Skip to content

Priorities for Strategy 3.2

Updated: January 2019

Develop ways to tailor existing and new interventions to optimize outcomes

Clinical trials have traditionally focused on diagnostic status and symptom severity. Inattention to the complex topography of treatment targets, broadly defined, and to individual differences in psychopathology and treatment, can limit the explanatory value of findings and their potential uptake in clinical practice. Two related priorities are important to addressing these issues.

First, more precise measures are needed for targeted intervention strategies to succeed. NIMH encourages applications aimed at developing more thorough assessment of clinical outcomes – particularly those that can enrich understanding of treatment efficacy and individual variation in response, and that might have utility as surrogate markers for more rapidly identifying potential responders. Assessments could include biological measures and markers, behavioral measures, and psychometric instruments. Studies that seek to optimize interventions for different age groups or disease stages, for different subgroups within diagnoses (e.g., biotypes related to specific biomarkers that are predictive of differential response), for trans-diagnostic clinical phenomena, or to identify moderators that help to account for demonstrated disparities in treatment outcomes associated with group membership (e.g., ethnicity, race, gender), are encouraged. To be responsive to this objective, applications must characterize participants according to individual variables that mediate/moderate treatment response, and not solely according to group membership.

Second, personalization of mental health care means that patients should receive treatment that is optimally matched to their individual needs. Optimizing interventions for individuals will not only require consideration of the individual’s clinical presentation, but also a broader consideration of facets concerning the individual, such as genetic, environmental, developmental, and cultural factors and, as data become available, other RDoC domains. Ultimately, precise, personalized care will also require consideration of the characteristics of the candidate interventions themselves and characteristics of providers and settings. NIMH encourages efficient research designs that can be used to examine prescriptive approaches for matching individuals to optimal care. To be responsive to this objective, studies must take into account mediators/moderators of treatment response that have implications for personalizing care; studies that involve randomization of large samples in pursuit of incremental gains in effect sizes are not a high priority.

Research Priorities

  1. Develop valid and innovative biomarkers to detect subgroups of individuals sharing common etiologies – whether within or across traditional diagnostic categories – as well as aspects of emotion, cognition, and social behavior that predict clinical response.

    Priority areas include:

    1. Developing psychometrically sophisticated assessments that provide sensitive, quantitative, and objective measures of specific domains of function for use in RDoC-informed treatment research. These measures should be useful across the full range (typical to atypical) of the dimension and may include adaptation of measures originally developed for basic research for use in clinical studies.
    2. Developing developmentally sensitive, quantitative, and objective measures of specific domains of function for use in RDoC-informed treatment research in children and adolescents.
    3. Developing and validating objective methods for stratifying subjects to optimally match individuals to interventions.  This may include age-appropriate biological measures, neurocognitive measures, and deep phenotyping via passively sensed naturalistic behaviors as well as actively sampled ecologically valid measures.
    4. Developing and validating biomarker(s) which can meaningfully inform clinical response to different treatments. It will be important to evaluate not just whether the biomarker(s) predicts response to a single treatment, but whether the biomarker(s) can inform selection between two or more interventions. This may include objective measures that predict the likelihood of responding or not responding to a targeted intervention.

  2. Foster personalized interventions and strategies for sequencing or combining existing and novel interventions that are optimal for specific phases of disease progression (e.g., prodromal, initial-onset, chronic), different stages of development (e.g., early childhood, adolescence, adulthood, late life), and other individual characteristics.

    Priority areas include:

    1. Developing novel therapeutic agents and approaches that have effectiveness at specific, targeted stages of CNS development and aging.
    2. Developing context-dependent interventions that target core mechanisms, including those that are common across mental illnesses, related to non-specific risk states as well as the prodromal, recent onset, and chronic stages of mental illness, including interventions that target modifiable risk factors in the service of preventing downstream morbidity and associated disability.
    3. Studying behavioral plasticity, alone or in conjunction with therapeutics, designed to address neurodevelopmental abnormalities by targeting identifiable mechanisms pertinent to risk, etiology, or maintenance of mental illnesses.
    4. Establishing the safety and efficacy of therapeutic interventions developed for adult populations in children and the elderly, as well as women at various phases of the reproductive cycle, while testing targets and target engagement specific to these populations.
    5. Analyzing heterogeneity of treatment effects in existing clinical trial datasets to identify predictors/moderators of treatment response that can be used to construct algorithms for initial treatment selection or sequencing.
    6. Conducting prospective studies to identify moderator variables and objective biomarkers, composite biomarkers, and/or multi-modality derived “biotypes,” (i.e., a group of individuals with the same genotype, including “digital phenotypes”–the trail of health information created through interactions with the internet, social media, and other technologies), that can be used to characterize young and adult patients more likely to benefit from a specific treatment modality, such as pharmacological intervention, device-based somatic treatments, specific psychotherapies, combination treatments, or augmentation strategies.
    7. Developing multi-modal intervention strategies that combine the simultaneous application of pharmacological, psychosocial, biologic, and/or neuromodulation interventions to selectively access specific therapeutic targets which are engaged through synergistic action across modalities.  This may include simultaneous circuit engagement via a cognitive intervention during targeted plasticity enhancement via pharmacological or neuromodulation interventions.

  3. Develop and refine alternative research designs and analytic approaches that can be used to test precise interventions.

    Priority areas include:

    1. Constructing integrated datasets with data from multiple trials collected in a manner to promote sharing and integration (e.g., using common data elements to capture clinical information, biospecimens, biomarker data, temporally dense behavioral capture from remote sensors) to promote reanalysis or meta-analyses and identification of moderators or tailoring variables for precise, prescriptive intervention approaches. Wide sharing of such data through the NIMH Data Archive is expected; registering with the Agency for Healthcare Research and Quality’s Registry of Patient Registries is encouraged when appropriate.
    2. Developing and/or employing alternative dynamic designs that maximize benefit to participants and arrive at definitive outcomes in an efficient manner (e.g., use of interim analyses to adjust randomization schemes or sample sizes to expedite conclusions regarding efficacy or safety, application of Bayesian approaches).
    3. Utilizing innovative computation approaches (e.g., machine learning, artificial intelligence, pattern classification techniques, predictive analytics) that can be applied to multiple streams of data (e.g., routinely collected electronic health records, sensor-based data, social media/device use metrics) to inform targets and timing for interventions and to facilitate clinical decision-making (e.g., gauging clinical deterioration or response to treatment).
    4. Creating designs and conducting studies that prospectively incorporate tailoring variables (e.g., clinical data, biomarkers, behavioral markers derived from passive sensing of naturalistic behaviors, patient response history) into participant assignment algorithms and employ sequential randomization and iterative evaluation of treatment effects to test prescriptive intervention approaches (e.g., MOST/SMART adaptive treatment designs).
    5. Apply novel trial designs to predict the functional impact of treatments on neurodevelopmental trajectories and long-term function.