The Jevons Paradox and Digital Health: Why Efficiency Can Increase Demand

For more than a decade, digital transformation has been seen as the solution to health system pressures. Artificial intelligence, automation, and predictive analytics promise efficiency, speed, and population-wide scale. Yet new research published in The Lancet Digital Health challenges this, arguing that efficiency gains in digital health do not necessarily reduce cost or workload. Instead, they can expand demand across a system already stretched to its limits.

In radiology, the Jevons Paradox is already visible. Artificial intelligence tools can now triage and report scans faster than ever, reducing turnaround times and lowering the per-image cost. Yet this very efficiency encourages clinicians to order more scans and expand screening criteria. What was once a scarce diagnostic resource becomes a routine step in patient diagnosis. Total imaging volume rises, not falls. And indeed, the NHS England's Diagnostic Imaging Dataset shows that total imaging activity continues to rise year on year. Between 2016 and 2023, CT activity increased by roughly 60% and MRI scans by more than 30%.

Each efficiency gain expands new clinical questions, findings, and follow-up activity. The result is a system that is technically faster but collectively busier, with value realised only if capacity, prioritisation, and workforce planning evolve alongside it.

The Jevons Paradox presents us with a different rationale to examine how technological efficiency interacts with human behaviour, system capacity, and social need. And to question...

What if efficiency alone does not save money or time, but simply reveals how much more we have to do?

The Jevons Paradox originates from nineteenth-century economics. When steam engines became more efficient, coal use did not decline. It grew. Cheaper and more efficient engines made new industries viable and increased overall consumption of coal. Efficiency created expansion, not conservation.

Reid and Mateen's Lancet Digital Health paper applies this historical observation to modern global health. They argue that the same principle applies to artificial intelligence and digital technologies. As digital tools make it cheaper and faster to detect, monitor, and manage health problems, overall demand for care rises. The unit cost of a service decreases, but total consumption increases because need is surfaced earlier, expectations rise, and new kinds of demand appear.

When technology makes it easier to access or identify health needs, people use more of it. When algorithms detect risk with greater accuracy, clinicians intervene more often. When care becomes more convenient, more people seek it. The paradox is not a flaw in the technology but a feature of how systems respond to efficiency.

The traditional model of digital investment in the NHS assumes that greater efficiency equals lower cost.

• Automation will reduce administrative workload.
• Remote monitoring will reduce hospital admissions.
• Predictive analytics will reduce emergency activity.

These assumptions are embedded in many digital business cases.

However, when viewed through the Jevons lens, these expectations can be misleading. Efficiency gains often redistribute work rather than remove it.

Digital triage may shorten initial contact times but generate more follow-up appointments. Automated documentation may speed up note taking but increase the time needed for review, validation, and supervision. Predictive models that identify new risk groups can increase overall clinical activity rather than reduce it.

If we equate efficiency with savings, we risk under-estimating total system demand. Efficiency is only beneficial if the value it creates can be absorbed and equitably distributed. Without this, efficiency can generate new pressures that compromise quality or widen inequality.

Digital tools can create immense value: earlier detection, better coordination, and improved experience. But their true success depends on what the system does with that value.

For commissioners, this means redefining digital health goals. Instead of promising that technology will save money, we should ask whether it improves outcomes and equity.

Instead of claiming that automation will reduce our workforce, we should consider how it changes the nature of work and allow us to plan accordingly.

Instead of measuring throughput, we should measure how digital innovation supports more accessible, high-quality, and fair care.

A mature approach treats efficiency as one input to a larger equation of value. The output must be measured in population health gain, workforce sustainability, and equity of access.

Recent deployments illustrate the pattern. Predictive risk algorithms have been introduced to identify patients most likely to deteriorate or be readmitted. These tools are accurate and clinically useful. Yet they also increase the volume of interventions triggered. Each prediction carries an expectation of action. More accurate detection means more activity, not less.

Similarly, digital symptom checkers and online access routes have made entry into primary care more convenient for many patients. But they have also changed demand dynamics.

General practices report that while digital access reduces waiting times for some, it can increase the total number of contacts. Lower friction creates higher utilisation. ‌In both cases, the technology works as intended. The paradox emerges because systems are not prepared for the new pattern of demand. Efficiency creates visibility, and visibility creates pressure.

Recognising the Jevons Paradox within health systems means we need to consider governing innovation differently.

1. Commissioning models should account for the demand effects of efficiency

When a digital intervention promises to make a process faster or cheaper, the business case must also model the potential increase in volume that follows. Demand modelling should be part of every investment decision.

2. Success metrics should focus on outcomes rather than inputs

Reducing appointment length or staff time per case is not an end in itself. The more meaningful question is whether the intervention improves health outcomes, experience, and equity.

3. Digital strategy must be integrated with workforce planning

Efficiency changes the kind of work clinicians do. As some tasks are automated, shifting roles and requiring new responsibilities for oversight, supervision, and patient engagement. Training and staffing models must evolve to reflect these changes.

4. Inclusion must be built in

The effects of efficiency are not evenly distributed. Populations with lower digital access or confidence may benefit least from the expanded opportunities that efficiency creates. Without targeted inclusion efforts, efficiency can widen inequities.

5. Evaluation frameworks should measure the whole system effect

This includes total workload, health outcomes, and variation in benefit across population groups. Only by tracking these dimensions can we understand whether efficiency has created net value or just shifted demand elsewhere.

The NHS operates under sustained fiscal and workforce pressure, which makes every investment accountable for demonstrable value. As digital tools become more sophisticated and more deeply embedded in care pathways, the Jevons Paradox reminds us that efficiency is not the same as progress.

Digital innovation can make care faster, smarter, and more personalised. It can also expand what we regard as clinically possible and publicly expected.

The paradox cautions that efficiency alone does not deliver value. Digital strategies must be built around outcomes, capacity, and fairness.

True success comes when we measure progress by outcomes, equity, and sustainability rather than by throughput or speed, delivering the purpose of digital health which is not to do more for less, but to do better for all.