Patients deserve better – why AI alone won’t solve pharma’s fragmented safety data problem

A major shift in approach is needed if patient-reported adverse events are to enable clearer safety insights and better safety outcomes.

Simply throwing AI at dispersed and incomplete safety data cannot achieve that. Progress starts with better data capture, and that requires a stronger mandate and guided digital experiences for those tasked with collecting it.

The ambitions of patient safety and ongoing pharmacovigilance are very clear – making drugs and therapies as safe as they can be so that they deliver ever better outcomes. Yet this entire endeavour is threatened by the enormous complexity in capturing meaningful, high-quality insights and acting on them swiftly.

Unless the pharma industry gets a grip on this problem, it risks missing a huge opportunity. That includes the potential to harness AI to derive better intelligence.
There is a fine distinction here. While AI technology promises much, it can only work with what it has, so trying to get it to turn bad or incomplete data into something of greater value is futile. AI’s role in transforming safety data is rather more nuanced.

Too often the safety data (adverse event reports) gleaned from patients and healthcare providers is patchy, incomplete and difficult to follow up. Diverse channels for reporting, and inconsistency in what and how much is captured, render findings hard to combine in a meaningful way – e.g. as the basis for actionable intelligence, blended with data combed from scientific journals, online forums, etc.

Today, technology vendors are bold in their claims about AI. Too often they imply that companies that may be grappling with fragmented and incomplete data can simply leave the chaos as it is and let AI make sense of it.

But it is not that straightforward. The first priority must be to get better data in when the opportunity is greatest; at the point when reporting is being done intentionally. Particularly where the facilitator is a paid third party contracted to deliver a patient support programme (a PSP vendor).

There is an argument that regulators should be exerting more influence around original patient safety data capture, to build the richest possible understanding of each patient’s experience. Until the authorities mandate that better data is captured at source wherever possible, and until the entire industry understands that AI is not a simple fix to the complexity problem, safety functions and their patients will be no better off.

Just as there is a ‘golden hour’ in critical medical interventions, there is a golden opportunity to capture maximum patient AE insights, that is at the time of initial reporting.

Anecdotally, large pharma organisations report that barely just 10% of attempts at information follow-up (to fill in gaps in the narrative) are successful once initial details of any side-effects have been reported. So, to miss this window is to forego those insights altogether.

Professionals tasked with capturing and collating patient safety data need better incentives and facilities, then, to capture a more rounded picture of an individual’s experience and wider health profile up front. It may be that the data collectors have a wider remit as patient support programme managers or as speciality pharmacies.

If their main vehicles for receiving and registering adverse event notifications are an email address, paper form or Word document, they will remain compromised in their ability to share meaningful insights.

‘Every safety data point is precious and needs to be treated as such from the moment of capture through to analysis’

With so much claimed in this industry about improving patient-centricity, it is puzzling that only some pharma companies (that are paying for the data) and regulators (with a remit of upholding quality and safety) are going all out to drive and enforce the capture of complete and high-quality data first time.

Better data would provide a much clearer picture of adverse events and what may be contributing to them (such as drug interactions, pre-existing conditions).

Consider the high volumes of incoming data ready to be recorded around weight loss drugs traditionally associated with diabetes treatment – those targeting GLP-1 and/or GIP receptors to control appetite. Side effects may range from digestive issues to reduced muscle and bone mass.

The opportunity to capture this information widely and draw trend information from it is rich and important, and arguably there should be a mandate for assigned professionals to improve the consistency and value of this activity.

Ultimately, this is about empowering respective actors with the right tools for the job, as well as a sense of accountability for the quality and onward value of the data that is captured.

Part of the challenge is the perceived business change associated with overcoming fragmented safety data. Today’s reality for pharma companies is a complex, multichannel landscape through that relevant data can flow.

When technology companies tell them that AI can help pull together all of the insights, then, without the need for any reformatting or data work, the prospect is very tempting. Yet AI can’t fill in the gaps, and strategies and approaches for improvements must provide for that.

Generally, it will be more effective to use AI to prompt good and comprehensive data capture up front (e.g. by guiding the user to provide additional information), for instance, than to apply tools later in the process once the opportunity has passed – although this too is a valid option if initial attempts were not possible or were unsuccessful.

AI could also be used to tailor and optimise the digital experience, e.g. for each inputter’s persona (e.g. patient vs HCP/pharmacist or CRO), likely medical knowledge, native language, the device being used, and so on.

Pharma’s aim, longer term, must be to break down safety data complexity and enable a more holistic, richer picture of a drug and its impact once in the market – starting with capturing more via the earliest patient feedback.

In the 2020s, sorting through reams of data sent by email and then trying to chase down missing details is not a fit-for-purpose system when better alternatives are available. It is inefficient, ineffective and costly, and it serves no one.

An optimised digital experience for the reporter, with pertinent questions or prompts to capture all of the preferred detail, has been shown in pharma company deployments to enable 70% overall improved efficiency, including reduced follow-up.

With the current pace of drug innovation, every safety data point is precious and needs to be treated as such, from the moment of capture through to signal detection and analysis. Transformation requires that the pharma industry demystifies the complexity and deploys the right tools for the given situation – including AI, where appropriate.

Companies must also join up systems and overcome data silos, allowing more insights to flow into all the relevant downstream systems – where they can be analysed and actioned without the need for manual data re-entry.

All of this will support more accurate triaging and onward decision-making, simultaneously boosting productivity and elevating patient outcomes.

As personalised medicine continues to grow as a proportion of pharma pipelines, and as smart devices do more to track individuals’ health, the data collected will inevitably become more patient-centric.

Developing better practices now will set the pharma industry in good stead for what’s to come.