What is Therapy Worth?

The price of therapy can range widely, from ~$60 per session up to a couple hundred. Usually insurance will foot some or most of that, but how did we arrive at that number? And more to the point, is it worth the price? You’ve got to have some way of calculating whether or not therapy is worth it to you, so I decided to take a deep dive into the data and give my best estimate of how much therapy is worth. 

The Cost of Therapy

First, let’s start off by setting a bar for therapy’s cost. Back in the 1960s, which is the earliest reference I could find, therapy cost $150 – $380 per session, adjusted for inflation. Today the cost still varies, but I think a good average can be set by how much the country’s biggest insurer – Blue Cross Blue Shield – reimburses. Looks like this ranges from around $65 to around $80, so I’ll split the difference and say we’ll set therapy cost at $75. So for therapy to be worth it for you, it has to be worth at least $75. Research says that generally you need between 8 and 26 sessions for therapy to be effective, so let’s be conservative and say 26 sessions worth is a solid dose of therapy. 

Value ≥ $75 x 26 sessions = $1950

Will Therapy Even Work?

The next thing is to assign a number for the effectiveness of therapy. After all, if therapy doesn’t do anything for you, it’s not worth anything to you. So really, the formula should be:

(Likelihood of improvement) x (Value) ≥ $1950.

How to account for likelihood of improvement? What does improvement even mean in psychotherapy? We could gaze deeply into our navels about what therapy is supposed to accomplish and talk about how some treatments are more geared to symptom reduction and others are more about personal growth, but the truth is there’s more data on the symptom reduction side of things, so that’s what we’re going to go with. This brings us into the world of effectiveness studies.

Way back in the 1950s, Eysenck ran a big study that showed that, basically, most people who get better in psychotherapy would have gotten better anyway, so there’s no reason to think that psychotherapy works at all. He may or may not have had an axe to grind with the dominant form of psychotherapy at the time, but I digress. 35ish years later, after CBT came around, Consumer Reports ran another big study that showed that no, psychotherapy is not just effective, it’s super-effective.

Now that we’ve got some even hotter therapies, like Mindfulness based treatments, I wanted to look at the best data available, which led me to this study, which basically showed how most effectiveness studies are fundamentally flawed and limits how much we can draw conclusions for them. So then I wanted to look at the okayest data available, which led me to a study that I don’t even want to cite because of how horribly I’m about to treat their data.

Image result for torture rack
 ‘If you torture the data long enough, it will confess.’ – Ronald Coase.

Basically what all of these studies do is look at a bunch of other psychotherapy studies and crunch their results down to one number that we call an effect size. This is oversimplifying to an absurd degree, but basically they look at the level of depression or anxiety or whatever in the group before the treatment and compare that to the level of whatever after the treatment. Some people get a lot better, some people get a little better, some people get no better at all, and some people might even get worse. That means an effect size is one measurement that combines both the likelihood of getter better, and the degree of improvement.

The nice thing is that they broke it down by condition, so I can say that treatment for anxiety has an effect size of 0.73, treatment for depression has an effect size of 0.83, etc. I made a table, and keep in mind how coy I’m being about my source because you have no reason to believe that I’m not making this up entirely.

Anxiety0.73
Depression0.83
Eating Disorders0.87
OCD1.23
Social Anxiety0.84
Smoking cessation1.96
PTSD1.57
Substance use0.53
Schizophrenia0.54
Alcohol Consumption0.2
Sleep0.57

Keep in mind how these crunch down the likelihood and amount numbers, because looking at this you’d think that we have a way better handle on PTSD than we do on sleep, but really it’s that there’s way more room for improvement on PTSD. My guess, without looking at the numbers, is that way more people get better in insomnia treatment, but the people who get better get way better in PTSD treatment. But I’m not too surprised that the best bang for your buck comes from smoking cessation.

These effect sizes are in a format we call Cohen’s d, but a better format for this number for our purposes is binomial effect size display (BESD). Just as I’m being open about how many liberties I’m taking with the data above, I’ll tell you I’ve never used BESD, but I found an article that said you just plug the d into this formula: BESD = d/sqrt(d x d + 4). Any statisticians reading this article, feel free to call me out, but I believe that the following numbers represent the % improvement due to treatment. Any teenagers reading the article, feel free to snicker at the phrase “plug the d into…”

Anxiety34
Depression38
Eating Disorders40
OCD52
Social Anxiety39
Smoking cessation70
PTSD62
Substance use26
Schizophrenia26
Alcohol Consumption10
Sleep27

So pick the appropriate number up there and plug it into this equation:

(%Improvement/100) x value ≥ $1950

So this number tells us how much we’re likely to improve (how likely we are to improve) due to therapy, but it still doesn’t assign a number value to that improvement. Now I have to give my third and most powerful caveat yet: I was feeling out of my depth on the psychological science so far, but now we’re diving into economics.

It’s fine. Don’t worry about it.

The Cost of Mental Illness

I’m going to look at two ways of calculating the cost of illness: Disability Adjusted Life Years (DALYs) and Value of Statistical Life (VSL). DALYs are actually pretty easy to conceptualize – these are the number of years of your productive life you lose to a disease, either due to early death, disability, or loss of productivity. This website, which is really worth a full read, says that not only is mental illness the leading cause of disease burden in America (!), the average DALYs or years lost due to mental illness was 3. So for a purely economic argument, you can plug 3 years of your salary into our formula and get your answer:

(%Improvement/100) x 3 years of your salary ≥ $1950

Let’s say you have depression and make $50,000 per year. So (38/100) x $150,000 = $57,000, which means that therapy is an absolute BARGAIN for you! But wait, there’s more.

I don’t think there’s anyone who would say the primary problem with mental illness is the harm it does to your earning potential. Really, the bigger problem is the pain that it causes in your life, and how do you assign a number value to that? The economic answer is Value of Statistical Life. As I understand it, this amount is calculated by asking people “How much would you spend on a program or insurance policy that would limit your risk of X?” or, more to the point “How much more would a job have to pay you to experience an increased risk of X?”

Applied to depression, this would mean “How much more would a job have to pay you if it increases your risk of depression by 5%?” I think it’s an interesting way to think about it, although I imagine people significantly underestimate their own value this way, but it’s a good starting point. ThThis article, another fascinating read, goes into the global cost of mental illness this way. Unfortunately, it doesn’t go into a lot of detail about how they get their numbers, which is too bad because wording probably matters a lot, but they estimate that the global VSL number associated with mental illness in 2010 is – holy shit – $8.5 TRILLION.

Image result for dr. evil

Now again I wish I had better information about this number because it’s probably very different for the United States versus Djibouti, but failing that we can use simple arithmetic to divide that number by the global population in 2010 (7 billionish) to get $1214.28. So that number really represents an underestimate of an underestimate, but we’ll plug it into our equation anyway.

Value of therapy = (%Improvement due to therapy/100) x (3 years of your salary + 1214.28).

The Verdict

So with that, you can do the math for yourself. If we want to cap this off, let’s talk about that mysterious economical figure: the Average American.

Median income in USA is $31,000, and the most common diagnosis is anxiety, so for the Average American, therapy is worth (34/100) X ($93000+1214.28) = $32000/(Average 26 sessions) = $1232 per session.

I just walked you through my research and calculation strategy, but I wanted to evaluate those numbers based on my own experience and intuition. The DALY number is probably based on the assumption of a long term mental illness. I don’t think minor PTSD symptoms from a car accident that last for less than 8 months are going to wipe 3 years worth of earnings potential off your life. On the other hand, the VSL number, as I said, is probably a severe underestimate.

Overall, when I arrived at $1232, I was shocked mostly because it’s SO MUCH MORE than what we charge. As I’ve chewed it over, though I don’t think my logic and my math are fundamentally flawed, so I’m willing to conclude that yeah, therapy is the absolute deal of a lifetime!

And just as a sad little epilogue, when I was starting out I got paid $30 per session.

“That’s why they pay me the big bucks.” – Every therapists’ favorite joke.

1 thought on “What is Therapy Worth?”

  1. Might want to add up the total knock on effect from physical health as well because (for example) depression correlates with heart disease, cancer and other lovely things like that.

    All of which suck and all of which cost a fortune themselves.

Leave a Reply

Your email address will not be published. Required fields are marked *