Learnings from DIA Global Annual Meeting 2023 – Boston, MA
As a leading provider of PV services to industry partners in Canada, we are often consulted on setting up modern PV systems. No discussion is complete without understanding the role of automation, artificial intelligence (AI) and machine learning (ML) in case processing; from initial case intake to submission to health authorities.
Regulators worldwide are grappling with this rapidly evolving field. Automating intake and workflows can reduce the resource pressure on drug safety professionals who are constantly asked to do more with less. The ability to streamline operations and reduce cost of monitoring the safety of drugs remains a major motivator.
During the DIA Global Annual Meeting 2023 in Boston, MA, Jane Carroll – VP, Global Commercial and Medical Operations at Moderna, summarized their experience with the introduction of AI into their PV systems during the pandemic as being beneficial for providing remarkable cost savings and improving regulatory and compliance metrics.
The common issues identified with the use of AI/ML in drug safety automation are rooted in the inconsistencies and insufficient details of how the automated systems are validated, used, trained and how their performance is monitored. The health authority guidelines and expectations in this regard are being developed and they might address some of the concerns above.
As with all AI/ML algorithms, the question of bias in how they are developed and implemented is also central to the discussion. What is set-up as “standard” or “good” remains debatable, especially as there is the possible intent to rely solely on AI for decision-making in the future.
Human oversight will still be critical to ensure the AI is not perpetuating a bias from inadequate learning. While those automations can be programmed to flag any safety signals and potential risks in response to anomalies, human input is required to further analyze and assess required actions to be taken to ensure ultimate patient safety based on known and emerging risks of the treatment in question. In addition, according to a Health Canada pilot project conducted between Mar to Apr 2021, human input is required to assess the seriousness and expectedness of cases (DIA Canada 2022).
At ZENITH PV, we appreciate the value AI brings to PV case processing and we firmly believe AI/ML will gradually play a critical role in streamlining PV operations. However, human intervention will remain paramount to interpreting safety signals and critical analysis required to make the determinations on the continued safety of a drug.
Call us for a consultation or attend one of our upcoming live seminars on hot PV topics.