Although second-generation antidepressants are more-cost effective than cognitive behavioral therapy (CBT) for the initial treatment of depression at 1 year, psychotherapy demonstrates better cost-effectiveness at 5 years, according to a study published in the Annals of Internal Medicine.1

Lead study author Eric L. Ross, MD, of McLean Hospital in Belmont, Massachusetts, and colleagues conducted a decision analytic modeling evaluation of the cost-effectiveness of the 2 treatments based on data from a meta-analysis of randomized controlled trials, as well as clinical and economic data from other publications. The study examined cost-effectiveness from health care sector and societal perspectives. Outcome measures included costs in 2014 US dollars, quality adjusted life-years, and incremental cost-effectiveness ratios.

The investigators noted that neither treatment modality demonstrated consistent, superior cost-effectiveness, emphasizing that physicians should incorporate other factors into their decision-making processes with patients. Indeed, although 70% of patients with major depressive disorder prefer psychotherapy to pharmacotherapy, <25% receive any form of psychotherapy. Given the large gap between patient preference and rates of psychotherapy provision, the authors note, access to therapies such as CBT should be expanded.

Limitations of the study included reliance on cost data obtained more than 10 years earlier, as well as the potential variability of CBT interventions.

The investigators wrote, “Our findings lend economic support to the American College of Physicians’ conclusion that either [second-generation antidepressants] or CBT is a reasonable initial treatment for [major depressive disorder].” They also noted, “Moving from current (<25%) to patient-preferred (70%) levels of CBT use could thus save more than $1.5 billion after 5 years” in terms of health care sector costs.

Related Articles

Psychiatry Advisor contacted Dr Ross to discuss his research and the state of the field.

Psychiatry Advisor: What motivated you to compare the cost-effectiveness of antidepressants and CBT?

This is an issue that my senior author, Kara Zivin, PhD, had been interested in for quite a while. We are actually targeting a primary care audience with this research. We’re asking a question about the initial treatment of depression, which is important because it affects so many people. Some of the other work that I had done with Dr Zivin looked at things like electroconvulsive therapy and other treatments for treatment-resistant depression, which is often managed by psychiatrists. But the research had the potential to address a very large population: essentially every patient who comes into a primary care physician’s office and brings up that they’ve been feeling depressed.

Psychiatry Advisor: What are the most critical implications of your findings?

The key finding was that from a cost-effectiveness perspective, either CBT or antidepressants would be a reasonable option for the initial treatment of depression. We have some data on the clinical efficacy of these 2 treatments and there is a lot of uncertainty around that data. There are wide confidence intervals around our estimates of how effective each treatment is relative to the other. And that’s OK; it’s an honest reflection where we stand with the clinical research right now. We’re uncertain about the cost-effectiveness of either of them.

We were actually surprised because we know with a fair amount of confidence that CBT costs more than antidepressants. That’s not something where there’s much uncertainty with the data. Given that certainty, I would have guessed going in that antidepressants would be much more cost-effective, but that’s not what we found. We found that the two are kind of equivalent in terms of their cost-effectiveness and either of them would be a reasonable option.

Based on our results, we don’t have enough confidence in either to say that we would dictate what providers should be doing based on these findings. What’s nice about that is it lets physicians move on to other things besides cost-effectiveness in decision making. It opens the door to all of the patient-specific factors physicians should consider when making decisions. In the absence of a clear winner from a cost-effectiveness perspective, what physicians should be thinking about are those things that should be considered in shared decision making: what patients prefer, what their values are, what they are more concerned about, and what they are able to do in terms of their time.

Psychiatry Advisor: You mentioned that 70% of patients favor CBT, but <25% of patients are actually receiving psychotherapy. In that context, how do physicians meet patients’ needs and keep in mind all of the other things they need to be thinking about, such as fluctuations in CBT delivery or medication adherence?

That’s where the rubber hits the road. We can conduct this research and make this recommendation that CBT is a reasonable option, but at the end of the day, if a patient lives in an area where there aren’t CBT providers closer than an hour’s drive, it’s tough for them to engage in CBT. I live in Boston where there are many CBT providers around here — we have a glut of physicians, therapists, and other medical professionals —and yet it can be challenging to find a therapist who works with a patient’s schedule and has a good fit with them personality-wise. Our research wasn’t focused on how to solve that particular issue. There are interesting things going on in the field that potentially offer some solutions.

People are showing interest in internet-based CBT or smartphone app CBT. A recent meta-analysis in JAMA Psychiatry that compared different delivery methods for therapy, specifically focused on CBT, came to the conclusion that the evidence was pretty similar for all of them.2 There wasn’t evidence that internet-based CBT was any worse than in-person or group CBT. We didn’t have enough data on different types of CBT to include that kind of granular comparison in our analysis, but I think that is interesting research. There’s also interesting implementation work involving health systems trying to incorporate CBT more easily into clinicians’ work flows to make it easy to refer patients to a group or to connect them with internet-based CBT.

Psychiatry Advisor: Over the past few decades psychiatrists have seen a shift from providing direct psychotherapy to prescribing pharmacotherapy. How might your research and findings affect the general landscape and the role of psychiatrists in care provision?

I don’t know if our research has much to say about how psychiatrists specifically should be practicing. The audience most affected by these recommendations would be primary care practitioners, who are often the ones first seeing patients when they present with depression. As psychiatrists, we’re often not seeing these patients until they’ve tried multiple treatments without success. At that point, if you’re somebody who offers CBT, you may be getting more patients referred to you for CBT if we were to adhere more to patients’ preferences.

Psychiatry Advisor: Can you comment on combined antidepressant-CBT treatment, which you mentioned in your research but did not directly analyze?

We didn’t look at it directly because there was a lot more data comparing one to the other individually than studies that directly compared the combination to either individually. We can hypothesize about what the cost-effectiveness might be. From one perspective, CBT alone is fairly expensive and the cost of an antidepressant is minimal, so adding pharmacotherapy is probably not going to cost much more and could offer some additional benefit.

There have been studies in other conditions. One trial for nonepileptic seizures randomly assigned people to CBT and antidepressant or both.3 In some cases, they found that people did worse with both than with CBT alone, and they chalked that up to the potential for more adverse effects when you introduce an antidepressant. It wasn’t something that we looked at in our research, and you could come up with hypotheses either way.

Psychiatry Advisor: What are the next steps you see in your research and in this line of inquiry? How might costs and treatment modalities change in the next 5 to 10 years?

I’m not currently working on other studies related to this issue specifically right now. I’m continuing to conduct other cost-effectiveness analyses within psychiatry. There are areas within psychiatry where the cost-effectiveness of certain treatments hasn’t been explored yet. 

With regard to what the next steps are for this particular question — the initial treatment of depression and the benefits and risks of antidepressants vs CBT — I think the next step is now that we know that each of these is a reasonable option, and that patients would prefer CBT at much higher rates than they’re currently receiving it, how do we go about implementing that? Another avenue for future research that we highlighted in our analysis is that there is still uncertainty in the relative cost-effectiveness for these 2 treatments. We did some analysis to figure out what is driving that uncertainty — which clinical factor is the biggest source of uncertainty in cost-effectiveness. What we settled on is uncertainty in the long-term effectiveness of the 2 treatments, and how well each treatment prevents relapse over the course of years.  Additional evidence would be most valuable in helping to refine our understanding of the cost-effectiveness of these treatments and providers’ decision making about which treatment they should use initially.


1. Ross EL, Vijan S, Miller EM, Valenstein M, Zivin K. The cost-effectiveness of cognitive behavioral therapy versus second-generation antidepressants for initial treatment of major depressive disorder in the United States: a decision analytic model [published October 29, 2019]. Ann Intern Med. 2019. doi:10.7326/M18-1480

2. Cuijpers P, Noma H, Karyotaki E, Cipriani A, Furukawa TA. Effectiveness and acceptability of cognitive behavior therapy delivery formats in adults with depression: a network meta-analysisJAMA Psychiatry. 2019;76(7):700-707.

3. LaFrance WC Jr, Baird GL, Barry JJ, et al. Multicenter pilot treatment trial for psychogenic nonepileptic seizures: a randomized clinical trial. JAMA Psychiatry. 2014;71(9):997-1005.