Why Sleep Apps Don't Work

Hello friends,

Sleep apps satisfy my inner nerd. I like being able to see every statistic about myself. In fact, during my A-levels I became obsessed with tracking myself on excel spreadsheets and biohacking. It didn’t end well.

Nerds across the world self-experiment, hoping to improve their cognitive performance. Some microdose LSD — taking a small dose everyday — just enough to increase creativity, but not enough to get ‘high’.

Billionaire PayPal co-founder Peter Thiel reportedly has blood transfusions with younger people – believing it to have regenerative life extension properties. This has led to interesting headlines such as Peter Thiel denies being a vampire.

Sounds dumb (and dystopian), but this procedure has a name in Medicine: heterochronic parabiosis. In studies, older mice have had their circulation connected with younger ones — showing life extension and even neurogenesis (for the older mouse).

Fortunately, the Scandinavian Blood Transfusion Database answered this question. Thousands of patients regularly receive blood transfusions. You can compare the age of their blood donors to their lifespan. On average, people who receive blood from younger donors should live longer. They don’t.

Tinkering with and measuring every bodily function sounds nice in theory, but sometimes science shows that it doesn’t work. This week, I want to talk about sleep apps.

I once asked a sleep consultant what she thought of Matthew Walker’s best seller: Why We Sleep. She scrunched up her face and said it was a bit… dramatic. Apparently it had caused a ‘middle-class run’ on her clinic. If you’ve read the book, here’s a great long read on everything incorrect/made up in just the first chapter.

The gold standard sleep test is the ‘sleep study’ (polysomnography). Sometimes a small extra test is done called a wrist actigraph. A small device on the wrist measures movement, tracking sleep/wake.

But people move in their sleep — so how does it differentiate between ‘sleep movement’ and ‘awake movement’? It doesn’t, it’s shown to have sub-50% specificity. As good as flipping a coin.

Sleep apps try to emulate this method, which isn’t very useful to begin with (but does have specific uses). Then they use a proprietary algorithm to estimate ‘sleep quality’ — a made up term based on no evidence.

When sleep apps are compared with the gold standard ‘sleep study’ — they show poor correlation. Sleep apps might be the new horoscope.

The next question is, if people like seeing their sleep quality (even though it’s made up) — then what’s the problem?

🔒 It gives a false sense of security. Someone with OSA who wakes up tired everyday might not present to their doctor because their sleep app gives them a thumbs up.

🏥 It increases workload for already oversubscribed healthcare systems. Completely healthy people go to their doctor because their sleep app reports sub-optimal sleep quality.

⚠️ It’s dangerous. Some tests are designed as screening tools — they’re meant to be performed on large populations to pick up underlying issues (e.g. atrial fibrillation).

Other tests are designed as specific tools to be used in the context of a full clinical picture. For example, an MRI is only recommended if there is a reason to perform one. This is because an MRI will always find an abnormality.

Once a ‘incidental’ abnormality is found however, it needs to be investigated. Cascades of these follow-up investigations cause patient harm. Even if it was possible to perform an MRI on everyone in the country — counterintuitively — it would do more harm than good. More tests does not always = more good.

Current sleep investigations are not designed to be used indiscriminately.

Regulating ‘medical devices’ falls under the remit of the MHRA. Not all medical apps count as ‘medical devices’ though. This makes sense. For example, the calorie tracking app myfitnesspal is intuitively not a medical device.

Broadly speaking, if a medical app is used in your medical management or affects your medical treatment — it’s probably a medical device.

Sleep apps fall outside of this remit because they don’t claim to diagnose you. They sit in a regulatory limbo.

Not all sleep apps are snake oil, some can work:

🌊 Sleep aids like Calm

☀️ Blue light dimmers like f.lux and Apple’s Night Shift feature

😴 Snore tracking apps like SnoreLab

🥼 CBT for Insomnia apps like Sleepio or Sleepstation

Experimental cognitive enhancers are called nootropics. I’d started taking a caffeine tablet and an over the counter supplement called L-theanine (found in green tea).

Evidence shows that mixing the two increases attention, slow-wave brain activity (‘relaxed learning’) and increases brain-derived growth factor (BDGF) and nerve growth factor (NGF) in the hippocampus (the ‘learning centre’).

I was following this philosophy. I thought that small increases in my brain-derived growth factor would lead to academic success and medical school admission (it didn’t).

Dave Asprey is going to live to 180. Either that or he’s full of ****. He’s a biohacker who writes about human longevity and has built a $500,000 at-home lab for his experiments. Whilst he recommends taking 100 supplements a day, injecting stem cells into your spinal cord and sleeping in a hyperbaric oxygen chamber — the core of his teaching can be summarised as:

Even Asprey admits that avoiding the big four will do 90% of the work on the road to longevity. That means getting the basics right through appropriate sleep, nutrition and exercise.

I think obsessive tracking and tinkering has a place for elite athletes, but the lesson for me was:

All the best,

Mustafa