Woman holding coffee mug and staring out window.

Across the world, there is a mental health treatment shortage. This problem is not just hinged on labor shortages - it is caused by barriers to efficiently matching patients to the right resources.

Across the world, there is a mental health care treatment shortage; we are in a crisis. Mental health care has never been in such high demand. In fact, by 2025, shortages are projected to get worse for psychiatrists, clinicians, counselors, and psychologists. Today, in the US alone, thousands of people are sitting on waiting lists for evaluation, treatment, and just someone to talk to. Clinical depression is one of the most common mental health diagnoses in the US, with more than 21 million adults reporting a clinical episode in 20201. In 2022, 9 out of 10 therapists reported “the number of clients seeking care is on the rise”1. This demand is magnified in rural areas. A family therapist in Georgia reported “I live in a rural town, but I still get approximately seven to 10 inquiries a week that I have to turn away”1. This is one of many other examples that show the gap in the mental health care system is increasing every day.

It is important to remember this gap in supply and demand can not be solely addressed through increased accessibility, which has been the focus of most innovative solutions over the last few years. While this “Room to Zoom'' approach, widely adopted recently, would potentially provide access to care for someone living in rural areas, it would not guarantee care availability. Too often patients and clinicians are restrained by factors out of their control: time, money, and policies. Even when a patient can see a mental healthcare provider, their insurance coverage, time, and geographic location impact their path to treatment. These factors are more critical when social determinants are also considered, making marginalized groups at higher risk for long-term mental health concerns. The current health care system keeps patients, and clinicians, trapped in a circle of inaccessibility.

Problems in Care Model And Resource Allocation in Mental Health Care

During the COVID-19 lockdowns, clinicians heavily relied on virtual communication to connect with their clients. While using these virtual communication tools made it easier for clients to access their clinicians, what limited their access was clinicians time, leading to longer wait times. In fact, in 2021, 75% of clinicians reported an increase in wait times for basic mental health services1. That is, the technological access innovation did not solve clinician availability. That is because our industry has not been optimized for scale.

The question then becomes, how can we make mental healthcare more scalable?

Furthermore, while there has been an increase in differentiation of roles for managing patient interactions, we need to reimagine these roles by equipping individual clinicians with the right tools to effectively manage the population waiting for care. For instance, there is only 1 psychiatrist for each 1,960 patients in need of mental healthcare. Given that a psychiatrist can only handle 200-400 patients per year, inappropriate referral of all patients to a psychiatrist would cause extensive backlogs. In fact, in the United States the average wait time to meet with a psychiatrist, for an initial evaluation, is 50 days1, almost two months. Research has shown that longer waiting time for treatment leads to poorer health outcomes1. This means that when patients are waiting for treatment, their mental health concern, no matter how minor, could become more severe and therefore more difficult and expensive to treat. Nevertheless, not all patients need to see a psychiatrist. According to a clinical trial we are performing at OPTT Health, ~85% of the patients could be handled by a combination of supervised online CBT plus weekly engagements by lower cost personnel like social workers and mental health coaches. The remaining 15%, which translates to 294 patients out of 1,960, would definitely benefit from psychiatrist intervention. This model of assigning different resources to different patients can perfectly work in the new collaborative care models1

But how can we measure each individual patients’ needs for such optimized resource allocation? 

The Answer: AI-Powered, Digital-First Approach

There is not one solution to solve this macro problem, however many experts suggest that well developed and clinically validated digital care can be a key part of the solution set. There has been a shift towards virtual care, telehealth, and asynchronous digital care. Research shows that this trend has been expedited by the pressures of the COVID-19 pandemic. Digital care reduces patient and provider costs, allows care to reach those living far from resource centers, and connects patients to clinicians in a timely manner. Also machine learning and AI (ML/AI) algorithms have been developed to support clinicians in making efficient and effective decisions in their practice. The use of technology drives a better experience for both patients and clinicians, and in fact has been shown to improve professional engagement and job satisfaction for practitioners.

There are a few cautionary notes in using digital technology for care delivery though. First, is the question of clinical validation. It is important to remember that mental healthcare, like any other field of healthcare, needs to follow a rigorous validation process. Not every solution that works in person, works digitally, and not every solution actually helps. Any digital solution therefore should be clinically validated in proper clinical trials and meticulously adjusted to meet patients’ diverse needs. Second, is the challenge with care adherence. This means that, even if a solution has been shown to reduce symptoms in a clinical trial, patients should actually go through the process to see the outcomes. Many digital mental health solutions have +80% patient attrition past sessions 3-4 of their care plans. This means that most of the patients would not benefit from them, even though they have been validated.  

On the analytical and ML/AI side, there is always the question of whether these algorithms can be trusted. Most of the ML/AI algorithms are black boxes, that crunch in an array of variables and spit out a decision. As such, it is not clear for a clinician what this decision was based on and how they can trust its outcome. This lack of transparency and accountability therefore can hinder their use in the clinical process. Furthermore, many clinicians consider such use of AI in decision making as a competing factor for their job security.   

I firmly believe the answer to both these concerns is technology backed care delivery by clinicians. A hybrid model of care delivery in which digital technology is used to transfer the main time consuming parts of therapy to the patients and to save clinicians for short, personalized feedback, could both scale up the number of the patients each clinician can handle, and keep patients accountable and adherent to treatment. Additionally, ML/AI algorithms should become like any other medical tool, that while they might be technologically advanced, they provide intuitive information that are easily comprehended by clinicians (e.g. MRI imaging). Therefore, any ML/AI in mental health should also follow an explainable ML design so their functionality is transparent and trustable. For instance, algorithms could compile a big range of unstructured data (including patients’ use of language and speech, their activity and sleep patterns, etc.) to produce a few clinically relevant variables that help clinicians in their decision, rather than making the decision for them. At the end of the day, tools are not replacing clinicians, but empower them. 

Dr. Mohsen Omrani is Co-Founder and CEO at OPTT Health. OPTT is a provider of comprehensive hybrid digital care plans augmented by AI for proactive triage and monitoring of patients to simplify the digital-first delivery of mental health services for care teams. Learn more at www.OPTT.Health