When you hear hoofbeats, sometimes it really is a zebra.

An enormous amount of research and EHR patient data already exists. Predictive algorithms are built from it, published in academic journals, and then never used in clinical practice. That was the fate awaiting our own algorithm for the prediction of Acute Hepatic Porphyria. After the study was published, we asked ourselves: what happens next? All this work, for what?

So we founded zebraMD. The backbone is a platform that pulls rare disease data and literature into one place, connects it to EHR systems and patient portals, and uses generative AI to look through millions of sources at once, connecting the dots between papers, genome databases, and individual patient records to produce clinically actionable insights.

Beyond our own models, we host and update predictive algorithms built by other physicians, retrofitting them to raw EHR data so they can finally be used in practice, and improving them over time with the new data they encounter. For physicians, zebraMD works with all EHRs through the national Health Exchange. Patients can directly connect their own patient portals such as MyChart, upload medical records, and link wearables and fitness and health apps. A manual symptom search offers a quick view of potential differential diagnoses.

Together this creates a 360 degree view of the patient, putting together the pieces clinical medicine has traditionally neglected: home wearable data, patient journals, videos and photos, Google and Apple Health data, fitness and meal trackers, environmental exposures and public health data, occupational health risks, and caregiver observations and care logs.

Get people diagnosed sooner, and get them the care they need, wherever they are.

A patient should never be disadvantaged because specialty care is out of reach due to geography, cost, or long wait times. zebraMD delivers specialty-level insight tailored to whatever department the patient is seen in, automatically, at the point of care.

zebraMD empowers patients to take control of their own health (controversial, we know). Patients can use zebraMD just like physicians can: connect their medical records and put our algorithms to work on their own rare disease questions.

Our original predictive model for Acute Hepatic Porphyria grew out of an academic collaboration between UCLA and UCSF. It is public domain, and we invite everyone to find the errors and iron out the kinks. Every model we develop will also be public domain, free to use for patients and providers, forever.

Academia moves faster with industry expertise and resources. Anyone who has attempted a multicenter research project knows the struggle and red tape. This work needs as many data sources as possible to build unbiased algorithms that work in any patient population and any EHR structure, so in 2023 we founded zebraMD Inc. to take on funding and partnerships and move far faster than our daytime roles as academic physicians would allow. zebraMD is a Public Benefit Corporation, not just a C-corp: our mission is written into our legal charter. Incorporating was also the only path to funding actual clinical integration, since typical grants will not pay for implementation.

We are proudly supported by the National Science Foundation through its Small Business Innovation Research program.

We are always transparent about our funding, contracts, data sources, and outputs. The original AHP model was developed with NIH research funding and industry sponsorship by Alnylam Pharmaceuticals through UCSF and UCLA. The zebraMD platform and its implementation into clinical practice are supported by Alnylam, our partner since the original project in 2020, and by the National Science Foundation. New predictive algorithms are funded through NSF grants, and we pursue disease foundation funding for each individual disease.

Grants alone are too slow when millions of rare disease patients remain undiagnosed every day, so we also partner with orphan drug companies developing therapies for rare and genetic disease. The interest is mutual: we want patients diagnosed and managed appropriately, and partners gain deidentified analyses of therapy efficacy and safety in real-world care, outside sterile clinical trials. In the future we may partner with the Centers for Medicare and Medicaid Services to deploy our algorithms nationwide.

We are also developing a shared revenue model with our patients. We inform patients of studies, collaborations, and opportunities where they can choose to contribute their anonymized, deidentified data and receive a share of the revenue from that contract. We tell patients exactly what is being done with their data, and we only take on contracts that benefit our patients, such as finding rare disease patients for a clinical trial or a cohort study that advances earlier diagnosis and new therapies. Patients choose what to participate in, stay in charge of their own data, and benefit from it both financially and health-wise.

Know of a suitable grant or partnership? We would love to hear from you.

We never share patient data with anyone unless the patient explicitly asks us to, such as sharing with another physician or health system. Deidentified analytics, combined with existing NIH research databases, power algorithms whose only purpose is to benefit patients directly. Patients can withdraw research consent at any time and can delete their account at any time, and all their data is deleted with it. No patient data is stored on our platform: clinical use happens ad hoc when a record is uploaded or a portal is connected. Full details are in our privacy policy.

Concerned about genetic testing? The US GINA law (Genetic Information Nondiscrimination Act) prevents insurance discrimination based on genetic test results. Physicians: every physician can order a genetic test, Medicare parts A and B cover testing and counseling, and panels from companies like Illumina and Invitae include free genetic counseling to help interpret findings. Genetic testing is no different from any other part of the workup. Just because it is difficult does not mean we should not do it.

Our secret sauce is an end-to-end app that can actually be used in clinical practice: a cloud platform with proprietary technology connecting any EHR, medical records, and research databases to retrofitted, self-learning predictive and management algorithms. Every point-of-care recommendation carries academic references so the reader can see exactly where it came from. With this technology we are building the largest rare disease co-op in the world.