Although the Biden-Harris administration is coming to an end, some of its work will continue. The administration has invested tens of billions of dollars to expand federal and state mental health and substance use services through the Safer Communities Act, the American Rescue Plan, and other legislation. In the fall, they significantly extended these efforts by finalizing rules to extend mental health parity laws.
This amounts to one of the most significant federal government investments in the behavioral health sector in history. For those of us who have toiled in the mental health sector for years (in my case decades), these developments are exciting. Behavioral health care has always been massively underfunded relative to the burden that these disorders have on people, families, and communities. To have the federal government leaning into these challenges with serious policy and funding initiatives can only be described as laudable. It’s about time.
But there is a potential problem. There is significant evidence that increasing access to behavioral health treatments does not, in and of itself, reduce the burden of mental disorders in the community. In August, Time magazine reported data showing that although more Americans are receiving mental health care, the number of those reporting excellent mental health has fallen. This conundrum, known as the “treatment-prevalence paradox,” is well known to researchers. For example, a study that examined access to mental health care across the U.S., U.K., Canada, and Australia concluded, “All four countries have had increases in rates of treatment. … Despite these changes, none of the four countries had any evidence for a reduction in the prevalence of disorders or symptoms over the period. If anything, there were indications of changes in the opposite direction.”
So, what gives? Why is providing greater access to behavioral health services not making a dent in the overall burden of behavioral health problems?
One important possibility is that our attempts to improve the quality of services have lagged behind increases in access. When we primarily focus on the goal of increasing access, then the number of people receiving services becomes our measure of success, irrespective of the outcomes achieved. The “quality gap” refers to the failure to provide routine clinical care consistent with best practices, such as those used in clinical trials. A recent study found that there has been no increase in the effectiveness of routine clinical care for depression over the past 30 years, and moreover, only about 20% of people who receive treatment recover. These outcomes are much worse than those typically observed in clinical trials where the recovery rate is roughly 60%-80%. Together, these findings suggest that we urgently need to implement systems to support high-quality evidenced-based practices in routine care.
There are many potential reasons for this quality gap. Practitioners delivering routine care in real-world settings have a difficult job where they often face overwhelming demand for their services that they struggle to meet. High-priority patients often have more complex problems that take more time and effort to treat, with a lower likelihood of success.
However, there are other factors. Some practitioners are not well trained in evidenced-based treatments, and others can experience a phenomenon known as “therapist drift” — the tendency for practitioners to reduce their adherence to evidence-based practices over time. This “drift” is most notable in more experienced therapists and those who practice in more isolated contexts — the latter being particularly concerning given that the vast majority of behavioral health practitioners in the U.S. work in small or solo practices. Moreover, when practitioners do drift away from evidence-based practices, they often shift from “doing” therapies to “talking” therapies such as non-directive supportive counselling. While most established approaches to psychotherapy have similar outcomes, this type of intervention is often less effective. So, the challenge is that when therapists “drift,” they tend to drift toward approaches that are less likely to work well.
When I started running clinical trials, I began to understand why they might see more success than we do in real-world practice. Clinical trials always include some procedures that are often not followed in routine practice. For example, patient outcome measurement is obsessively followed up, and practitioners are closely supervised to ensure treatment fidelity (often using video or audio recordings of their practice). These recordings ensured that patients receive the treatment as intended, and the follow-up allow us to know their outcomes — both essential for publishing the results of the trials.
However, these processes require a lot of time and effort, two things that are in short supply in a busy clinical practice. Recording treatment sessions offers supervisors valuable opportunities to give trainees targeted feedback, but who has the bandwidth to do that consistently? Instead, we use unstructured clinical notes that (unlike direct observation and video recordings) primarily reflect the practitioner’s perspective on patient progress, rather than detailed and objective assessment of the therapist’s skill or client outcomes. Indeed, research has shown that practitioners who rely solely on clinical judgment are often not able to reliably predict patient outcomes in the absence of standardized measurement.
What we need are solutions that allow us to scale up not only access to clinical services, but also rigorous quality control and outcome measurement processes. These methods need to be practical to implement in routine practice contexts, where time is scarce, and profit margins are low. Just as the invention of the Sony U-matic videotape recorder in 1969 ushered in a new way of recording clinical practice, more advanced technological solutions are now at hand that can provide a breakthrough in the feasibility of ensuring quality standards in routine practice — at scale.
In our university research, my colleagues and I have developed methods for collecting continuous objective measures of health behaviors (e.g., sleep, physical activity, mobility, mood, cognition) using the smartphone that is already in everyone’s pocket. However, we didn’t want this to just be a research tool — we wanted these exciting new capabilities to have an impact on clinical practice. So, we started a company, Ksana Health, to build tools that use these new developments in ubiquitous computing and artificial intelligence to help behavioral health practitioners get their job done more easily, while also supporting them to deliver high-quality care. When patients and their practitioners use these tools, they create an objective data set that reflects both patient progress and practitioner actions — all collected in the clinical workflow with little to no extra effort. Artificial intelligence, particularly large language models, can analyze these complex data sets and provide patient-specific feedback and guidance on best practices — just like we do in training clinics and clinical trials, but at a fraction of the effort.
This is just one example among a number of promising new approaches, where the combination of scalable approaches to objective measurement of treatment processes (such as audio recordings and transcripts of treatment sessions) can be combined with advanced AI capabilities to monitor and improve the quality of behavioral health services.
Despite the promise of these approaches, there are also some potential barriers to scaling them up. First, patients will have to become comfortable with sharing their health care data, although given that people are generally willing to share digital data if they receive benefits for doing so, that is probably a tractable issue.
Perhaps a greater hurdle will be for practitioners to accept the need for their work to be observed and evaluated to ensure quality standards. There is a long history of treating the psychotherapy room as a “black box” that is purely the purview of the practitioner. However, given that data on treatment processes are systematically collected in other areas of health care, there does not seem to be a justification for not doing so in behavioral health care. Our patients deserve nothing less.
Importantly, improving the quality of routine clinical care can make a difference. Research has found that treatment integrity in behavioral health care (i.e., adherence to treatment procedures with skillful delivery) significantly improves outcomes. Indeed, other areas of health care have also shown that focusing on improving quality standards results in better outcomes. One example is the use of checklists in surgical and ICU contexts, where this simple intervention has been shown to increase adherence to best practices and improve patient outcomes. If appropriately deployed, these new, scalable approaches to ensuring the quality of services will ensure that the impact of every dollar invested in increasing access to behavioral health care is maximized by ensuring that those in need receive high-quality care, irrespective of who they are or where they live.
Nick Allen is professor of clinical psychology at the University of Oregon and the director of the Center for Digital Mental Health. He is also the CEO and co-founder of Ksana Health, a digital behavioral health company.