The prevalence and burden of mental illness.
Mental disorders are amongst the largest chronic disease groups and represent a major public health concern, posing a huge financial burden on the European health care system.
The total cost of mental disorders has estimated at more than 4% of GDP – or over €600 billion – across the 28 European countries in 2018, meaning these disorders pose a heavier financial burden on Europe than, for instance, cancer (€199 billion, 2018). Overall, individuals with severe mental disorders have not benefited from the general advances in health care in the last decades, largely due to a lack of progress in pharmacological treatment and disease management. Moreover, the presence of multimorbidities, including both comorbid mental and somatic diseases, adds to the suffering and large reduction in quality of life (QoL) experienced by this patient group.
Why treatment is so challenging
A fundamental challenge in psychiatry is treatment of psychotic and affective symptoms, which are core characteristics of the severe mental disorders schizophrenia (SCZ), bipolar disorder (BIP), and major depressive disorder (MDD).
While current medication for psychosis (antipsychotics) and depression (antidepressants) are effective for the majority of patients, there is a large variation in efficacy and adverse effects, while non-response to medication is a significant clinical problem, with failure rates above 30% in SCZ and similar rates in BIP and MDD. In addition, adverse effects are common, and often cause non-compliance.
Despite this, treatment decisions are currently made on a trial-and-error approach, as the understanding of the underlying disease mechanisms is still limited.
Further, many pharmaceutical companies have left the mental health field due to lack of novel drug targets. This makes it vital to optimise treatment outcomes using existing pharmaceutical alternatives.
The promise of ‘personalised medicine’
Due to recent progress in psychiatric genetics, there is now great potential for improving treatment outcomes (reducing adverse effects and increasing treatment response) by optimising use of existing medication based on the patient’s genetic profile. However, the relevant datasets necessary to develop and validate individualised treatment have not previously been available.
Real-world data (RWD) provides a unique opportunity to obtain massive datasets with sufficient statistical power to leverage novel ‘big data’ analytical methods. Gathering this data will enable the development of prediction and stratification tools with the precision required for translation into clinically useful decision support tools for ‘personalised’ psychiatric treatment.
REALMENT brings personalised medicine interventions to psychiatry across Europe. Building on RWD, we developing: i) a sustainable data infrastructure (REAL-WD platform) to facilitate access to RWD data across Europe; ii) novel artificial intelligence (AI) and machine learning (ML) tools to exploit large datasets/volumes of data; iii) a clinical management platform (4MENT) with prediction algorithms for medication response and adverse effects to significantly improve patient outcomes and quality of life.
Mental disorders with affective and psychotic characteristics (MDD, BIP and SCZ) are complex chronic conditions (CCC) with high degrees of heritability (40-80%).
Recent advances in psychiatric genetics have relied on many thousands of samples, and similar numbers are likely to be needed for genetic association studies of treatment response and of adverse effects.
A further complication is that many patients with mental disorders receive several drugs because of a complex, clinical picture with multiple disorders, including other psychiatric conditions (e.g. anxiety) and/or somatic diseases (e.g. cardiovascular, autoimmune diseases and cancer).
This makes multimorbidities an important clinical challenge. However, most current approaches are unable to include this aspect in the analysis. RWD from biobanks and registries can provide such data and the sample sizes needed to reach adequate power for discovering genetic factors associated with treatment outcome and multimorbidities within a reasonable timeframe and budget.
Further, emerging knowledge on variation in treatment outcome and comorbidity trajectories suggests stratification and prediction, and thus improved clinical management, are now achievable goals. However, novel analytical approaches are required to successfully integrate information into actionable disease management.
How REALMENT will help
REALMENT focuses on optimisation of existing treatments for SCZ, BIP and MDD, due to their high prevalence, overlapping symptoms and similar medication regimens, incorporating artificial intelligence (AI) and machine learning (ML) algorithms to increase analytical performance, building on methods from REALMENT partners.
Image: The REALMENT approach will integrate RCT data with cross-border RWD from multiple sources, develop methods for prevention and treatment stratification to optimize outcomes and quality of life in people with mental disorders.
Randomized clinical trial (RCT) data consists of well-controlled and ‘clean’ data. Typical RCTs follow up patients for less than six months. This short period prohibits assessment of long-term treatment outcomes. Likewise, patients with multimorbid diseases are excluded in RCT studies, since they are at greater risk of developing adverse effects and requiring more than one drug (polypharmacy).
The lack of data on multimorbidity impairs the broader application of RCT data to the evaluation of efficacy and the risk of developing adverse effects. Furthermore, treatment adherence is better in RCTs than in the real-world. In order to fill this knowledge gap, REALMENT will collect, collate and exploit one of the world’s largest collections of RWD from multiple sources including registries, biobanks and medical records.
REALMENT provides a unique opportunity to validate findings from RCTs (short term, monotherapy) in RWD (long term, polypharmacy) and vice versa. REALMENT will curate and combine cross-European RWD available to our partners and answer questions related to the long-term effects of psychopharmacological treatment in psychiatric patients, the development of multimorbidity and the effect of polypharmacy.
We will determine the clinical trajectories of patients who suffer from e.g. cardiovascular disease. In addition, we will characterise different treatment trajectories by combining prescription registries and electronic health records (eHR).
Since RWD is subject to high variability and confounders, RWD requires careful curation and validation. REALMENT will combine information from RWD data and RCTs for integrated analysis to ensure the necessary quality of the measures related to treatment efficacy and adverse effects. Further quality control will be performed using different types of RWD, from registries and biobanks (Nordic/Estonian), medical health records, as well as large clinical research data on mental disorder cohorts and data collected through interviews, mobile app data or by online questionnaires.
Validation will be done by an intervention study to re-phenotype (recall of participants) with ethics and privacy-compliant protocols. This is crucial for identifying the genetic architecture underlying treatment outcomes and multimorbidities, thereby enabling useful clinical application.
The information from integrated data will be included into new prediction algorithms and used to transform psychiatric treatment towards ‘precision medicine’ to improve treatment outcomes and patient wellbeing.
The discoveries and AI and ML algorithms from REALMENT will be used to develop the “4MENT” clinical management platform to support clinical decision making as applied in other CCCs15-17.
The 4MENT platform will ensure clinical exploitation and optimise treatment efficiency, reduce adverse effects and provide better long-term outcomes and quality of life (QoL) for psychiatric patients by incorporating genotypes and personal clinical traits into treatment strategies, taking multimorbidities into account.
This integration will lead to major benefits for patients and the European health care system at large.
Challenges to be solved by REALMENT
Large scale data capture of novel RWD linked to RCT data.
REALMENT will apply time-efficient deep learning/ML algorithms to optimise RWD collection and curation from eHR and registries about treatment and adverse effects to bypass inefficient, time-consuming methods unsuitable for integrating large sample sizes. The findings from RWD will be validated in available RCT data for the specific outcomes and drug types.
Collating large enough samples with relevant data to make use of big data algorithms.
Data from thousands of samples with treatment and adverse effect and multimorbidities are needed to release the potential of ML and AI, through streamlined computation pipelines and frontline analytical tools with proven effect in similar use cases.
Information about the timing of event is crucial for prediction. While data on disease onset and multi-morbidity is unavailable in most studies, including genome-wide association studies (GWAS), this is included in REALMENT.
Providing a sustainable, European-wide data analytics infrastructure.
REALMENT partners have developed the Tryggve infrastructure to analyse large samples of sensitive genetic, registry and medical records data in a secure, distributed manner using software containers across the Nordic countries.
The Tryggve solution will be integrated with other nodes within the European life-sciences infrastructure for biological information (Elixir) (section 1.3), to make a European-wide system, named the “REAL-WD” platform.
This secure data infrastructure and distributed analytical framework will solve ethical and GDPR aspects, and form the basis for further large-scale initiatives and exploitation.
Algorithms for prediction of drug response and adverse effects in individual patients.
REALMENT will connect genetic information with clinical and outcome data to develop novel algorithms to predict treatment response and development of multimorbidities, and to stratify patients accordingly – thereby enabling improved treatment decisions.
Previous results suggest that genetics together with clinical factors have the potential to predict response and adverse effects of medication, with a large potential for major psychopharmacological agents.
Validation of algorithms with intervention in patients.
We will validate the prediction/stratification algorithms in clinical settings, to ensure real-world value, uptake and adherence to privacy protection guidelines.
Clinical management platform for individual patients.
The algorithms, individual patient’s genetic profile and baseline clinical data will be incorporated into a clinical management platform (“4MENT”) that will function as a clinical decision support tool to predict disease development, and improve outcomes and disease trajectories.
REALMENT will develop and validate digital tools and data platforms to capture and analyse RWD from patients with mental disorders and associated comorbidities.
These tools, based on RCT and RWD, will form the basis for a new treatment strategy for mental disorder patients where use of available pharmaceutical agents will be optimized and lead to considerable benefits for patients, their families and society at large.
Furthermore, REALMENT will implement a programme to make the infrastructure, tools and platforms sustainable for the future and will perform intensive communication, dissemination and exploitation activities to ensure wide implementation across Europe.
The REALMENT partners are uniquely positioned to undertake these activities as they have collaborated extensively in the past 10 years resulting in numerous ground-breaking scientific discoveries.