We have built modern medicine on evidence. But what happens when the evidence doesn’t fit the person in front of us?

Evidence-Based Medicine (EBM) shaped my training and early clinical thinking. We were taught that research, data, and guidelines provided the safest route to good care. Follow the evidence, reduce variation, improve outcomes. It offered clarity in environments filled with uncertainty.

But over time, a difficult question emerged: what happens when people do not behave like the averages on which evidence is based?

Healthcare systems increasingly rely on processes and guidelines derived from large-scale studies. These frameworks may have contributed to improved safety and consistency. Yet, in creating them, we may also have built rigid structures that struggle to accommodate real human complexity.

Recent academic critiques suggest that EBM, once revolutionary, may now be facing a crisis of relevance. Papers such as The Emperor’s New Clothes: A Critical Appraisal of Evidence-Based Medicine and Evidence-Based Medicine: A Movement in Crisis? argue that medicine risks mistaking measurable certainty for meaningful understanding.

EBM underpins guidance from NICE and the Royal Colleges, shaping national standards of care. Evidence hierarchies place systematic reviews and randomised trials at the top, while qualitative research, clinical judgement, and lived experience sit near the bottom.

But this hierarchy carries unintended consequences. By privileging what can be measured across populations, we risk overlooking those who do not fit expected patterns. Outliers become statistical noise rather than signals that our models may be incomplete.

Qualitative research captures patient and family experience, offering depth and context numbers alone cannot provide. Yet such work is often treated as weaker evidence rather than evidence answering different questions. Similarly, individual case experiences are frequently dismissed, despite the fact that many advances begin by noticing when something does not fit.

Guidelines, built from aggregated evidence, increasingly shape decisions. While they support safety and consistency, they can also become perceived as unquestionable truths. Clinicians risk becoming passive implementers rather than thoughtful practitioners, with fear of deviation replacing confidence in professional judgement.

If EBM alone delivered transformative health gains, we might expect improvements comparable to sanitation, vaccination, or smoking reduction. While progress has been made, many of today’s greatest challenges: intesectionality, trauma, neurodevelopmental vulnerability, social complexity, long-term conditions and resist protocol-driven solutions.

This is not an argument to abandon EBM. Evidence and guidelines remain important. But medicine now faces a landscape defined by complexity, not simplicity.

We need to expand what we mean by evidence.

We must value lived experience, include families and carers, and recognise that context shapes outcomes. Philosophy, systems thinking, and relational care must sit alongside science to help us navigate uncertainty rather than simply trying to eliminate it. Improving outcomes in the real world requires moving beyond rigid hierarchies of evidence and rediscovering judgement, humility, and curiosity in clinical practice.

A Call to Action

Healthcare leaders, clinicians, commissioners, researchers, and educators now face a choice.

Do we continue refining systems designed for predictable problems, or do we build new approaches capable of holding complexity?

We need to:

• Treat lived experience as essential evidence, not an optional addition.
• Support clinicians to use judgement alongside guidelines.
• Design services for people who do not fit standard pathways.
• Learn from outliers rather than exclude them.
• Create systems that value reflection and learning over blame and compliance.

Evidence should guide us but it should not blind us.

The future of medicine will belong not to those who abandon evidence, but to those willing to broaden what counts as evidence in the first place.

And that work starts now