Increasing data volumes and increasing data complexity, surge of work for safety teams, continuously changing regulations worldwide – these are just few examples of challenges that Life Sciences industry faces in the field of vigilance. Can technology help the businesses overcome these challenges?
We spoke to Michael Braun-Boghos, Senior Director Safety Strategy at Oracle Health Sciences, to learn more about the key vigilance challenges in 2021, trends that will disrupt the industry in the future, and the role technology will play in it.
What do you believe to be the main challenges the Life Sciences industry in the field of vigilance will be facing in 2021?
I see three main challenges this year. The first is the continuation of a phenomenon that has already been evident over the last few years – the enormously high annual growth rate in the number of incoming AE reports, which has been estimated at 30-50%. This increase has been exacerbated at companies that are developing COVID-19 vaccines as the total number of doses given worldwide has topped the 1 billion mark – with a commensurate rise in AE reporting. This surge in the workload for safety teams has motivated the development of artificial intelligence (AI) solutions because just throwing more human resources at the problem seems to be reaching its limits.
The second challenge also has its origins in COVID, as both R&D and market approvals have been massively accelerated in order to halt the pandemic. With this expedited process, the safety department is under more pressure than ever to ensure that products are safe before they become available to the public. Since there may be less clinical trial data available to them than for past approvals, predictive signal detection (again using AI technology) will start to become an invaluable tool for safety evaluators.
The third challenge is related to the second. As safety teams begin to expand the number and types of data sets in which they conduct signal detection – in order to shorten the time needed to build accurate benefit-risk profiles – they will increasingly need to make sense of different signal scores across multiple data sets and how to aggregate those scores. So-called multimodal methods are being developed to address this problem.
What are the most challenging regulations that should be supported by the safety solutions available on the market?
In addition to the traditional Big Three Regulators – FDA, EMA/EC, and PMDA – MHRA (after Brexit) as well as health authorities in China, Taiwan, South Korea, Russia, Canada, Mexico, and Brazil are starting to publish more requirements that can often be unique to those countries. In addition, regulations for medical devices and combination products are currently experiencing a quick rate of change.
Finally, a new and somewhat concerning trend is for regulators to revise their rules much more frequently than ever before – often after the regulations have already been finalized and in many circumstances with short notice to the industry.
This leaves companies scrambling to reconfigure or upgrade their systems and validate those changes on time, all the while wondering whether another surprise update from the regulator can alter everything again. Safety systems need to be mature, robust, and global in order to agilely adapt to this landscape, and it’s a good idea to work with software developers that have decades of experience in the field.
What are the main processes impacted by these regulations?
Perhaps the main process affected by these regulatory changes is the software upgrade cycle. In order to keep up with the swiftly changing rules, companies can no longer afford to stay on the same software version for five years or more. Some companies have even decided that the least disruptive way to remain up to date is with annual upgrades, which ensure that each version jump is small in scope and therefore in validation effort.
How can an average company balance with minimum disruption existing technology assets with the new regulations?
Many companies that have traditionally managed safety systems on premise are now moving to the cloud as a result of the fast pace of regulatory change and correspondingly frequent upgrade cycle. SaaS services can take over a large number of upgrade tasks and thus significantly reduce the burden on safety and IT teams. In addition, some companies have moved to an annual upgrade strategy, which has several advantages over the traditional “upgrade project” approach:
- the company is always up to date with regulatory compliance, new features, and software support windows
- the scope of changes is small because no (or very few) software releases are skipped
- a small, dedicated team can be put into place to manage the annual upgrades
- a predictable, relatively small upgrade expense can be built into the department’s running costs rather than having to go through a long and difficult budget justification and approval effort for every upgrade.
We are now witnessing the trend of safety caseload increase. What are the new data sources in 2021 that will contribute to detecting safety issues earlier?
Recently a research paper (1) was published that demonstrated how machine learning methods can be used with data from in vitro pharmacology assays to predict adverse reactions before drug candidates reach human clinical trials or enter the market as approved medicines. I believe this is a very promising avenue for further study.
What do you believe to be the main triggers of this growth?
One of the triggers is certainly an increase in regulatory requirements. Unfortunately I don’t believe the huge growth in incoming cases necessarily translates to a better understanding of product safety profiles. Often a large percentage of the cases are non-serious reports, consumer reports, reports from patient support programs, and reports on very mature products – many of which can be of low PV value, but still have to be processed.
Contrast this to new data sources being used for signal detection, based on a voluntary decision made by the safety team in order to enhance their understanding of benefit-risk. Such expansions of safety data for signaling may be more effective from a scientific point of view than the processing of an ever-ballooning number of individual cases.
How can the use of technology help companies to cope with this growing caseload?
Next to the cloud, the technology that will significantly transform PV and multivigilance is AI, which is quickly evolving from a dream to reality. In the area of case intake, AI can be used today to automatically extract and structure information in incoming documents and thus avoid the need for manual data entry, which is one of the biggest time sinks in case processing. In future, AI will eventually be used to automate most or all of the other workflow steps in case management, and will have a similar positive impact on signal detection and management.
Do you see the growth in the pharmacovigilance outsourcing?
There will always be a need for CROs, as there will always be companies that are either too understaffed or underequipped to deal with the workload, especially with the current case growth rates.
Companies now have a range of options spanning both AI and outsourcing, and everyone will choose a mix that makes sense for them and that they are comfortable with. Nonetheless there is a finite number of safety specialists in the world, whether they are employed by pharmas or CROs. In effect, outsourcing is just pushing this problem downstream, and CROs will also have to adopt AI solutions in order to solve it.
What are the activities that are outsourced?
In general, individual case management is outsourced more frequently than other activities such as periodic reporting or signal management. And within case processing, if any workflow step is kept in-house, it’s medical review.
Often companies want to keep full control of tasks that involve medical judgment or risk assessment. Now, with AI for case intake available for the first time as an alternative to outsourcing, it will be interesting to see which strategy prevails. It’s truly a fascinating time to work in safety as the industry evolves into its next phase!
Manage adverse event cases more accurately and more efficiently whilst ensuring regulatory compliance!
In Arithmos, our mission is to support our clients in embracing digital transformation that allows them to develop therapeutics for patients in a smart, innovative, and cost-efficient manner.
To help our customers effectively harness technology and facilitate and optimise their research, in 2011 we joined Oracle PartnerNetwork – a unique network designed to accelerate value realisation for companies.
Thanks to Oracle partnership, we are able to offer you a powerful tool, Argus, that was designed to solve the pharmaceutical industry’s biggest regulatory challenges.
Reference
(1) Machine Learning Guided Association of Adverse Drug Reactions with in vitro Target-Based Pharmacology. Ietswaarta R, Arat S, Chen AX, Farahmand S, Kim B, DuMouchel W, Armstrong D, Fekete A, Sutherland JJ, Urban L. EBioMedicine 2020.
