A therapy chatbot and app for depression and anxiety
businessinsider.com
businessinsider.com
The evidence suggests otherwise. Internet-based CBT does not seem to be markedly inferior to face-to-face CBT, which is the current gold-standard talk therapy. We really don't have any evidence that the human contact element of psychotherapy is integral or even significant to the effectiveness of the therapy.
We should also be careful to avoid letting the perfect be the enemy of the good. There are many barriers to psychological care, from cost to stigma to simple availability. For the overwhelming majority of people worldwide, it isn't a choice between internet-based CBT or face-to-face psychotherapy - it's a choice between internet-based CBT or no treatment at all. Even if internet-based CBT is significantly inferior, it still has the potential to prevent immense amounts of suffering and death.
Woebot helped me change thought patterns that kept me from getting help and it could do so at any time of the day or night. When going to see your counsellor is too hard Woebot is only a ping away and is free.
Now not to discount your point that indeed sometimes you need words of encouragement or humans to understand complex situations but there is room for both in the grand scheme of things.
There's no reason small children should feel comforted by a stuffed animal, but they do, and it can provide an experience very much like human contact for them.
I went through several bouts of extreme depression, and I might only be alive because I discovered the joy of breaking ceramic plates. I have no idea why the destruction of dinnerware gives me so much joy, but if things ever get too much for me, I just go to target and buy a few dozen dishes. I then take them to my local dump, which has a cement cliff you throw your trash over into the pit. I toss the plates like frisbees, and watch them sale 20 or 30 feet to the cement floor. When they crash they make a tremendous noise, like when a server accidentally drops a bunch of dishes in the kitchen at a restaurant. Except out at the dump the sound is much much louder and cleaner. Sometimes I throw one plate up really high like a clay pigeon and try to hit it with a second plate, which I only have a couple times. I've probably wasted several thousand dollars doing this.
The point of my long ramble is that strange things can save you from depression. There is no reason a chat bot can't work for some people.
Anger always finds a target, and when you don't allow yourself to be angry at someone or something the target becomes yourself and (mostly mental) illness results.
Clearly.
I don't think we'll see effective talk therapy chat bots for quite some time.
I'm very skeptical of apps and products that claim to be based on CBT or "mindfulness" and/or claim to cure various mental illnesses (mostly because these commercial actors use buzzwords rather than conform with the formal definitions used in research), but I don't think we should offhandedly discount all possibilities.
Another possibility is that you can have thoughts which are absolutely reflective of reality but which are causing you distress. These thoughts are not at all useful to your well being, yet you are spending an inordinate amount of time thinking about them anyway. Learning how to become more aware of these thought patterns and how to steer away from them in to more useful thought patterns will help improve your well being over time.
The key in both of these scenarios is learning how to be more aware of your own thoughts and feelings and manage them appropriately.
Maybe a sophisticated bot/system might, generally, be better at rooting an individual's issues. But I would fear that a human experiencing deep emotional stress would be much safer in the psychotherapy environment v.s. with a bot in any random place where that person is most likely alone.
I believe that depression and anxiety may need more unbiased research for the best solutions to be found. Between therapists and pharmaceuticals, there's a lot of vested interest in the status quo.
Therapists can be hugely beneficial. But they are also human, and are subject to their own timidity, attitude, perspective and theories. With books, these are laid bare and are unashamed. An author is sometimes more willing to get in your face about something than a therapist ever would dare.
And for me, that's what I needed most. New ideas, new perspectives, new ways of looking at myself, my emotions, and my relationship with others.
Not knocking seeing a therapist, but one can get comfortable with them and not make the progress you need. Likewise, I know a great deal of people who have had repeated bad experiences with therapists, sometimes to a criminal degree.
However, it doesn't take that many machine learning algorithms or Nvidia GPUs to send someone an email/facebook message/SMS "in the morning" asking them to be in a good mood.
haha oh man that's actually my side project!
I disagree with this, at least for myself. For me, talk therapy is essentially an alternative to 'writing therapy,' where the act of just somehow collecting my thoughts and talking verbally through them (rather than them flashing about in my mind) has more of an impact than the other person accepting whatever it is I am rambling about.
That's not to say anxiety and depression are always lifestyle issues, for many people there is physiological cause that can be treated with medication, but a lot of psychological problems are really life problems.
So it sounds like they're doing the right thing. They're helping people with mental health issues so I want to believe good privacy practices will continue.
Time will tell. Until then, this sounds like a wonderful project.
[UPDATE] Okay, I just tried the app. The first thing it asks for is your name. Fine. Pseudonym, no problem. The second request is, "I need to send you a code. What's your phone number?"
So much for "we don't know who you are". >:
[SOLUTION/PSA] Use a service like Twilio for signup. For $1 you can pull a random SMS-compatible phone number out of a pool, get your signup code from the Twilio web console, and then release the phone number. Not anonymous but obscured identity, at least.
At least with a name (maybe) and a number they have a way to contact you. Or the police. Or anything that will take you off the brink.
Also, poor sap down the road who wants to try it but gets your random number will be devastated. Hypothetically, could have been their last resort.
It sounds interesting. It also reminds me of ELIZA[1] a chat bot designed in the 60s to demonstrate the superficiality of human conversation which ended up being a surprisingly effective therapist to its users.
Woebot: Write a negative thought
Me: Testing abc
Woebot: Would you like to work on Testing abc?
Woebot: Can you think of a time when...
Me: Well, so...
Woebot: *Blindly replies with the next line in the script.*Quote from your wiki-source:
> Lay responses to ELIZA were disturbing to Weizenbaum and motivated him to write his book Computer Power and Human Reason: From Judgment to Calculation, in which he explains the limits of computers, as he wants to make clear in people's minds his opinion that the anthropomorphic views of computers are just a reduction of the human being and any life form for that matter. [...]
EDIT: I later learned that I'm friends with Weizenbaum's grand-niece (or something like that).
Like with everything, if you can't tell it's working, it's not working.
Stop guessing. Go talk to your doctor.
Your doctor will also guess. There are about 20 drug treatments and a dozen or so talk therapies that are all about equally effective at treating depression and anxiety. We have no evidence-based way of figuring out which treatment will work better for which patient. It's a total crapshoot, even for psychiatrists.
Anecdotally, escitalopram worked for you. A thirty second Google search will turn up hundreds of anecdotal reports saying that escitalopram is worse than the devil incarnate. The evidence says that it's no better or worse than a laundry-list of other treatments. You got lucky.
With that said, talking to your doctor is a sensible decision.
Besides, there's family history to consider, which may help rule out some drugs. The doc isn't always going in totally blind.
Depression and anxiety are complex problems that are incredibly personal (in the sense that they are unique to the individual). More so than most other disorders. Ergo:
1) No off the shelf solution like popping an ibuprofen for body aches
2) what works for others may not work for you
3) one needs to try multiple solutions
4) and most important, have a professional at your side to guide you!
You may not necessarily figure out how to navigate this on your own.
Which isn't to say that drugs can't help, just that they're very much treating the symptoms and not at all the underlying cause.
This is obviously not true for some people. If you have a genetic condition that means you cannot produce certain essential chemicals (let's say Vitamin D) and your deficiency is causing your brain to function inefficiently making you feel depressed then the genetic condition is the cause and the drugs are the cure. You seem to be refuting "what works for you might not work for others" by saying that drugs cannot treat the underlying cause of depression.
And, there is an argument that is made that such drugs aren't effective anyway (or no more than placebo).
However, I think that's an over simplification [2]. And I don't necessarily accept the contention that SSRIs / SNRIs and others just treat symptoms.
This goes back to what I was saying earlier, that what works for person X may not work for person Y, not because the treatment isn't effective by itself, but because the etiologies may be different. All the way from a single SNP (single nucleotide polymorphism)) introducing a subtle amino acid substitution in a receptor on the cell surface to the cause being due to a vitamin deficiency (B-complex, for instance, or vit D, as another comment mentioned, or low Mg levels!) to psychological trauma. There are so many reasons that can lead to depression & anxiety.
It's like fever. Some fevers can be treated with Tylenol. Others will need quinine, and Tylenol in such a case would be ineffective.
What it also means is that we need to find better therapeutic targets in addition to seretonin / dopamine receptors, etc., and maybe move to a combinatorial drug approach once the targets are identified & validated. (like is used in oncology).
So please do not dismiss drugs. Or therapy. Or CBT. Or lifestyle changes (diet, exercise). Environmental changes (job, social circle). It could be any of these that can help someone deal with, navigate through and resolve depression. And it can be frustrating to try and narrow down among these. But that's the nature of the beast :/.
[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4471964/
[2] http://slatestarcodex.com/2014/07/07/ssris-much-more-than-yo...
It's possible I'm just bitter from my experience with doctors over the years.
So you are absolutely correct that it can be, and often is, a crapshoot in terms of figuring out what treatment / approach would work for a person.
It is frustrating. And I speak from experience too! It is hard to keep telling yourself, the next one will work, in the middle of depression when making the effort just to stick to the current plan seems like an thankless pointless task. Almost seems sisyphean. But., what we know today about the disorder is tons more than what we knew 10 years back. And treatment approaches are improving. Don't give up is the only useful thing I can say, having walked some of the same paths as you have.
[1] https://qz.com/1190151/why-am-i-unhappy-a-new-study-explains...
Also, going through comments, it seems to be an Eliza-like chatbot. I.e. not very complicated on algorithmic front. I wonder, what it would take to have it instead of M-x doctor in my Emacs?
EDIT: trying out the Messenger bot. God, I'm already in love with the humour and the flow. Nicely done!
If you make it more convenient to do this or encourage people to participate then having a nice UI would probably work wonders. Also, if the user believes that the system on the other end is human enough I think you could see better results.
I was pretty bored with the app ecosystem in 2017, but Calm Harm gave me hope that there are still some people out there innovating and actually working to use our existing technology to help people for good rather than just making some money off advertising or Yet Another To Do App.
The only thing I get concerned about with these apps is the storage of this data and how it can be made personally identifiable. Hopefully these angles are tackled with people's privacy in mind.
Err, so why do I have to register my phone number before I can do anything with the app?
I look forward to computational therapists, it should help with a lot of the oddities with traditional therapists.
Love to see it turned into a non-cloud service though. Mental health stuff can be incredibly personally sensitive, and a non-cloud solution would be far preferable.
There is a vibe of "i'm one of the cool kids", like when parents try to "level" with their teenagers. And the machine learning appears rudimentary or perhaps tastelessly implemented.
But it gets CBT in the hands of more people. Maybe people wouldn't pickup a book but are willing to checkout an app. So i think thats an important win, even if the app itself isn't all that.
I tinkered with it some more. it is a good idea, and a tasteful medium to do it in. However, the implementation is half-assed. And - I don't doubt that they put a lot of thought into it, but because your only service complexity IS this dialog, you have to nail it. It has to be really good. As is, it's about at the level of a poorly done IVR for a corporation. And you're expecting people to use this for any kind of therapy?
I'm just along for the ride, playing through the options. But there are clearly cliffs where, one moment it asks you to tell you about something personal, and the next it goes "ok see ya!" - i guess in a way it's good they managed to replicate the emotional distance you can feel from therapists in real life, too.
Perhaps it'll improve over time. If they want this to be good, they'll have to.
Now, let's chat up that bot...
Citation: Fitzpatrick KK, Darcy A, Vierhile M. JMIR Ment Health 2017; 4(2):e19.
Link: https://doi.org/10.2196/mental.7785
DOI: 10.2196/mental.7785
Abstract:
Background: Web-based cognitive-behavioral therapeutic (CBT) apps have demonstrated efficacy but are characterized by poor adherence. Conversational agents may offer a convenient, engaging way of getting support at any time.
Objective: The objective of the study was to determine the feasibility, acceptability, and preliminary efficacy of a fully automated conversational agent to deliver a self-help program for college students who self-identify as having symptoms of anxiety and depression.
Methods: In an unblinded trial, 70 individuals age 18-28 years were recruited online from a university community social media site and were randomized to receive either 2 weeks (up to 20 sessions) of self-help content derived from CBT principles in a conversational format with a text-based conversational agent (Woebot) (n=34) or were directed to the National Institute of Mental Health ebook, “Depression in College Students,” as an information-only control group (n=36). All participants completed Web-based versions of the 9-item Patient Health Questionnaire (PHQ-9), the 7-item Generalized Anxiety Disorder scale (GAD-7), and the Positive and Negative Affect Scale at baseline and 2-3 weeks later (T2).
Results: Participants were on average 22.2 years old (SD 2.33), 67% female (47/70), mostly non-Hispanic (93%, 54/58), and Caucasian (79%, 46/58). Participants in the Woebot group engaged with the conversational agent an average of 12.14 (SD 2.23) times over the study period. No significant differences existed between the groups at baseline, and 83% (58/70) of participants provided data at T2 (17% attrition). Intent-to-treat univariate analysis of covariance revealed a significant group difference on depression such that those in the Woebot group significantly reduced their symptoms of depression over the study period as measured by the PHQ-9 (F=6.47; P=.01) while those in the information control group did not. In an analysis of completers, participants in both groups significantly reduced anxiety as measured by the GAD-7 (F1,54= 9.24; P=.004). Participants’ comments suggest that process factors were more influential on their acceptability of the program than content factors mirroring traditional therapy.
Conclusions: Conversational agents appear to be a feasible, engaging, and effective way to deliver CBT.