A new universal language, based on data and real-world evidence, will transform behavioral health care and make reliance on trial and error a relic of the past
By Nawal Roy, Founder and CEO, Holmusk
Imagine for a moment that you don’t feel well. You grab your thermometer and a minute later you read the results: 101.2°F. You instantly know that this is higher than the average resting human body temperature of 98.6°, but not as concerning as a temperature of 106.3°. Based on this knowledge, you take action: swallow a dose of your fever-reducing medication of choice, drink some fluids, have a rest, and take your temperature again in a few hours.
Now imagine that you live in a world where medicine has not yet developed standards to measure, diagnose, and treat people based on changes in their body temperatures. In such a world, there would be no universally accepted set of endpoints, no understanding of the difference between 98.6° versus 106.3°, and no clear guidelines on what to do if you have a temperature of 101.2° or 106.3°. What would a temperature of 101.2° signify in such a world, and what actions would you take? Would you take 400mg of ibuprofen, rush to the E.R., or do nothing at all?
This scenario highlights the unfortunate reality of behavioral healthcare — we don’t yet know the real difference between a 98.6°, a 101.2°, or a 106.3°, or how to effectively treat each specific case. In the absence of a standardized measurement scale, scientific knowledge about behavioral health is in a deplorable state. Meanwhile, 970 million people globally suffer from mental illness and substance use disorders, resulting in an economic burden of $1 trillion annually. The increased prevalence of these conditions, along with the devastating impact on public health, the economy, and individual lives, underscores the urgent need to address behavioral health as one of the most pressing crises of our time.
I founded Holmusk with one overarching mission: to facilitate the use of objective knowledge to learn from each patient’s experience and to improve the next patient’s journey through the behavioral healthcare system. The first step is to create that measurement scale, and Holmusk is well on its way. We are building the world’s largest real-world evidence (RWE) platform for behavioral health. The goal is to use real-world data and digital solutions to become the gold standard for data and evidence on which everyone — clinicians, researchers, life sciences companies, and payers — relies to make decisions. By enabling measurement in behavioral healthcare, we will unlock deeper understanding, allow for clearer communication, and see better outcomes.
Trial & Error: Symptom of A Broken System
To understand how Holmusk is transforming the field of behavioral health, let’s take a closer look at the problem we seek to solve. The current mental health system is riddled with issues, perhaps none more pressing than the reliance on trial and error when it comes to treating mental health conditions.
The typical patient does not seek professional help for symptoms right away. In fact, patients often wait an average of 11 years between the onset of symptoms and their first treatment. Seeking out a mental health provider is only just the beginning of a long and often arduous process. A typical patient will go through months — even years — of trying numerous medications, at different dosages, before finding a course of treatment that effectively helps manage the condition.
The reasons for this are multifaceted and complex, including a limited understanding of many conditions, a lack of choice and innovation in medications, and a broken payment model. As a result, providers often have to rely solely on their own training, opinions, and clinical experiences — without benefiting from the vast knowledge which is contained within the experiences of countless other providers and patients who have traveled the same road.
In behavioral health, different treatments may be known, and may have been tried, but they are hidden within the individual experiences of patients and doctors. They are not part of a system in which they can be recorded, shared, and be readily available for review and research. That’s why data is so important.
The opinions of care providers are invaluable — but think of how much more powerful they would be if they were informed by a measurement-based system that allowed for much clearer determinations on treatment options. This would create a significant ripple effect in the way care is delivered, and begin to move opinions away from “me” toward “we,” from single-person decision making to a collectively informed effort.
Let’s go back to our fever scenario. What if your doctor, upon learning your temperature was 101.2°, first advised you to try taking a nap because that had seemingly worked for some of his previous patients? Then, as your fever climbed, he recommended a cool bath; later he might suggest you try a bowl of chicken soup plus two aspirin. What if it took days, weeks, months, or years for you and your doctor to land on the treatment that would effectively reduce your fever and restore your health?
We wouldn’t accept this reliance on trial and error as a standard part of treatment in other fields of medicine like oncology, obstetrics, anesthesiology, or cardiology. For those with behavioral health needs, trial and error has become a widely accepted part of the treatment journey. It doesn’t always need to be. Nor should it.
The Data Problem
Behavioral healthcare is facing a data problem, but it’s certainly not due to a shortage. In fact, data is ubiquitous in healthcare, as anyone who’s ever been to a doctor can attest — it’s in patient records, doctors' notes, insurance codes, test results, bills, and so on. However, the issue is that vast amounts of data are collected but rarely stored or organized in a way that is easily comprehensible and accessible. Therefore the power of data cannot be harnessed and the true benefit of data cannot emerge.
The lack of access to actionable data creates an avalanche of issues. For example, take the woefully limited selection of medications traditionally used to treat mental health conditions. Due to the high failure rate of clinical trials and widespread availability of generics for the most prevalent disorders, new therapeutics have been challenging to introduce. What is required is a comprehensive data system to guide research and generate the evidence necessary to break this dysfunctional cycle.
For example, there has been astoundingly little innovation in the drug classes for major depressive disorder, with the overwhelming majority of current medications targeting the serotonin pathway despite nearly one-third of patients failing to respond to these treatments. Additionally, even though an estimated 24 million people worldwide are affected by schizophrenia — that’s 1 in 300 — there are still no drugs available that can alleviate all symptoms and improve quality of life for all patients.
Without robust evidence, pharmaceutical companies cannot show the value of new therapeutics on patient outcomes required to justify the cost-benefit of new products. When new treatments are introduced, their usage is limited due to a scarcity of evidence. The lack of consensus on the best treatment regimens means there are insufficient evidence-based practices to guide clinicians. We must catch up to other fields of medicine and expand the range of available therapeutic options — and the data that tells the story of each patient's experience, starting with their electronic health record (EHR), will make this possible.
Cleaning 'Dirty' Data
Holmusk's mission is dependent on the wealth of information stored within the electronic health record, commonly referred to as clinical data. This data is essential for comprehending the intricate details of diseases, including their progression and outcomes. Unfortunately, most health systems fail to utilize this data, primarily due to the lack of accessibility. This is especially true for the field of behavioral health, where the potential for clinical data to advance patient care remains largely untapped. Roughly 40% of behavioral health data is objective: patient age, height, medication, dosages. This is structured data. The remaining 60% consists of subjective observations and provider opinions: How did the patient seem when they walked in? Did they report trouble sleeping or eating? Did their emotional state appear to have improved or declined since their last visit? This is known as unstructured data.
If you only have structured data, you will understand less than half of the patient’s profile. To better understand the patient’s condition, it is crucial to harness the unstructured data. If done successfully, we can combine new outputs with structured data and begin to build a truly comprehensive picture of the patient's health.
This is the biggest target in behavioral health — and Holmusk is the first company to be able to accomplish it. Think of Holmusk as a refinery that takes “dirty data” (unstructured and therefore unusable) and turns it into “clean data” that is meticulously curated by trained clinicians and made available for others to draw insights from in nearly infinite ways.
Each and every person who is treated for a behavioral health condition is a valuable source of data. Similarly, each of those individuals would also benefit from the valuable data of others who have been treated for similar conditions. When de-identified and aggregated with the experiences of millions of others, information on symptoms, diagnoses, treatments, and outcomes become fundamental building blocks that pave the way for a deeper understanding of behavioral health conditions.
In its current state, behavioral healthcare is somewhat comparable to trying to speak English before the existence of letters. At Holmusk, we are creating a universal language of standard measures that will enable us to move away from subjective opinions and fragmented systems, and towards a world where clinical research and clinical practice are in sync. With this new common language, the potential is limitless.
The journey to Holmusk’s inception traces back to my 20-year tenure in finance, starting at the Bombay Stock Exchange, where I traded in the pit using only my fingers, lips, and voice — the limited tools available to individual traders in making subjective decisions based on scant information. After I moved to the US in 1994, my fascination with financial markets led me to Wall Street, where I witnessed the transformative power of the Bloomberg Terminal.
This software system provided real-time analysis of financial market data, along with instant access to news, insights, and other crucial tools. With its ability to link collective data and facilitate objective real-time analysis, the Bloomberg Terminal enabled me to make informed decisions that were no longer based on guesswork, making it possible for me to trade without needing to be on the floor.
I've always been profoundly drawn to complex problems, but to devote my life’s work to them requires me to believe that they are solvable. There is no problem more complex, or more critically important, than our broken mental health system. It may not have been solvable in the past and with traditional methods, but I believe that with the power of data it can be. Once I learned of and acquired an invaluable asset, a vast dataset owned by Duke University with more than half a million de-identified patient profiles, I had a lightbulb moment. I knew that true innovation would be possible if I could organize this data and start building the equivalent of a Bloomberg Terminal for behavioral health.
What Holmusk Does
To facilitate this transition, we recognized the need to establish the foundation for future evidence and innovation by constructing the largest real-world evidence platform for behavioral health. Holmusk's flagship offering, NeuroBlu, is this evidence engine. Accessible through a single web-based platform for all behavioral health stakeholders to explore, learn, and innovate, NeuroBlu is powered by the world’s largest clinical behavioral health database of its kind. It contains over 20 years of research-grade, anonymized, longitudinal patient-level data including demographics, medications, and categorized clinician notes for over 1.4 million patients (and counting).
The Lancet Psychiatry recently published a peer-reviewed study showing that illness severity and instability are key factors in predicting future risk of hospitalization — a study that was made possible by NeuroBlu’s real-world data. This study was just one of the 44 scientific papers and presentations that Holmusk has produced in the last year. As our partners continue to find innovative ways to utilize our real-world data to solve pressing real-world problems and our database grows in size, we expect this number to increase exponentially. At Holmusk, we are committed to leveraging our cutting-edge technology to generate actionable insights that drive meaningful change in the behavioral health industry.
When NeuroBlu launched in 2020, its primary goal was to promote the use of real-world-evidence (RWE) for pharmaceutical development and drug commercialization. However, it has become clear that other stakeholders, such as doctors, insurance providers, start-ups, and policy makers, can derive just as much — or more — value from this evidence engine. We envision a future where NeuroBlu is an indispensable part of behavioral health research, innovation, and care delivery.
A Brighter Future
We live in a world where ChatGPT can write plays in Shakespeare’s style and cars can drive themselves, yet most players in the behavioral health care ecosystem don’t have access to troves of valuable data despite its constant collection. Holmusk is committed to fixing this.
Harnessing that data and turning it into real-world evidence will reduce the reliance on trial and error, enable better therapies, enhance our understanding of common conditions, and create a new fundamental architecture for mental healthcare delivery, payment, quality, and outcomes based on actual measurement.
We recognize that our partnerships are essential to achieving this goal. That’s why we have established strong collaborations with multinational organizations, including life science companies, payers, academic institutions, and government agencies. Our most recent partnership with Veradigm is an excellent example. By adding tens of millions of de-identified patient profiles to the NeuroBlu Database, this collaboration will significantly expand our RWD dataset, empowering us to achieve our mission of transforming mental healthcare.
To demonstrate the untapped potential of data, we can look at an organization like the Department of Veterans Affairs (VA), which cares for millions of people suffering from behavioral health conditions. With Holmusk’s support, every veteran’s experience and treatment could become part of a larger dataset that the VA could leverage to make a meaningful impact: For instance, what can experience tell us about which patients are likely to die by suicide one year from now? 30 days from now? How can we predict which patients will return to the hospital within 60 days of their release, and which ones will flourish? The VA’s own data, once “cleaned” and organized, could be unleashed to help understand patients in a new way and offer them better, more individualized treatment.
To improve behavioral health, we need to be able to quantitatively assess health outcomes and disease severity. The entire measurement industrial complex by which care currently gets delivered and paid is based on this fundamental architecture of measurement. This is absolutely crucial, as we move toward a value-based payment system, in understanding total cost of care and accurately developing risk stratification and risk adjustment models. However, due to the fractured nature of our healthcare system, this has been difficult to accomplish, especially in the behavioral health field, for two primary reasons: A lack of consistent measures across healthcare sites, and existing symptom measures are often subjective to the provider, leading to reduced consistency.
To mitigate this issue, we need to create standard quality behavioral health metrics with clearly defined scoring guidelines. These metrics should allow for valid assessment of both disease severity and outcomes — which we can use to analyze efficacy of treatment interventions and improve patient well-being. Without an objective, measurement-based system to assess whether doctors are prescribing the “right” or “wrong” therapies, we will continue to see claim rejections, or worse, the all-too-common reality of behavioral healthcare being limited or excluded from coverage.
At Holmusk, we understand that there is no magic bullet to eliminate behavioral health conditions. However, we believe the next best thing is to bring to the field an accessible superstructure of behavioral health knowledge in order to understand and develop the tools required to inform the improvement of care so many people need. Better evidence will drive value across the ecosystem. Standardized measures in behavioral healthcare will benefit patients suffering from anxiety, depression, bipolar disorder, OCD, PTSD, and other related conditions, making them as well-understood and treatable as a common fever.