Solving Data Management Problems Through Unified Solutions

Transcription

1 A White Paper Presented By ArisGlobal Solving Data Management Problems Through Unified Solutions Simon Wilson Head, Clinical Program Management and Training, ArisGlobal

2 ABSTRACT Slow and costly study builds continue to be a headache in even the most modern electronic data capture (EDC) platforms. This is for many reasons, such as the need for expert programmers, building in the back end without instant visibility of forms and edit check outputs, inefficient spec, build, review processes, and lengthy iterative cycles. Building a study across EDC, randomization and trial supply management (RTSM), and/or electronic patient reported outcomes (epro) often means that study-by-study integrations need to be built, adding considerable time, cost, and risk to study preparation. Cleaning data across 3 systems still tends to occur in batch mode and requires spikes in activity ahead of interim analysis and database lock. Ad hoc analysis of any complexity across three databases remain nigh-on impossible. This whitepaper will examine the reasons for these problems and how a unified solution can help contract research organizations (CROs) to meet client deliverables on time and at reduced cost while providing increasing data visibility for both users and clients needs. 2

3 INTRODUCTION Clinical data management (CDM) is an essential function in clinical trials and it is therefore critical to ensure high quality data is captured by site staff - whether through paper based case report forms (CRFs) or electronic case report forms (ecrfs) - and available for timely review. The integrity and quality of data being captured or collected and transferred from subjects to clinical data management systems should be carefully monitored, maintained, and quantified to ensure a reliable and effective base for new submissions and subject safety decisions. Relevant data and metrics are also required to make better and timely decisions about trials conduct, for example dose escalation or adaptive design decisions. The use of technology to collect and access data in clinical trials has grown in recent years; this has affected the activities of clinical research for multiple stakeholders including sponsors, CROs, and sites. Although technology and systems have been installed to enhance various aspects of the clinical research and development process, implementation has not been without difficulty and still largely fails to deliver on promised efficiencies. There are also challenges in customized configurations and improved technology requirements from sponsors. These clinical data management applications need to be better integrated with clinical operations so that the data can be more effective when investigating and resolving potential issues. THE CHALLENGES The competitive pressures in today s market are forcing the industry to find better ways of reducing drug development times and increase productivity. The adoption of technology has created new demands for further improvements and features, which has led to improved efficiencies in operations and study execution timelines. This has driven the industry to choose a technology environment that provides and promotes robust data management, real time data analysis, and global trial management and study capabilities. Shortening the clinical trial lifecycle by delivering good quality data faster and making the data available and easy to access are clear solutions to the challenges that the industry has been experiencing for years. Adopting new data management technologies and systems in the clinical trial process offers a solution with success stories. This has led to a growth of vendors providing solutions and systems in the market place. Despite the growth in the industry, there are still key challenges faced by clinical data management departments. The following statements from CDM teams help to highlight common pain points: 1. Would like to build studies faster and less expensively than currently. 2. Need to build or adapt integrations between EDC, IWRS, and other applications for each study. 3. Spending an inordinate amount of time accessing multiple data sources to get the data and metrics required to make good, timely decisions about trials. 4. Looking to reduce time from protocol approval to go live. 5. To roll out protocol amendments without EDC/ePRO/RTSM downtime. 3

4 6. Ad hoc data analysis continues to prove difficult. 7. Looking for continuous data consolidation with EDC/ePRO/RTSM. 8. Struggling with setting up local and central lab normal ranges. Let us take a few of these problems and examine them in detail. SLOW, COSTLY STUDY BUILDS Study builds continue to be a headache with even the most modern EDC platforms. This is for numerous reasons, such as the requirement for expert programmers to build the back end without instant visibility of forms and edit check outputs, inefficient specifications, build, review processes, and lengthy review cycles. It often takes two or three months from final protocol to go live. While these durations can be reduced by using now commonplace interactive review cycles, most EDC systems require the programmer to have significant training and expertise. An EDC platform with simple drag-and-drop form, dynamics, and calculation build capabilities allow the designer to see the form and validation check output as it s being built. This saves weeks of study build time and requires significantly fewer and cheaper resources. Add to this easily configurable, study, or URL-level configurations and notifications, and even more time, labor, and expense is saved. This is just one element of the study. If EDC is to be used alongside epro and/or RTSM, builds in each system need to be staggered; interdependencies between the design of epro and EDC, and RTSM and EDC elements mean that progress towards going live is limited by the speed of the slowest development. Very often, this approach will involve three vendors with disparate processes and review cycles, communication issues, and complex vendor management and governance. STUDY-BY-STUDY INTEGRATION EFFORT Building a study across EDC, RTSM, and/or epro often means study-by-study integrations need to be built, adding considerable time, cost, and risk to study preparation. These integration efforts are likely to be complicated, making it difficult to get the relevant data and metrics from multiple data sources, slowing down decision making and increasing complexity fatigue. The need for enterprise level or study-by-study integration between EDC, epro, and RTSM is common. Large elements of EDC, epro, and RTSM databases are shared and integrations are simply used to transfer data from a source system to a target system. Such integrations often add weeks to study build timelines at considerable expense and are of limited value unless very well specified and tested. All three applications need to use open standards and architectures in order to transmit data effectively. Even now, CDISC ODM and web service adoption is not universal. Even with standardization and repeated use of the same solutions, this is never simple. Additionally, even with CDISC standards, definitions of context (e.g., site, country, and region) can differ between the three systems, so the value of data exchanged is limited. 4

5 HIGH SPEND ON TRIAL SUPPLIES Clinical trial managers are faced with a perpetual conundrum with regard to clinical trial supplies. Based upon country and site-specific subject recruitment estimates, they need to provide an overage of supplies or risk stockouts; both equally costly options. Stockouts can impact on subject safety and mean significant delays in subject recruitment and ultimately late closure of studies. Overages of supplies (often in excess of 50%) mean that expensive trial medications, lab kits, giving sets, etc., are procured securely in the knowledge that this overage will not be used. In a late Phase III or Phase IV study, this may actually divert significant quantities of investigational medicinal product (IMP) from sales and an additional cost of lost sales is incurred. To avoid this, it is critical that an RTSM solution with sophisticated randomization and resupply algorithms is fed with accurate subject recruitment and status data. This is only possible through tight integration with EDC. However, such integrations are expensive, risky and, by definition, require additional time between final protocol and first patient first visit. BATCH DATA REVIEWS AND DATA CLEANING OUTSIDE SYSTEM: DELAYED DATABASE LOCK CROs tend to be in control of data cleaning tasks. Since they need to work with multiple EDC applications and because recruitment and data entry rates are not generally linear, they tend to export datasets into programs like SAS for batch data reviews, cleaning, and coding outside the EDC platform. This means that sophisticated workflows available within the EDC application cannot be exploited: because workflow is broken, the status of individual datapoints is unknown to the EDC system. If some data needs to be consolidated between EDC, epro, and RTSM, a significant burden is placed on clinical development teams to ensure data currency and reconciliation between these 3 applications before any data reviews, cleaning, and coding can commence. While such processes have developed for understandable reasons, they inhibit the realization of eclinical systems delivering their true potential in effect, paper-based processes are still commonplace. Such processes give rise to resource spikes, lengthy database lock cycles, and prevent ad hoc analysis and real-time decision making. However if a unified platform is used, datasets are created in a cumulative, cross-application manner with no need for painful data consolidation and reconciliation effort. Flags can be assigned to datapoints to enable continuous data reviews, cleaning, coding, and ultimately datapoint rather than database locking. Such processes allow powerful ad hoc analysis, enable adaptive design, and mean that the database can be locked within days of last patient last visit. 5

6 DISPARATE LANDSCAPE There are many different applications used by clinical data managers. Even if they only use one system comprised of different applications from multiple vendors, there is still the considerable burden of vendor management and complexity fatigue. Users have multiple log-in credentials to remember and there is the cost of duplicated user admin to consider and users need to hop between applications and use spreadsheets to gain visibility of data status across these applications. This results in a loss of or very limited cross-application visibility and oversight. The lack of data accuracy and consistency, due to incompatible business status definitions between disparate systems of different vendors, adds to this lack of visibility and limits the value of data that is integrated between systems. Each application comes with its own infrastructure and support, so hosting, licensing, and support fees are multiplied. This also creates a greater sponsor IT burden, to manage vendors, and support a plethora of applications. The result is that despite the considerable expense and time consumption, data managers still struggle with an inaccurate view of data brought about by a lack of full data consolidation and reconciliation. 6

7 ATTEMPTS AT INTEGRATION Given this complexity, life science companies are seeking solutions that enhance decision making through greater efficiency and knowledge sharing. This requires a level of integration that spans across critical departments and systems. There are many possible points of integration between these various applications and systems. In the past, integration of these systems was difficult; attempts at integration would typically be point to point, with integration layers developed between each application. There are also the limitations of SOA and web services, as optimal utilization leads to higher costs to handle the additional development requirements. More often, this, combined with future upgrades, impact on duration, risk and cost, turns into an extensive and painful integration process that is complex to implement, often prone to error, and requires substantially more staff resources to achieve. With a standard model study build, we see the development of studies at the application level; each application build had dependencies on other applications. Traditionally, organizations had to increase their internal infrastructure to support additional applications. They also had to ensure that any data that is transferred or integrated between applications met the need of the business and that the business definitions match. They would want to ensure that they are communicating with other internal resource teams on those business definitions, such as subject and CRF statuses. This was not only time consuming and costly, but also ran the risk of the data not meeting the needs of the business. With these interdependencies, the speed of the study build is dependent and interdependent on these application builds and the slowest application that you have in terms of EDC, RTSM, and epro. Lack of context is another issue to consider with integration; context refers to definitions of particular data, such as a visit window or a subject status. Unless very carefully specified in the build effort, the different definitions of these data and that particular context can be different in each application. Unless that context is common across all applications, then the data that is exchanged is of limited value. The applications in an integration build are commonly traditional, and are based on the spec build test rather than an agile interactive type methodology that we have today. Unification gives you the ability to take an agile approach and agile methodology. 7

8 Moving away from the standard model to the unification model study build, we see true building of studies. VALUE OF A UNIFIED CDM PLATFORM For some time now, we have seen the unification of randomization with supply technology; rather than having IWRS/ IVRS purely for randomization, we have seen this combined with supply management systems and we now have RTSM. This can be taken a step further by unifying this with EDC and epro. The result is one database that surfaces EDC, epro, and RTSM applications. The database could have selected elements of combined functionality depending on the study; for example, you may have a study that has EDC with RTSM, another study that has EDC and epro, and a third study could have all three. We have reached the point where we have eliminated the need for integration with the unified platform approach; the modules and functions can be configured based on the user s requirements. The unified data management platform above is delivered through single sign-on (SSO) with a common user interface (UI) and cross module dashboards and reporting. On the right, we can see that it can support web services and ODM based integration with applications outside of data management, such as CTMS, pharmacovigilance, and SAS. 8

9 Returning to the routine clinical trial world, what we have seen through this kind of capability is an optimized EDC option for paper or hybrid studies that allows a selectable approach to study design, eliminating a once very rigid approach where everything is either going to be a paper study or an EDC study. We now can do either, and that means all kinds of design options are open to a protocol stage. It also allows companies to have a flexible site and country selection, so no longer must we only work in certain sites or certain countries with an EDC solution, we can open it up to any site or geography that would like to use that particular study, also opening up all sorts of realms with regard to regulatory and sales opportunities. This on-going consolidation of data within EDC means we do not have to worry about back-end consolidation with a clinical data management system (CDMS) solution and have much more rapid availability of datasets for analysis. Finally, with this reduced programming burden we also have a reduced cost, meaning that with an EDC solution that supports paper and hybrid studies, we can see accelerated clinical development programs at a reduced cost. The benefits of a unified platform include: DEPLOYING TRIALS FASTER WITH RAPID AND CHEAPER STUDY BUILDS As mentioned in challenges, if EDC is to be used alongside epro and/or RTSM, builds in each system need to be staggered; this approach will involve three vendors with disparate processes and review cycles, communication issues, and complex vendor management and governance. Even with standardization and repeated use of the same solutions, this is never simple. Once the system is ready to go live, the sites, user accounts, and training still need to be created and provided, which will require even more time and effort. 9

10 When these applications are all built on the same database, this need for integration evaporates, as well as duplicated site and user management, three sets of different login credentials, multiple training programs, and differentiated UI and navigation options. One solution built on one set of data structures and one architecture that includes EDC, epro and RTSM finally realizes the potential of clinical data management technologies. A fully configurable, unified EDC, epro, and RTSM platform in which studies are easily built on can reduce weeks off study build time across all these modules, achieve instant and full data consolidation with zero effort, and eliminate duplicated site and user administration and training. FASTER DECISION MAKING THROUGH INSTANT, REAL TIME DATA CONSOLIDATION Integration is simply used to transfer data from a source system to a target system. These can add weeks to study build timelines at considerable expense and are of limited value unless very well specified and tested. It would also need to include the necessary standards and architectures in order to transmit data effectively, such as CDISC ODM and web services, which are still not universally adopted. With a unified solution, end-to-end supply tracking with full accountability can be achieved in real time, improving accuracy whilst not requiring any integration effort which would involve unnecessary cost and risk. The EDC solution can be used to remind subjects to enter data into epro, to attend a rescheduled visit, and to take their medications, all with no integration requirement. Subjects are immediately randomized in EDC and epro data appears in EDC automatically and instantly. Having continuous data consolidation between these three core CDM applications makes it feasible to continuously review, code, and clean this data. An easily configurable, unified CDM solution allows workflows to be quickly constructed at enterprise, TA, sponsor, or study levels. The potential of EDC can be experienced as reviewing and cleaning of data actually takes place within the system so that the status of each field is known. This means that the bucket of clean, usable data gradually fills up, in turn allowing powerful analysis and decision making with very high levels of confidence. REDUCES SPEND ON TRIAL SUPPLIES The site workflow is optimized through the immediate flow of data from EDC to RTSM. There is no need to go into EDC to create a subject or to update subject s visit status, and then go back into RTSM to perform the randomization there. The user can randomize subjects directly in EDC without having to access the RTSM module. This eliminates redundant data entry and potential transcription errors. It also provides real time subject status because there is an immediate flow of that data; the user does not have to wait for somebody to get round to updating RTSM. Even today, it is still often required that somebody throughout the study has to reconcile the data that has been entered into the randomization tool against data that is entered into EDC. The instant flow of unification eliminates this whole reconciliation process. 10

11 Because of the immediate randomization of subjects in EDC and immediate flows, there is also an immediate update to supply inventories. Randomizing dispensing events in EDC do trigger immediate updates at site level inventories and, depending on the way that the re-supply levels have been configured, those will trigger automated re-supply requests from the site to the depot, and from the depot to the warehouse. It is this kind of supply mechanism and this workflow throughout the supply chain that reduces the impact of overages and stockouts because of this instant flow of information. All of this is possible through integration solutions, but it is simpler, cheaper, and quicker to achieve through a unified CDM solution with in-built RTSM capabilities. RAPID AND COST-EFFECTIVE DATABASE LOCK With unification, workflow can be exploited so that data points can be locked one at a time as the system and individual people viewing the system know when each data point is ready to be locked. Typically, there are many data points that are never touched after data entry. Those points can go very rapidly through this process, get locked, and not touched again. This means that at any point in the study, the vast majority of datapoints have been entered, cleaned, reviewed, and locked so that the full database can be locked within a few days of last patient last visit. Such a workflow enables the smooth utilization of resources throughout the study. This contrasts sharply with the common experience of a 4 to 6 week database lock phase during which there is a huge resource spike at sponsor, CRO, and site levels. Unified CDM platform with EDC, epro, and IWRS/RTSM modules 11

12 SAMPLE CASE STUDY Current Situation This case study is of a mid-tier CRO looking to improve data management efficiencies with a global client focus. Their goals were: To improve time of study builds from study start-up through close out. To provide competitive pricing to clients. To ensure data integrity through eliminating multiple complex point-to-point data integrated systems. Study builds that include multiple integrations often cause late study commencement at additional cost. For a CRO, costly builds mean a difficult sell and delays in client deliverables mean a lack of return in business. Without integration, the currency of data and its status is unclear. Clients become frustrated when clinical data metrics and reports are inaccurate, and therefore unable to reflect the true status of study protocols; the measurement of KPIs becomes a problem and mid study decisions are delayed. Company Profile : Mid-tier CRO with US, EU, and APAC presence Goal : More rapid study builds and study completion More competitive pricing Tight integration with RTSM and epro Frustrated Customers: Slow, complex, and expensive study builds Delayed go live due to integration effort Inability to perform ad hoc analysis Delayed database lock AFTER ACHIEVING UNIFIED SOLUTION The CRO implemented a unified EDC, epro, and RTSM; one database with modules in which each module s functionality are surfaced. This ensured data consolidation and allowed for on-going data cleaning and reconciliation while eliminating the cost and delays that point-to-point integration entail. This CRO is now able to meet client deliverables on time and at reduced cost while providing increased data visibility for both users and client needs. Users are now able to see patient status across all studies, real-time patient diary information, and efficient patient randomization. Time of go live date to study close-out is decreased and cost is significantly reduced from using paper to ecrfs. 12

13 Real time decision making, such as for event based monitoring, dose escalations, and safety assessments are suddenly a reality. Drug shipments are much more efficient, so that expensive overages or stockouts are avoided. At the same time, full drug accountability is enabled. Satisfied Users: Significant reduction in study build effort & time in EDC Cost-effective replacement of paper based CDMS Integration effort and delays eliminated Real-time data consolidation and decision making Continuous data cleaning and review Faster database lock without spike in activity Satisfied Customers: On time go live and study closure with lower investment Ad hoc and interim analysis Repeated business for CRO CONCLUSION As we have discussed, standard integration seems to fall short of handling data management problems in the past and today; although integration fixed some of the inherent problems of clinical data management applications, pain points such as duplicated user admin, many process, and multiple user training and vendor management still persist. These issues are likely to lead to greater costs and time delays, which is not ideal in an industry looking to accelerate clinical development programs at lower cost. Instead of having multiple applications running on different platforms and accessing multiple databases, a unified data management solution can offer rapid study deployment, eliminate integration cost, and minimize vendor management. By only having one database across all modules, this optimizes resource management, allows ad hoc analysis, and rapid early database lock. The result is capturing and accessing data quicker and being able to deploy studies faster. 13

14 About ArisGlobal ArisGlobal ( is a leading provider of integrated solutions for pharmacovigilance and safety, regulatory affairs, clinical development and medical affairs and compliance. Hundreds of life science companies rely on ArisGlobal s advance solutions for maintaining regulatory compliance, workflow automation, improving operational efficiency and easily sharing information around the globe. Specific to clinical development, Total Clinical is ArisGlobal s integrated platform that gives life science organizations instant access to all core eclinical applications, including third-party vendor applications, from agworld, a central clinical portal with single sign-on. Our Total Clinical Data Management solution is a fully-unified solution that includes EDC, IWRS/Supply, epro and medical coding. Built on a single database layer, its advanced architecture and feature set dramatically reduces study development times and easily supports the implementation of changes to live studies without disrupting study progress. agcapture is a cost-effective, comprehensive EDC system to capture, clean and report clinical trial data. agoutcomes is an advanced electronic patient reported outcome (epro) system that can run standalone or fully integrated with a traditional EDC study. agsupply is an integrated interactive web response system (IWRS), randomization and logistics management system that automates the tracking of clinical trial supplies throughout the supply chain. agencoder is a centralized dictionary coding tool that enables fast and consistent coding of medical terms. For more information on ArisGlobal solutions, visit and About the Author Simon Wilson is Head of Clinical Program Management and Training, and a subject matter expert at ArisGlobal. Simon started his career in clinical research at a site and lab level before progressing though data management, clinical operations, clinical technology and process implementation leadership posts at Bayer, PAREXEL and CSL. Over 15 years experience in head office and affiliate Pharma and CRO settings have provided Simon with a thorough understanding of this topic. Simon has a keen interest and expertise in unlocking the potential of eclinical applications. WP-CDM ArisGlobal LLC. All rights reserved. Total Clinical, agcapture, agoutcomes, agsupply, agencoder the ArisGlobal logo are trademarks or registered trademarks of ArisGlobal in the US Patent and Trademark Office and other jurisdictions. Other marks may be trademarks or registered trademarks of their respective owners and are acknowledged as such. The information you see, hear or read on the pages within this presentation, as well as the presentation s form and substance, are subject to copyright protection. In no event, may you use, distribute, copy, reproduce, modify, distort, or transmit the information or any of its elements, such as text, images or concepts, without the prior written permission of ArisGlobal. Trademarks, service marks, and logos (collectively the trademarks ) displayed on the pages within this booklet, are the property of their respective owner. No license or right pertaining to any of these trademarks shall be granted without the written permission of ArisGlobal, LLC (and any of its global offices and/or affiliates). ArisGlobal reserves the right to legally enforce any infringement of its intellectual property, copyright and trademark rights. 14