VPH-DARE@IT D3.4 1v2 Final
Contents
1. eo DO YU VPH DARE 011 DELIVERABLE 3 4 Database infrastructure for enabling studies based on prospective data in WP and 2 Grant agreement no 601055 FP7 ICT 2011 9 Project acronym VPH DARE IT Project title Dementia Research Enabled by IT Funding Scheme Collaborative Project Project co ordinator Prof Alejandro Frangi University of Sheffield Tel 44 114 22 20153 Fax 44 114 22 27890 E mail a frangi sheffield ac uk Project web site address http www vph dare eu Work Package amp Task WP 3 Task 3 2 Lead beneficiary STH Editor PMO Author s Kevin Teh Steven Wood Quality reviewer Annalena Venneri Sebastien Kozerke Dissemination level PP Restricted to other programme participants including Commission Services __ Restricted to a group specific by the consortium including Commission m Services O Confidential only for members of the consortium including Commission Services FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 Issue Record Author s Reason for modification 01 03 15 1 1 26 04 15 S Wood Final version following Complete review comments 8 5 515 Final Check Copyright Notice Copyright 2013 VPH DARE IT Consortium Partners All rights reserved VPH DARE IT is an FP7 Project supported by the European Union under grant agreement no 601055 For more information on the project2. lemme ro le bata esmero oM he TABLE MODALITY Ta iE ItemGroupRepeatKey 1 TransactionType Insert gt Itenbata Item TABLE MODAL DATAN ne oU Tq uM gt lt ItemData ItemOID I TABLE MODAL TYPE TABLE Value 1 ItemGroupData lt FormData gt FIGURE 5 XML FRAGMENET SHOWING THE IMAGING STUDY INFORMATION zT s FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 The ItemOID and Value of ItemData as shown in Figure 5 are required The value of I TABLE MODAL TYPE TABLE will be programmatically replaced to be a type of scan modality in this case 1 represents MRI scans These fields will make the basis of experiment creation in XNAT via secure SSH This is done by using a plugin provided by XNAT called XNATRestClient over SSH to the XNAT virtual machine This provides an easy way to call the XNAT REST API via the Java command line tool Some of the example calls that we used are as shown below XNATRestClient host https localhost u username p password m PUT remote E e quive o Ee E Gs exeo sU menes E e e erue O Oan a t mrSessionData date 03 09 14 For this example a subject called man88 will be created with an experiment name called mr_030914_man88 The experiment convention is chosen to represent modality_date_subjectname to ease data entry in XNAT Modality type and date was obtained from a REST query to CRF1 Figure 2 Schematic overview o
3. v13 0 V2 0 V5 0 v7 0 v5 0 11 0 v3 0 v3 0 V3 0 Neuropsychological and Neuropsychiatric Tests 11 0 Table of Modality Add Subject To XNAT Images Upload Completed Study Summary v1 0 v1 0 V3 0 V2 0 6 El root E Cl 8 3 Gl 3 E 3 8 E View this Subject s Record SA EE Workflow Study Event Overview Data Entry Mark Event CRF Complete SESEESSSSSESEEES cuounnaaaman GECECEEEECEEEEEE root Data Manager en Log Out te y FIGURE 7 OPENCLINICA SUBJECT DATA ENTRY PAGE Once the subject is created the user then selects from the range of CRFs available which data entry tasks to perform The subject level display summary shows the status of each CRF for this visit or event and who performed it The CRF data pages themselves have a simple layout and a typical example is shown in Figure 8 this example contains a variety of input controls like radio button lists drop down lists dates etc 16 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 CRF Header Info Subject 0 33 Title Subject Eligibility Page UO mark CRF Complete a AUS 5 Ja Age range P Subject is 50 85 years old inclusion criteria For all Inclusion 4 Study group he OTF HC or MCI selected Press ENTER after making selection Q mci O ad ftd O vad Exclusion 15 Exlusion Cr
4. its partners and contributors please see http www vph dare eu You are permitted to copy and distribute verbatim copies of this document containing this copyright notice but modifying this document is not allowed All contents are reserved by default and may not be disclosed to third parties without the prior written consent of the VPH DARE IT consortium except as mandated by the grant agreement with the European Commission for reviewing and dissemination purposes All trademarks and other rights on third party products mentioned in this document are acknowledged and owned by the respective holders The information contained in this document represents the views of VPH DARE IT members as of the date of its publication and should not be taken as representing the view of the European Commission The VPH DARE IT consortium does not guarantee that any information contained herein is error free or up to date nor makes warranties express implied or statutory by publishing this document Author s for Correspondence Lead author name Correspondence details Dr Steven Wood T 44 114 271 3649 E steven wood G sth nhs uk FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 Table of Contents INTRODUCTION 5 EL BACKGROUND e s 5 2 SYSTEM DESIGN AND RATIONALE cccccccccccccccccccccccccccccccccccccccccccccccccccccccccce
5. 08 05 2015 white bread How much do you on average eat slices of white bread or baguette per day Porridge P How many deciliters do you on average eat porridge per day Cereals P8 How many deciliters do you on average eat low fibre breakfast cereals per day Muesli P How many deciliters do you on average eat muesli per day Sweet bread P How many deciliters do you on average eat sweet bread per day SUT Tarr wilt GU OU vo Lal Vegetable sterol margarine Butter vegetable oil mixture Butter 7 Spread a PB What kind of spread do you usually use on your bread v PB How much do you eat sweet patisseries ice cream puddings or chocolate One portion is e g a piece or Danish pastry 3 4 cookies ice cream cornet pudding chocolate bar v How much do you eat sugar honey or sweets One portion is e g 2 teaspoons of sugar or hone Beverage FIGURE 25 MULTI SELECT EXAMPLE Also for any calculations score 1 e Neuropsychological and Neuropsychiatric Tests CRF the scores results will only be calculated once the CRF is saved Medical History For partial dates key in manually 1 e Feb 2015 for onset and start time in the medical History and Medications table respectively Neuropsychiatric Test NPI Total scores will only be calculated when you press ENTER or SAVE the CRF If SAVED to view it go back into the CRF again Image Upload in XNAT The site for XNAT is https vphdare xnat shef ac uk In order to uplo
6. If you need to alter a field at any point after marking the CRF as complete you will have to fill in a discrepancy form by clicking on the flag next to the field you have just altered and explaining why you altered it 6 After clicking Save you will be returned to the enter data page again Continue doing this until you reach the last 3 CRFs Table of Modality Add Subject To XNAT and Images Upload Completed these have to be completed in sequence to create a subject in XNAT 7 Firstly complete the Table of Modality CRF Here it is important to accurately enter firstly the modality Mod Type and then the Date Performed of that imaging modality as shown in Fig 5 below If your subjects have had more than one imaging scan click the Add tab below the modality pull down menu and repeat the process again Click Save when you have added the subject modalities and date scanned again shown in the Figure 20 example 728 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 Table of Modality v1 0 amp STHManual Y CRF Header Info ModTable 0 2 Title ModTable Page CI markcrr complete iC OMT 3 Table of Modality Mod Type Date Performed w 03 Sep 2014 Ci p ES rr 3 ES Add O mark cer Complete TT GEN xS 6 FIGURE 20 TABLE OF MODALITY CRF 8 This will return you to the enter data page Now you will need to fill the Add Su
7. OpenClinica will provide the information necessary to create the relevant experiment using an XNAT plug in via SSH These RESTful queries are used to obtain subject specific information The database status queue will also then be updated to Completed once the XNAT plug in returns a response confirming subject creation These will be described in more detail in the next sections Event Driven Thread Timer Driven Thread Wait for SMTP P Check Queue Message for messages Messages exist Place Message on Queue Request additional Information from OC Create imaging episode in XNAT More Messages exist FIGURE 3 FLOWCHART FOR OCX INTERFACE A detailed description of the webs services used for the interfacing of this service into OpenClinica is described in section 6 2 and the same code base was used in both systems However there is one specific issue did affect the OCX service that was not relevant in the context of the research platform interfacing and this is the fact that OpenClinica modifies the internal subject ID from that specified by the user Whilst testing the OpenClinica and XNAT subject creation link we discovered that OpenClinica only supports up to 8 unique characters This could potentially be a problem as any repeating strings larger than 8 characters will trigger an automatically generated unique SubjectKey sometimes referred to as key mangling This 1s an 1ssue because of the way in which the RES
8. This is because there are no classes or methods built in natively supporting the NET architecture supporting java spring web service security encryption WSSE 19 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 After further investigation the correct login header information as shown in Figure 10 is needed for OpenClinica SOAP authentication We found out that the Dot NET version used does not produce a SOAP authentication header which includes UsernameToken One solution that we tried was to use a Dot NET username token manager WSE 3 0 to generate the missing UsernameToken However this solution also produces additional information that is not needed by the java spring security check of the OpenClinica machine The additional information is shown in Figure 11 giving two additional fields Nonce and Created causing login authentication to fail lt soapenv Header gt wsse Security soapenv mustUnderstand 1 xmlns wsse http docs oasis Open Oro wes 2004 7017 oas15 20040l_ wse wesecurity secext l 0 xsd gt lt wsse UsernameToken wsu Id UsernameToken 27777511 xmlnsewsuc hEtos docs cases open org es 20047 Ol 4042539200408 ws ws securiry Utility 1 0 xsd lt wsse Username gt user lt wsse Username gt wsse Password Type http docs oasis open org wss 2004 01 o0asis 200401 wss username token profile 1 0 PasswordText gt password lt wsse P
9. gt EN lt ClinicalData StudyOID gt lt SubjectData gt Item lt StudyEventData gt Date lt FormData gt ItemGroupData gt lt ltemData ltemOlD Value gt lt ltemGroupData gt lt FormData gt lt Study EventData gt lt Subject Data gt lt ClinicalData gt FIGURE 12 THE MAPPING MODEL FOR THE OPENCLINCAL METADATA SCHEMA ONTO THE UNDERLYING DATABASE SCHEMA CDISC ODM XML File OpenClinica Study OID gt OIL FormRef OID gt lt StudyEventDef gt lt ltemGroupRef OID gt lt FormDef gt lt ItemRef ItemOID lt ltemGroupDef gt lt ltemDef OID gt CodeListRef CodeListOID gt lt ltemDef gt lt Study gt n Q 3 o Y ClinicalData StudyO SubjectData gt eStudyEventData gt lt FormData gt lt ltemGroupData gt lt ltemGroupData gt lt FormData gt lt Study EventData gt lt Subject Data gt lt ClinicalData gt FIGURE 13 THE MAPPING MODEL FOR THE OPENCLINICA DATA SCHEMA ONTO THE UNDERLYING DATABASE SCHEMA Crudely speaking the response is split into two parts the schema definition containing information on the data groups types and lookup value lists etc along with the presentation state metadata such as the question text which appears next to the data entry box The second block actually contains the clinical data entered where each data item is referenced back to the
10. saving the aggregated data document for all subjects and CRFs in a single XML file One additional benefit of this feature is that we can now allow users to publish data from OpenClinica without ever having to have an account or connect live to the services because the format of the manual data export from the web page 1s exactly the same as that from the web service This would allow an OpenClinica administrator to perform some predefined query on the system and pass it to the DPS without having to create special accounts etc which is often quite desirable m iG ADDSU UNGROUPED A ES ARENA G_INCLU_UNGROUPED IG GENER UNGROUPED 3664 AECE IG LIFES UNGROUPED Be G_CSF_UNGROUPED DS Activities of Daily Living Inventory ADCS ADL o0 o H BE e StudyEventlD m d tatus StartDate 2d ae Alcohol EnrollementDate Status Ae Pr Sri g Smoke Project StudyEventKey Smoke ALAN A C RST i i Computer based memory test Muistikko 1G_MEDIC_UNGROUPED_3771 IZ Diagnosed or treated diseases Table patent medicines dietary supplements vitamins etc 13 Table prescription and over the counter drugs MO 16 STUDY UNGROUPED IG_CONSE_UNGROUPED FIGURE 15 PROSPECTIVE TRIAL SCHEMA WHEN PUBLISHED 294 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling s
11. schema using a unique ID This model in fact maps directly onto the functionality needed by the DPS during its ETL process First it requests the data schema and then iterates through the data groups within it requesting the data az FP7 601055 VPH DARE IT 08 05 2015 D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and 2 6 3 DPS PLUGIN DEVELOPMENT The plugin for connection to OpenClinica needs some core information regarding the location of the application and it also needs login credentials to access those services These are contained in the Connection group on the interface As with the XNAT browser we have made available a facility for browsing the data within the system as well This feature 1s particularly useful when trying to visualise what the final data structures might look like when the export 1s completed In this instance the user can decide which attributes of the data in OpenClinica should be used to generate the table and field names that will eventually be presented to the researchers In this particular case the options are OpenClinica ID Name OpenClinica ID amp Name Comments Question This may seem like an unusual list of options when you understand the internal workings of OpenClinica but in fact where the meaningful information concerning a data item is held depends entirely on the person who designed the CRFs in the Excel template shown in Appendix 2 Having reviewed
12. the patient in XNAT Alerts amp Messages Instructions Info Study Events Study Events 3 Initial Visit Status data entry started Hal Subjects Eligibility HB study visits B 18 Basic Information He General Physical Examination 18 Neurological Examination Medical Questionaire HE Lifestyle Health Questionaire HB Blood Sample A8 Lab Analyses blood sample HE csr 18 Neuropsychological and Neuropsychiatric Tests HE Table of Modality 18 Add Subject To XNAT 18 Images Upload Completed 18 Study Summary E FollowUp Visit El Imaging OpenClinica Commun fy Ed tion v y v Home Subject Matrix Notes amp Discrepancies Study Audit Log Tasks v Enter or Validate Data for CRFs in Initial Visit 9 El VPH Dare Clinical Data Repo VPHDare Change Study Site Edit Study Event Study Subject ID abc4444 Study Event Initial Visit Location N A po Study Subject OID 55 ABC4444 Start Date 07 Nov 2014 m End Date Time m Subject Event Status Last Updated by CRFs in this Study Event CRF Name Subjects Eligibility Study Visits Basic Information General Physical Examination Neurological Examination Lifestyle Health Questionaire Blood Sample Lab Analyses blood sample CSF data entry started root 11 Dec 2014 E study Subject ID Go Version Status Initial Data Entry Double Data Entry Actions
13. two other project configurations during the development of this code we decided to keep the options for naming as open as possible The interface shows the Create column names from and Create descriptions from drop down controls which the users select Figure 14 shows the interface design gt vphdare oc Data Publication Suite cx J fm File Edit Tools Window Help 23H S Ax vphdare Properties ERE S XS F 3 3 MINI MENTAL STATE EXAMINATION MMSE A Source name 63 Modality Table ci f MODIFIED FRONTAL BEHAVIORAL INVENTORY FBl mod Proj S VPH SMA Create column names from Create description names from LE Mri comp URL https vphdare oc shef ac uk OpenClinica Name Comments z 7 Mri_complete JE Mi date Usemame root 3 ParentKey L Password 4S 43 08 Neuropsychiatric Inventory NPI xd am FILE ONLY IMPORT f PATIENT HEALTH QUESTIONNAIRE 9 PHQ 9 3 8 PHONEMIC FLUENCY Projects Connect Local file path C Users smwood Work wphdare_oc_extract xml m 8 Smoke Local chache expiry days 0 Use Cookies T1 Life cigars E fF Life cigquan Browse dat L bie toes Browse data E Life_setfroliquan Subjects EN E GE FT ParentKey STHJessSmith Initial Visit 1 MINI MENTAL STATE EXAMINATION MMSE 1 Stroop _ STHTest Msse_date 2015 01 05 A wood1 M mhil Pros0000 2 2 ig Destinations nx ProsKuopio test Msse writesent 1 E Y gt 4 Pros
14. A MODEL USED BY OPENCLINCA 6 2 1 OpenClinica Web Service Integration 6 2 1 1 SOAP Simple Object Access Protocol OpenClinica implements a SOAP web service interface which allows a significant amount of functionality to be executed without any user interaction Experience in together data integration applications has shown us that given the choice between SOAP and REST discussed in section 6 2 1 2 the development time and reliability of the former is much higher due to the WSDL Web Services Description Language definitions that are provided by SOAP interfaces This allows a large amount of automated code generation to be provided to the developers and also gives a very robust way of updating and finding errors 1f the definitions from the third party system change for any reason Some of the key calls that are supported by OpenClinica are below and it was clear that for the interaction required for the project this would be more than sufficient create creates a new study subject listAllByStudy lists study subjects in a study schedule schedule a visit event import imports data getMetadata returns study metadata Due to the feature rich SOAP web service interface this was our first approach to obtain CRF item data for a specific subject We encountered issues using the OC SOAP service because the default mechanisms provided by the Dot NET framework did not authenticate to the java spring security framework that OpenClinica uses
15. ENDIX 3 EXAMPLE OF CDISC ODM DATA DOCUMENT e eee eee eee eee eee ee ee oeeoeue 38 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 Table of Figures Figure 1 Structure of data collection system ooocccccccnnnonnnnnonnnnnnnnnnnnnnnnnnnonnnnnononononinoninanoss 8 Figure 2 Schematic overview of OCX Integration service ooccccccccnnnnnnnnnnnnnnnnnnnoninoninininonoss 11 Figure 3 Flowchart for OCX Interface ooccccccccccccnnnononnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnss 12 Figure 4 An XML fragment showing the subject key mangling for IDs longer than 8 characters A 13 Figure 5 XML fragmenet showing the imaging study information sss 13 Figure 6 Diagram showing the CRFs for a subject volunteer for different visits 15 Figure 7 OpenClinica subject data entry page sssssssseee nnn 16 Figure 8 A typical CRF data entry pPage ooccccccccccnnnnnnononnnnnnnnnnnnnonnnonnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnss 17 Figure 9 Data model used by OpenClincCa occcccccccnnnnnnnnnnnnnnnnnnnnnnnnnnnonononnnnnnnnnnnnnnnononinininos 19 Figure 10 Correct authentication header required by OpenClinica 20 Figure 11 Extra information sent by wse3 0 net sess 20 Figure 12 The mapping model for the OpenC
16. EST web service 1s rather limited to filtering based on a RESTful URL This will return either a json html or XML format A standard RESTful URL looks like http localhost 00 OpenClMinica rest clinmcaldata format mode StudyOID StudySubjectKey StudyEventDefOID StudyEvent RepeatKey FormDefOID FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 From the generic URL we can see that the user can specify output format json html or xml mode view or print and also all the other subject specific information lt soapenv Header gt wsse Security soapenv mustUnderstand 1 xmlns wsse http docs oasis opencorg wes 2002 401 54881795 200420 w55 wssecupitv secext l0cx5d wsse UsernameToken wsu Id UsernameToken 27777511 xdmbinsesweuc tor docs ode 1s OPen org wss 2004 01 039818 200401 ws8s8 ws security utility 1 0 xsd lt wsse Username gt user lt wsse Username gt lt wsse Password Type http docs oasis open org wss 2004 01 o0asis 200401 wss username token profile 1 0 PasswordText gt password lt wsse Password gt lt wsse UsernameToken gt lt wsse Security gt soapenv Header An example of a typical REST call to the OpenClinica services is the URL is shown below this will return all subjects in a specific study in this case with study OID called S VPH Note that a useful feature built into OpenClinica s RE
17. Kuopio Test2 J Msse_readobey 1 TestSth Msse followcmmd 3 Data Destinations n GEA ltest5 Msse_nameobj 2 ltest6 Msse_recallwords 1 Msse_count 5 Msse_regwords 3 Msse_locate 5 Initial Visit 1 IG CSF UNGROUPED 1 Initial Visit 1 Stroop 1 Initial Visit 1 Computer based memory test Muistikko 1 Initial Visit 1 WAIS III 1 Initial Visit 1 Digit span 1 Initial Vist 1 Beck Depression Inventory 21 BDI 21 1 Initial Vist 1 MADRS 1 Initial Visit 1 MODIFIED FRONTAL BEHAVIORAL INVENT Initial Visit 1 LOGICAL MEMORY 1 Initial Visit 1 PHONEMIC FLUENCY 1 Initial Visi 1 Generalized Aroiety Disorder 7 GAD 7 _1 Msse time 5 Msse performed KK Msse complete Yes Msse comm FIGURE 14 SCREENSHOT OF USER INTERFACE FOR DPS DATA EXTRACTION AND EXPLORATION Nm FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 At any point the user can change these options and see the effect on the data structures that will be produced by browsing for subjects and their CRF data The final feature of the interface 1s that 1t will cache the patient data for a period of time This is certainly desirable during the configuration of the system but will also ensure that the live server is only polled on a periodic basis no matter how frequently the extract 1s requested to run from within the DPS This functionality is handled by
18. ST interface is the use of as a wildcard call Nets vondare o siae sees Cpene linea eese abwarten all low SL S tens This REST call is used to obtain all subjects in the study and all of their completed CRF records A snippet of the folded XML output from the above call is as shown in Appendix 3 6 2 2 OpenClinica data model In the XML from Appendix 3 we see how the XML in OpenClinica is structured for a REST query to obtain all the data in a study For this example the case study site is VPH Study OID 5 VPH OpenClinica generates its XML files based on the CDISC ODM standard http www cdisc org odm XML convention A more constructive description of the document format is provided in Figure 12 and Figure 13 which describes both the structure of the XML response document and also the way in which these elements map onto the underlying OpenClinica database schema these images were copied from https docs openclinica com 3 1 technical documents openclinica 3 1 database model zT s FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 CDISC ODM XML File OpenClinica Study OID gt lt ltemGroupRef OID gt lt FormDef gt lt ItemRef ItemOID lt ltemGroupDef gt CodeListRef CodeListSlD gt lt ltemDef gt lt Study gt OIL FormRef OID gt lt StudyEventDef gt FormDef OID
19. T services rely on the SubjectKey to access data on a specific subject The core URL for obtaining all subject level data in the system has the following pattern EIS FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 https ServerName OpenClinica rest clinicaldata xml view Project Sub jectKey A specific example for the deployed system and test subject being HEtoSs 7 7 vondare ass seis ce le oscila fase Cline lolis vainly A V2 SG SUB UME SE es As an example we previously named one of our subjects STHTestBoy and then added a subject called STHTestBot shown by the XML excerpt in Figure 4 The subject STHTestBoy will have a SubjectKey of SS_STHTESTB whereas STHTestBot has a SubjectKey of SS_STHTESTB_7448 Here we can see how OpenClinica only takes 8 characters of the string STHTestBot and then randomly generates a 4 digit number after the 8 characters This behaviour needed to be managed reliably as OpenClinica only accepts REST queries based on SubjectKeys and not StudySubjectID as shown from the subject specific RESTful URL above If we indeed used StudySubjectID the REST query would return nothing for the subject Most partner sites will have subject IDs longer than the 8 unique character length restriction placed by OpenClinica sonogeerpata m subyecthev so Ta IB 446 OpenClinica StudySubjectrD STHTestBot OpenClinica Status available Op
20. Tip In 17 Inc_VADcriteria VAD inclusion criteria fulfilled VAD inclusion criteria fulfilled 18 Inc_FTDdiagnosis FTD diagnosis FTD diagnosis a href javascript void 0 onmouseover Tip Di 19 Inc_FTO criteria FTD inclusion criteria fulfilled FTD inclusion criteria fulfilled 20 Inc_criterial Exlusion Criteria 1 Exlusion Criteria 1 a href javascript void 0 onmouseover Tip O 21 Inc_define1 Exclusion Criteria 1 define Exclusion Criteria 1 define Diagnosis SNOMED 22 Inc criteria2 Exlusion Criteria 2 Exlusion Criteria 2 Other causes of dementia 23 Inc_define2 Exclusion Criteria 2 define Exclusion Criteria 2 define Diagnosis SNOMED Inc criteria2 Exlusion Criteria 3 Exlusion Criteria 2 a href javascript void 0 onmouseover Tip Ca 25 Inc define Exclusion Criteria 2 define Exclusion Criteria 2 define Diagnosis SNOMED 26 Inc MSSECompleted MSSE Completed MSSE Completed 27 Inc_MSSEDate MSSE Date MSSE Date 28 Inc_MSSEOverallScore MSSE Overall Score MSSE Overall Score 29 Inc CDRCompleted CDR Completed CDR Completed 30 Inc CDRDate CDR Date CDR Date Y M 4 gt M CRE Sections Groups y Items Instructions ESE gt FIGURE 30 CRF EXPORT TO SPREADSHEET The CRF platform is both feature rich and simple To hide or show items is as simple as setting conditional flags on a excel column We have also explored using javascript for calculation to hide long strings of texts labels using the onmouseove
21. a ecg 010914 STHV ct 030914 STHVoO mr 030914 STHVo FIGURE 26 XNAT HOMEPAGE SELECT PATIENT TO ADD IMAGING SCANS 3 This will take you to the page for Session mr 030914 STHTestBoy mr 301014 testsub Click on Upload then upload scans under Actions this is shown in Figure 27 E Advanced Home New Upload Tools Help gt PROJECT VPH gt SUBJECT STHTestBoy gt mr_030914_STHTestBoy MR Session mr_030914_STHTestBoy Edit View gt Accession XNAT_E00048 Subject STHTestBoy Date Added 09 26 2014 11 10 05 admin Gender Date 09 03 2014 Handedness Upload gt Upload Scans Age HERE Download gt Tagged Upload Email Manage Files Series Desc FIGURE 27 XNAT UPLOAD SCANS 4 After clicking Upload Scans a compressed upload screen will pop up Click on the Prearchive radio button then click Choose File as shown in Figure 28 Now select the imaging files that you want to store and click Upload When Jay FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 completed 1t will give you a Successfully Uploaded 1 Session into the Prearchive message below the upload status bar when it is completed as shown in Figure 29 Compressed upload Raw image files can be zipped zip or tar gz and uploaded using the form This tool currently supports DICOM and ECAT files Selecting Prearchive will place your images into a tempor
22. a lasting and open platform which may well desire additional data capture after the funded period we decided that pursuing a proprietary solution which would create a kind of vendor lock in was not desirable Following these considerations it was decided that ArQ would only be considered as the data management platform for the project once all other viable options had been ruled out From an open source and community supported software perspective this leaves us with two primary systems for consideration OpenClinica https www openclinica com and XNAT http www xnat org These choices come from the almost equal demand in the project for an imaging based data collection which is the primary function of the XNAT system and the need to collect detailed clinical and phenotypical information FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 of the subjects for which OpenClinica is the largest community based offering We decided to evaluate both these systems for use in VPH DARE IT and the following describes the implementation of the final system design 2 SYSTEM DESIGN AND RATIONALE 2 1 OPENCLINICA EVALUATION OpenClinica has excellent facilities for creating and managing electronic case report forms eCRF s and it enables compliance with Good Clinical Practice GCP and regulatory guidelines such as 21 CFR Part 11 via differentiated user roles and privilege
23. ad an imaging study the only acceptable image is a DICOM study folder compressed in a zip file This is easy to achieve and if you do not have a client already installed a good open source application for this 1s 7 Zip which can be downloaded from http www 7 zip org Login into XNAT using the username and password provided On the Recent Data Activity Tab you should be able to see your newly created subject in OpenClinica The subject label created for your convenience is split into modality date d m yrj subjected created in OpenClinica 1 e mr 030914 STHTestBoy Hover over the label to get more information if only partially displayed Firstly click on the subject label as shown in Figure 26 Ne Ads FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 YXNAT Em aen Home New Upload Tools Help XNAT currently contains 4 projects 29 subjects and 47 imaging sessions Projects Recent Subjects MR PET CT Favorite My projects m ID Name Description Other projects 4 Stored Searches Keywords Investigator SELECT Y Data Submit VPH mr 030914 STHTe Project ID VPH mr 040914 ST mr 030914 STHTestBoy VPH Data Repository mr 090514 STHTe You are a member for this project mr 010714 STHPa mr 040811 STHTest mr 030914 STHHe mr 080914 STHJa ct 030914 STHJa ct 070809 STHJa mr 020914 STHHe mr 010914 STHJ
24. aging based project is a significant impairment Since the evaluation of the system a community extension has been released for the system from University of Aachen called OC Big which now supports large file transfers into the system https community openclinica com extension openclinica big data oc big This system was only released in late December 2014 and whilst very interesting it is still quite immature We have therefore decided not to deviate from our initial planning to use this as the majority of the internal implementation work had already been done towards the solution described We will however continue to monitor this development and should another requirement for prospective data capture present itself from within the project or as a result of external engagement we will revisit this decision since a single platform solution would be desirable 2 2 XNAT EVALUATION FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 XNAT is primarily an image sharing and management portal It was designed explicitly to support research and one can think of it as a PACS Picture Archiving and Communication System with extended components to facilitate the demands of clinical research The features above those that we would expect in a clinical system are e Web based upload and download of DICOM studies recent PACS systems support the WADO protocol now as an add
25. ary holding space You will then have the ability to review the details and match the data to the proper subject amp session ID If you are confident the data will be mapped properly you can directly Archive the files and specify whether the resulting session should go into a quarantine state Project VPH Session mr 030814 STHTestBoy Destination Prearchive Archive in quarantine O Archive no quarantine File Choose File No file chosen No file chosen at FIGURE 28 FILE UPLOAD POP UP PAGE 5 Lastly return to step 7 for Patient entry in OpenClinica and complete the Images Upload Completed CRF Project STH Session mr 020914 trytest Destination Prearchive Archive in quarantine O Archive no quarantine File Choose File WRIXzip Upload Loading File Upload 100 Extraci Review 10096 Successfully uploaded 1 sessions to the prearchive The following sessions have been uploaded 1 sessions s has been moved to the pre archive FIGURE 29 FILE UPLOAD COMPLETED Housekeeping If you lose your OpenClinica or XNAT passwords email kevin teh Osth nhs uk for a new password 36 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and 2 08 05 2015 APPENDIX 2 EXCEL SPREADSHEET REPRESENTATION OF A CRE ES Microsoft Excel Subjects Eligibility xls ia Elle Edit View Insert Format Tools Data Window Help fe Inc_
26. assword gt lt wsse UsernameToken gt lt wsse Security gt soapenv Header FIGURE 10 CORRECT AUTHENTICATION HEADER REQUIRED B Y OPENCLINICA wsse Nonce EncodingType http docs oasis open org wss 2004 01 oasis 200401 wss soap message security 1 0 Base64Binary gt f8nUe3YupTU5ISACy3X9Gg lt wsse Nonce gt wsu Created 2011 05 04T19 01 40 981Z wsu Created FIGURE 11 EXTRA INFORMATION SENT BY WSE3 0 NET Not being able to negotiate a secure connection to the SOAP web services essentially meant we could not use them for the project This required an alternate approach discussed in the next section 6 2 1 2 REST Web Services Due to the incompatibility of the preferred OpenClinica SOAP solution with our Microsoft based development stack we decided to pursue the REST solution instead As described earlier we do not usually favour REST interfaces due to the lack of tooling to allow automated checking of interfaces etc This is also compounded by the immaturity of the current implementation in OpenClinica which means we would expect to have significant refactoring work to perform if we move to subsequent versions of the product However we do not anticipate an upgrade in the life of the project so the risk was deemed manageable The first issue we encountered was that the REST web services interface did not work before version 3 1 4 This required us to have to upgrade our OpenClinica version to 3 2 In OpenClinica the R
27. ated by General Inclusion CRFs in this Study Event Exctusion Criteria CRF Name Version Status Initial Data Entry Double Data Entry Actions General General v1 0 f Physical a Examination iteria 1 0 za Neurological xamination v3 0 lt a Exam c2 Medical Histo V4 0 m Lifestyle Health chin T c2 Qustionaire Lifesty ionaire V1 0 va Medical History Medical Medical Questionaire 1 0 lt 2 Questionaire 1 lood Si rpie V1 D lt 2 Blood Sample Lab Lab Analyses blood sample v1 0 lt 2 Analyses blood CS v1 0 lt 2 sample J csr Neuropsychological and Neuropsychiatric Tests v1 0 lt 2 Table of Modality v1 0 d Neuropsychological and Neuropsychiatric Images Upload Completed v3 0 a Tests Add Subject To XNAT 1 0 2 C39 E5 EB 20 53 Ey Ey EB La EA Ey 13 ES Table of View this Subject s Record Modalit y Workflow Add Subject To XNAT FIGURE 19 CRF DATA ENTRY PAGE 5 Fill in all the CRFs contained in the enter data page This is done by clicking on the pencil enter data tab on the same row as the CRF name you want to fill under the Actions column Most of them have validation rules associated with them and OpenClinica will throw up an error at the top of the CRF in red if 1 e a field was not filled When a CRF is completed click Save There is a Mark As Complete tab on top of the CRF next to the Save button Only do this if you are sure all the data you have entered is correct
28. bject To XNAT CRF as shown in Figure 21 For question 1 select Yes then for question 2 key in your initials and then for question 3 enter the date today Also 1f you want you may also upload your subject informed consent When you are done click Save On the first Save click there will be a red warning prompting you to create a new subject in XNAT circled in Figure 21 On seeing this click Save again and you will again return to the enter data page s FP7 601055 VPH DARE G IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 Add Subject To XNAT v1 0 STHManual Y CRF Header Info There are issue s with your submission The data has NOT been saved See below for details AddSubj 0 4 Title AddSubject Page C Mark CRF Complete TS Gi E Are pati B 1 Are you creating a I po subject in XNAT 2 Verification h Affirmation Initials t 3 Date 26 Sep 2014 Ea Please scan the signed informed consent and attach by uploading the PDF below Attached informed Click to upload file yo consent O mark RF Complete ETT GENE 0 FIGURE 21 ADD SUBJECT TO XNAT CRF 9 Lastly fill in the Images Upload Completed CRF Here there is a link that will take you to the XNAT site where the subject will have been created It may take up to a couple of minutes for your new subject to be created Click on the blue link called XNAT LINK and this will open a new tab wit
29. criteria3 Type a question for help m3j3 wo uide Inc DoB Date Of birth div ID DateOfBirth gt Date Of Birth lt div gt Inc VisitDate Visit Date div ID OtherDate Visit Date lt div gt Inc Age Age div ID CalculatedAge gt Age lt div gt lt script src ajax googleapis com ajax libs jquery 1 9 1 jquery min js gt lt script gt Eod _5 Inc ageRange Age range Age range Subject is 50 85 years old inclusion criteria for all 6 Inc Group Grouping Study group If HC or MCI selected Press ENTER after making _ 7 Inc_MSSE MSSE gt 27 MSSE gt 27 Inclusion criteria for controls 1 _8 Inc CDR CDR 0 CDR 0 Inclusion criteria for controls 2 9 Inc memory No memory complaints No memory complaints Inclusion criteria for controls 3 10 Inc HC HC inclusion criteria fulfilled HC inclusion criteria fulfilled Inc cognitive Referral because of cognitive Referral because of cognitive impairments Inclusion criteria for MCI 1 EN impairments _12 Inc MCIdiagnosis MCI diagnosis MCI diagnosis lt a href javascript ivoid 0 onmouseover Tip In 13 Inc MCIcriteria MCI inclusion criteria fulfilled MCI inclusion criteria fulfilled _14 Inc ADdiagnosis AD diagnosis AD diagnosis lt a href javascript ivoid 0 onmouseoverz Tip In _15 Inc AEcriteria AD inclusion criteria fulfilled AD inclusion criteria fulfilled 16 Inc VADdiagnosis VAD diagnosis VAD diagnosis a href javascript void 0 onmouseover
30. e and this deliverable will describe in detail the infrastructure components put in place to support these data collection activities and also their integration with the rest of the data network 1 BACKGROUND The proposal outlined two possible solutions for clinical data capture and management available to the project the first being a product produced and sold by Sheffield Teaching Hospital NHS Foundation Trust called ArQ This is a complete clinical application development system which can be used to produce a highly integrated patient management system within the clinical environment It is used extensively in Sheffield and is now also a component in several other clinical trials type scenarios several of which are EU funded projects There is no question that this system could be used to collect and manage the data associated with this project however it is not open source and it is not freely available Whilst STH would be happy to allow its use free of charge for the duration of the project there would have to be a funding stream to pick up the commercial aspects of continuing this system following the end of the project In addition the real advantages of the ArQ system 1 e its large collection of components for integrating data within the clinical environment were not really necessary when we analysed the processes that the research teams would put in place to collect the data Finally given the project philosophy of wanting to create
31. e contact kevin teh Osth nhs uk for an account Initial Visit 1 Login to the site using the username and password provided This will take you to the welcome page as shown in Figure 16 below On the menu bar click Tasks then under Submit Data click Add Subject VPH Dare Clinical Data Repo STH 1 y t root Home Subject Matrix Notes amp Discrepancies Study Audit Log Tasks ITA MES INS tudy SubjectID ICJ Submit Data Welcome to VPH Dare Clinical Data Repos Subject Matrix Schedule Event C Add Subject View Events OpencCiinica Alerts amp Messages Welcome to OpenClinica Root User You last Notes amp Discrepancies Assigned to Me Ht logged in on 26 Sep Notes amp Discrepancies Import Data 2014 Monitor and Manage Data Subject Enrollment By Site Source Data Verilication Site Enrolled Expected Enrollment Percentage ANM Study Audit Log STH 111 200 C E Instructions Extract Data If needed you may change View Datasets gt k Study Progress the study site or request Study Setup scheduled 3 2 Administration data entry started 126 90 x Studies oe 4 completed Users LRFs Other signed locked Update Profile skipped stoppe d FIGURE 16 HOME SCREEN AFTER LOGIN 2 Fill in the Study Subject ID it is very important to get this entry correct the Sex of the subject and the date of enrolment in the study Then click on Save and Assign Study Event as shown in Fi
32. elis ie ItemOID ItemOID uibem IE NND ItemGroupRepeatKey ItemOID ItemOID FormOID OpenClinica Version a ItemGroupOID OpenClinica Status ItemGroupRepeatKey ItemOID ItemOID ItemOID ItemOID SubjectKey OpenClinica StudySubjectID OpenClinica EnrollmentDate StudyEventOID StudyEventRepeatKey StudyEventOID StudyEventRepeatKey StudyEventOID StudyEventRepeatKey StudyEventOID StudyEventRepeatKey OpenClinica Status OpenClinica StartDate OpenClinica StartDate OpenClinica StartDate OpenClinica StartDate SubjectKey OpenClinica StudySubjectID OpenClinica EnrollmentDate StudyEventOID jene ZEE TONS EST S STA StudyEventOID ocudyEventRepeatKey StudyEventOID xs evil mss e pe cts EM StudyEventOID StudyEventRepeatKey OpenClinica Status OpenClinica StartDate OpenClinica StartDate OpenClinica StartDate OpenClinica StartDate SubjectKey OpenClinica Sex StudyEventOID StudyEventRepeatKey StudyEventOID DISUICK ENC Misc wc cheno StudyEventOID StudyEventRepeatKey StudyEventOID StudyEventRepeatKey OpenClinica StudySubjectID OpenClinica EnrollmentDate OpenClinica StartDate OpenClinica StartDate OpenClinica StartDate OpenClinica StartDate 38
33. enClinica SecondarylD 12344 OpenClinica Sex m OpenClinica EnrollmentDate 2014 11 04 gt lt StudyEventData Stu ucymyentOID SE_INITIALYISIT OpenCclrvnscaestuartDbdte 2014 11 04 00 01 0020 OpenClinica Stakus data entry Started StudyEventRepeatKey 1 gt lt FormData FormoOi ERAS UB OpenClinica Version v1 0 OpenClrnvca Status rnrtral data entry gt lt ItemGroupData CtemGSroupOTD rc ADDSU UNGROUPED ItemGroupRepeatKey 1 TransactionType Insert gt lt ItemData IcemOrD T ADDSU IMS bE valine 00 fe lt ItemData ItemOID I ADDSU IMS DATE Value 2014 11 05 lt ItemData ItemOID I ADDSU IMS INIT Value kt lt ItemData ItemOID I ADDSU IMS CONF Value 1 ItemGroupData FIGURE 4 AN XML FRAGMENT SHOWING THE SUBJECT KEY MANGLING FOR IDS LONGER THAN 8 CHARACTERS To resolve the issue of the 8 unique characters an additional REST query has to be made A REST query call to for all subjects in the study is shown below heros vendare Ore Slaven cle le CoemClitonee ese eli occi scully uo de a AES Here we searched the StudySubjectID for the entire study S VPH to retrieve their respective SubjectKey After obtaining the correct SubjectKey we then used that information to obtain values from a CRF called Table of Modality The event data specific XML is as shown below lt FormData RormOTD E I ABLEORMODAL WACO Open Limca Versiones vee Opencdunsesssbeabus unararsal data entry gt
34. equired field Study Subject ID STHManual Instructions Study Event Definition Initial Visit non repeatina vl Other Info efTime 26 2014 i w o v Study VPH Dare Cl nica A lt DD MMM YYYY HH MM Pe T w v End Date Time O DD MMM YYYY HH MM P Site Leave this field blank if the end date time is not applicable Study Subject ID THfManua Start Date N A End Date N A 5 edule Another Event opt 1 Proceed to Enter Data PI om Protocol Verification IRB Approval Date FIGURE 18 SCHEDULE STUDY EVENT 4 This will bring you to the Enter or Validate Data for CRFS in Imaging Visit enter data page page shown below in Figure 19 Now you have created a subject in OpenClinica Um FP7 601055 VPH DARE G IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 VPH Dare Clinical Data Repo STH OpenClinica 7 Home Subject Matrix Notes amp Discrepancies Study Audit Log Tasks Coat tT Study Subject ID CJ Alerts amp Messages The study event with Enter or Validate Data for CRFs in Initial Visit O definition Initial Visit and subject STHManual was created successfully Study Subject ID Study Event Instructions Location N A Info Study Subject OID SS STHMANUA Study Events Start Date 26 Sep 2014 Study Events 1 3 Initial Visit End Date Time Subject Event Status scheduled Status Last Upd
35. er sitting in front of Tomcat and is proxied using the standard jk connector components that ship with apache A CSR was created using OpenSSL allowing another Terena SSL certificate allowing for secure server site connections The images are stored off the server on a local NAS drive which currently has 12TB of storage configured as a single RAID6 volume with two hot spare disks note this is not the same storage as 1s used by the VMWare environment for hosting the virtual servers described above 3 1 4 Backups The backups for these two servers are purely for the databases given that the actual imagestore for XNAT is not on the virtual machine itself anyway Bash scripts were produced to run the backups nightly via a cron job and the backed up files are shipped off server to a local NAS on the local network In addition these backups located physically in the same room are replicated to the central university data centre 1km away from the operational servers In addition we keep a rolling backup on a nightly basis for 7 days so if necessary we can restore to any point in the last 7 days The images stored by XNAT are also backed up to this data centre although it is not possible to keep rolling snapshots of a storage volume this big so it is simply a pure married backup of the file system that 1s replicated The only step we have not implemented yet is transaction log shipping off site so that in the event of a catastrophic database event
36. es 6 ZN OPENCTINICA EVALUATION curacion alison lomera 6 2 2 ANNATE ROA du LPs ea a iiem 6 2 9 ARCHITECTURE SOLUTION succionar eiii 7 3 SYSTEM IMPLEMENTATION coito 8 3 1 SS cena leuneaawhecceaceieuaaoneenasenulecmeaneiecsnas 8 3 1 1 VMWare environment is S 2 4 2 OpenClinica S CIV EP ias 9 3 1 3 ANAT Aids 9 3 1 4 A e A 10 3 2 OCX INTERFACE IMPLEMENTATION cccsceccsceccececcececcsceccecscescscescscecescecescscescecscscesescusences 11 4 OPENCLINICA DATA ENTRY AND CONFIGURATION coccccccccoccnconccnccoccnccoconconcccncciconos 14 5 XNAT DATA ENTRY AND CONFIGURATION cccccoccocccccccccconccnccnccocononococoncnccnococcnccccnoconcnos 18 6 INTEGRATION WITH RESEARCH PLATFORM ui ccccccsccccccccccccccccccccccccsscccccccscees 18 6 1 REQUIREMENTS FOR A DATA SOURCE PLUGIN ccccsccsscsscccccsccsscsccescscescessescessesccscescesceses 18 6 2 OPENCLINICA INTERFACE cccsceccsceccscescsceccsceccsceccsceccecscscscscscescscesescecescusescecescscescscesences 19 6 2 1 OpenClinica Web Service Integration eese nennen nennen eene 19 6 2 2 OpenC nica data modo 21 6 3 DPS PLUGIN DEVELOPMENT 5 erre omro sra Ev or ninr anniari Po RU PER oC Suo PP FS PEDPePrE SE VUPP KU PER POP ia 23 FEE D S LE MU e E 25 APPENDIX 1 USER MANUAL FOR PROSPECTIVE CLINICAL TRIAL eee eee eese 26 APPENDIX 2 EXCEL SPREADSHEET REPRESENTATION OF A CRPF 00 37 APP
37. f OCX Integration service fields whilst project name VPH and subject name ID man88 were obtained from the database queue After creating a patient and the associated experiments we also did a verification check if this step has been successful This is done before the database queue gets updated to Completed by using a call as shown below XNATRestClient host https localhost u username p password m GET remote data archive projects VPH subjects man88 experiments format csv If the subject exists in XNAT the OS X integration service would flag the transaction as completed otherwise it would set a Failed status in the queue Active monitoring of the queue is provided by the nagios system http www nagios org which checks the queue every three minutes and will email the systems administrators 1f any failed messages occur These emails will then be actioned and the users contacted to ensure the case can progress unimpeded 4 OPENCLINICA DATA ENTRY AND CONFIGURATION Following the integration of OpenClinica and XNAT we focused our attention on the creation of CRFs for clinical data entry Data entry for most prospective data collection groups will fall into clinical CRF data entry and imaging data entry Using OpenClinica the clinical team will register the subject and fill in all the clinical CREs ranging from subjects eligibility to neurological test results The imaging team will then fill in the last 3 CREs of course in man
38. ge 7 NEXT STEPS The platform built is very flexible and as is often the case we anticipate a number of developments to take place over the coming 12 months as people start to use it In particular we expect that the XNAT system will need configuring to take more complex data types than the simple DICOM used for the prospective trials outlined in WPl These developments will be taken on a case by case basis once representative data is available We will continue our open dialogues with WP1 and WP2 in order to meet their needs and do not anticipate any significant further development of this service just small reconfigurations as the project requirements evolve 2952 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 APPENDIX 1 USER MANUAL FOR PROSPECTIVE CLINICAL TRIAL Introduction This manual describes the process of registering a subject in OpenClinica and then uploading subsequent images via XNAT Prospective imaging data collection is a two step process Firstly a subject has to be created in OpenClinica which will also host all the clinical data entry CRFs Next imaging data upload will be hosted in XNAT after the patient creation in OpenClinica Patient Creation and CRF input in OpenClinica The site for OpenClinica is https vphdare oc shef ac uk OpenClinica You should have already been sent credentials for accessing the system if you have not pleas
39. gure 17 96 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 VPH Dare Clinical Data Repo STH 12345 root Data Manager er OpenClinica n Community Edition Home Subject Matrix Notes amp Discrepancies Study Audit Log Tasks Study Subject ID Go Alerts Messages M Instructions STH Add Subject O Other Info Study VPH Dar indicates required field Study Subject ID STHManual me Start Date N A Secondary ID End Date N A Date of Enrollment for Study VPH Dare Clinical 26 Sep 2014 PI Teht Data Repository Protocol Verification IRB Sex Mie Approval Date Po ELITLTTESULTTS LIE ATI TI Workflow Add New Subject FIGURE 17 ADDING A SUBJECT 3 This will bring you to the Schedule study for subject page as shown in Figure 18 Study Event Definition should be set to Initial Visit non repeating Set the Start Date Time as the time of data entry date and time now Do not fill the field End Time as this 1s not required Then click on Proceed to Enter Data of VPH Dare Clinical Data Repo STH 12345 OpenClinica ty Edition C Commun Home Subject Matrix Notes amp Discrepancies Study Audit Log Tasks v jig GT eat Study Subject ID Go Alerts amp Messages The Subject with unique Schedule Study Event for STHManual 9 identifier STHManual was created successfully indicates r
40. h the XNAT login page Next do the XNAT steps explained below and return to this step to verify that you have finished image upload by answering the two questions in this CRF 30 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 Add Subject To XNAT v1 0 STHManual CRE Header Info There are issue s with your submission The data has NOT been saved See below for details Create New Subject in XNAT AddSubj 0 4 Title AddSubject Page Ol mark cer Complete MENTIS a Gy LAA iu subject in XNAT 2 Verfication KT h Affrmation Iretisis gt Dake 26 Sep 2014 a Please scan the signed informed consent and attach by uploading the PDF below Attached informed Click to upload file consent O mark CAF Completo TT NEN x OG FIGURE 22 ADD SUBJECT TO XNAT CRF WITH HYPERLINK TO THE XNAT WEB PAGE Follow up Visits and Subsequent Follow up Visits 1 Again login to OpenClinica with your credentials 2 Next click on Subject matrix under Tasks and Submit Data as shown in Figure 23 This will take you to a page displaying the entire subject in this study site Search for the patient you want to enter Follow up data for This 1s done either by scrolling through them or search for the name by typing keywords below the Study Subject ID tab BL FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on p
41. her Occurrence as shown in Figure 24 us FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 Scheduled sth6829 Data Entry Started sth9102 Stopped sthbbc123 Skipped STHJaneSmith Completed STHJoeSmith signed STHJohnSmith Locked sthmale199 Invalid STHManual 22222020222 BBBBBEBSBS E a Ee Ea Actions sTHTestYou Subject STHManual Event FollowUp Yisit View STHWolTest Edit Occurrences1 of 1 stream123 29 Sep 2014 Status data entry sub0001 started Remove Restore cm sub00016 Reassign sub00017 5go Sign sub00018 1 E E sub00019 FIGURE 24 ADD ANOTHER OCCURRENCE FOR FOLLOW UP VISIT Also if for whatever reason you were timed out idle timeout 30mins do steps 1 and 2 as described by the Follow up Visits and Subsequent Follow up Visits instructions above And to continue where you were select and click on the appropriate visit column and then click on View Enter Data OpenClinica Quirks CRF Subjects Eligibility After keying in date of birth and visit date press TAB CTRL to automatically calculate the age and fill the Age field If someone intends to enter a multi select value as shown in Figure 25 below hold ctrl and click on the next option Ja s FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2
42. implement data workflows in the system 1 e users are presented with filtered lists of eCRFs for completion based on data already entered into the system Finally whilst the interfaces for image access are based on well established standards the model for data collected into the forms engine is not and proved very difficult to produce a standardised solution for A complete description of this issue can be found in Deliverable 3 1 as this problem needed to be addressed as part of the retrospective data provisioning The solution presented there whilst effective is not one we would choose to pursue on a routine basis 2 3 ARCHITECTURE SOLUTION The two systems evaluated are excellent in their intended domains and indeed both offer some functionality across the full solution space the project needs to cover but neither would be sufficient in their own right as the final application stack For this reason we decided to couple the two systems together to achieve a best of breed solution for the clinical users FP7 601055 VPH DARE G IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 The task of coupling these two systems together was not trivial Whilst they are both open source solutions 1f we modified any of the code base especially within OpenClinica we would invalidate all the testing and certification the products have gone through At the same time the workflow in a clinical trial data e
43. iteria 1 T W Label Hover Mouse here 19 Exlusion Criteria 2 T P Other causes of dementia MSSE 21 MSSE Completed v fe CDR 24 CDR Completed Pu Revie 27 All inclusian and m exclusion criteria reviewed 38 Subject Fulfills all Y m inclusion and exclusion criteria 29 Date of inclusion and fs n exclusion review Return to top L Mark CRF Complete Save O i3 FIGURE 8 A TYPICAL CRF DATA ENTRY PAGE This section 1s intended to give an overview of what has been delivered to the project Appendix 1 has the full user manual provided to the clinical teams which contains a Er zn FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 very detailed view of the CRF s within the system and also the user view of uploading the images into XNAT 5 XNAT DATA ENTRY AND CONFIGURATION Since XNAT is essentially a sub system to OpenClinica for user interactions and our configuration challenges are significantly less than those with the OpenClinica system however there are still some constraints on its use The user interaction and interface is described in the prospective data entry manual Appendix 1 at the end of this document We have chosen the upload format to be DICOM This is because DICOM is a common format easily obtained from most imaging modalities indeed the only format available on most clinical imaging modali
44. ition to the standard DICOM transfers but these are not used extensively at present e Segregation of images within the system e g image collections are separated by project and each project has its own set of users who can only see the relevant images This kind of role based access to studies in a PACS system is not present in most systems e It supports access to the image file collections through protocols like WebDAV which makes it simpler for researchers to access use the data than the traditional DICOM interfaces which are very complex e Has an extended metadata model which is focussed on curation and validation so checks of image quality and suitability for analysis tasks can be stored alongside the images in the database e Has the ability to store arbitrary file collections attached to the subjects record and whilst these are not indexed in to the same level as the imaging studies this feature provides a useful addition for non standard file storage e Itnow has a feature to allow administrators to create data collection forms The last feature is where we found the system s main shortfall in fact it does have the facility to extend the standard metadata model but this is very complex to implement and not easy to update In addition it does not have some of the more advanced features of OpenClinica allowing record locking and sign off which are often key requirements in clinical trial scenarios Also we did not find a way to
45. lincal metadata schema onto the underlying database SOMA ea 22 Figure 13 The mapping model for the OpenClinica data schema onto the underlying database SCIO AE UU q IO E Mp MM ER dee EM MIN EE EIE EE PEE 22 Figure 14 Screenshot of user interface for DPS data extraction and explorati0n 23 Figure 15 Prospective trial schema when published ococcccccncnnnoncnnnnnnnnnnnnnnninnnonininoss 24 Figure 16 Home screen after login oooccccccnnnnnnnnnnnnnnnnnononnnnnnnononnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnonoss 26 Figure 17 Adding SUD Elina 21 Fionre 15 Schedule Study CV ene p 27 Figure I9 CRF data entry page serenidad 28 Figure 20 Table of Modality CRF susi ia 29 Figure 21 Add Subject TOXNAT CRPE nuneiesecten eet sit ereere nAn ESEE n EAS ENa Ei 30 Figure 22 Add subject to XNAT CRF with hyperlink to the XNAT web page 3 Figure 23 Subject matrix follow up data entry oooccccccnnnnnnnnnnnnnnnnnnnnnnnnnononononnnnnnnnnnnnnnnnnnnss 32 Figure 24 Add another occurrence for follow up VISIt cocccccccccnnnnnnnnnnnnnnnnnnnononononinininnnos 33 Figure 25 Multi Select example oococcccccnnnnnnnonnnnnonononnnnnnnononononnnnnnnnnnnnnnnnnnnnnnnnnnnnnnninanoss 34 Figure 26 XNAT homepage select patient to add imaging SCanS cccccccccnnnnnnnonnnnnnnononononos 35 AAA 35 Figure 28 File upload pop up PAC cccccccccnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnonn
46. nment based on the VMWare ES Xi 5 5 0 product This hypervisor 1s freely available although not open source and 1s installed on the following Hardware Model DELL PowerEdge R815 Processors 32 CPU x 2 6 GHz Processor Type AMD Opteron TM Processor 6212 Hyperthreading Inactive Total Memory 255 97 GB FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 Number of NICs Disks BigDisk Capacity 10 91 TB 8 22 TB free datastorel Capacity 128 50 GB 126 60 GB free The system has two disks one is the internal 128GB capacity RAIDI mirrored operating system disk The second is the external direct attached storage array which comprises 12 x 1 5TB disks configured in a RAID6 configuration across 10 disks with two hot spares This server 1s hosted in a server room in Sheffield Teaching Hospitals but connected to the University of Sheffield s network so it has a direct link to the UK academic high bandwidth network JANET which should ensure that hosting connectivity is not a bottle neck for large volume data transfers 3 1 2 OpenClinica Server This is a virtual machine running Ubuntu 14 01 and the virtual resources allocated to it are as follows Processors 4 CPU x 2 6 GHz Processor Type Virtual Hyperthreading Inactive Total Memory 5 12 GB Number of NICs 1 Disks Disk1 Capacity 100GB The OS installed is Ubuntu 64 bit 14 01 platfo
47. nnectors and go onto the specifics of how this was implemented for the OpenClinica system 6 1 REQUIREMENTS FOR A DATA SOURCE PLUGIN Since any data source used in the platform must ultimately be converted to a relational database structure MySQL the nature of any plugin is to handle this transformation where necessary To this end the two core requests a plugin must handle are e GetSchema Contains a list of tables with their fields and data types plus any relationships that exist between them e GetData TableName Returns a DataTable with the contents of the requested table Each table is called sequentially until the whole set is produced In most sources this 1s a natural trivial mapping onto the underlying data structures but in the case of OpenClinica this 1s not the case A further desirable property 1s that the GetData method returns quickly as this will often be used extensively in the interactive development of the de identification template 18 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 6 2 OPENCLINICA INTERFACE Fortunately the data model implemented in OpenClinica 1s based on a snowflake model as are most automatically generated database applications and this maps naturally onto the research platform hosting and query services again these are defined in detail in D3 1 Figure 9 shows the design of this model FIGURE 9 DAT
48. nnnnnininoss 36 Figure 29 File upload completed ococcccccccnnnnnnonnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnononnnnnnnnnnnoss 36 Figure 30 CRF Export to spreadsheet ooocccccccccccnnnnnnnnnnnnnnnnnnnnnonnnnnononononnnnnnnnnnnnnonnnnnnnnnss 37 FP7 601055 VPH DARE G IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 INTRODUCTION VPH DARE IT is a hugely ambitious scientific endeavour seeking to integrate information from across many domains to provide a systematic multifactorial and multi scale modelling approach to understanding dementia onset and progression and enable more objective earlier predictive and individualised diagnoses and prognoses of dementias to cope with the challenge of an ageing European society The number of participants involved in the project and their disparate needs for scientific analysis results in a very complex collection of requirements The two key components in an infrastructure to support such an activity are data storage and compute services and these two facets of the platform need to be integrated seamlessly to deliver a coherent solution A further separation in the data provision for the project 1s that of retrospective data collection available both publically and through consortium partners and prospective data collected during a targeted clinical trial executed within the project Deliverable 3 deals with the former subtyp
49. ntry process is often such that the more common integration strategies of scheduled tasks checking for new data and updating another system are not viable The solution needed has real time updates or very close to 1t in terms of data content but does not require any modification of the core products on either side Fortunately there 1s a standard feature of OpenClinica that can be attached to the completion of any eCRF in the system and that is an email notification to someone typically the study co ordinator on subject recruitment or completion It seemed possible to trap these triggered events from OpenClinica use the content of the messages to interrogate the system and extract the data necessary to create an imaging record in XNAT which the researchers can then upload images into This process will be elaborated on in more detail in section 0 but in summary it gives us the ability to slave the XNAT system to OpenClinica in near time to reduce the errors of managing data in two separate systems We have called this component the OCX Integration Service The high level structure of the system 1s shown in Figure 1 OpenClinica SMTP Message Sere Create Imaging Integ ration record Information request Service Clinical data entr Image upload FIGURE 1 STRUCTURE OF DATA COLLECTION SYSTEM 3 SYSTEM IMPLEMENTATION 3 1 HOSTING 3 1 1 VMWare environment The hosting for the data collection platform is on a virtualised enviro
50. r tip and untip javascript function and also putting the XNAT link using html tag 37 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and 2 08 05 2015 APPENDIX 3 EXAMPLE OF CDISC ODM DATA DOCUMENT version FileOrD xmins FileType encoding Description CreationDateTime ODMVersion xmins OpenClinica xmins OpenClinicaRules KENS AX SA OID OpenClinica OpenClinica OpenClaimacas OpenClinica OpenClinica xsi schemaLocation StudyOID StudyOID 2 yIE Heel ois ci IS Scarus Status 3 SEE EDI TransactionType Transa Cu one TransactionType OpenClinica Opene linea OpenClinica OpenClinica OpenClinica OpenClinica OpenClinica OpenClinica OpenClinica OpenClinica OpenClinica OpenClinica OpenClinica OpenClinica OpenClinica Sex otarus Status BSEC Status Sex 3 SEEDS otacus rte cites BSEC Status Status OLETUS Status Status MetaDataVersionOID SubjectKey 9j em elm ates cos StudyEventOID StudyEventRepeatKey StudyEventOID StudyEventRepeatKey StudyEventOID StudyEventRepeatKey StudyEventOID StudyEventRepeatKey OpenClinica Version ItemGroupOID OpenClinica StudySubjectID OpenClinica EnrollmentDate OpenClinica StartDate OpenClinica StartDate OpenClinica StartDate OpenClinica StartDate OpenClinica Status HOMO Tl amp e Im ess OS ENG c
51. rm This OS has also been patched for heartbleed and updated against any other bugs in the Linux system 4 CPUs have been provisioned for it with a total hard disk space of 100GB OC was installed with Tomcat version 7 0 53 and Java7u65 The database engine used was PostgreSQL 9 3 5 This machine also has firewall exemption rules written to allow connections on port 443 and 22 only A local Certificate Signing Request CSR was created using the standard tomcat keytool and a Terena SSL certificate was imported allowing secure connection to the OpenClinica site 3 1 3 XNAT Server This is a virtual machine running Ubuntu 14 04 and the virtual resources allocated to it are as follows Processors 2 CPU x 2 6 GHz Processor Type Virtual Hyperthreading Inactive Total Memory 5 12 GB Number of NICs 1 Disks Disk1 Capacity 120GB FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 For this component we decided to use a standard virtual XNAT machine provided by the www xnat org site and changed all of the security settings This server has a 64 bit Ubuntu 12 04 OS and was installed with Tomcat version 7 0 26 and Java7u51 Again this machine also has all the relevant updates and patches for security updates The database engine used was PostgreSQL 9 1 14 The server also has firewall exemption rules opening only port 443 and 22 to clients XNAT has the Apache web serv
52. rospective data in WP1 and2 08 05 2015 VPH Dare Clinical Data Repo STH 1 OpenClinica E Community Edition BITE ETA co Report Issue Support EVEN co Alerts amp Messages Instructions Subject Matrix for STH O Info Icon Key COEBESCB is 9 show More Select An Event Add Ne Reb Study Subject ID Initial Visit FollowUp Visit Actions 3l Not Started slasher08 N Scheduled Data Entry Started sth0006 j slasher09 Stopped sth1029 3 Skipped sth1580 i Completed sth6666 3 signed sth6829 Locked sth9102 3 Invalid sthbbc123 a 00802809 2 A STHJaneSmith 5 View STHJoeSmith 3 Edit STHJohnSmith Remove sthmale199 Restore BGGGegagcgaagaagagea STHManual Gl BESSGEEEEEEEESS JBBBBEBESBBEBSB Reassign eBgg60enE STHTestYou Subject STHManual Sign Event FollowUp Visit Results 46 60 of 113 Occurrence 1 of 1 Status not scheduled FIGURE 23 SUBJECT MATRIX FOLLOW UP DATA ENTRY 3 Once you have found your subject click the blue icon Not Started on the Follow up Visit column then click schedule Now do steps 3 9 from the Initial Visit steps above One important difference is to setting Study Event Definition should be set to Follow up Visit non repeating in step 3 of Initial Visit instructions above 4 For subsequent Follow up Visits do steps 1 3 above but instead of a blue icon click on the yellow Data Entry Started icon then click on Add Anot
53. s password and user authentication security electronic signatures SSL encryption de identification of Protected Health Information PHI and comprehensive auditing to record and monitor access and data changes It also has a fully validated software development lifecycle SDLC These features distinguished the system over all other open source solutions for web based data collection we considered There is an extensive feature list of the system located in the following URL if further information is required https www openclinica com product features The way in which the forms are designed is intended to be simple enough for non technical people to design a system To achieve this the platform uses a Microsoft Excel based spreadsheet model where the clinician user simply adds data elements with some simple definitions of the data types and description etc and this when deployed produces a set of web pages to capture each of these data items Many examples of such forms will be shown throughout the rest of this document Key features of OpenClinica are Open Source license Web based Supports all types of clinical studies No programming IT knowledge needed for CRF design Built on leading independent standards Significant support for data validation and sign off The main deficiency of OpenClinica is that at the time of evaluation it did not have the capacity or infrastructure components to manage large file uploads which on an im
54. this instance the service is installed on the same server as OpenClinica but this is not a requirement and could run on an independent server if required Relational Database MySQL Ubuntu 14 04 Ubuntu 12 04 ESXi Server FIGURE 2 SCHEMATIC OVERVIEW OF OCX INTEGRATION SERVICE The first stage of the OCX data integration service was to obtain the SMTP message using an email listener in MONO This is then placed into a MySQL database queue This database records the following e ncremental number Auto generated for processing e study name From email e sitename From email e subjectname From email e eventname From email e current status Current processing status of this transaction Subject registration in XNAT only utilises the sitename 1 e 5 VPH and subjectname i e MAN88 When a new SMTP message is received the relevant components of this message are parsed and placed into the database queue with a status of New When the message processing thread next queries the queue the oldest item with the New status will be elle FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 changed to processing This state is where most of the OCX integration service work takes place In summary as shown in Figure 3 relevant data will be fetched from the database queue for the currently processing transaction These together with a further web services call to
55. ties DICOM also has a significant amount of metadata associated with 1t which makes it possible to perform automated validation of the uploaded 1mages against the expected clinical records Hence images will first be stored in XNAT prearchive before being archived fully in to the imaging repository This is done so that images can be verified by a person before being fully committed into the database This final validation step helps ensure the correct imaging has been attached to the correct patient since this 1s the most likely cause of data error in the collection system 6 INTEGRATION WITH RESEARCH PLATFORM The data publication suite DPS is the name of the software component used to connect to a data source process the data in some way annotate the data and then publish the results to an internet accessible server for consumption by the research community The tool is designed to be used by scientists as opposed to data managers in the hope that they can take responsibility for managing their own data resources There is also a significant advancement in this software over other tools which perform a simple ETL Extract Transform Load function and that is the integration of semantic annotation into the published data framework A detailed description of this software can be found in Deliverable 3 1 so it will not be repeated here but to give this development work some context we will describe the process for creating new data co
56. tudies based on prospective data in WP1 and2 08 05 2015 Following the schema analysis phase we are left with the relationship model shown in Figure 15 This is quite a lot larger than initially anticipated comprising of 47 individual tables 789 individual data items Whilst the size 1s not a problem technically the concern is that with an information model this complicated the researchers who wish to find and use the data within the system will not be able to find it effectively At the point of writing there is not a significant amount of data entered into the electronic system as it 1s still undergoing the final iterations in terms of CRF design However we do have several techniques for dealing with these issues firstly by default any tables or fields which are empty do not get published to the researchers and whilst this should not be the case in a clinical trial we do see sometimes that operational issues lead to some items being dropped once recruitment has reached full pace Secondly and probably more valuable is the feature of the DPS and research platform that allows only a subset of the source data to be published for any specific use So we might imagine that there would be several different versions of this data containing different classes of data e g genetic lifestyle or biochemistry These datasets would have their own access policies and allow individuals to have a tailored view of the data to ease their exploration and usa
57. we will be able to restore data back to within a few minutes of the failure as opposed to the previous night as it is the case now The reason for this is primarily the cost of implementing such systems and maintaining them is often quite high and the risks of data loss are very low This is due to the fact that all the clinical data in OpenClinica is captured to paper CRF in the clinical service and then transcribed into the database so there 1s always the source data available if it needs to be re entered This along with the fact that the recruitment rate in this project is not high means that the manual recovery procedure is far more resource efficient The XNAT database can be reconstructed from the images so this 1s not a high risk either We are comfortable for the moment that the processes we have in place are commensurate with the risks we have identified but we will periodically review this situation and may choose to implement more sophisticated protocols in the future 10 FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 3 2 OCX INTERFACE IMPLEMENTATION Here we will describe in a bit more detail the interface service produced to couple OpenClinica with XNAT Figure 2 shows the schematic overview of how the service is implemented The service itself is written in Microsoft Dot NET and runs under the mono http www mono project com framework on Linux In
58. y places this might be the same team but it does not need to be A FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 OpenChnica Subject FIGURE 6 DIAGRAM SHOWING THE CRFS FOR A SUBJECT VOLUNTEER FOR DIFFERENT VISITS In the design we used a feature in OpenClinica allowing for repeating and non repeating events Here we set the initial visit to occur once non repeating and set up the subsequent Follow up visits repeating to be continuous There are initially 14 CRFs to be completed one of which is shown in Figure 8 These are comprised of 11 clinical based CRFs and 3 imaging CRFs These are titled Clinical CRFs Subject Eligibility Study Visits Basic Information General Physical Examination Neurological Examination Lifestyle Health Questionnaire Blood Sample Lab Analyses CSF Neurological and Neuropsychiatric Test Study Summary Imaging CRFs e Table of Modality e Add Subject to XNAT e Image Upload Completed A I FP7 601055 VPH DARE IT D3 2 Database infrastructure for enabling studies based on prospective data in WP1 and2 08 05 2015 Figure 7 below shows the OpenClinica user interface for subject data entry After the registration of a subject data entry a person completes the first initial visit CRF Then the imaging team will fill in the imaging CRFs and as described in section Appendix 1 this will trigger the creation of
Download Pdf Manuals
Related Search