‘Treatment-resistant depression’ (TRD) can be defined as a severe form of depression that does not improve after first-line treatments, such as antidepressant medication or psychotherapy. It frequently substantially decreases the quality of life of individuals affected by it and places a high socioeconomic burden on society at large due to increased healthcare service utilisation and increased rates of unemployment.
In order to improve the prognosis of people with ‘treatment resistant depression’, it is important to consider alternative pathways and augmentations to currently available treatments. For instance, a previous randomised controlled trial (RCT), the Tavistock Adult Depression Study (TADS; Fonagy et al., 2015) found that treatment-as-usual (TAU) combined with long-term psychoanalytic psychotherapy (LTPP) was better than TAU alone in improving the long-term outcomes and reducing the risk of relapse.
A previous blog on the original TADS study upon its publication, pointed out that a cost-effectiveness analysis could have been helpful to include to deliver a more realistic picture of the decision-making in psychological healthcare (McDonald, 2016). Koeser et al. (2023) have now released a new peer-reviewed paper utilising the TADS RCT data to compare the cost-effectiveness of LTPP and TAU together versus TAU alone.
All participants had a minimum of a 2-year history of depression and at least two previous unsuccessful treatment attempts.
- The treatment as usual (TAU) is comprised of various recommended short-term treatments as indicated by national guidelines (NICE, 2010).
- The long-term psychoanalytic psychotherapy LPTT+TAU group was additionally offered 60 weekly sessions of individual psychoanalytic psychotherapy over an 18-month period.
In both groups, 70-80% of participants were also receiving antidepressant treatment during the treatment and the follow-up periods.
TADS was conducted over a 3.5 year time period with the treatment period lasting 18 months. It used the CORE-OM (Evans, 2000) and the HRSD-17 (Hamilton, 1960) measures. Data was collected at baseline, 3-month intervals during the treatment, end of treatment, and then at 24, 30, and 42 months after treatment. For the present study, quality-adjusted life years (QALYs), which are generally the preferred outcome measure for economic evaluations, were calculated from CORE-OM scores using a previously established algorithm (Mavranezouli et al., 2012).
A comprehensive account of different healthcare costs was accounted for including the cost of the treatments being assessed, as well as visits to the GP, hospital, and other healthcare costs. The cost of antidepressants was not included, but was said to be small. All the costs were adjusted to 2014 prices using the hospital and community service index with costs and outcomes discounted at 3.5%.
The incremental cost-effectiveness ratio (ICER) between the two treatment groups was estimated by dividing the difference in mean costs by the difference in mean QALYs.
The main sensitivity analyses conducted were the effects of altering the cost of the intervention by +/-25% and using minimum rather than average wages to cost lost employment and receipt of unpaid care.
On average, participants in the LTPP+TAU group attended 48 sessions of psychoanalytic psychotherapy, three assessment sessions, and three review sessions. Additionally, the LTPP+TAU group received an average of four sessions of individual non-psychoanalytic therapy. In the TAU-only group, on the other hand, participants received on average 12 sessions of non-psychoanalytic individual psychotherapy.
The LTPP+TAU group received informal care for 2.7 hours per week on average, compared with 3.4 hours reported by those in TAU. There were small changes in employment status across the groups during the trial. Slightly more individuals from the LTPP group were in employment compared with those receiving TAU. Both groups had similar working days lost due to illness over the trial period, resulting in 9 days in both groups.
The average cost of LTPP was estimated to be £5,468 per person. GP costs were similar in the two groups, but the TAU group had more contacts with the GP, which was a result of the differences in the length and type of the contacts. The groups differed in two measurements: a) the cost of community care which was substantially lower in the LTPP + TAU group and b) the cost of hospital care, which was higher in the LTPP + TAU. However, estimates across these two measurements were associated with a large degree of uncertainty. The overall cost differences between the two groups were mainly associated with the cost of treatment. The mean difference in healthcare costs between the two conditions was £5,227, with the LTPP+TAU group being more expensive.
As for the QALYs and HRSD-17 scores measuring the effectiveness of the interventions, the LTPP+TAU had higher average QALYs indicating higher quality of life (mean different 0.16) and a lower weighted average HRSD-17 score indicating reduced levels of depression (mean difference -1.46). Additionally, one patient in the TAU-only group died by suicide, which led to the estimation of suicide risk being 0.5% higher in that group.
In terms of cost-effectiveness, regardless of the outcome measure used, LTPP+TAU was found to be >90% more likely to be both more effective and more expensive than TAU alone. The ICER was £33,000 per QALY in the LTPP+TAU group. As such, the likelihood that LTPP+TAU was cost-effective compared to TAU alone was 18%, assuming a QALY gain of £20,000. The £20,000 estimate is based on the newest NICE guidelines (NICE, 2022) according to which the most plausible ICER is below £20,000/QALY. According to the estimates in the present study, LTPP gives an ICER of £33,000/QALY gain, which places it in the >30,000/QALY category, where the case for other factors than cost-effectiveness needs to be increasingly strong.
Finally, these results were highly sensitive to variations in the cost of LTPP. Reducing its cost by 25% dropped the ICER to approximately £24,500, whereas raising it by 25% increase it to around £41,800.
Although this study suggested that long-term psychoanalytic psychotherapy in combination with treatment as usual was not cost-effective compared with TAU, there may be more explanations to this. Given the sustained long-term effects over the follow-up period and the trajectory of treatment benefits in the LTPP+TAU group, it seems plausible that the timescale of this study was too short to accurately portray the long-term benefits of LTPP. Furthermore, the authors point out that QALYs estimates are sensitive to the statistical approach used. Additionally, the cost of the intervention and the parameter uncertainty in the current study was quite large. Therefore, a more comprehensive exploration into the long-term clinical effects and cost-effectiveness LTPP augmented treatment is warranted.
Strengths and limitations
The study presents a few strengths to take into consideration, including:
- Health economic data collection was comprehensive, as it was based on the pragmatic TADS RCT. The original RCT had a well-defined sample of individuals with a TRD diagnosis.
- Outcome measures in the TADS RCT included functioning and quality of life measures. Moreover, the measures were reliable and rated by independent raters blind to participants’ treatment allocation.
- Cost-effectiveness results could have high external validity due to the TAU control condition being a real-world comparison group.
However, there were certain limitations to point out:
- The sample size was relatively small, and the study was conducted in a single clinical setting, the Tavistock Clinic in London. This service has expertise in the delivery of psychodynamic approaches and psychodynamically informed interventions. This only leaves the readers with questions about the generalisability of the findings to populations with different demographic characteristics and clinical settings.
- QALYs were calculated with a measure designed for depression research (CORE-OM) and little research has been conducted to date on its properties.
- Some of the service-use data was missing which may have impacted the pattern of costs in each group. Furthermore, it may have been helpful to identify a value for the cost of the antidepressants as well as to support the conclusions and interpretation of the findings.
- The potential benefits, including all cost benefits of LTPP on patients’ families and carers were not accounted for in this study.
- Primary care reimbursement rates were used as the method to estimate the cost of LTPP sessions which was justified as providing good face validity. Yet, the cost per session was in the higher range when compared to, for instance, CBT session rates. This may have implications for the ICERs given that these were highly sensitive to variation in the price of LTPP.
- What happens to outcomes over a longer timescale? Understanding the durability of treatment effects beyond the study period would enhance the practical implications. If a longer course of psychoanalytic psychotherapy helped address treatment resistant depression more permanently above and beyond 3.5 years, how might this impact its cost-effectiveness?
Implications for practice
The treatments were based on NICE guidelines so the costs are likely to be more or less translatable across different NHS trusts even if some variation might exist. However, the provision of public healthcare for depression is of variable quality. Particularly the care for treatment resistant depression (TRD) is frequently not optimal, hence there is a strong need for improvement in establishing and providing effective treatment options.
As the authors point out, it may be justified to value each QALY gain in this particular population greater than in those with less severe conditions. This could have implications for areas that fall outside the scope of this cost-effectiveness assessment, such as suicide risk. Nevertheless, this is difficult to assess even in currently widely prescribed treatments such as antidepressants. Future directions could also include identifying other clinical subgroups for whom LTPP might be particularly effective.
I want to leave you with two questions to reflect on:
- Is the current model of addressing relapse after relapse for those with persisting depression truly sustainable in the long run?
- Is it indeed more costly to deliver a year of therapy to help people with treatment resistant depression gain the insight and tools to achieve a longer-lasting reduction in symptoms and an increase in quality of life?
I know what sounds like a more sustainable approach to me. What about you?
Statement of interests
Ella Tuominen is a PhD student at UCL and her research interests include the cognitive mechanisms of psychological therapy modalities and their impact on anxiety symptoms, such as avoidance behaviour. Ella has no conflicts of interests with the study reviewed in this blog.
Koeser, L., Rost, F., Gabrio, A., Booker, T., Taylor, D., Fonagy, P., … & McCrone, P. (2023). Cost-effectiveness of long-term psychoanalytic psychotherapy for treatment-resistant depression: RCT evidence from the Tavistock Adult Depression Study (TADS). Journal of Affective Disorders, 335, 313-321.
Evans, J.M.-C., 2000. CORE: clinical outcomes in routine evaluation. J. Ment. Health 9, 247–255.
Fonagy, P., Rost, F., Carlyle, J. A., McPherson, S., Thomas, R., Pasco Fearon, R. M., … & Taylor, D. (2015). Pragmatic randomized controlled trial of long‐term psychoanalytic psychotherapy for treatment‐resistant depression: the Tavistock Adult Depression Study (TADS). World Psychiatry, 14(3), 312-321.
Hamilton, M., 1960. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry 23, 56.
Mavranezouli, I., Brazier, J.E., Rowen, D., Barkham, M., 2012. Estimating a preference- based index from the clinical outcomes in routine evaluation-outcome measure (CORE-OM) valuation of CORE-6D. Med. Decis. Mak. 33 (3), 381–395, 0272989X12464431.
NICE, 2010. Depression – The NICE Guideline on the Treatment and Management of Depression in Adults: Updated Edition.
NICE, 2022. NICE Health Technology Evaluations: The Manual.