902
User Centred Networked Health Care A. Moen et al. (Eds.) IOS Press, 2011 © 2011 European Federation for Medical Informatics. All rights reserved. doi:10.3233/978-1-60750-806-9-902
The Single Source Architecture x4T to Connect Medical Documentation and Clinical Research Philipp DZIUBALLEa,1, Christian FORSTERb, Bernhard BREILa, Volker THIEMANNa, Fleur FRITZa, Jens LECHTENBÖRGERb, Gottfried VOSSENb, Martin DUGASa a Department of Medical Informatics, University of Münster, Germany b Department of Information Systems, University of Münster, Germany
Abstract. Clinical trials often require large and redundant documentation efforts, because information systems in patient care and research are separated. In two clinical trials we have assessed the number of study items available in the clinical information system for re-use in clinical research. We have analysed common standards such as HL7, IHE RFD and CDISC ODM, regulatory constraints and the documentation process. Based on this analysis we have designed and implemented an architecture for an integrated clinical trial documentation workflow. Key aspects are the re-use of existing medical routine data and the integration into current documentation workflows. Keywords. Clinical information systems, EHR re-use, single source, system architecture, clinical data management system
1. Introduction Clinical trials require extensive documentation efforts as they often include hundreds to thousands of attributes per patient. These data are commonly captured twice in two independent information systems. Daily routine documentation is entered into a clinical information system (CIS). Study documentation occurs on Case Report Forms (CRFs) and is stored in dedicated research databases (Clinical Data Management System, CDMS). Double data entry entails occupation of time and impacts negatively on the documentation behaviour of physicians and nurses. Recent studies show that clinicians spend about a quarter to one third of their daily working time for routine documentation [1, 2] and study documentation is added on top. Many patients have at least a basic electronic medical record [3] and routine data is available in CIS, eligible for re-use [4, 5] in clinical research. Although re-use applies to 11% - 69% of data items [6, 7], trial documentation processes connecting patient care with clinical research rarely exist. To address this issue, the eSoure Data Interchange (eSDI) Initiative of the Clinical Data Interchange Standards Consortium (CDISC) has promoted the eSDI Document [8] to analyse the use of electronic technology in the context of eSource data 1
Corresponding author: Philipp Dziuballe, Institute of Medical Informatics, University of Münster; E-Mail:
[email protected]
P. Dziuballe et al. / The Single Source Architecture x4T
903
regulations in clinical trials. In this document the “single source” scenario is envisioned among others, as a promising concept that implies to capture medical data at one single point reducing double data entry and promoting secondary use [5]. A first prototype of this concept was successfully implemented by Kush et al. [9]. Our own research projects show the feasibility of this approach [10, 11]. Another scenario of the eSDI paper is the “Extraction and Investigator Verification” solution, where documentation occurs within the CIS. With regard to this concept, the Retrieve Form for Data Capture (RFD) [12] profile was jointly developed by CDISC and the Integrating the Healthcare Enterprise (IHE) Initiative, to enable data capture for clinical research and other purposes within a CIS session. It defines four actors who participate in specific transactions based on web services. This profile is extended in the REUSE project [6], through the integration of forms towards a profile called “Retrieve & Integrate Forms for data capture” to enable direct study documentation within the CIS reusing present medical routine data. Furthermore, the procedure of clinical trials is strictly regulated by law and supervised by authorities like European Medicines Agency and U.S. Food and Drug Administration. These international regulations have to be respected while connecting routine and research documentation. In this paper, we pursue the following objectives: We intend to identify the amount of available routine data in CIS eligible for study documentation. After that, we design and implement an architecture to facilitate the re-use of available medical routine data for clinical trial documentation with due regard to regulatory constraints and the usage of established international standards.
2. Materials and Methods We have analysed CRF items of two currently conducted multicentre trials at the University Hospital Münster (UKM) concerning the availability and manifestation in CIS ORBIS from Agfa Healthcare [13]. To assess the secondary use potential in our approach, we have identified the amount of CIS data suitable for re-use in two studies conducted at the UKM; their CRFs contain 278 and 318 items, respectively. This occurs through a manually review of all implemented CIS forms available in the respective clinical department by a medical informatics professional. We have also analysed clinical trial documentation workflows through interviews with employees of the Centre of Clinical Trials in Münster and with research physicians at the UKM. After that, we have conducted a literature search and identified communication standards, established in the healthcare and clinical research domain. We have selected the Operational Data Model (ODM) published by the CDISC because of its ability to archive trials [14] and exchange metadata definitions and data. We have also reviewed the IHE RFD profile to exchange documentation forms between CDMS and investigators’ CIS. With respect to the CIS features, we have identified the existing information system architecture at the UKM. Regarding interfaces we have analysed the Health Level 7 (HL7) communication standards. Based on our analyses, we have designed and implemented a system architecture. The clinical trial metadata was processed in ODM format.
904
P. Dziuballe et al. / The Single Source Architecture x4T
3. Results 3.1. Analysis To identify the amount of re-usable CIS data and to examine the desirability of a single source approach, we have manually mapped 596 study items (278 items from the first trial, 318 from the second) to the CIS documentation. We assigned three categories: “"identified in CIS,” “require modification” and “not found in CIS”. Figure 1 shows that 47% of the CRF items were found within the available routine patient care documentation. For about 11%, the item values cannot be used directly because they are free text or contain only similar information. In this case, a modification of the item value is necessary. About 42% of the CRF items are not found in CIS. This result shows that the re-use of CIS data values is a rewarding step for study documentation, as almost half of the required items (47%) for study documentation are available. 47%
Identified in CIS 11%
Require modification
42%
Not found in CIS 0%
20%
40%
Amount of items 60%
Figure 1. Amount of identified CRF items in CIS based on an analysis of two clinical trials.
Concerning the RFD profile, the analysis of the present CIS architecture shows that a direct implementation within the CIS is limited due to technical restrictions (proprietary system structure) as well as license issues (restrictions regarding new interfaces). Specifically, direct import of CRFs is not available in the current CIS version. ORBIS neither supports Web services nor XForms and pre-filling of CRF items as described in RFD. 3.2. Architecture To overcome these drawbacks and make use of existing CIS data, we developed an architecture to create an interface between CIS and CDMS, compliant with regulatory constraints and recommendations (GCP, eSDI, Title 21 CFR Part 11) and applying established standards. RFD was extended and refined, which broadly resulted in a combination of RFD and the single source concept of the eSDI document. We developed an integrated documentation process based on identified regulatory principles. Certainly, t is not feasible to directly implement this process in current CIS solutions and due to their limitations we designed a middleware component (Figure 2) to connect it to clinical research systems. This mediator – called x4T (exchange for Trials) – is hosted in the hospital environment to establish the integrated documentation process.
CIS
x4T
CDMS
Figure 2. Single source architecture with the middleware component x4T between the CIS and the CDMS.
x4T enables the exchange of forms and medical routine data and also enables to send notifications to study physicians. Single source eCRFs will also be prepared for
P. Dziuballe et al. / The Single Source Architecture x4T
905
presentation and data input. The interfaces of x4T and the CDMS are able to exchange eCRF definitions and completed ones. x4T consists of the following modules: • Interface management Due to heterogeneity in CIS landscape, the CIS-x4T connection needs to be adapted to the communication interface provided by the CIS. Standards for healthcare communication such as HL7 messages and Clinical Document Architecture are preferred. For systems that do not serve those interfaces XML and a specific wrapper are used. The CDMS-x4T communication relies on ODM wherever supported by the CDMS; otherwise it is adaptable to the specific protocol. • User management Configuration of sites including users and their associated roles is done within the user management. The completed CRFs need to be signed by the user. • Form mediator and database It is currently not possible to directly store and display eCRFs inside the CIS. Therefore a separate form database is needed to temporarily store eCRFs. The mediator transforms these eCRFs into displayable XForms and enables prefilling of items. With regard to regulatory requirements pre-filled items need to be verified by the user and after documentation a copy of the CRF will be archived in CIS. If there are several item occurrences available, for instance blood pressure values at many points in time, the correct value needs to be selected and confirmed by responsible study physicians. • Ontology matching This module handles the eCRF pre-filling with routine patient care data. Eligible data items have to be verified and mapped with controlled vocabularies. A semantic layer is currently missing in ORBIS so that data annotation occurs externally in x4T. CRF items also need to be matched with this vocabulary. Pre-filling is possible in case of overlapping semantic concepts. A conversion engine translates measurement units or calculates item values in case of different data types. For instance, patient age for a CRF is calculated from date of birth and date of visit provided by the CIS. • Notification management To support clinical workflow, the CIS user has to be notified about new eCRFs to be filled.
4. Discussion and Future Work In this paper, we have proposed a system architecture to support integrated clinical trial documentation workflows. The central component of this architecture is the middleware server x4T to establish the connection between patient care and research systems. Due to the current gap of standards, a direct link from hospital to study systems would require an adapter for every CIS and CDMS in a full mesh topology for multicentre trials. With the mediator approach, only one interface per system is required. According to our architecture, study forms are documented in a single system with the advantage that documentation forms have not to be built in any CIS. In case of CRF updates, implemented CIS forms as realised in [10, 11] would cause huge maintenance costs. x4T enables re-use of medical data and pre-filling of study forms.
906
P. Dziuballe et al. / The Single Source Architecture x4T
Avoiding discrepancies in the validation of eligible CIS data, clinicians or other experts need to be consulted to reach consensus. The whole amount of re-usable CIS data may differ from study to study and can finally be calculated in a real trial setting. Pre-filling is only possible if CIS data are well-structured. In order to utilize our implementation, the x4T interfaces have to be adjusted. A proof-of-concept of x4T is planned for a clinical study in dermatology.
5. Conclusion Pre-population of eCRFs with CIS data is a promising approach to avoid redundant data entry because of a quite large overlap between CDMS and CIS items. Due to limitations of current CIS and regulatory constraints, the exchange of data between CIS and CDMS should be enabled by a mediator.
References [1] [2] [3]
[4]
[5] [6] [7] [8] [9]
[10]
[11] [12] [13] [14]
Ammenwerth, E. Spötl, H.P. The time needed for clinical documentation versus direct patient care. A work-sampling analysis of physicians' activities, Methods Inf Med. 2009;48(1):84-91. Tipping, M.D. Forth, V.E. O'Leary, K.J. Malkenson, D.M.. Magill, D.B Englert, K. Williams, M.V. Where did the day go?-a time-motion study of hospitalists, J Hosp Med. 2010 Jul-Aug;5(6):323-8. Jha, A.K. DesRoches, C.M. Campbell, E.G. Donelan, K. Rao, S.R. Ferris, T.G. Shields, A. Rosenbaum, S. Blumenthal, D. Use of electronic health records in U.S. hospitals, N Engl J Med. 2009 Apr 16;360(16):1628-38. Williams, J.G. Cheung, W.Y. Cohen, D.R. Hutchings, H.A. Longo, I M.F.. Russell, T. Can randomised trials rely on existing electronic data? A feasibility study to explore the value of routine data in health technology assessment. Health Technol Assess. 2003;7(26):iii, v-x, 1-117. Prokosch, H.U. Ganslandt, T. Perspectives for medical informatics. Reusing the electronic medical record for clinical research. Methods Inf Med. 2009;48(1):38-44. El Fadly, A. Lucas, N. Rance, B. Verplancke, P. Lastic, P.Y. Daniel, C. The REUSE project: EHR as single datasource for biomedical research, Stud Health Technol Inform. 2010;160(Pt 2):1324-8 Zahlmann, G. Harzendorf, N. Shwarz-Boegner, U. Paepke, S. Schmidt, M. Harbeck, N. Kiechle, M. EHR and EDC Integration in Reality. Applied Clinical Trials 2009. eSDI Initiative. [http://www.cdisc.org/esdi-document] Kush, R. Alschuler, L. Ruggeri, R. Cassells, S. Gupta, N. Bain, L. Claise, K. Shah, M. Nahm, M. Implementing Single Source: the STARBRITE proof-of-concept study, J Am Med Inform Assoc. 2007 Sep-Oct;14(5):662-73. Breil, B. Semjonow, A. Dugas, M. HIS-based electronic documentation can significantly reduce the time from biopsy to final report for prostate tumours and supports quality management as well as clinical research. BMC Med Inform Decis Mak. 2009 Jan 20;9:5. Fritz, F. Ständer, S. Breil, B. Dugas, M. Steps towards single source--collecting data about quality of life within clinical information systems. Stud Health Technol Inform. 2010;160(Pt 1):188-92. IHE International: IHE ITI Technical Framework Supplement: Retrieve Form for Data Capture (RFD). [http://www.ihe.net/Technical_Framework/upload/IHE_ITI_Suppl_RFD_Rev2-1_TI_2010-08-10.pdf] Agfa Healthcare. [http://www.agfahealthcare.com/] Kuchinke, W. Aerts, J. Semler, S.C. Ohmann, C. CDISC standard-based electronic archiving of clinical trials, Methods Inf Med. 2009;48(5):408-13.
