Machine Learning Series: Reinventing Psychiatry

This article is part of the FreeV Spotlight. Subscribe for monthly highlights of our portfolio and Berkeley startup activity!

Almost 1/5 of people in the United States currently have mental health issues. Over 1/2 of adults with a mental illness don’t receive treatment. Demand for psychiatrists is increasing every year but patients are faced with exorbitant fees and long wait times.

Innovation in psychiatry has empirically been slow, but two students from UC Berkeley are set on reinventing the patient journey.

Matt and Doug co-founded Foresight Mental Health (Batch VIII), a mental health and wellness clinic in Berkeley. They create customized genetic profiles through genetic testing, symptom profiles, and brain-chemical balance estimates. They factor in mental health history, family history, and lifestyle as well. The Foresight Platform analyzes clinical research and the patient profile to assist your doctor in accurately identifying the optimal treatment. Weekly surveys and practice-issued wearable technology monitor patient health to ensure treatment is effective.

(Co-Founders Matthew Milford & Doug Hapemen)

We recently sat down with Matt and asked him about why he started the company, the tech behind it, and his journey on campus.


Part 1: The Idea 💡

Q: What are current problems with patient care in the mental health space?

Matt: “There’s a lot; it’s really a mess. Let’s start by looking at a treatment. There are currently 30 different anti-depressants. Sometimes they work and other times they don’t — and it takes a month to tell if it is even effective. You can imagine a psychiatrist nodding as you talk about your life and sending you out with a script saying ‘see you next month’. You have no idea if it’s going to work; neither do most psychiatrists because of the degree of trial and error psychiatry is forced to use right now. We’re trying to leverage data to improve this process. It is also very difficult to track side effects because of the lack of monitoring done on patients outside of the office. The prescribing method is very detached where psychiatrists check in on patients once a month and hastily prescribe medication based on surface level interactions. There is no effective method to regularly track results.”

Q: Why are you interested in improving mental health?

Matt: “My mom was a doctor so I was always interested in medicine. I didn’t want to be a doctor but I wanted to use engineering in healthcare. It feels like one of the slowest fields to adopt technology — but one of the best for it. It’s so inefficient right now. A field ripe for tech disruption. Some of my closest friends have been afflicted by mental illnesses. It’s been a personal mission to try and fix this problem.”


(Clinic Font Desk)

Part 2: Tech Advancements

Q: How do you provide specific data driven solutions to patients?

Matt: “Remember, right now psychiatrists are often just picking medications out of a hat when determining which medication is best. We spent a lot of time tweaking our algorithm. We interviewed close to 200 psychiatrists trying to figure out how they make prescribing decisions. We built our model on medical guidelines and a database of information each medication. The data includes different factors that influence it, genetics, neurotransmitter estimates, biometrics, etc. Our software can sift through this in real-time. Next, we’re going to integrate a machine learning model into our precision medicine platform to add predictive power as we collect more patient data.”

Q: How do you collect data to test your algorithm?

Matt: “We realized that working with psychiatrists in practice today would not be helpful. They’re getting paid $400–1k per hour and demand for them will always be high. There’s little incentive to get better at their job. We built an actual clinic to not only fund engineering development, but also data collection and testing. Having real-time data on patients is the best scenario. We collect data through diagnosis, treatment, and monitoring daily with real patients. We have reputable psychiatrists and therapists that treat patients at our office every day.”

Q: Where do you see your product moving forward?

Matt: “Imagine as a patient, before you even enter the building, you are diagnosed through an AI-driven chat bot. Your treatment plan is already constructed with sections dedicated to medications and therapy, in addition to exercise, sleep, and diet. When you walk int the office for the first time, you are given a genetic test and your doctor discusses the plan with you for treatment and monitoring.”


Part 3: Campus Journey

Q: What resources did you use on campus?

Matt: “When first starting the company, we went to many different lab directors and professors in business, biotech, genetics, and medicine trying to learn as much as we could. Through this, we came up with a good way to monetize our idea. After this, we came to Free Ventures and met many helpful people while talking to mentors and investors. After hotdesking at SkyDeck for almost a year, we met Ken Singer and Stephen Torres from SCET. The SCET has been extremely helpful to us with advice and their network.”

(Clinic Lobby)

Q: Why did you decide to pivot your idea after starting the company?

Matt: “Finding product market fit is extremely tough. Initially, our company was a genetic testing company that would tell patients what medications would give them adverse effects. We kept developing new tech that would drastically change things but we saw a huge resistance to change. We realized 2 things: 1) Our initial solution was too niche and didn’t solve the root cause of why the industry wasn’t improving. 2) Every step of patient care had to be reinvented, which is why we started the clinic. ”

Q: Why build a mental health center in Berkeley?

Matt: “Having gone to Berkeley and seeing all the anxiety from students, it’s been great to be able to give back and help out the Berkeley community. We analyzed 10 locations in California and Berkeley was the best. We looked at psychiatrist to population ratios, population density, income levels, incidences of mental health issues, and proximity to potential patients. We decided to build our innovation lab in Berkeley and will be opening our flagship location in 9–12 months most likely in San Jose.”


If you’d like to personally get in touch with Matt, feel free to reach out to me at pratik@freeventures.org and I can help set up an introduction.

Visit the Free Ventures website to learn more about our current and past batches, and subscribe to FreeV Spotlight for monthly updates and articles about our amazing founders!

The FreeV Spotlight and article were written and edited by Pratik Bhat, from the FreeV Internal Team.