Healthcare matters – and the current system is terrible at using data to improve care. Brad and Craig (and all of the physicians we have interviewed and surveyed) have firsthand experience of the overwhelming amount of data doctors face each day, and how poorly it is used as a result. The idea for Herald came directly out of their experience as doctors, and is something they want to use personally. All of us are passionate about healthcare. Matt worked for 5 years in the business arm of the largest health system in New York and understands our customers. Andrew does research in healthcare economics and likes using data to solve big problems.
What it does
Physicians currently spend about 12% of their time reviewing test results and records--roughly the same amount of time spent face-to-face with patients! Much of this time is spent on manual, tedious, repetitious "pulls" from the EMR. Doctors often handwrite to-do lists to check for test results.
Our product introduces (1) a push model (2) customizability and (3) a crowd-sourced model for how physicians receive important updates from clinical data. Some administrators in large hospitals have created alert systems to ensure physicians see important updates, but these are clunky and one-size-fits all solutions (most doctors simply ignore these alerts). Finally, in lieu of a crowd-sourced model, doctors and administrators tend to have ad hoc rules about notifications, sacrificing quality of care.
Herald will significantly reduce the 12% by delivering exactly the right clinical information in real-time.
Challenges we ran into
1) Identifying a pilot site for Herald has been a major challenge, and it underscored the cost and difficulty of one-off negotiations with major institutions such as hospitals and health systems. For this reason, we have been drawn to platforms such as MDP solutions, where we can get assistance with distribution and the ability to reach a much broader base than we could otherwise.
2) Another challenge has been balancing ease of integration with our pilot site with the desire to build something scalable for the future. One of our major goals is to build an interoperable product that can easily be installed at sites running different EMRs, but this adds a lot of complexity to the current phase, in which we are trying to validate product market fit with an initial set of users.
How we built it
The first iteration of Herald was nothing more than a concept on a piece of paper that we started presenting to doctors for feedback. Once we became more confident in our product approach, we created a clickable prototype for users to experiment. We were amazed at how much user research we were able to accomplish with these very lean methods.
We've now translated into a full-scale product development mode, and we're fortunate to have three of four founding members with deep programming experience. Craig leads the product development and assigns tasks in a weekly sprint model. We are developing on the MEAN stack, use JIRA for product management, and have been diligent around properly modularizing our code with unit tests throughout in order to avoid significant code debt as we scale up Herald. We're proud of the development processes we've put in place and believe this will be a strength as we reach a live pilot and beyond.
Accomplishments that we're proud of
The Herald team came together last fall. We started with an idea, which through intensive user research evolved into a clearer vision of the product. We organized several focus groups among doctors, and then conducted visits to outpatient and inpatient settings to develop use cases and learn more about clinician needs. In November, we competed in the Brigham and Women’s Hospital Hackathon, where we developed our first prototype of the product and won the competition.
Coming out of the hackathon, we pursued three tasks. The first was to find customers. A major local academic medical center was interested in having our product live in the hospital, and after several rounds of negotiation we have agreed in principle to the terms of a pilot. In the meantime, we have begun reaching out to other providers and built a website (http://www.heraldhealth.com) to provide potential customers with more information about the product.
The second task was to further develop our prototype into a full MVP that could be deployed in the pilot. Through December and January, we restructured our front-end web interface and held meetings with technical teams familiar with the back end services available at our partner hospital.
The final task was to begin formally building as an organization, including incorporation in January. We’ve also participated in a number of area competitions. We are a finalist in the Harvard Deans' Health & Life Sciences Challenge and the HBS Rock Accelerator, a semifinalist in the MIT 100K Accelerate challenge, a finalist in the MIT Sloan Healthcare Case competition and were selected into Harvard innovation lab’s Venture Incubation Program.
What's next for Herald Health
We are each planning to work on Herald Health over the coming summer and have applied to a number of accelerators. We will be spending most of our time (and any funds raised) doing two things: 1) ensuring that we are ready for a successful pilot and 2) beginning our efforts to commercialize Herald.