Although automation and artificial intelligence were yet again some of the buzz words at this pharmacovigilance conference, I was even more interested to hear other perspectives on how the industry is changing and what initiatives are being implemented in order to adapt to this environment. The World Drug Safety Congress US took place in Philadelphia on 10-11 April with over 350 attendees and over 100 speakers.
The first session, presented by Jose Vega from Merck, highlighted some of the recent collaboration activities under the umbrella of the TransCelerate organisation. This includes an initiative to harmonise the interpretation of pharmacovigilance regulations, which I am sure many would highly appreciate to enable compliance, patient safety and regulatory efficiency. More interesting was the initial results of a project to evaluate the value of different data sources for case reports. In particular, they found that whilst social media was a vast source of data, the value of the data was very low – after analysing over 180,000 records (by NLP and machine learning), they identified just two valid adverse events and zero safety signals. I was somewhat surprised by these results given many other studies have suggested that there is benefit to monitoring social media (for example see my previous blog post.
However, the value of social media in signal detection was also questioned by Felix Arellano (Roche). He claimed that the key driver for using technology was not to identify more signals but to work with data that is free of false positives. Otherwise we lose the meaning of pharmacovigilance amongst all the non-valid data. With that in mind, do we really want to add social media into the mix when there is little value to the data?
Nevertheless, whether you agree with the potential value of social media data or not, there seems to be a general consensus that using smart technologies is the only way to find the colloquial needle in the haystack in this domain. I for one believe that information from social media can enrich our knowledge of drug safety, as it brings the patient perspective into the equation. Whilst this information may not lead to identification of a new signal, it may provide important information about the impact of adverse effects on quality of life which could potentially tip the risk-benefit balance.
Fabio Lievano from AbbVie gave us a whistle stop tour of many innovations underway to drive the future of pharmacovigilance. This included the integration of pharmacoepidemiology throughout the product lifecycle using techniques such as precedence analysis, synthetic controls, and a patient integration lab. The adoption of artificial intelligence was evident in projects to establish a multi-tenant database platform as well as automation of case assignment. As we have heard before, it is inevitable that those working in pharmacovigilance will need to adapt and become a ‘new generation of PV specialist’.
A highly engaging session was that from Vivek Ahuja at ArisGlobal. He presented his vision of artificial intelligence as being the 4th industrial revolution, after steam, electricity and electronics. It is no wonder then that AI is a continued hot topic for pharmacovigilance. Moreover, in the PV world the progressive revolutions have been from Excel to software, followed by outsourcing, now automation and eventually the Internet of Things. His leading question about automation, ‘Is your organisation ready?’, certainly had many of the delegates wondering if they truly were prepared for this impending technological revolution.
One session that left me with much food for thought was ‘Why reciprocity is the Key to the Future of Pharmacovigilance’, presented by Edward Mingle and colleagues from Celgene. Edward challenged the status quo of the current operating models used in pharmacovigilance as these are unsustainable. So is it possible for us all to work together as we evolve in the technology space? Can we share internal processes, successes and failures, collaborate on pilots and proof of concepts? Because ultimately we are all trying to achieve the same outcomes, and transformation isn’t the responsibility of a sole company.
Personally I believe that those in the industry are willing to share their experiences in the technology space – at least a high level through forums such as the World Drug Safety Congress and similar events. We at Adis Pharmacovigilance have been evaluating the use of automation in our processes, and it has been encouraging to hear that others are taking a similar approach to the application of these technologies. But could we all go one step further and share the specific details of these innovations? Let’s see who puts their hand up first.