Will apps solve the problem of insufficient specialist resources in the NHS? A case study from children's insomnia

Can a smartphone replace a specialist? That’s the question many health services, including the NHS, are grappling with. In the face of growing demand and scarce specialist staff, particularly in paediatrics, digital health interventions promise scalable, cost-effective solutions. But as a new UK trial on sleep problems in children with epilepsy shows, apps alone may not be the silver bullet.

Children with epilepsy are 12 times more likely to have sleep problems than their peers. Sleep issues affect learning, behaviour, seizure control, and family wellbeing. Yet, most NHS epilepsy clinics lack the time or specialist staff to address sleep in routine care. With these constraints in mind, researchers developed COSI – the CASTLE Online Sleep Intervention – a self-guided behavioural sleep programme for parents, designed to be delivered entirely online, without professional support.

A multi-centre NHS trial tested whether adding COSI to standard care could improve sleep for children aged 4–12 with epilepsy. The trial, conducted in 26 NHS paediatric epilepsy clinics across the UK, asked a bold question: could a web-based tool, accessed by parents at home, replace the need for specialist sleep support?

The results were mixed. On the one hand, COSI showed potential. Children in the intervention group fell asleep on average 16 minutes faster, as measured by actigraphy (a form of wearable sleep tracking). Parents who engaged with the programme reported increased knowledge about child sleep and overwhelmingly said they would recommend the tool to others.

On the other hand, the primary measure – a parent questionnaire about child sleep problems – showed no significant improvement. And only around half of families engaged with the intervention’s core behaviour-change content. From an economic perspective, COSI was not cost-effective under standard NHS thresholds.

So, what went wrong?

First, COSI, while accessible, was also isolating. Unlike traditional therapy, there was no therapist, nurse, or educator to motivate or troubleshoot. Behaviour change science tells us that human contact matters – not just for emotional support, but to encourage sustained action. Engagement dropped off quickly, a well-documented phenomenon in digital health known as "Eysenbach’s Law of Attrition".

Second, the tool required parents to be both patient and practitioner. This is a high ask – especially for parents already managing a chronic condition like epilepsy, which often carries with it the burden of stress, sleep deprivation, and anxiety about night-time seizures or sudden death. Without external prompting or accountability, many simply didn’t use the tool enough to benefit from it.

This doesn’t mean the concept failed. Rather, the trial offers a powerful insight into the limits of purely digital healthcare. While COSI’s sleep strategies are evidence-based and have worked in other settings (like autism and ADHD), success often depends on how interventions are delivered, not just what they contain.

A key lesson from the CASTLE Sleep-E study is that digital tools alone cannot bridge the gap left by workforce shortages. But hybrid models might. Imagine a paediatric epilepsy nurse trained in basic sleep support, signposting the COSI platform while providing encouragement and personalised feedback. This kind of "blended care" could balance cost-efficiency with the relational support needed to achieve real-world impact.

The trial’s authors suggest exactly this: that online interventions like COSI hold promise but need human scaffolding to succeed. It’s a vision increasingly backed by global evidence, which shows that digital interventions are most effective when paired with professional guidance – even in small doses.

As the NHS looks to scale digital health solutions across its services, the message is clear: technology can be an enabler, but not a replacement, for human care. Apps may help ease the burden on overstretched clinics – but to truly change lives, they need to be part of a broader, blended system that empowers both patients and the professionals who support them.

In the case of childhood sleep problems in epilepsy, the path forward may lie not in choosing between digital or specialist care – but in combining both.

Read more about the results of the trial here.