PRE-ACT: From prediction to prevention & making risk intelligence actionable in primary care
Health systems are becoming increasingly good at predicting risk. Routinely collected health data, machine learning, clinical risk scores, remote monitoring and digital technologies can identify people who may be at increased risk of disease or adverse health outcomes – sometimes well before those outcomes occur.
But prediction is not prevention.
Knowing that someone is at increased risk only creates value if that information can be translated into an appropriate response: something understandable to the person, clinically meaningful to professionals, feasible within routine care, and capable of supporting action. This is the problem that the Prediction-Enabled Action for Self-care in Primary Care (PRE-ACT) consortium sets out to address.
The PRE-ACT consortium is a collaboration between Imperial College London Self-Care Academic Research Unit (SCARU) and the Research Unit OPEN, University of Southern Denmark, bringing together researchers, clinicians and implementation partners with expertise spanning primary care, public health, risk prediction, digital health, behavioural science and self-care.
Closing the gap between risk and action
Much of the innovation in predictive healthcare has understandably concentrated on improving model performance: identifying the right variables, improving discrimination and calibration, and determining whether an algorithm can accurately classify risk. These are necessary questions. But they are not sufficient:
Between a risk estimate and an improved health outcome lies an important translational pathway. Risk must be interpreted. It must be communicated appropriately. Decisions need to follow. Patients may need support to act. Primary care teams need workable pathways through which to respond. And all of this must occur without widening existing inequalities.
PRE-ACT starts from the proposition that prediction should be viewed as the beginning of a preventive pathway, rather than its endpoint.
The initiative therefore brings together predictive analytics, decision support, primary care and self-care within a single translational framework. Its focus is not simply on whether we can identify risk, but on what should happen next.
Why self-care matters
This question is particularly important as healthcare moves towards earlier intervention and greater participation by patients in managing their own health.
The 2018 Declaration of Astana on Primary Health Care provides an important foundation. It renewed the global commitment to primary health care as a route to universal health coverage and emphasised, among other principles, the importance of prevention and health promotion, empowering individuals and communities, and enabling people to acquire the knowledge, skills and resources required to maintain their health.
PRE-ACT takes this principle into an increasingly data-driven healthcare environment. If predictive technologies can identify an opportunity to prevent deterioration, then we should also ask how individuals can be meaningfully supported to respond. That may involve self-monitoring, behavioural change, appropriate use of digital tools, supported self-management, or timely engagement with healthcare services.
Self-care in this context does not mean shifting responsibility from health systems to individuals. It means designing systems in which people are equipped and supported to participate in prevention, with appropriate professional oversight and clear routes into care.
Prediction-enabled prevention must also be responsible
There are important safeguards. A technically sophisticated predictive system may still have limited value if clinicians cannot interpret its output, patients do not understand what the result means, actionable services are unavailable, or those with lower digital access are systematically disadvantaged. PRE-ACT therefore places issues such as equity, human oversight, transparency, evidence, privacy, clinical governance and implementation readiness alongside predictive performance.
These considerations matter because the future of prevention will increasingly involve interactions between people, professionals, health systems and intelligent technologies. Success cannot be judged solely by the accuracy of an algorithm. We also need to ask whether an intervention is understandable, actionable, acceptable, equitable and capable of being integrated into real-world care.
From predicting risk to changing outcomes
One way of expressing the PRE-ACT pathway is: Prediction → Interpretation → Decision support → Activation → Self-care & clinical action → Outcomes
Each transition matters. A failure at any point can break the chain between recognising risk and preventing harm. Conversely, designing these elements together creates an opportunity to move predictive technologies away from passive risk stratification and towards genuinely preventive healthcare. This is the broader ambition of PRE-ACT: to help define what responsible, human-centred and actionable prediction should look like in primary care.
As predictive technologies become more powerful, the central question may therefore become less “How accurately can we predict what happens next?” and more: “What can we enable people and health systems to do differently because we know?” That is the point at which prediction becomes prevention.