PAT: An intelligent Authoring Tool for facilitating Clinical Trial - PDF document

Institute of Communication and Computer Systems (ICCS), National Technical University of Athens, (NTUA), Greece PAT: An intelligent Authoring Tool for facilitating Clinical Trial Design Anastasios TAGARIS, Vassiliki ANDRONIKOU, Efstathios KARANASTASIS, Efthymios CHONDROGIANNIS, Charalambos TSIRMPAS, Theodora VARVARIGOU and Dimitris KOUTSOURIS

PONTE Project Overview • Funding Scheme: FP7 – ICT Programme • Duration: 3 years (2010- 2013) • Purpose/Idea: – Efficient Patient Recruitment for Innovative Clinical Trials of Existing Drugs to other Indications . PONTE is about… • Facilitating clinical trial design for drug repositioning – Facilitating Clinical trial design ● Platform to support CT design and evaluation of I/E criteria against EHR’s ● Assist in evaluating the trial’s fundamental hypothesis using linkage to literature data bases – Drug Repositioning ● Drug repositioning (e.g. Drug repurposing, Drug re-profiling, Therapeutic Switching and Drug re-tasking) is the application of known drugs and compounds to new indications , i.e. new diseases

CU Institute of Psychophysiology & Rehabilitation IOPR of the Kaunas University of Medicine (IOPR) H • R&D Providers (CETIC, ICCS, TUD) • Technology Providers (SMI) • Clinical Sites/Users (CNR-IFC, IOPR, CUH, NKUA) • Legal Consultants (LUH) 7 countries TU 1 company D SM 2 research LU I institutes H 6 universities ICCS/NTUA NKU A CNR/IFC National Research Council Institute of Clinical Physiology (CNR-IFC)

Some facts about CR… • At present, only about 9% of drugs entering clinical trials succeed in being approved by the regulatory bodies. • In addition, many of those being approved don’t succeed commercially upon introduction in the marketplace because of side-effects. • Key issues which need to be addressed in clinical trials include incompleteness, ambiguity and inconsistency with most errors being identified within the protocol writing process.

Some more facts (about set of EC) • Major determinant for the success of a CT is the set of Eligibility Criteria (quite often poorly justified, vague, ambiguous or overly rigid). • Vague and/or ambiguous criteria may: – Increase patient risks and weaken the ability to utilize the inclusion/exclusion criteria to establish appropriate labeling • Overly rigid criteria may lead to: – Failure to represent the real patient population and mimic clinical practice – Limited generalizability – Efficacy remaining hidden – Increased study complexity & costs – Limited patient accrual affecting study feasibility, duration and costs

PAT – CTP Design Tool • PAT offers the GUI that integrates and visualizes a series of novel mechanisms aiming at improving several important processes (and their corresponding parameters) in clinical research: 1. CTP Management & Design 2. Specification of study parameters, (duration, no. of patients, …) 3. Description of the eligible patient population (I/E Criteria Specification) 4. Links to Intelligent search through predefined queries and background knowledge 5. Suggestion of I/E criteria based on investigational drug, study disease and published research findings 6. Patient safety and Study efficacy-oriented validation checks on I/E criteria 7. Estimation of number of patients available at cooperating hospitals allowing for checking on study feasibility and for evaluating the recruitment sites 8. Translation of the I/E criteria into the PONTE EHR model parameters

PAT in the PONTE platform architecture

Design & Implementation Methodology (1) • Interviews with Clinical Research Experts – Scope: sketch out the CTP structure, parameter dependencies and semantic relations – Materials used: Existing CTProtocols ! static & dynamic Mockups – Outcome: – CTP Structure, – specification of parameter dependencies – identification of the informational elements which could benefit from decision support – predefined queries per protocol section

Design & Implementation Methodology (2) • Thorough analysis of the information available at the clinicaltrials.gov – Scope: Understand the way that PIs use to express the EC, discover the common criteria categories and corresponding structures to build user friendly GUIs. – Materials used: XML description of studies => Downloaded in a local DB – Analysed types of study parameters (study type, blinding methods, control types, end points, etc) – Analysed the EC structure, categorization, detail levels; Analysis was based on ~1000 criteria selected across all studies of the portal, where a wide spectrum of disorders and actives substances was covered (intentionally) – Outcome: – Eligibility Criteria categorisation and corresponding structures – In parallel this procedure helped us capture the semantics of EC and develop the Eligibility Criteria Ontology (EC-O) that allows for semantic inference.

Design & Implementation Methodology (3) • Translation of User Requirements into Technical and Functional Specs to be followed for implementation – “access from anywhere”, – “collaboration among experts” – “diversity in the structure of data to handle” • Implement the PAT as a Web tool following the Model View Controller (MVC) approach – MVC separates the representation of information from the user's interaction with it ! code reusability and separation of concerns.

evaluate$the$user$creden-als invoked$to$access$the$CTP$Repository PAT$Internal$Architecture$&$Interac1ons provided$through$the$PAT$in the$user$wishes$to$perform$a in$order$to$ store$and$retrieve$the$CTP order$to$ grant$or$deny$the$user manual$search$in$the$underlying instances$ edited$by$the$users. access $to$the$PONTE$plaQorm. is$called$by$the$PAT$to$ retrieve$the$CTP data$sources structure ,$as$well$as$its$ sec-on$and seman-c$dependencies receives$requests$ from$the$browser, Provides$persistence$services$for invokes$the$appropriate$ac1ons$on$the implements$the$core$applica1on$logic is$also$responsible$for$ communica-ng the$needs$of$the$PAT.$All$CTP and$is$responsible$for$ interac-ng$with PAT$ Business&Logic&component ,$ creates parameters$are$persisted$on$the with$the$Authen-ca-on$module$ in the$corresponding$UI$elements $from PONTE$components$ (DSS,$OM).$It$also local$storage$throughout$the order$to$establish$a$ security$context$ for the$ UI&Controls&Library&component $and uses$the$ Local&Storage&component $to the$execu1on$of$the$request authoring$lifecycle renders$the$response$back$to$the temporarily$store$the$CTP$parameters Provides$ rich$user filled$in$by$the$users$and$other browser Runs$on$ any$web$browser . interac-on$capabili-es applica1on$state$informa1on. Contains$a$JavaScript$library and$communicates$with component$(PAT$Web$Client the$serverFside Library)$implemen1ng$the components presenta-on$logic$ of$the applica1on CTP$parameter$dependency checks,$pa1ent$popula1on informa1on,$or$subjects$eligible for$recruitment$along$with profiling implements$the$ core func-onality$ of$the$PAT.$It encapsulates$the$ user$interface rendering$logic$ and$the orchestra1on$of$ business workflows • PAT Web Client ! JavaScript • PAT Web Client Library ! jQuery library and Ajax . 11 • PAT Web Server Application ! JAVA EE ! runs on Apache

Clinical Trial “Steps” Incl./Excl. Hypothesis Clinical Trial Final Trial Patient Criteria Evaluation Design Protocol Selection Evaluation Step I Step II Step III • Unbiased • Efficient and validated specification of study • Determination hypothesis parameters of a evaluation representative • Well-justified inclusion / exclusion criteria - target patient • High quality assesed and evaluated against real-world population data - which determine the right patient evidence • Acceptable about benefit population for the study of intervention patient accrual and relevance in terms of to clinical size and time practice

PAT delivers the GUI for • CTP Management and Editing – 3-way navigation through the study parameters – Saving different CTP versions with different levels of access – CTP status management • Links to GoPONTE for automatic user-driven predefined questions • I/E criteria representation • Request Patient Selection • View Eligible Population Size per Healthcare Entity • CTP-based Healthcare Entity Management • …and is integrated with the PONTE security mechanisms

Integrates with GoPonte

Structu CTP De Methodology

Conclusions - Discussion • PAT encapsulates intelligent, semantically-assisted mechanisms for study parameter and EC specification. It organizes CTPs around 3 different views: the hierarchy of its parameters, dependencies among them and their semantic linking with the CTP three key concepts • These views were considered as “very helpful” during trial design by the experts, as they allowed for “easy and fast detection of inconsistencies” and “less time consuming parameter specification” although the dependencies should be further enriched while the 3 “key concepts” could be expanded to also include “clinical trials” and “patient” concepts.