Evaluating a Personalised Online Program With Human Support for Women With Depression and Its Impact on Mental Health and Biological Changes

OBJECTIVES Primary objective: To assess the impact of the 3D programme on depressive symptom severity compared to a psychoeducational control condition.

Secondary objectives: To evaluate improvements in overall functioning, quality of life, perceived stress, and menstruation-related distress. Additionally, the study will investigate metabolic signatures associated with treatment response, focusing on tryptophan and steroid hormone pathways.

HYPOTESES Main hypothesis: Participation in the 3D programme will lead to greater improvements in depression severity (measured by HDRS-17) compared to receiving only informational materials when used as an add-on to Treatment As Usual (TAU) in women experiencing moderate depressive episodes.

Secondary hypotheses:

1. Participation in the 3D programme will improve overall functioning in women with moderate depressive episodes.
2. Participation in the 3D programme will improve health-related quality of life in women with moderate depressive episodes.
3. Participants in the 3D programme with higher perceived stress levels will show less improvement in depression severity compared to those with lower stress levels.
4. Participants in the 3D programme with higher menstruation-related distress will show greater improvement in depression severity than those with lower levels of distress.
5. Participants in the 3D programme with more gynaecological problems will experience greater improvement in depression severity than those with fewer gynaecological issues.
6. A metabolic signature related to tryptophan metabolism and steroid hormones will predict the response to the 3D intervention.
7. Participation in the 3D programme will lead to changes in hormone production and/or balance, with the persistence of these hormonal changes correlating with the long-term effects of the intervention.

SAMPLE SIZE ESTIMATION AND RECRUITMENT The optimal sample size is estimated to be 60 participants per group. This calculation is based on a Cohen's d effect size of 0.2, a 95% confidence interval, a statistical power of 90%, and an anticipated attrition rate of 20%. Sample size estimation was conducted using G* Power 3.1.7 software.

All patients meeting the inclusion and exclusion criteria will be invited to participate in the study by their treating team at the outpatient mental health facilities within Hospital del Mar. Additionally, gynaecologists at Hospital del Mar will be encouraged to refer patients suspected of experiencing depressive symptoms. In such cases, patients will be screened by the study psychologist/psychiatrist to confirm these symptoms. Recruitment will conclude upon reaching the target of 120 participants with complete data.

RANDOMISATION AND BLINDING Participants will be randomised in a 1:1 ratio using the REDCap randomisation module, which will automatically assign participants to either the experimental or control group through simple randomisation with equal allocation.

This study will implement single blinding. Due to the nature of the intervention, participants will be aware of their group assignment (experimental or control) following randomisation. Likewise, therapists delivering the intervention will be aware of the assigned group, as is standard practice in psychological intervention trials. To uphold the integrity of the blinding process, the following measures will be applied:

1. Personnel responsible for data analysis will remain blinded to group assignments.
2. Independent raters conducting outcome assessments will be blinded to participant group allocation.
3. Other stakeholders, including clinicians, statisticians, and the principal investigator, will remain unaware of the randomisation process and group assignments to minimise potential bias.

STUDY DESING Part 1. Randomised controlled trial The study follows a naturalistic treatment approach, ensuring that all participants continue their usual care. The study will be conducted in accordance with the CONSORT guidelines.

PART 2. Metabolic signature of treatment response The metabolic signature of treatment response will be characterised through the analysis of tryptophan and steroid-related pathways. All analyses will be conducted using methods developed and validated by the Applied Metabolomic Research Group at Hospital del Mar Research Institute, ensuring proper sample handling protocols as outlined in previous studies.

Tryptophan Metabolism:

Tryptophan pathway metabolites-including tryptophan, kynurenine, serotonin, 3-hydroxykynurenine, 5-hydroxyindoleacetic acid, and kynurenic acid-will be quantified using liquid chromatography-tandem mass spectrometry (LC-MS/MS). The analysis will follow the protocol established by Marcos et al.

Analysis will focus on the kynurenine/tryptophan (K/T) ratio, a marker of indoleamine-2,3-dioxygenase (IDO) activity, which plays a key role in immune regulation and inflammatory responses.

Additionally, serotonin synthesis and its relationship with tryptophan availability will be assessed, alongside the tryptophan/large neutral amino acid (LNAA) ratio, which serves as an indicator of tryptophan's availability for serotonin production. Data will be integrated using network analysis to identify potential associations between these metabolites and treatment response.

Steroid Hormones:

The steroid profile will be analysed to monitor key sex hormones (e.g., estradiol, testosterone, progesterone) and glucocorticoids (e.g., cortisol, 20α-DHE, and 20β-DHE), along with their metabolites. These hormones play a pivotal role in modulating inflammation and metabolic pathways, with altered levels linked to physiological stress responses and treatment outcomes. These measurements will be performed by LC-MS/MS, following the method previously outlined by the research group.

Biological Samples:

Analysis will be conducted using various biological matrices, including blood, saliva, urine, nails, and hair, to provide complementary insights. Saliva, blood, and urine samples will capture acute metabolic changes at the time of the visit, while nails and hair will reflect chronic metabolite production.

ASSESSMENTS Assessments will be carried out at baseline (T0), immediately following the intervention (3 months after baseline, T1), and at a follow-up assessment (6 months after baseline, T2).

The baseline assessment will include:

• Collection of socio-demographic and clinical data, with self-reported information gathered via a secure, online-based assessment system (REDCap).
• Collection of biological samples (blood, saliva, urine, nails, and hair), with sample collection times recorded to account for circadian-sensitive metabolites. These samples will be collected exclusively from cisgender women.
• A gynaecology consultation to assess hormonal and menstrual complications. These will be evaluated using a 5-point Likert scale, ranging from "1 - No complication" to "5 - Severe complication requiring significant gynaecological intervention or further medical examination." Patients who require additional medical evaluations will be referred for appropriate follow-up care.
• Metabolic analysis, focusing on tryptophan metabolism and steroid hormone regulation, with the goal of identifying biomarkers predictive of therapeutic outcomes.

Clinical data and biological samples will be collected at all time points.

DATA MANAGEMENT All the data will be text-based (.docx) or numeric format (.xlsx). For all published files, a document record and change track will be included (author contact information, status, version, change reason and date, contents" description, title, origin of the data including a description of the measurement and/or experiment setup) in a separate metadata file for each characterization action called METADATA.ODS.

Data will be stored on a secure server at the Hospital del Mar Research Institute, in compliance with EU and Spanish data protection regulations (General Data Protection Regulation, GDPR, of May 25, 2018; and Organic Law 3/2018 of December 5, on the Protection of Personal Data and Guarantee of Digital Rights).

STATISTICAL ANALYSIS

Primary and Secondary Outcomes:

Analyses will follow an intention-to-treat (ITT) approach, including all participants according to their randomized allocation, regardless of adherence to the intervention.group. Linear mixed-effects models (LMMs) will be used to examine changes over time and between groups for the HDRS-17, WHO-DAS 2.0, EQ-5D-5L, PSS, and MEDI-Q scores. These models will include fixed effects for group, time, and their interaction, and a random intercept for participants to account for within-subject correlation over time. Where appropriate, subscale-level analyses and models using change scores will also be conducted for secondary outcomes. To further explore the temporal dynamics of the intervention's effects, post hoc simple effects analyses will be performed to probe significant interactions.

Additional Analyses:

LLMs and multiple linear regression models will be employed to identify predictors of clinical improvement. Additionally, moderator analyses using linear mixed models will assess whether levels of perceived stress (PSS), or menstrual distress (MEDI-Q) influence treatment response as measured by changes in HDRS-17 scores over time.

Exploratory Biomarker Analyses:

Partial Least Squares Regression (PLSR) and multiple regression will be conducted to model associations between metabolite profiles, their changes over time, and changes in depressive symptoms. Variable Importance in Projection (VIP) scores and cross-validation techniques will be applied to evaluate model performance and identify key predictive biomarkers. Additionally, metabolomic evolution will be assessed through exploratory factor analysis, grouping variables into latent factors.