The Therapy Elite Won’t Like This

I feel shook.

I attended a departmental colloquium earlier this month by the very impressive Dr. Daisy Singla, a clinical psychologist and senior research scientist at Toronto’s Centre for Addictions and Mental Health (CAMH). I loved Dr. Singla’s talk and walked away having learned something important. Though perhaps not what she intended.

New information has come to light, man.

Her talk, “Scaling up access to patient-centered psychotherapies,” focused on large-scale trials she’s run around the globe showing that non-specialists and video-conference delivery of mental health support are effective and scalable. While she talked about a few papers with samples from Uganda and Bangladesh, I want to focus on one published in Nature Medicine a few months ago.

The paper itself is impressive. It reports the results of the “Scaling up maternal mental healthcare by increasing access to treatment” (SUMMIT) trial. SUMMIT is a large, multisite randomized trial conducted across hospital networks in Canada and the United States, pulling in 1,230 pregnant or postpartum women between 2020 and 2023 who were suffering from depression.

Everyone in the trial received the same behavioral activation therapy, a brief, structured, evidence-based treatment for depression. What varied was who delivered the therapy and how. Patients were randomized to see either a specialist (licensed clinicians) or a nonspecialist (doulas, community health workers, peer supporters, or nonclinical hospital staff). These are not randos off the street, but hospital staff who have some perinatal experience but no formal mental-health training or licensure. Therapy was delivered in-person or via videoconferencing with technology like Zoom.

The nonspecialists weren’t thrown in cold. They received structured training in behavioral activation following a standardized manual that included learning basics, observation, role-play, games, and homework. After this initial workshop, they entered an eight-week supervised internship, where they delivered BA to one or two real patients from start to finish while being monitored for competence. During the trial itself, they continued to receive weekly to biweekly supervision...