Psychiatry · Medical Literature
Cost-effectiveness analysis and cost-utility analysis of a model of integrated mental health video consultations (PROVIDE-C randomised controlled trial) for people with depression and anxiety.
Abstract
We conducted a health economic evaluation alongside the PROVIDE-C randomised controlled trial which analysed the effectiveness of a model of integrated mental health video consultations for people with depression and anxiety in primary care in Germany in comparison to treatment as usual (TAU). A cost-effectiveness (CEA) and cost-utility analysis (CUA) were performed six and twelve months after baseline from a societal perspective. Bootstrap procedure was applied and for each bootstrap replication, multiple imputation was conducted, and incremental costs and effects were estimated using generalised linear models. In total 376 participants (18 to 81 years) with depression and/or anxiety were included in the analysis. We found that €3662 would be required for an additional person who experienced a clinically minimum important reduction of 5 points at the Patient Health Questionnaire Anxiety and Depression Scale (PHQ-ADS) and €1622 would be required for an additional person with a PHQ-ADS score below 10 points (persons with not more than mild symptom severity) six months after baseline. At twelve months after baseline, PROVIDE-C was dominant compared to TAU. While 6 months after baseline the probability of cost-effectiveness was approximately 36%, 40% and 56% in terms of quality adjusted life years (QALYs) for the respective willingness-to-pay threshold of €0, €20,000, and €100,000, at 12 months after baseline the respective probabilities of cost-effectiveness increased to approximately 71%, 78%, and 86%. While the PROVIDE-C intervention is less likely cost-effective 6 months after baseline, it was likely cost-effective 12 months after baseline.
Clinical relevance
In adults with depression and/or anxiety treated in German primary care, is an integrated mental health video consultation model (PROVIDE) cost-effective and cost-utility-effective compared with treatment as usual (TAU) at six and twelve months from a societal perspective? __SNAPSHOT__ {"studyDesign":"Health economic evaluation (cost-effectiveness analysis [CEA] and cost-utility analysis [CUA]) conducted alongside an assessor-masked, multicentre, parallel-group randomised controlled trial (PROVIDE-C)","setting":"29 primary care practices in Germany; participants seen in-practice with remote mental health specialist video link to a single trial hub","sampleSize":"n=376 (intervention: n=187, control [TAU]: n=189)","population":"Adults aged 18–81 years presenting to their GP with depression and/or anxiety; mean age 45 years (SD 14), 63% female, mean PHQ-ADS score 26 points (SD 7.6) at baseline","intervention":"PROVIDE model: five 50-minute real-time transdiagnostic video sessions between the patient at the primary care practice and an offsite mental health specialist, delivered over eight weeks, focused on diagnostics, treatment planning, and brief psychotherapy","comparator":"Treatment as usual (TAU) by the GP, including brief counselling, psychotropic medication prescriptions, and possible referrals to mental health specialists","primaryOutcome":"CEA primary endpoint: proportion of participants achieving a clinically minimum important reduction of ≥5 points on the PHQ-ADS (Patient Health Questionnaire Anxiety and Depression Scale), and proportion with a PHQ-ADS score <10 (not more than mild symptom severity), at 6 and 12 months. CUA endpoint: quality-adjusted life years (QALYs); specific QALY instrument not reported in abstract.","followUp":"6 months and 12 months after baseline","fundingSource":"German Federal Ministry of Education and Research (BMBF); no competing interests declared by authors per published trial protocol"}
Key findings
From a societal perspective, the incremental cost-effectiveness ratio (ICER) for the PROVIDE intervention at 6 months was €3,662 per additional patient achieving a clinically minimum important reduction of ≥5 points on the PHQ-ADS, and €1,622 per additional patient reaching a PHQ-ADS score <10 (mild or no symptom severity). At 12 months, PROVIDE was dominant over TAU — meaning the intervention was both more effective and less costly — in the primary CEA analysis. The abstract indicates results extending to QALY-based cost-utility data and acceptability curve analyses, but specific ICER values per QALY, confidence intervals, p-values for cost differences, and probabilistic sensitivity analysis thresholds at which PROVIDE remained cost-effective were not fully reported in the provided abstract (abstract appears truncated after '6&'). No serious adverse events were reported in either group in the parent RCT. Bootstrap procedures with multiple imputation were applied to handle uncertainty and missing data; incremental costs and effects were estimated using generalised linear models. NNT for each CEA outcome cannot be independently calculated from the abstract data provided.
Practice implications
This study establishes that the PROVIDE integrated mental health video consultation model is potentially cost-effective compared with TAU at 6 months, with modest per-patient incremental costs to achieve clinically meaningful symptom reduction, and appears cost-saving (dominant) at 12 months from a societal perspective. These findings are consistent with the parent RCT's effectiveness results, which demonstrated statistically significant but small absolute improvements in PHQ-ADS scores (−2.4 points at 6 months, −2.9 points at 12 months), raising questions about the magnitude of clinical benefit even if cost-effectiveness thresholds are met. The study does not establish whether PROVIDE should replace TAU across all primary care settings in Germany; it demonstrates economic plausibility within the specific German statutory health insurance and GP-referral context. Key populations not well represented include patients with severe mental illness, those requiring crisis intervention, non-German-speaking patients, and those without access to suitable in-practice technology. The 12-month dominance finding is clinically important for healthcare planners, but should be interpreted cautiously as only an abstract is available, preventing full appraisal of the cost-utility results, sensitivity analyses, and the completeness of the societal cost components included. Practice change is premature pending full publication review, independent replication in other healthcare systems, and assessment of implementation costs at scale. __LIMITATIONS__ Only the abstract is available; the full cost-utility (QALY-based) results, probabilistic sensitivity analysis outputs, and cost-effectiveness acceptability curves at various willingness-to-pay thresholds cannot be fully evaluated.,The abstract appears truncated; data on 12-month CUA results and acceptability curve thresholds are not reported here.,Single-country study (Germany) with a specific GP-centric, statutory insurance-funded healthcare model; generalisability to other health systems (e.g., UK NHS, US fee-for-service) is uncertain.,Relatively small sample size (n=376) may limit the precision of bootstrapped cost estimates and the robustness of subgroup economic inferences.,The clinical effect size in the parent RCT was small (−2.4 to −2.9 PHQ-ADS points); whether this translates to meaningful QALY gains depends on the mapping algorithm used, which is not specified in the abstract.,Societal perspective includes productivity costs (absenteeism, presenteeism), which are subject to estimation variability and may not reflect payer (insurer) perspective relevant for reimbursement decisions.,Multiple imputation for missing cost data introduces assumptions that may affect reliability, particularly if missingness is not at random.,The 12-month dominance finding may be influenced by reduced healthcare utilisation in the PROVIDE group over time; the mechanism (intervention effect vs. regression to the mean) is not determinable from the abstract.,Selection bias in practice-based recruitment (noted as a limitation in the parent RCT) may limit the representativeness of the sample.,No information on implementation costs at scale (training, infrastructure, IT) that would be required for a real-world economic assessment. __BOTTOM_LINE_CATEGORY__ potentially_practice_informing __BOTTOM_LINE__ The PROVIDE integrated mental health video consultation model appears economically favourable compared with TAU in German primary care: at 6 months, each additional patient achieving clinically meaningful symptom reduction cost approximately €1,622–€3,662 (depending on the threshold used), and at 12 months the intervention was dominant (more effective and less costly from a societal perspective). These findings are promising for healthcare planners considering telepsychiatry-integrated primary care models, but full publication review is required before conclusions about QALY-based cost-utility and sensitivity analysis robustness can be drawn, and generalisability beyond the German health system is uncertain.