What access and adherence barriers distinguish real-world transcranial magnetic stimulation use from efficacy trials in treatment-resistant depression?

The first barrier is getting referred at all

Before a patient ever sits in a TMS chair, someone has to recognize treatment-resistant depression and send them for the therapy. That step is where access often breaks down. A quality improvement project at a Florida integrative medicine practice found that TMS and intranasal esketamine use was limited by non-standardized referral processes, gaps in provider knowledge, and inefficient workflows, all of which delayed treatment and restricted access [4]. When the practice gave seven psychiatric providers a short educational module plus a standardized referral workflow, their knowledge scores rose substantially and they reported more confidence, fewer perceived barriers, and a universal intention to adopt the new workflow [4]. In other words, a large share of the access problem sits with referring clinicians' knowledge and process, not with patients' willingness.

A six-week course is a real commitment — and the sickest patients leave first

Conventional TMS is not a single visit. In the rural case series, all three patients completed 30 sessions over six weeks [1], and that schedule is typical of conventional protocols. Accelerated schedules compress treatment into fewer days — one real-world program used 10 daily applications across 5 days [3] — which could ease the time burden, though that study is designed to compare effectiveness rather than to measure adherence barriers [3].

The adherence picture from a large real-world program is telling. At a Quebec tertiary referral center, 247 patients received accelerated TMS, but only 147 completed treatment; the 100 who did not complete had generally more severe depressive and anxious symptoms and greater treatment resistance [2]. Among completers, 46.3% met response criteria and 36.1% achieved remission [2]. The pattern suggests that the patients with the greatest need are also the ones most likely to fall out of care — a selection effect that efficacy trials, which typically enroll and support patients through a fixed protocol, are not built to reveal.

Real-world results already fold in the barriers

When studies report real-world TMS outcomes, those numbers describe a filtered population — people who were referred, could travel, and stayed in treatment. A Japanese multisite observational study of 497 patients treated under public insurance found HAMD-17 scores fell from a mean of 18.9 to 9.7, with 53.5% response and 42.8% remission, and a dropout rate due to adverse effects of only 4.2% [5]. That low adverse-effect dropout is reassuring about tolerability, but it does not mean adherence is easy: it counts only people who left because of side effects, not those lost to logistics, cost, or worsening illness [5].

The Quebec data make the distinction concrete: non-completion there was linked to greater baseline severity and treatment resistance rather than to the device itself [2]. So a clinic's headline response rate depends heavily on who makes it through the door and through the course. Comparing real-world response rates to trial efficacy without accounting for that filtering can overstate how well a program is reaching the patients who need it most.

What would actually narrow the gap

The evidence points to two practical levers. The first is upstream: standardizing referral workflows and educating providers measurably improved knowledge, confidence, and readiness to refer in one small quality improvement project [4], and rural implementation appears feasible when access barriers are reduced [1]. The second is schedule and support: accelerated protocols may reduce the number of clinic days a patient must commit to [3], though whether that improves completion in routine care is still an open question, since the accelerated-versus-conventional comparison is designed to measure symptom response, not adherence [3].

What remains genuinely uncertain is scale. The rural feasibility evidence comes from three patients [1], and the referral project involved seven providers [4], so both are signals rather than settled answers. Larger implementation studies tracking who is referred, who starts, and who finishes — not just who responds — would show whether these fixes hold outside small pilots.

Sources used in this answer

Related research questions