At-home, sublingual ketamine telehealth is a safe and effective treatment for moderate to severe anxiety and depression: Findings from a large, prospective, open-label effectiveness trial

Depressive and anxiety disorders are major and growing contributors to disability, and existing treatment capacity and effectiveness are increasingly seen as inadequate. Earlier research has documented rapid symptom reductions with compounds that have dissociative or psychedelic properties, and ketamine in particular has substantial evidence from intravenous (IV) and intramuscular (IM) studies; sublingual and intranasal forms are emerging because they are easier to deliver. At the same time, telemedicine and remote monitoring have expanded access to mental healthcare, but large-scale evidence is needed on safety and effectiveness when ketamine is delivered at home with psychosocial support via telehealth. Hull and colleagues set out to evaluate the real-world safety and effectiveness of an at-home, sublingual ketamine-assisted therapy (KAT) model delivered through a telemedicine platform. The prospective, open-label study aimed to characterise symptom change in depression and anxiety over four weeks, document adverse events and dissociative reactions, and identify distinct patient response trajectories and their predictors using observational clinical records and routine outcome measures.

The evaluation used de-identified routine clinical and quality-management data from a telemedicine service that provided KAT to patients in 15 US states between 21 January 2021 and 30 November 2021. Patients completed eligibility screening online, then a video-based intake with a prescribing psychiatric clinician who confirmed diagnosis and safety for at-home sublingual ketamine. Analytic procedures were approved as exempt by an institutional review board. Inclusion required age ≥18, reliable internet access, a clinician-assigned diagnosis of depression or anxiety via video intake, and baseline scores ≥10 on the PHQ-9 (depression) and/or GAD-7 (anxiety). Exclusion criteria included ketamine allergy, current substance dependence, history of opioid use disorder, active or past psychosis or mania, uncontrolled cardiac or other serious medical problems, pregnancy or nursing, recent suicidal intent or attempt, or any other clinical concern rendering at-home treatment unsafe. The extracted text reports that 2,848 patient records were reviewed under these criteria; of those, 28 dropped prior to first medication session and 1,247 had sufficient outcome data for analysis. (The Discussion later refers to a chart review of 4,334 cases; this discrepancy is present in the extracted text and is not resolved here.) Treatment involved an initial 40–60 minute video intake, mailing a single-dose pack of 300 mg to 450 mg rapidly dissolving sublingual ketamine tablets, and provision of ancillary materials (ondansetron, a digital blood pressure cuff, eye mask, music, and written/video instructions). Dosing aimed to compensate for sublingual bioavailability and to achieve clinician-observed dissociative effects; the initial dose was described as approximately 5 mg/kg for the first session with individualised adjustments thereafter. A trained non-licensed guide supported each patient with a pre-session 30 minute video to set context and confirm a physically present peer monitor, real-time text support during treatment, a 30 minute guide-led meeting after the session, and daily text check-ins. Clinicians conducted follow-up video consultations 1–2 days after the first session and as needed thereafter. Safety rules prevented treatment if the peer monitor was absent or vital signs exceeded thresholds. Outcome assessments included the PHQ-9 and GAD-7 at baseline, after Session 2 (week 2) and after Session 4 (week 4). Suicide screening (C-SSRS), alcohol (AUDIT) and drug use (DAST-10) were completed at baseline. Side effects and adverse events were captured via a single-item checklist after sessions 2 and 4 plus EHR-recorded clinician or guide reports. Dissociation was measured with three adapted items from the Clinician Administered Dissociative States Scale using a 5-point Likert scale. Primary clinical definitions were: response = ≥50% reduction in score; clinically significant change = ≥5-point reduction crossing below the clinical threshold; remission = follow-up score <5 after baseline ≥10; deterioration = ≥5-point reliable increase. Analyses were performed in R 4.1, with Growth Mixture Modeling (GMM) in Mplus 8 used to identify latent symptom trajectories for PHQ-9 and GAD-7 modelled as parallel processes. Missing predictor values were imputed with the missForest random forest algorithm (100 trees, 10 iterations). After trajectory identification, multinomial logistic regression (3-step) tested baseline and treatment-related predictors including dissociation and side effects.

The analysed cohort comprised 1,247 patients with at least two outcome measure completions. Patients were aged 19 to 76 (mean 40.0, SD 9.1); 54.6% were women and 5.0% lived in CMS-designated rural zip codes. Baseline measures indicated moderate-to-severe depression and predominantly severe anxiety. The baseline C-SSRS mean was 0.54 (SD 0.92), with 408 patients (33.4%) reporting some suicidal ideation. Baseline substance and alcohol use scores indicated that 51.2% reported some drug use in the prior 12 months and 80.9% reported some alcohol use. Primary clinical outcomes over the four-week, four-session protocol showed large group-level improvements. Response rates were reported as 62.8% for PHQ-9 (depression) and 62.9% for GAD-7 (anxiety), with effect sizes of d = 1.61 and d = 1.56 respectively. Reported remission rates were 32.6% for PHQ-9 and 31.3% for GAD-7. Deterioration was uncommon: 0.9% on the PHQ-9 and 0.6% on the GAD-7 as reported in the abstract; the main text also describes overall deterioration as <1%. Adverse events and tolerability metrics were favourable in this sample. Side effects were reported by 59 patients (4.7%) after Session 2 and 27 patients (3.8%) after Session 4. Four patients had treatment discontinued due to adverse events: one for elevated heart rate, one for worsening depressive symptoms, one for increased urinary pressure with haematuria 24 hours after the fourth session (subsequently resolved by CT and ultrasound), and one who presented to an emergency department for worsening anxiety and depression and was admitted for psychiatric care. Two additional patients were removed from treatment by clinicians for noncompliance; the authors report a total of six patients dropped or removed for these reasons. Dissociative responses were common: after Session 2 the mean dissociation score was 3.4 (SD 2.8) with 82.9% (1,034/1,247) reporting any dissociation; after Session 4 the mean was 3.7 (SD 2.9) with 87.3% (618/708) reporting any dissociation. Growth Mixture Modeling identified three distinct concurrent PHQ-9/GAD-7 trajectory classes. The largest class (Improvement, 79.3%) began with lower baseline scores and showed steady symptom reduction across four weeks. The Chronic class (11.4%) had higher baseline symptoms and, despite some reduction, remained clinically elevated. The Delayed Improvement class (9.3%) started with elevated symptoms that remained relatively unchanged across the first two sessions and then declined sharply in the final two weeks. Predictor analyses found that reporting side effects at Session 2 was associated with membership in the Delayed Improvement group, and Chronic patients were more likely to exhibit dissociation at Session 4 and to endorse more baseline suicidal ideation or behaviour than the Improvement group. No other significant baseline predictors were reported in the extracted text.

Hull and colleagues interpret these data as evidence that at-home, sublingual KAT delivered with clinician prescribing and guide support via telehealth produced rapid, clinically meaningful reductions in depression and anxiety with a low incidence of adverse events. The authors note that response and remission rates are similar to or slightly stronger than those reported for clinic- or laboratory-administered ketamine treatments, and that the sustained administration schedule (repeated doses) used here may contribute to improved outcomes compared with single-infusion protocols. They position their findings relative to prior work by highlighting the notable anxiolytic effects observed, an area with fewer prior studies, and by drawing attention to reductions in suicidal ideation consistent with other ketamine research. The higher-than-typical proportion of men in this sample is discussed as a potential indicator that this treatment model may overcome some barriers to care for male patients; conditional models did not show outcome differences by gender in the extracted text. The authors acknowledge several limitations. Missing follow-up symptom measures at week 4 affected about half of records and could bias outcome estimates upward despite sensitivity analyses that detected no systematic differences on observed variables between completers and non-completers. The study’s open-label, observational design lacked a control group and involved self-selected patients, limiting causal claims and generalisability. Short follow-up (four weeks) precludes conclusions about durability of benefit. The authors also note complexities in the role of dissociation: higher dissociation at the end of treatment was associated with reduced likelihood of improvement, while mid-treatment side effects related to delayed improvement, suggesting timing and individual differences are important areas for further research. Finally, they suggest that telehealth-enabled, at-home sublingual KAT combined with remote monitoring and guide support may be a scalable approach to increase access to rapid-acting treatments for depression and anxiety, but emphasise the need for future research to assess long-term outcomes and to compare this model directly with other ketamine delivery methods and non-ketamine treatments.

Across 1,247 patients with sufficient follow-up data, four at-home sublingual KAT sessions over four weeks were associated with response rates of 62.8% for depression and 62.9% for anxiety and large effect sizes (d = 1.61 and d = 1.56). Remission occurred in roughly one-third of patients, deterioration was rare (<1%), and six patients discontinued or were removed from treatment for adverse events or noncompliance. The authors conclude that at-home sublingual KAT with telehealth-based clinical oversight and guide support demonstrated rapid, substantial clinical benefits with a favourable safety profile, and recommend studies to evaluate the durability of these effects.