Quality over Quantity: Alleviating the Therapist Shortage with Better Mental Health Care

Towards high-fidelity evidence-based practice in therapy

Gaga ooh la la…it’s bad therapy

Mental illness is among the most prevalent health issues in the U.S., and there aren’t enough qualified care providers to help those in need. The National Institute of Mental Health estimates that 1 in 5 Americans suffer from mental illness, and Kaiser Family Foundation data suggests over 115M Americans live in areas with in-person mental health care provider shortages (defined as areas with more than 30k people per available provider).¹ The American Psychological Association reported in 2022 that psychologists have seen unprecedented demand for their services.²

While many virtual care companies are helping to expand therapy access to areas with shortages, the overall number of mental health care providers still hasn’t met the national need. Companies are trying to bridge this gap in a variety of ways, such as:

• New ways of delivering therapy content: Offering on-demand content (e.g., Real), group therapy sessions (e.g., Lifestance Health, Soulside), or automated therapy (e.g., Woebot)

• Increasing therapist productivity: Building technological tools that automate tasks like clinical documentation and session planning (e.g., OPTT)

• Task shifting: Leveraging less highly trained care providers, where appropriate, in order to most efficiently leverage all providers (more on mental health provider types and training in the table below). This can include, for example, employing health coaches (e.g., Modern Health, Spring Health) for lower-acuity needs or peer support (e.g., Marigold Health, HearMe) where peers with similar life experiences can be a uniquely valuable resource.

Table: Overview of Mental Health Provider Types & Training

Sources include Bureau of Labor Statistics, professional clinical associations, and others (can share upon request)

Each of these is an important part of the solution, but they don’t obviate the need for clinically trained professionals who can deliver evidence-based mental health care. While some have suggested simply growing the clinical workforce in mental health - perhaps similar to how Kaiser Permanente launched its own School of Medicine to address gaps in the clinician workforce - I believe this doesn’t address the root of the problem. Increasing the quantity of care leaves untouched the critical problem of quality of care.

Even for people who have relatively easy, affordable access to therapy through their health insurance networks or employer benefits, finding the “right” therapist is a well-known challenge. While this is often chalked up to issues of “fit”, it is much rarely discussed that many therapists simply are not practicing evidence-based medicine. Or, as this 2022 Wired article states more plainly: some therapy sessions are just bad therapy.

If a patient were getting “bad” (non-evidence-based) care for another disease – say, diabetes – we wouldn’t expect their health to improve. Similarly, if patients are not getting good care for mental illness, we can expect to see continued high rates of mental illness and correspondingly high rates of care utilization as people try to get better. Conversely, if we get people the right care – based on evidence-based clinical practice that is appropriate for their specific illness – we can expect people to get better sooner and care utilization to drop. In a nutshell, delivering consistently high-quality mental health care could ultimately reduce demand for therapy and thereby help to alleviate the therapist shortage.

Quality over quantity

Companies providing virtual therapy and/or psychiatry have become household names in the last few years – think Headspace, Talkspace, BetterHelp (by Teladoc), Talkiatry, and many more. While some of these companies sell to different types of customers (employers, health systems, payers, directly to consumers), their services otherwise appear to be relatively undifferentiated. They are all hiring from a limited pool of care providers and claim to offer “high-quality therapy” that will win payers’ approbation and lead to increases in insurance reimbursement rates. Theoretically, higher quality care should lead to better patient outcomes – potentially including fewer therapy sessions and less medication use in the long term – and as a result, lower total costs of care for payers. But in speaking with clinicians and patients, I’ve observed that “quality” in psychotherapy isn’t exactly a well-defined concept.

The American Psychological Association acknowledges the importance of measurement-based care, but payers and providers haven't established a clear set of measures for gauging the effectiveness of a therapy session. Therapy companies typically make claims about their quality by citing changes in patient outcomes via common assessments such as the GAD-7 (anxiety) and PHQ-9 (depression) questionnaires. But these are outcome metrics, not process metrics, and can be impacted by many factors outside of a therapy session. A patient’s PHQ-9 score could suggest they have severe depression today and mild depression next month, but this says nothing about the quality of the therapy sessions they received.

So what is quality? When it comes to measuring quality of care for other illnesses, it is typical to measure adherence to the standard of care – guidelines set out by respected medical bodies based on rigorous clinical evidence. For example, for patients with diabetes, clinicians are expected to routinely measure a patient’s blood sugar, prescribe the appropriate medications recommended by the standard of care, and advise about certain lifestyle modifications that can help. 

In behavioral health, evidence-based standards of care were adopted widely in the 1990s. This blog post explains it well: “Evidence-based psychotherapy (EBP) refers to a psychotherapeutic intervention that has been demonstrated by research to be effective in the treatment of one or more mental health issues. Many studies have found that EBPs are more effective than the typical care that people would receive from providers in their community.” Types of evidence-based psychotherapy include cognitive behavioral therapy (CBT), dialectical behavior therapy (DBT), and mindfulness-based cognitive therapy (MBCT), among others. Each of these is characterized by certain best practices; for instance, case formulation is a key technique in CBT. Just “doing” CBT, DBT, or MBCT isn’t enough – it is adhering to the principles and protocols of treatment with fidelity and skill that matters.

Given that the standard of care in psychotherapy involves a conversation rather than discrete actions such as ordering a test or medication, it has historically been a bit more challenging to measure adherence to the standard of care. The gold standard in measurement is direct observation of psychotherapy – i.e., a skilled supervisor watching a video or listening to audio of a therapy session – and providing feedback on adherence to evidence-based treatment protocols. This is difficult to implement in practice because it requires experts and training human coding teams, which is slow, labor intensive, and expensive. As a psychologist friend who managed a large health system’s DBT program and supervised several therapists-in-training shared with me, this type of feedback was the most valuable part of her own training, but it was impossible to review all, or even most, of her supervisees’ sessions. While this type of performance feedback does occur to some extent in therapy training and supervision programs, it is extremely rare once providers have completed these programs and are in full-time clinical practice. This can contribute to therapist drift, which occurs when therapists fail to deliver the optimum evidence-based treatment. With this context, it’s not so surprising that patients seeking therapy for the same types of conditions often have very different experiences of care, and that at least some patients are getting “bad therapy.”

This suggests the barrage of news articles about America’s therapist shortage may be missing the point – or at least some of it. Given the paucity of measurement and feedback for therapists, it is likely that therapy quality, on average, is not as high as it could be. Raising the bar on quality could alleviate at least some of the access problem by helping patients to achieve better outcomes, faster, and without bouncing around from therapist to therapist as is often the case in today’s system. 

Lyssn up: new approaches to measuring and improving care quality

Advances in technological capabilities have given rise to more data-driven and automated training and feedback approaches in other medical fields. For example, Theator collects and analyzes data on surgeries to provide surgeons with specific feedback about their adherence to evidence-based standards of care, ultimately reducing variability in surgical outcomes. Could automation similarly make it possible to deliver faster, cheaper, and more standardized feedback to therapists?

(Da ba dee da ba die)

Emerging companies like Eleos Health, which recently raised a ~$40M Series B round, and Lyssn, a company that has been predominantly funded by SBIR grants, are paving a promising new path to improve quality in behavioral health care. According to its website, Lyssn’s proprietary software can take an audio recording of a therapy session and rate it on several clinical quality metrics in about 5 minutes, a process that normally takes a highly-trained human rater several hours. The company’s transformer-based models, which the company says are built on the largest database of therapeutic interactions of its kind (1.91M total sessions/calls covering >46M minutes of interaction and >4.3B unique words), are trained to code sentences/phrases in a conversation with the behaviors they represent. For example, the model can rate a therapist’s adherence to motivational interviewing behaviors, including demonstrating empathy (deep understanding of what the client stated, including reflecting back content not explicitly stated by the client), collaboration (a highly collaborative clinician encourages collaboration and power sharing regarding what’s discussed in the session), and other evidence-based metrics. Lyssn’s model achieves >90% accuracy in evaluation of metrics. 

The company’s CTO articulated Lyssn’s value proposition well in a 2021 blog post:

“If you are a clinician practicing CBT, you would be lucky if one or two sessions in your entire career were scored for CBT fidelity by an expert rater. Now, with Lyssn you can score every single session over the course of a week or month. If you are a supervisor, you can view the CBT metrics of your entire clinic on a daily, weekly or monthly basis. You can examine each of the 11 [Cognitive Therapy Rating Scale] criteria to assess what aspects of your practice are hitting the mark and what might need improvement. This is the type of reflective practice we hope to facilitate at Lyssn.” 

Lyssn’s website states that it has over 70 institutional customers including government agencies, digital behavioral health and coaching companies, academic medical centers, 988 crisis support services, and Certified Community Behavioral Health Clinics. The company recently announced a partnership with Centerstone, a large provider of behavioral health care.

On the for-profit side, Eleos Health is building a broader software solution for behavioral health care providers, which includes providing feedback on their adherence to evidence-based practices. The company’s website states that “Research shows that providers using Eleos achieve 2x higher client engagement, 3–4x better symptom improvement, and 36% greater usage of evidence-based techniques.” Driving better symptom improvement for patients could unlock a path to materially reducing demand for therapy. While still early, the emergence of companies facilitating the delivery of more data-driven, evidence-based care is a step in the right direction for better quality – and, ultimately, access – in mental health care.

Evidence-based care: it’s a no-brainer

In the long run, it’s hard to see how any therapy training program or practice will be able to maintain credibility – or payer reimbursement – without implementing these readily available, low-cost tools that can improve the quality of care. Not only can these solutions make therapy more effective, but they can also improve therapists’ satisfaction with their work – after all, most people enter the profession because they genuinely want to make a positive difference in their patients’ lives, and adhering to evidence-based practice is the best known way to do so.

Objectively measuring and managing therapy quality can also add value to virtually every other stakeholder group impacted by the mental health crisis. Patients will be more likely to receive high-quality therapy that adheres to the standard of care – and potentially spend less time in therapy and more time living their lives. Therapy companies can more easily identify the highest-quality therapists and support therapists with ongoing, objective performance feedback and professional development opportunities. Health plans can more easily direct patients to truly high-quality therapists and, theoretically, save money if patient outcomes improve more quickly under evidence-based care practices. As a society, we can benefit tremendously from these tools. More people can get better care, potentially at a lower cost. It’s a no-brainer: high-fidelity evidence-based mental health care is the future.

A big thank you to my dear friend Liza Pincus, a practicing clinical psychologist and co-founder of Azalea Psychology Group, for her thought partnership on this piece.

Update on 1/12/24: A previous version of this blog post incorrectly stated that Lyssn is a non-profit, which it is not.

1. Shortages may be somewhat alleviated with the growing adoption of telemedicine. For example, in recent years many states have adopted PSYPACT, legislation that came about in 2019 allowing licensed psychologists to practice across state lines.

2. While this post focuses on therapy as a treatment for mental illness, there are many other possible treatments, including medications and prescription digital therapeutics. On the pharmacological side, numerous studies have demonstrated that antidepressants are more effective on average than a placebo in treating depression, but not all antidepressants are equally effective and different medications work better for some populations than others. Some companies have developed tests to help identify the best medications for mental health conditions based on a person’s genetics (though these haven’t yet proven they are efficacious); others are collecting and analyzing real-world data to identify statistical relationships between patient demographics/key characteristics and therapeutic outcomes, with the goal of applying the insights to personalize care. Researchers and companies are also exploring new potential therapies like psychedelics and other types of drug candidates to treat mental illness (e.g., Karuna Therapeutics’ schizophrenia candidate KAR-XT), as well as digital solutions (e.g., Akili for ADHD, Arcade Therapeutics for anxiety and depression).