The Future of Pharmacovigilance: Proposals for More Efficient and Effective Systems-Based Approaches
- 2.7k Downloads
Clinical research is at a cross-roads of sustainability due to the current cost of development, amount of regulation, and irrational use of medicine. There are also too many silos in clinical research and development, post-marketing safety, as well as a lack of standards, and effective communication. This paper covers the future of pharmacovigilance and argues that it needs to re-focus and not be a silo in the last phase of clinical research. In fact, the integration of pharmacovigilance is an integral part of much needed change in clinical research and development based on the application of systems thinking. Clinical research and development must improve as a whole system through the synergy of all parts.
KeywordsSpontaneous Report Electronic Data Capture Safety Database Unify Database Uppsala Monitoring Centre
The need for bringing pharmacovigilance back to its original focus.
The need for a comprehensive systems approach.
Having a unified drug safety-clinical database for improved data capture.
Need for regulatory harmonisation.
1 The Need for Bringing Pharmacovigilance Back to its Original Focus
The human relationship with medicines is as dynamic as it is complex and our understanding of effective use is continually developing. The traditional view was that licensed medicines were consistently beneficial and completely safe. However, as medical advancements were made and non-communicable diseases, along with their corresponding treatments increased, it became clear that there are not just risks associated with medicines themselves but also in the systematic context in which they are taken. These include not only side effects but also interactions between medicines (both licensed and unlicensed) and other products (devices, food and environmental chemicals), individual variation in responses (both genetic and co-morbidity of underlying disease) as well as the ubiquitous potential for medication errors. All these forms of adverse events (AEs) carry varying and detrimental individual, societal and economic impact, so that life-cycle safety surveillance of a medicine—commonly referred to as pharmacovigilance—needs to be better appreciated as a critical part of our holistic understanding of not only the medicine but also the systematic influences on how a medicine is used, rather than just a regulatory requirement where compliance is all that matters.
The aim of pharmacovigilance is that of identifying drug-induced adverse reactions as soon as possible in a medicine’s lifecycle so as to take appropriate measures to minimise their impact on patients. In fact, WHO defines pharmacovigilance as “the science and activities relating to the detection, assessment, understanding and prevention of adverse effects or any other possible drug-related problems” . Unfortunately, the focus of pharmacovigilance has shifted from being science driven to a regulatory exercise. The reason is that the pharmaceutical industry is a highly regulated industry (although implemented and enforced variably), especially the drug safety sector, which must adhere to a tremendous number of laws and guidelines, many of which have a weak evidence base. Usually, regulation is based on consensus derived from the International Council on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH, formerly known as International Conference on Harmonisation) guidelines, but the differences on how these guidelines are implemented varies widely from one country to the other. Unfortunately, there is no independent mechanism for assessing whether current regulations are really improving patient safety or if they are creating confusion, unneeded complexity and leading to excessive burdens for regulators and the industry. Instead, there is evidence showing that companies struggle to comply with regulations, even the most basic ones, such as those regarding expedited reporting . It has to be acknowledged that the European Medicines Agency has recently launched an initiative  to measure the efficacy and effectiveness of some pharmacovigilance processes with the aim of improving them. However, this initiative is limited to specific aspects of pharmacovigilance such as the time to identify a new risk or the time from its identification to action. A systematic analysis of the overlap, increased workload and impact of the multitude of non-harmonised regulations and guidelines on drug safety departments seems to be missing. The risk of this initiative, as with the new European pharmacovigilance legislation that was prepared with the aim of reducing the administrative burden, will further increase the administrative burden instead of reducing it. Therefore, we are left with the view that pharmacovigilance is predominantly a set of regulatory requirements, even though this is a misconception, since it is a fundamental step to help society develop our understanding of medicines.
2 The Need for a Comprehensive Systems Approach
The need to develop a more complete and holistic understanding of the transformational role of pharmacovigilance as part of a complex, interconnected clinical research and development system with feedback loops, rather than a transactional process between pharmaceutical companies and regulatory agencies, is long overdue . Therefore, there should be a system-based approach  to pharmacovigilance regulations: each regulation should not be viewed in isolation, but all the regulations and guidelines (including their interactions and overlaps) should be assessed, together with the impact they have on all the concerned stakeholders. The assessment of regulations’ efficacy and effectiveness should take into account feedback from all stakeholders. This is the vision of the Alliance for Clinical Research Excellence and Safety (ACRES) and a core of one of its main initiatives, the Product Safety Culture Initiative (PSCI), which intends to explore application of systems or organisational sciences to designing safe systems for clinical research. The re-framing of pharmacovigilance—defined as the safety activities encompassing the whole life cycle of a medicinal product—is very much part of this larger discussion on developing systematic integrated information technologies, interoperable standards, policies and practices to enhance clinical research safety, quality and efficacy worldwide to benefit all stakeholders, especially patients. To that there are two areas requiring urgent improvement and change.
3 The First Area of Improvement: a Unified Drug Safety-Clinical Database
With the increasing volume of information about medicines now available (so-called ‘Big Data’), there is an urgent need to develop and implement technology for pharmaceutical companies to have a single repository of safety data, coupled with the correct levels of security and access, for clinical, post-market observational studies and spontaneous reports. The parallel co-existence of clinical and post-market observational databases in almost all pharmaceutical companies creates confusion and additional work as trial-related data and (often) data from post-market observational studies are sent to the clinical database and (in part) to the drug safety database, while those from spontaneous reports are only entered in the drug safety database. There is neither consensus about structure of and access to these databases nor uniform approaches to reconciliation, despite the need to access all data to continuously search for signals. Therefore, there is a need for having all AEs, regardless of their type, source and seriousness stored in one single repository, the unified drug safety-clinical database. Furthermore, this database should also contain laboratory, clinical and vital signs data originating from any study.
Ironically, pharmaceutical companies have clinical trial and post-market observational studies that not only share similar methodologies for collecting data, but they also have comparable (even if not identical) individual case safety report (ICSR) reporting requirements, which are also similar to the ones for spontaneous reports. Globally, the responsibility for a common repository containing homogenous safety data for better understanding of the adverse reaction profile of a medicinal product has been allocated by the WHO to the Uppsala Monitoring Centre (UMC), which established and maintains society’s only global post-market reporting safety database, called VigiBase, which is largely based on post-marketing cases with a small percentage of clinical trial and observational study data. The UMC has made advances to streamline the collection of ICSRs from member countries in the WHO Programme for International Drug Monitoring and proved the effectiveness of incorporating standardisation through the use of ICH-E2B, MedDRA, WHO-ART and WHO Drug Dictionary structures and standards. These are essential for the ability to effectively collect and analyse such a large quantity of data. The inclusion of related cases from clinical trials and non-interventional studies would further improve this valuable data set. Although it is critical that post-market reporting continues to be reported by companies to VigiBase, via regulatory authorities, it is not feasible at this point for companies to report all safety data from all sources (clinical, post-market observational studies, spontaneous reports, laboratory, and vital signs data). Therefore, it is of paramount importance that companies perform their signal detection work on the totality of safety data they have, since a significant part of these data are not included in any other database.
A contributing cause to the confusion and inefficiency deriving from pharmaceutical companies having partially overlapping safety data included in two databases, comes from the fact that clinical, observational and post-market safety databases evolved separately and thus use different formats. For example, the standard for reporting clinical trial data has been developed by the Clinical Data Interchange Standards Consortium (CDISC), a non-profit organisation that mainly addresses the requirements for data submission to the FDA as part of a new drug application, which includes electronic Submission Data Tabulations and Analysis Data Sets. For collecting clinical trial data (which include AEs) CDISC has developed the Clinical Data Acquisition Standards Harmonisation (CDASH) . In parallel, the standard for electronic submission of post-market ICSRs to regulatory authorities (named ICH E2B ) was developed by the ICH. CDISC has acknowledged that the unification of formats and standards is essential for a single-sourced capture of safety data. Therefore, the CDISC CDASH-E2B project team has mapped these two standards . The analysis of the differences between CDASH and E2B (version R2) has shown that they are purely a matter of man-made convention and are not driven by scientific evidence required for clinical research and drug safety.
The mapping of these data conventions is the foundation for capturing AE information for clinical, observational data and spontaneous reporting needs and, since it resolves the discrepancies between the clinical and drug safety standards, it makes the use of electronic data capture (EDC) more attractive for real-time safety surveillance. The mapping can be used as a basis for developing a single interface for capturing clinical and observational study data together with spontaneous ICSRs. Software vendors could develop functionalities that capture data according to a standard that meets the needs of clinical and drug safety, which ultimately are the same. It starts with creating fields large enough to accommodate case narratives and introducing alerts for serious AEs/adverse drug reactions (SAEs/ADRs), and would extend to include SAE/ADR capture in its workflow. Both SAEs and ADRs are being mentioned since there is no reason why the benefits of single-source data capture cannot be extended to observational studies and to spontaneous reports. The only difference between capturing the data for an observational study as compared to a clinical trial would be that for an observational study (and for a spontaneous report) there may (but not necessarily) be less detailed information and therefore the interface fields that need to be used and completed for an observational study would need to be defined in advance, prior to beginning the study.
4 A Unified Database Will Drive Improved Data Capture
In addition to standards and database unification, there are also clear opportunities to repurpose tools, methods, and technology between clinical trials, spontaneous reports, and observational studies. Technological advancement  has led to more mature processes in clinical trials as is seen through the speed with which clinical trial data can be collected and instantaneously validated. Of course the nature of post-market observational studies is not identical to the “controlled” environments of clinical trials where the regulations based on Good Clinical Practices  (GCP) require sponsors to actively intervene to obtain better quality data. However, the advent of electronic health records and personal health data via mobile apps provide an opportunity for improved data capture. Once sources of post-market data reach critical mass in terms of size and granularity then clinical trial tools such as EDC , electronic patient diaries (eDiaries), and eReporting can be better utilised for collection. Today, many prospective organised data collection systems frequently rely upon manual intervention and to expedite the collection and improve analysis of the data, automation will be required. Therefore, web-based solutions such as EDC-based tools should provide efficient ways for collecting and analysing data. The success of EDC depends upon limiting the requirements at each investigator site and this can be achieved with a thin client solution. Thin client means that no additional hardware at the investigator’s site is necessary to support software and only minor applications are installed locally on existing site PCs. The core functionality resides on the central server.
Once EDC systems are implemented, separate sub-databases can be retained to meet independent reporting requirements. However, a single aggregate database jointly shared by the clinical and drug safety departments is essential to advancing pharmacovigilance . In practice, the idea is that once keyed in through EDC, all the safety information that is currently separately directed and stored in the clinical and safety database is stored in one single repository. A concern that may impede the implementation of EDC for studies other than clinical trials is the difference in the quality and quantity of information available for AEs originating from clinical trials as compared to those originating from spontaneous reports and observational studies. This can be easily accommodated by customising the EDC, eDiary or drug safety case processing personnel user interface (i.e. it only needs to be defined which fields will be available to the various users). Processes can be redesigned to allow different end users to have access to predefined datasets: for example, drug safety would have access to the datasets pertaining to all AEs and to summary reports on laboratory and clinical data, but not to datasets pertaining to the efficacy of each single patient. In addition, clinical personnel would remain blinded to drugs taken by patients experiencing Suspected Unexpected Adverse Reactions or clusters of SAEs requiring signal detection and evaluation . In this way users would feel they are looking at “their database” despite the data not being permanently stored in an ad hoc format, but remaining in the unified database. It is evident that merging data acquisition is the first step towards combining the two databases.
The need for industry pharmacovigilance professionals to access all types of data (currently stored in separate company databases to which they do not have routine access, or that are difficult for them to access) to understand benefit/risk is critical for the advancement of pharmacovigilance as an integrated service that informs all stakeholders and supports effective regulatory decisions. It is clear that having independent processes in silos to manage clinical trials, non-interventional organised data collection systems, and spontaneous safety data is inefficient, costly and risks resulting in non-compliance. A single system for capturing all information from the study sites assures sponsors of data consistency and facilitates pharmacovigilance access to all data. It is also the foundation for establishing a single data repository that would offer a convenient and more compliant single source for the preparation of aggregate reports and continuous access to all clinical, observational, and spontaneous report data for signal detection, validation and evaluation activities. This means that drug safety personnel would easily have access to all AE information (not only SAEs, as it is now for clinical trials) and would be able to compare the incidence of clinical trial AE reports occurring with the investigational drug to those occurring with placebo or comparators for signal detection purposes. Similarly, drug safety personnel will be able to easily verify if a signal originating from spontaneous reports is confirmed or disproved by the alterations of pertinent laboratory data originating from interventional or non-interventional studies performed in relevant populations. At present drug safety personnel do not have access to this information since it is stored in the study database, not in the clinical one. With the implementation of the unified database, drug safety personnel will have access to data from all studies (in addition to spontaneous AE data), but, of course, once a signal is identified, they will need to select which studies and which information is relevant for signal work-out purposes.
As the “Big Data” story unfolds with growing financial and regulatory pressures, comprehensive access to all data is essential. The unified safety database concept is the ultimate solution that can deliver total data access to pharmaceutical companies and make it easier to protect patient safety and free up resources for achieving this goal. The dilemma is whether the pharmaceutical industry should wait for software vendors to grasp the concept and create a solution or if the industry will take the lead by specifying the fundamental features of a unified database.
5 The Second Area of Improvement: Regulatory Harmonisation
The second key for holistic pharmacovigilance is adhering to consistent formats for submitting regulatory information along with agreed upon data requirements. Despite the existence of international standards and formats for aggregate and single-case reports agreed through ICH, these are not always implemented by all countries. An example is the Periodic Safety Update Report (PSUR); in 2012 the ICH released a guideline describing a new format for PSURs (E2C R2). Unfortunately, this guideline was not implemented by all countries in the same manner. Some countries, such as China, still require PSURs in the old format, others, such as Brazil, require the PSUR in old format in addition to a cover letter, which is a summary of the PSUR and of the Risk Management Plan. In the USA, both PSURs (in the new format) and Periodic Adverse Drug Event Experiences (PADERs) are accepted. However, the periodicity with which PADERs (or PSURs) need to be prepared in the USA (quarterly for the first two years since product launch and then annually) is different from that of the PSUR periodicity in the European Union (EU). Unfortunately, the inconsistencies of document format and submission timelines are not limited to PSURs: ICSRs are another example of these different requirements. The international standard for submitting ICSRs to the authorities is the form developed by Council for International Organisations of Medical Sciences, known as CIOMS I form, if the report is submitted on paper and the ICH E2B format if the report is submitted electronically to the authority. However, in China, there are three templates for submitting ICSRs on paper and none of them include the CIOMS I template. The differences apply not only to the ICSRs submitted on paper, but also to those submitted electronically. In fact, it is sufficient to compare the number of pages of the document on the regional technical implementation of electronic reporting requirements in the EU to those in the USA (more than 100 pages and 15, respectively), to understand how the same requirements have been differently implemented in these two regions. Even the ICH definition for ADRs has not been universally adopted.
As a result, the current differing regulatory arrangements requires pharmaceutical companies marketing a drug across the world to prepare country-specific documents for the same drug without any evidence that patients benefit in that country because formats differ from ICH. More importantly, the need to present the same safety information in different formats increases the bureaucratic burden on regulators, leads to varying opinions on the same datasets, and impedes industry’s ability to effectively monitor the use and adverse reactions of medicinal products in their respective populations. In fact, having different formats in different countries (apart from the European Economic Area ones) does not permit work-sharing among the authorities, but each authority needs to separately assess the same information.
Thus, the lack of harmonisation between the requirements of different countries seriously limits effectiveness of pharmacovigilance by wasting valuable resources. Therefore, there is a need to have a mechanism for a global regulatory system, based on existing organisations such as, WHO, CDISC, and/or ICH, that can bring together regulators of every country to discuss and agree to a minimum set of guiding principles for pharmacovigilance and drug safety, along with formats, to be adopted by the greatest number of countries as possible. Unluckily, at present, not even the countries that are part of ICH completely embrace these guidelines, either because they have different interpretation of the guidelines, (such as is the case of ICH E2B for individual case electronic reporting) or because they have different aggregate report requirements and timelines (as is the case of PADERs) or because they have many additional requirements (as is for European Good Vigilance Practices). Realistically, complete harmonisation would be very hard to achieve but these differences should be administrative in nature around the submission process, while the reports’ format and content should be harmonised. We accept that when there is public health evidence supporting different regulatory approaches, then countries should be free to add additional justifiable national requirements but the reasons for these differences must be published and transparent. This will enable debate about these differences in a neutral forum such as the International Society of Pharmacovigilance.
By creating a comprehensive vision of the interconnected nature of pharmacovigilance, health-related organisations, patients, and health professionals can better observe the entire product life-cycle, including clinical data, observational data and spontaneous reports. It is during the analysis of these data, together, that we develop a better understanding of the full effects of medicines. Not only does the intended and actual use of medicines differ between clinical trials and real world use, but the diversified pool of patients also differs, as the real world use involves pregnant women, children, elderly, disabled, and people with co-morbidities. To create such a diversified patient group during development is prohibitively expensive. The need for active post-market surveillance of real-life therapeutic performance is critical to our understanding of medicine and the data collected needs to be more widespread and further utilised across all spectrums of health professionals. More importantly, the analysis of these data needs to provide useful feedback to those who use, and those who are responsible for advising on the use of medicines. Combining these different sources of data, has the potential to dramatically improve patients’ outcome because the knowledge of this data involves the development, administration, and consumption of medicine. This knowledge can be used to better meet the needs of all people taking medicines, not just the target population in a clinical trial.
Ultimately, effective pharmacovigilance will enable the efficient collection of information from all sources, conversion of meaningful information into data with empowerment of countries to use this information to better understand the trends of medicine use in their countries, and ultimately enable them to make wise therapeutic decisions on the use of a medicine. The evolution of pharmacovigilance cannot happen in isolation, rather it must be part of a larger effort to improve global clinical research and development and reform the regulatory systems, such as that being led by ACRES  to bring together an alliance of stakeholders who share the belief that a high-performing global system that provides people access to safe, effective, dependable and affordable medical treatments, medicines and devices is an essential societal good that benefits all stakeholders.
Compliance with Ethical Standards
No funding was received specifically for the publication of this article.
Conflicts of interest
Giovanni Furlan is actively involved in supporting the ACRES vision and mission as a not for profit organisation. He is employed by a pharmaceutical company. Greg Koski serves as the Chairman, President and CEO of ACRES. Antonio Mastroianni servers as a strategic adviser in communications and partnership alliance to ACRES. He served as the COO of the Uppsala Monitoring Centre from June 2013 to December 2015, and currently works for a company providing management consulting services to healthcare-related organisations. Brian Edwards serves as a vice president, safety for ACRES. He works as a consultant to the pharmaceutical industry.
- 1.WHO. Pharmacovigilance: ensuring the safe use of medicines. Geneva: WHO; 2004.Google Scholar
- 2.Archaedon P, Grandinetti C, Vega JM, et al. Optimizing expedited safety reporting for drugs and biologic subject to an investigational new drug application. Ther Innov Regul Sci. 2013. doi: 10.1177/2168479013509382.
- 3.European Medicines Agency Pharmacovigilance Risk Assessment Committee. PRAC strategy on measuring the impact of pharmacovigilance activities. EMA/790863/2015. London, 11 January 2016.Google Scholar
- 5.Leveson N. Engineering a safer world. Cambridge: The MIT Press; 2011.Google Scholar
- 6.CDISC. CDASH Core Clinical Data Acquisition Standards Harmonisation (CDASH). Clinical Data Interchange Consortium 2016. http://www.cdisc.org/cdash. Accessed 06 Jan 2016.
- 7.International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH). Clinical safety data: data elements for the transmission of individual case safety reports E2B (R2). Step 4 dated 05 Feb 2001. http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Efficacy/E2B/Step4/E2B_R2__Guideline.pdf. Accessed 06 Jan 2016.
- 8.CDASH-E2B Project team. CDASH Serious Adverse Event Supplement. 25 Nov 2013 CDISC. http://www.cdisc.org/system/files/all/standard_category/application/pdf/cdash_sae_supplement_v1___3_.pdf. Accessed 06 Jan 2016.
- 10.International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH). Guideline for Good Clinical Practice E6 R1. Step 4 dated 10 June 2996. http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Efficacy/E6/E6_R1_Guideline.pdf. Accessed 06 Jan 2016.
- 11.El Emam K, et al. The use of electronic data capture tools in clinical research. J Med Internet Res. 2009;11(1):e8. doi: 10.2196/jmir.1120.
- 13.US Department of Health and Human Services Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Center for Biologics Evaluation and Research (CBER). Safety Assessment for IND Safety Reporting. Guidance for Industry. Draft Guidance December 2015. http://www.fda.gov/ucm/groups/fdagov-public/@fdagov-drugs-gen/documents/document/ucm477584.pdf Accessed 06 Jan 2016.