Why We’re Bringing Precision to Mental Health

COVID-19 spotlighted the centrality of mental health to our overall well-being. For once, a bit of stigma slipped away during those months of social isolation. Waves of investment went to telehealth, and employers began including subscriptions to mindfulness apps and access to mental health services in their pandemic benefits packages.

After more than a decade in mental health tech, I was heartened by the trend. But at the same time, the tech-fueled innovation I see today seems all too focused on addressing downstream effects rather than root causes.

We’re hyper-focused on the supply side of mental health: increasing patient access to, and experience with, mental health services, while keeping clinicians afloat with their massive caseloads. It’s led to a rush of consumer-grade tech with risky business models - the very thing we’ve witnessed in other industries and have been clamoring for in healthcare.  

Careful What you Wish For

With the hyper-focus on the supply side, the motivation is to keep up demand, which in this case means sick patients.  Reducing the number of sick patients is, simply put, bad for business. So despite the gains in care access and patient experience, how good can we make that patient experience if, as the data shows, fewer than half of them actually get better?  

It’s cliche, but I’ll echo what’s been said many times before: the system is broken.   We’ve become all too accustomed to pointing fingers and assigning blame. We start by blaming the clinicians, saying they lack incentives to get patients better or are too lazy to accept insurance. In my experience, that couldn’t be further from the truth. The clinicians I speak with every day are individuals motivated by a genuine desire to help patients get better, so much so that they spend many years in exhausting and expensive graduate education (and if they take insurance, they will be reimbursed at a rate that does not make their practice viable).

Many blame the insurers, claiming that low reimbursement leads to fewer people accessing care. In truth, it’s a complicated relationship: those insurers that reimburse at reasonable rates inevitably start bleeding cash and are forced to raise premiums to cover the spread in a vicious catch-22. 

In my experience, I see a system where motivations are mostly aligned across clinicians, insurers, and patients. Of course patients want to get better. Of course clinicians want to get their patients better. And yes, even those insurers benefit from their patients getting better.

The system may be broken, but the energy is finally here. It’s time we shift that energy from exploiting a broken system to fixing it. 

Our Care Quality Problem

Let’s be blunt here: we have a care quality problem. If we look beyond care access, we’ll see that clinical care itself hasn’t changed much in decades. When a patient walks (or logs) in for their first appointment with a clinician, that clinician knows nothing about the patient except what the patient can tell them. Patients have 30 minutes to share all the important details.  Can you remember the name and dosage of every medication you’ve ever taken?  Every diagnosis you’ve ever had?  Every feature of your life story you may or may not realize could be contributing to your condition?  This is the essence of care quality problem #1: the sick patient is the single source of information for the clinician to make a diagnosis and treatment plan.

Problem #2 is, even if the patient succeeds in conveying all of their relevant information to the clinician, the foundation we have in place to make a diagnosis itself is fraught with issues. We arrive at a diagnosis of depression, for example, based on a vague, subjective multiple choice questionnaire developed by pharma a quarter century ago. 

There are more than 10,000 ways one could answer the questionnaire to qualify for a depression diagnosis - quite a diverse diagnostic bucket to say the least.  Yet it’s this same questionnaire that pharma uses to find and assess new treatments.  And it’s the same questionnaire that insurers use to decide coverage.  

How can we ever optimize care for these patients if we look at them all the same?  At the end of the day, the data-out is only as good as the data-in, and the data-in is way too imprecise to lead us to any meaningful advances in care quality.

There’s Hope

Oncology had the same problem 30 years ago.  Back then, people would get diagnosed with things like “breast cancer” and that broad diagnosis translated to broad treatments like chemo, radiation, surgery or some combination of those – blunt tools with relatively poor patient outcomes.  

Fast-forward to today and no one gets diagnosed with just breast cancer.  Instead, we see diagnoses like “ER+/HER2- stage 3 invasive ductal carcinoma." And pharma, insurers and clinicians have studied these more precise patient populations and developed much more precise ways to treat them.  The end result? Drastically improved patient outcomes.  

There’s a playbook here.  Our ability to leverage patient data to more precisely segment patient conditions allows us to treat more precisely.  In oncology, it was largely biological data that led this segmentation.  We’ve tried the same in mental health, but there’s too much noise when you look at biology absent other factors (remember, there’s at least 10,000 different ways to qualify for depression so it should be no surprise that the population is too noisy to find biomarkers).  Instead, it will take more precise symptom and environmental segmentation first to enable a biology-driven precision psychiatry approach. 

I co-founded Headlamp Health because I believe in our collective ability to bring data-driven patient stratification to psychiatry.  I believe it will ignite a renaissance in clinical quality akin to what we’ve witnessed in oncology. And, once you have a way to stratify mental health conditions, the rest works itself out. 

With more specific patient populations, insurers can more precisely and efficiently determine coverage,  enabling the economics for clinicians to return to accepting insurance. 

With more specific patient populations, we can reignite our search for biomarkers and more precise treatments. 

And the stigma we talk about? That dissipates with clarity and precision of diagnosis.

With the deep data+AI expertise of the team at super{set}, we’re building a better way to map patient history, find the right path forward, and continue until remission is achieved. In our model, data-driven segmentation of mental health diagnoses will enable the phase change we’re after.

Join Us

Want to join us on this journey? If you’re a mental health clinician or want to partner in the space, I want to hear from you at andrew@headlamp.com. If you’re an engineer or technologist ready to revolutionize the mental health space, check out our site at headlamp.com for more.

Andrew Marshak is a co-founder of Headlamp Health and a passionate voice in the precision psychiatry movement. Previously, Andrew headed Product for Myriad Neuroscience, where he helped develop, launch, and grow GeneSight – the market-leading pharmacogenomic test in mental health.