As 2025 wraps up, it’s worth stepping back and looking at what actually moved the needle in sleep medicine this year. Not every development was flashy, but several will shape how we diagnose and treat sleep disorders for years to come. Here’s my honest assessment of the year’s most significant advances.

New Treatment Options Made Headlines

The biggest pharmacotherapy news was the continued expansion of orexin receptor antagonist (DORA) prescribing. Lemborexant and suvorexant have both seen significant uptake as clinicians grow more comfortable with the drug class. Real-world data published throughout 2025 has reinforced what the clinical trials showed — these medications maintain their effectiveness over longer treatment periods without the tolerance issues that plague older sleep drugs.

On the device front, hypoglossal nerve stimulation (HNS) for obstructive sleep apnea continued to gain ground. The Inspire system saw expanded insurance coverage across several major payers, making it accessible to more patients who can’t tolerate CPAP. Early data from next-generation implantable devices suggest even smaller profiles and bilateral stimulation capabilities that could improve response rates.

We also saw growing interest in oral appliance therapy refinements. Custom-fit mandibular advancement devices, guided by AI-analysed imaging of upper airway anatomy, are producing better first-fit success rates. This matters because a significant number of patients historically needed multiple adjustments before finding a therapeutic jaw position.

AI Got Real (Not Just Hype)

2025 was the year AI in sleep medicine started delivering tangible clinical value rather than just impressive conference demos. Several developments stood out:

Automated polysomnography scoring reached a level of accuracy that’s genuinely comparable to inter-scorer agreement among trained technologists. The American Academy of Sleep Medicine published updated guidance on the validation requirements for AI scoring systems, which should help standardise quality across vendors.

Home sleep testing got smarter. AI algorithms applied to single-channel and multi-channel home sleep apnea test (HSAT) data are improving diagnostic accuracy for conditions that traditionally required in-lab polysomnography. This is particularly important for expanding access in rural and underserved areas.

Predictive models for treatment adherence moved from research papers into clinical products. Several platforms now offer risk stratification that identifies patients likely to abandon CPAP therapy within the first 90 days, enabling targeted early intervention.

Companies like Team400 have been working with healthcare providers to bring AI implementations into clinical practice in ways that actually work — focusing on workflow integration rather than just algorithmic performance. That practical focus has made a real difference in adoption rates.

Research That Changed Our Understanding

Several studies published in 2025 reshaped how we think about sleep disorders:

A landmark longitudinal study in The Lancet Respiratory Medicine provided the strongest evidence yet linking untreated moderate-to-severe OSA with accelerated cognitive decline and dementia risk. The effect was independent of known confounders like BMI, hypertension, and diabetes. This gives clinicians much stronger ground to stand on when counselling reluctant patients about the importance of treatment.

Research on sleep and metabolic health also matured. The connection between short sleep duration, circadian disruption, and insulin resistance is now established enough that endocrinologists are increasingly screening for sleep disorders in patients with treatment-resistant type 2 diabetes.

Telehealth Continued to Reshape Access

The pandemic-era explosion of telehealth in sleep medicine didn’t just persist in 2025 — it matured. Virtual consultations for initial sleep evaluations, CPAP troubleshooting, and follow-up care are now standard at most sleep practices. The American Telemedicine Association reported continued growth in sleep-specific telehealth utilisation.

The most interesting evolution has been the integration of remote monitoring data into telehealth encounters. Instead of asking patients how their CPAP is going, clinicians can pull up 90 days of objective usage data, leak metrics, and residual AHI scores before the visit even starts. This makes virtual appointments substantially more productive than they used to be.

What Didn’t Work

Not everything was rosy. Direct-to-consumer sleep gadgets continued to proliferate with dubious clinical validation. Wearables that claim to diagnose sleep apnea remain unreliable compared to proper testing, and I’ve seen patients delay appropriate diagnosis because their watch told them their sleep was “fine.”

Consumer sleep supplements — particularly the branded melatonin gummy market — exploded without any meaningful regulatory oversight of dosing accuracy. Independent testing continued to find wildly inconsistent melatonin content across products, sometimes varying by 400% or more from what’s on the label.

And despite growing evidence for its effectiveness, access to cognitive behavioural therapy for insomnia (CBT-I) remained frustratingly limited. We know it works. We know it should be first-line treatment. We still don’t have enough trained providers to meet demand.

Looking to 2026

The trajectory is encouraging. Treatment options are expanding. Technology is becoming genuinely useful rather than just novel. And the broader medical community is taking sleep health more seriously than ever.

But the fundamental challenge remains: too many people with treatable sleep disorders go undiagnosed, and too many who are diagnosed don’t stay on therapy long enough to benefit. Solving those problems — the access gap and the adherence gap — is where the real progress needs to happen next year.

Here’s to sleeping better in 2026.