It’s becoming increasingly clear that minimal signal detection is not cutting it, when it comes to doing the best for patients and staying ahead of emerging issues. For smaller pharma organisations and MAHs lacking state-of-the-art systems there has to be a better way, says Qinecsa’s John Cogan.
It can be challenging to stay ahead of emerging safety issues and accurately maintain a drug’s safety profile, especially as a smaller pharma company without advanced systems or statistical expertise. For these entities, the process of post-marketing signal detection has traditionally been a manual and passive one geared to routine safety reporting. But that position is unsustainable now.
All major pharmacovigilance regulations and guidelines, such as those from the European Medicines Agency (EMA), the US Food and Drug Administration (FDA) and the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH), advocate the use of statistical methods in signal detection. They assume that monitoring and detection will take place at a suitable frequency for the given context.
There are reputational implications if companies do not employ some degree of statistical signal detection (SSD), too. If the health authorities know more about a company’s products and any emerging safety patterns before the organisation itself, this does not look good. Ideally, any anomalies would already be known internally, allowing the company to prepare a response and initiate action before they are audited.
Making statistical methods more accessible to all
Statistical methods are intended to supplement traditional PV methods. These methods, also collectively known as Quantitative Signal Detection (QSD), use mathematically-calculated scores to assess the relationship between a medicine and an adverse event. Minimum thresholds are then applied to highlight or raise alerts on specific drug-adverse event pairs, and avoid statistical noise or false positive findings.
There are three broad categories of statistical methods. Disproportionality methods are useful for finding new patterns in large data sets. Change detection methods, which pick up on increased frequency or severity, are useful to monitor known issues or identify manufacturing issues. Reference data methods match findings to identified medically important events, for example through comparison with Designated Medical Events (DME) lists maintained by the regulators.
Until now, the prospect of combining these methods, and applying appropriate thresholds and monitoring cadence, to optimise signal detection for a product, has felt unattainable for smaller safety teams. But more recently other options have emerged to fill the gap.
QSD services delivered by a trusted partner have levelled the playing field. For those with limited budgets or internal infrastructure, this has provided affordable access to highly specialised expertise and clear, real-time oversight of their products’ ongoing safety performance.
Outsourcing QSD brings access to the latest statistical algorithms, techniques and expert consultancy within the reach of all pharma companies, as on-demand resources. Meanwhile the ability to perform simulation studies on real data can help determine the minimum alert thresholds, as well as the number of times surveillance of a particular product will be needed, helping companies to strike a balance between cost and required vigilance.
Opening up new strategic opportunities
The opportunity is not just one of risk management; it is about building richer intelligence about a product and its relative impact in the world. The major health authorities, with their extensive databases about all products, are able to perform rich segmentation to understand expected health issues – such as the expected rate of heart attack – in geriatric populations, paediatric populations, and general populations, and note any significant changes linked to classes of drug or individual products over time.
It can then be determined whether the number of reports of particular adverse events are within expected parameters, or whether they are disproportionately greater in incidence for those taking a particular drug. In this way they can distil, for example, whether patients taking a particular nonsteroidal anti-inflammatory drug (NSAID) are statistically more prone to a heart attack or stroke than those taking an alternative. Being ahead of these trends means drug companies can perform an earlier analysis of, and investigation into, what may be causing those spikes.
Proactive change detection signal detection methods will also shine a light on any emerging issues with manufacturing, indicating possible contamination in one of the manufacturing pipelines for example. The earlier potential issues of this nature are detected, the lower the risk of patient harm, and of reputational damage.
Further value may come in detecting possible issues linked to complex products that require careful administration, such as once-a-day medicine formulations versus split-dosing. Known issues have emerged here with cancer-related pain treatment, involving the use of controlled-release opioid products. If the intended effects of these medications don’t last for the expected period, so that the patient goes on to experience pain, this should be picked up in the safety reporting. The solution to managing an identified signal of this nature may be better healthcare professional training, or patient information leaflet updates.
The biopharma evolution towards advanced therapeutics will further increase the need for statistically-robust safety signal detection. Take CAR-T cell therapy, which involves genetically engineering a patient’s own T cells to attack cancer cells. Data capture of adverse events on such products is enhanced due to the setting in which they are administered. The relatively high data volume per patient can quickly overwhelm smaller safety teams reliant on more manual detection methods. Yet regulator scrutiny of such products is high, so statistical rigour is essential.
Staying ahead of emerging issues
Whatever the make-up of companies’ respective portfolios, there is a risk to doing the bare minimum when it comes to post-marketing safety signal detection. Making statistical signal detection more accessible paves the way for all pharma organisations, whatever their size and scope, to be more proactive and strategic about their real-world safety monitoring and responsiveness.
Adopting QSD will stand companies in good stead for the long term, too. That’s as data volumes increase, and as continued scientific breakthrough extend the human lifespan – which could give rise to new health challenges, or adverse events not previously anticipated.
About the author
John Cogan, COO of Qinecsa, is a life sciences industry veteran with over 30 years’ experience in transforming strategy, organisations, processes, and systems across R&D. Qinecsa is a global provider of innovative, digital pharmacovigilance solutions, including cloud-based analytics solutions and services for medical research and healthcare delivery.