The Pharmacovigilance IT Ecosystem
A Key to Patient Safety
Vikalp Khare, Director, Otsuka Pharmaceutical Development & Commercialization
This article explores the details of a typical Pharmacovigilance (PV) Information Technology Ecosystem setup & maintenance which is essential for any pharmaceutical company to ensure the safety of the patients consuming their medicinal products. Article also touches upon a possible controlled approach of leveraging AI & ML to establish an efficient, effective & cost-benefiting PV-IT ecosystem.
As per European Medicines Agency (EMA), Pharmacovigilance (PV) is the science and activities relating to the detection, assessment, understanding and prevention of adverse effects or any other medicine-related problem. Any Pharmaceutical company that receives more than a handful of Adverse Events (AEs), whether it’s for their marketed products or for products only in clinical trials (CTs), they at least need a database to collect, assemble & report on these AEs. Regulations & Reporting requirements for these are not only voluminous but also do follow strict guidelines in terms of content, format & timing of these reports to be sent to Health Authorities (HAs), Contractual Licensed partners (LPs), Associated Clinical research Organizations (CROs), Ethics committees (EC), Investigation Research Board (IRBs), etc. A global centralized database to collect the incoming safety information, process the same to confirm the relation to company’s medicinal products & thereafter reporting the same to Health Authorities to ensure the safety of patients is what constitutes the PV Information Technology (IT) ecosystem for any pharmaceutical company. Needless to mention that the size of ecosystem setup is dependent on multiple factors like number of medicinal products, countries where the products is marketed or is undergoing CT, Patients enrolled within those studies, etc. Setting up and maintaining a Compliant, Efficient & Cost-Effective IT landscape for Global Pharmacovigilance operations involves a combination of robust infrastructure, Global data management/standardization, automation technologies (Artificial Intelligence & Robotic Process Automations), Integration gateways & a strong governance setup to tie it all together.
PV IT Ecosystem setup will primarily consist of Hosting platforms, PV applications & the Data Management procedures.
Pharmacovigilance Database & Applications
As briefly mentioned above, this database collects, processes, stores & reports the safety information (Adverse Drug Reactions (ADRs)) to Health Authorities. This application will act as your centralized engine for conducting all PV related activities within the company. Some companies customize their databases, though majority of companies opt for buying standard packages available from various vendors. Size of database is often found to be inversely proportional to ease of processing & transferring the data. Safety Database application should have the ease to be integrated with other existing IT applications within & outside the company. One will also need to frequently setup & maintain gateway connection for transferring data & reports to both internal & external stakeholders. Often for this, a separate gateway maintenance software is also recommended to be used in conjunction with safety database applications.
Many databases also have workflow and quality components built in with email and messaging functionality, which further helps in performing various data processing activities. One will have several different types of users including physicians who would be accessing this database to standardize & review the data confirming the relation with company’s medicinal products and will also further transform the data in required format for submissions to health authorities. Maintaining proper access to this global database is therefore important. Centralized database will also be accessed by your local & global users to perform needed business operations. If all users are not trained properly to enter & process data, it may lead to data duplication & other data integrity related issues.
Recently, pharmaceutical companies have also started plugging in an additional intake plug in front of Safety Database, which provides an added advantage to standardize the data before it enters the global safety database. Again, like Safety database, Intake tools will have the capability to provide local language support to diverse set of end users. Standardizing the data at intake stage itself has been a great benefit for the companies as they can now manage a core team working on central safety database, and the local data entry operations can be governed through these intake tools.
Lastly, all the application setup also needs to be supported through a robust back-up & restoration procedure to ensure business continuity in event of any data losses & disasters.
Hosting platform
Companies may have multiple preferences on the type of hosting infrastructure they want to operate on. This platform is accountable for providing the computing power needed to perform the operational PV activities & therefore is utmost important to have the right selections done here. Companies often prefer to also do a performance testing by replicating the user load on applications to be sure that there are no system slowness issues. There is also a need to develop a framework to have these tests performed on periodic intervals to proactively identify if there are any network latencies and local access issues are occurring. These days with evolution of Cloud based hosting, a diverse set of options are available for companies to pick & choose from. OnPrem versus Cloud hosting strategy has been debated in industry since last couple of decades and both have their own pros & cons. Often companies also prefer to opt for Software as a Service (SaaS) model where the maintenance of hosting platform is already taken care by the vendor providing the software. SaaS models often limits your capability to customize the databases beyond the permissible options of managing it through available configuration handles. However, the advantage of overall software maintenance is a great benefit which pharmaceutical companies may prefer.
Platform also needs to be secured & needs to ensure that there are no security vulnerabilities identified in any of the components installed on that infrastructure. Vulnerability identification exercise should always be done proactively & is therefore recommended to have a monitoring framework which is consistently monitoring the incoming & outgoing data network. Predefined data exchange points need to be secured as well as they will be subject to the hacking attacks the most. Also, it is recommended to perform at least yearly Penetration testing exercise to ensure that your infrastructure is secure.
Global Data Management & Reporting
Most critical element of maintaining global IT applications is to ensure consistent data entry across the regions where the users are accessing the applications to perform their routine operations. Systems should be able to manage data entry from a diverse teams of healthcare professional, patients, partners, etc. & should also be equipped with multi language support. Note that all global system will need to also interact with local systems as well to ensure that local compliance requirements are in check. Furthermore, the local system interaction with global systems involves exchanging Personal Health Information (PHI), which is also governed by local privacy requirements like General Data Protection Regulation (GDPR), etc. Any PV system which is being setup by any pharmaceutical company must comply with regulatory framework such as US Food & Drug Administration (FDA), EMA (EU), and UK’s Medicinal & Healthcare products Regulatory Agency (MHRA), Japan’s Pharmaceuticals & Medical Device Agency (PMDA), Health Canada, etc. In addition to the framework, the adverse information safety reports containing the safety information which are to be submitted to Health Authorities needs to also support the industry standard format like WHO Global Individual Case Safety Reports (ICSRs) R2/R3 format, MedWatch (FDA), CIOMS, etc.
AE Data received from different sources needs to be standardized using industry standard coding practices which may further entail use of entities like MedDRA medical dictionaries, Uppsala’s WHO drug codes, Japan (PMDA) specific drug codes, Korea (MFDS) drug codes, etc. Consolidating all this information together is critical and proper arrangements are to be done to ensure we have right software applications to ensure our business objectives are met.
Lastly, it is utmost important to track & record every activity/task performed on the data right from origination from patient to the end where it is reported to the needed destination. This ensures we have the data integrity maintained. It is amazing to see how the AE data originating from a patient in USA for a medicinal product needs to be reported back to PMDA HA in Japanese in a particular format compared to the same being reported to EMA in a completely different format.
Artificial Intelligence & Machine Learning
Today’s IT landscape for every company is changing at a rapid pace so that they can take the advantages of modern age technology around Artificial Intelligence (AI) & Machine Learning (ML). PV IT landscape can also leverage these to make the current procedures more efficient & cost effective. However, there has not yet been a detailed guidance provided by any Health Authority on how they would recommend using these in context of ensuring patient safety. Hence, the recommendation would be to go conservative and utilize these new-age ways of working to always provide finite/fixed outcomes. Decisions made through operations conducted using the PV applications will impact human lives & hence the recommendation for conservative approach. AI/ML can be leveraged at all levels within PV-IT application setup.
AI/ML can help you modify your cloud computing power based on the trends of systems utilization. This will help with cost savings on infrastructure hosting expenses. It can also be used to modify the way medical information is being entered in the intake applications or safety database directly. AI/ML can help you enable auto import of the data from a handwritten prescription or from a scanned pdf form. These auto imports need a thorough human review as the machine can make errors & we would not want that to change the context of entire operation. The review activity along with the changes done by humans can be utilized by machine learning algorithms to further tune the data import operations.
Lastly, the Data Management & Reporting will also benefit with AI/ML, as the medical review decisions, data standardizations, multi-language support, etc. all can be managed with ease. AI/ML can further simplify the report output creation for companies through use of multiple automated platforms.
It is critical that a proper governance model is established within each company to monitor the activities being performed by AI/ML platforms along with other operations. It is important that we expose the use of AI/ML technologies conservatively in PV operations by controlling the output of these technologies to be finite (locked outputs) for each type of input scenario. Having a fixed outcome for every scenario should be a good starting point to be able to adapt to the new changing world for PV based IT operations.
The intent of this article was to provide a high-level overview of how-to setup & maintain an IT landscape to support compliant PV operations. With proper due diligence, set framework & controlled AI/ML inclusion an efficient, effective & cost-benefiting IT ecosystem can be successfully established.
Author Bio
Vikalp Khare is a Director at Otsuka Pharmaceutical Development & Commercialization (OPDC). He has been implementing new age technology-based solutions to support Global Pharmacovigilance Operations. Has extensive experience in setting up Safety Databases, Intake solutions, Signalling solutions, etc. Which are compliant as per Global Health Authority regulations for ensuring Patient Safety. He is currently operating as US head of Safety Data Management, GPV and oversees Global pharmacovigilance technology & Business Intelligence operations at OPDC.
