Redefining Measurement-Based Care: Designing a System for Evidence-Based Care that Clinicians and Patients Will Actually Use

A defining limitation in the diagnosis and treatment of behavioral health disorders is the absence of clear biomarkers. Other illnesses have markers that are more directly able to identify and monitor therapies. Treatment effectiveness for a person with diabetes can be objectively measured by testing blood glucose levels. Hypertension treatment is informed by monitoring blood pressure readings. A broken bone can be visualized on an X-ray.

 

Unlike diabetes or a broken bone, however, there is no panel of lab tests or imaging alone that can definitively point to anxiety, depression, or autism. Instead, we diagnose and monitor behavioral health conditions using interviews and a collection of standardized measures, which are targeted questionnaires that screen and monitor symptoms of specific disorders. Administered at regular intervals, the standardized measures form the basis for measurement-based care (MBC), which is the evidence-based practice of using patient-reported progress data to inform treatment decisions and engage patients in their recovery journeys. 

 

Measurement-based care is the current gold standard, endorsed by organizations including the American Psychiatric Association and the American Psychological Association. Initially developed for use in clinical trials, these measures provide essential baselines to demonstrate drug and treatment efficacy. When applied consistently, MBC is associated with 40% to 60% improvement in clinical outcomes.

 

Still, only 46% of behavioral health care providers in a recent study reported using MBC with at least half their patients. We may understand why: Standardized measures are valuable, but they don’t go far enough to deliver meaningful change in behavioral health treatment. We are currently using standardized measures in psychiatry as if they were objective biomarkers, like those we use in physical medicine. In fact, we can get closer to effective biomarkers for behavioral health by looking at granular, symptom-level, and psychosocial data. To truly improve behavioral health outcomes, we must expand MBC to include the collection and measurement of symptom-level, psychosocial, and functional status data.

 

The Limitations of MBC

Before we attempt to build upon standardized measures, it is helpful to explore why they do not suffice as biomarkers.

 

Away from the controlled setting of a clinical trial, standardized measures do not meet the full definition of MBC. Too often, what we are calling MBC is really quality assurance. Standardized measures set baselines and track outcomes, but they are much less effective at informing treatment decisions and engaging patients. In fact, low perceived clinical utility is the barrier most strongly associated with less frequent use of MBC.

 

In clinical practice, patients resist the questionnaires. Returning to the lab panel metaphor, imagine if each individual blood test required a separate blood draw. A system that required patients to frequently submit to dozens of needle sticks would be challenging.

 

This is analogous to our current questionnaire-based measurement system. There’s a different questionnaire for each disorder—ADHD, depression, anxiety, etc.—and progress is measured by repeating the tests every few months. In practice, patients balk at answering dozens of questions, many of which are repetitive across questionnaires and ask about challenges they don’t have. As a result, clinicians reduce the number of standardized measurements given or abandon MBC altogether. Additionally, much can happen clinically in the time between the repetition of the assessments. Is a static snapshot taken every few months really providing clinicians with an accurate view of patient progress?

 

Further, clinicians recognize that standardized measures themselves offer limited insights. Current assessments measure the presence and frequency of symptoms with broad brush strokes: “always,” “never,” “sometimes.” However, these metrics are incredibly subjective. How often is “sometimes”? A weekly symptom may be manageable for one person but debilitating for another. We must track frequency in a more granular way, assessing how much each symptom impacts the individual’s quality of life, ability to function, and level of medical intervention required.

We must be tracking more symptoms, as well. A person’s behavioral health is, in fact, a detailed picture painted with numerous mental, physical, and environmental symptoms and factors: battles with a sibling, chronic stomach distress, inability to sit still in class. Those individual factors should also be considered in diagnosis and treatment plans, and making progress on individual symptoms could be quite meaningful in terms of patient outcomes.

 

To overcome the inherent limitations of MBC, we must not only collect and measure granular, symptom-level and environmental data, but we must actively use it as a mediator to improve treatment. Advances in cloud computation, analytic speed, and AI have created extraordinary opportunity to connect disparate systems and compile clinical patient data and electronic medical records in a secure, HIPAA-compliant manner. However, this granular, symptom-level behavioral health data has thus far not been part of the dataset.

 

Redefining Data Collection

Our mission is to change that. The Trayt technology platform was designed to support this new approach to data collection and analysis in day-to-day clinical practice. Standardized measures are incorporated and serve as a baseline for assessment, as is relevant data from the electronic medical record. However, we are also tracking and measuring from a set of 750 possible individual symptoms and factors self-reported onto the platform by patients or contributed by members of the patient’s care ecosystem. Leveraging advanced analytics, the platform has the capability to process this granular data to create a comprehensive 360-degree patient profile with a unique set of meaningful symptoms to measure and track on an ongoing basis. 

 

What do we mean by granular data? The following types of data are critical to include:

 

·       Between-visit symptom and experience data. Through a patient- and caregiver-facing application, we collect real-time insights from everyone in the patient’s ecosystem, such as parents, school counselors, and the patients themselves. Therapists can see day to day whether a student is, say, sitting still for longer periods of time or completing homework assignments more effectively. They can track dips in progress and look for correlations to explain them. Feedback from primary care physicians and other medical professionals is incorporated as well so that therapists have a complete view from which to make more accurate diagnoses and treatment plans.

 

·       Social Determinants of Health (SDOH) data. Factors such as housing insecurity, neighborhood food deserts, lack of reliable transportation or internet access, and community safety all significantly impact mental health and treatment success. When therapists can see this data as part of the overall profile, they can incorporate SDOH into diagnoses and treatment plans, connecting individuals to appropriate social services and removing barriers to better health.

 

·       Social Determinants of Mental Health (SDOMH) data. Traditional SDOH are broad building blocks that may impact mental health, and they affect everyone in a neighborhood or community in much the same way. We must also collect and measure what we refer to as the Social Determinants of Mental Health, which are the nonmedical drivers of health that combine to create an individual’s unique experience. Included in these are Adverse Childhood Experiences (ACEs), family dynamics, and factors such as cultural attitudes toward therapy.

 

Meaningful Improvement in Patient Outcomes

With a personalized set of symptoms to measure, clinicians can target granular issues and confirm with precision whether the patients are indeed improving.

 

The standardized measures can be insufficient in this regard. For example, an individual’s overall score for depression as measured by the PHQ-9 assessment may be improving, but other issues may be causing her anxiety to increase. If we are not also administering the HAM-A, it may go unnoticed until it reaches a crisis point.

 

A granular data-driven approach turns this paradigm on its head. Instead of beginning with a diagnosis and looking for symptoms, we start with symptoms and map them to more precise diagnoses. Clinicians then monitor those individual symptoms on an ongoing basis and use the results to inform treatment.

 

The Trayt software platform has primarily been deployed in state Psychiatry Access Programs, which expand access to behavioral health care by connecting pediatricians, OB/GYNs, family practitioners, and other primary care clinicians to remote psychiatry support at major state institutions. Every institution on the Trayt platform can use symptom-level data to inform better clinical decisions. Clinicians can track not just whether a PHQ-9 score is improving, but whether the patient is getting to work on time or getting better grades at school at school. If those symptoms are improving, they might have an indication that anxiety and depression are improving.

 

If the psychiatric community begins to embrace symptom-level, psychosocial, and environmental data as aggregate markers for behavioral health conditions, we fill in around those broad brush strokes and arrive at a much more precise definition of measurement-based care. Only then are we more effectively measuring patient outcomes on a wider scale.