WINK

View Original

#75 - Digital CBT for Insomnia: Will We Be Replaced By Robots?

If you don’t have time to read blogs, here’s the short story: No.

And here’s the Goldilocks story: No - but there’s potential for these Bots to cut a big chunk out of our salaries. There will be some people attracted to the idea of a convenient ‘quick fix’ - including doctors who will develop a preference for referring to these Bots. But the Bots aren’t fully trained in the complexities that humans present with.

So the full story explains the current flaws in the Bots - and the benefits of seeing a sleep expert who is a living breathing human.

What is Digital CBT for Insomnia?

Last week I wrote about the rise of the Sleep Coaches (sounds like a movie title) - which I proposed could meet the demand of sleeplessness - if - sleep coaches were properly trained and supervised.

Similarly, the concept of digital CBT for insomnia (dCBT-I) is built with the intention to meet the same demand.

The most successful dCBT-I programs are Sleepio and SHUTi.

And they happen to be developed in part by two of the biggest insomnia gurus in the world (Prof Colin Espie and Prof Charles Morin).

Prof Colin Espie has done some remarkable things when attempting to meet the need for people’s insomnia.

For example, he wrote the best-selling self-help book Overcoming Insomnia and Sleep Problems.

Prof Colin Espie … the one on the left …

Overcoming Insomnia was released in 2012, and back then, books were a great way to meet the demand of sleeplessness.

The main drawback of books though, is the lack of artificial intelligence (AI) - meaning, a book is not able to perform a proper assessment of the reader’s insomnia.

Even if someone meets the diagnosis Chronic Insomnia Disorder - there are many varieties of insomnia.

For example, I can recall the insomnia patient I assessed when I was fortunate enough to have Prof Espie in my clinic.

His insomnia felt unique. And a human was needed to deliver a sensitive nudge to some of his behaviours (eg, how does a book motivate an Englishman to reduce his daily consumption of 12 black teas?).

Don’t get me wrong though. CBT for insomnia will work for the majority of people. But as a human, I’m suggesting some varieties of insomnia require slight modifications to CBT-I that a book cannot provide.

And there’s 2 main modifications to CBT-I to consider, which are:

  1. Whether the person’s insomnia is also influenced by a mistiming of their 24-hour circadian rhythm, and

  2. Whether the person may have co-morbid sleep disordered breathing (when someone stops breathing in their sleep).

I may have already mentioned that early in my career I treated a 40-something year-old female for her insomnia with CBT-I … and the following week she had a car accident.

We later learned she had both insomnia and a shocking case of sleep apnea.

So us humans who use CBT-I for insomnia are now more aware of Point 2 above (I haven’t had another client since who had a car accident … well, at least whilst they were being treated by me … touch wood!).

And as a human who spent 100s of hours in a sleep laboratory directly measuring circadian rhythms, I can spot circadian rhythm misalignments from a mile away.

Unfortunately, a book isn’t the best at identifying sleep apnea.

Even when it’s written by one of the best insomnia gurus in the world.

And that’s not Prof Espie’s fault. The only way a book adapts over time is when a new edition is released.

But Prof Espie found a superior way to help those with insomnia. One that included AI …

Big Health and The Prof

In the same year Overcoming Insomnia was released, there was also another publication that attracted the interest of insomnia researchers ….

Prof Espie is probably the best to make a judgement about when Sleepio was born, but perhaps this is one of Sleepio’s milestones?

This study demonstrated that delivering CBT for Insomnia in an online format - and in a way that way engaging and interactive - could produce benefits to those experiencing insomnia, including:

  • taking less time to fall asleep (almost 30 min faster)

  • spending less time awake during the night (almost 50 min less)

  • getting more than 1/2 hr more sleep.

It was a great start to the evolution of meeting the demand of sleeplessness.

But it wasn’t actually the start…

The study that launched SHUTi was released in 2009 - 12 years ago - and 3 years before Sleepio’s study.

Nevertheless, the more the merrier.

For example, a study published last month in the journal Sleep showed that digital CBT for insomnia was more cost-effective than any other type of insomnia intervention, including (in ranked order):

  • CBT-I delivered by an expert to a group of insomnia patients

  • Pharmacotherapy (ie, sleep medications like Ambien and Temazepam)

  • CBT-I delivered by an expert to an individual insomnia patient

So digital CBT-I has the greatest potential to meet the demand of sleeplessness in our community, and it is also the most affordable.

This also means that CBT-I delivered 1-on-1 by an expert is the most expensive - and it’s also a method that is not so effective at getting to all those who need it.

Conclusion?

Will those of us who can provide CBT for Insomnia be replaced by robots?

No.

Why?

Well, perhaps it’s better to hear why from Prof Charles Morin (who is on the SHUTi side-of-the-fence).

I took the photo below at the Worldsleep2019 conference in Vancouver - where Prof Morin was finishing the part of his talk about dCBT-I …

Prof Morin’s talk was not the only one to demonstrate a high dropout rate (ie, patients who stop treatment early) from dCBT-i. A colleague of mine - Prof Børge Sivertsen also demonstrated this in another talk at the conference.

But such high dropout rates are common for many digitally delivered CBTs (eg, anxiety, depression, substance use, etc.).

As one client told me this week - some people wish to see a human.

And as Prof Morin’s slide above confirms, dCBT-I has the risk of misdiagnosis.

In my experience, I have seen people who have not benefited from dCBT-I. One of them had co-morbid sleep apnea. Another had a circadian misalignment.

These cases were then referred to a human … who knew how to deliver a tailored sleep treatment.

As per the solution in last week’s blog about Sleep Coaches, humans and robots can work together to the benefit of the many sleepless people in our society.

But - maybe us CBT-I experts need to rethink how expensive we are?

  • Prof Michael Gradisar

___________

p.s. I know the feeling Professors Espie and Morin had when thinking about how they could help more people in the world. In 2016 I began working with a company to develop an app that could help to translate my clinical trial and clinical experiences to families around the world who had an infant with a sleep problem.

It was very enlightening to learn about algorithms, decision trees, gamification (ie, people want a reward within minutes rather than waiting a week to see a health professional), marketing, business speak and the like.

But after almost every meeting with the company, I thought that this app won’t be able to do what I can do for a family.

To this date, that thought still remains true …

See this form in the original post