-studies, the outlook and processing of study centers and the start of the randomization process. With this interface realization of OpenClinica and RANDI2 it is now possible to implement adaptive procedures more easily, to simulate the randomization by given parameters, and to guarantee a high quality standard.

As an example the new system is evaluated by using the closed lung cancer study GemTax III with the comparison of two chemotherapies using study data of 339 patients. As in the original study a block randomization was applied with a balanced outcome in the two therapy arms, in each of the study centers (stratification parameter), as well as in the age distribution. Some comparisons of the RANDI2/OpenClinica realization are shown in detail in our paper [1] with the result of a better data handling and management, and an essential flexibility of our approach.

Discussion: The randomization process and basis documentation of patients’ data are pivotal in conducting a modern clinical study. Adaptive randomization plays a prominent role for a portion in oncological studies as shown, e.g. by the International Association for the Study of Lung Cancer [2]. For individualized therapy the adaptive design of new studies becomes now more common [3]. But it must be mentioned that not in every case adaptive procedures are the first choice [4]. Thus, RANDI2 in connection with OpenClinica opens the possibility making more profound decisions by simulating various study parameters. Additionally data handling and management is easier and more reliable.

The presented elaboration using the open source electronic randomization system RANDI2 and the newly incorporated interfaces to OpenClinica [5], [1] establishes the conduction of a variety of modern clinical studies with defined communication, the plug-in of new (adaptive) randomization systems and an easy to handle web-based user interface.


References

1.
Schrimpf D, Haag M, Pilz L. Possible combinations of electronic data capture and randomization systems: Principles and the realization with RANDI2 and OpenClinica. Methods Inf Med. 2014;53(3):202-207.
2.
Session PC01: Clinical Trial Design for Drug Development. 15th World Conference on Lung Cancer (WCLC); Sydney, Australia; 2013 October. p. 27-30.
3.
Berry DA, Herbst RS and Rubin EH. Reports from the 2010 Clinical and Translational Cancer Research Think Tank Meeting: Design Strategies for Personalized Therapy Trials. Clin Cancer Res. 2012;18:638-644.
4.
Freidlin B, Korn EL. Adaptive Randomization Versus Interim Monitoring. J Clin Oncol. 2013;31(7):969-970.
5.
Schrimpf D. Elektronische Randomisation einschließlich der Anbindung an eCRF’s in onkologischen Studien [dissertation]. Medical Informatics: University of Heidelberg; 2015.