Echidna Health: medical AI chatbot & blockchain incentives

Echidna Health accelerates healthcare and empowers patients. It interviews patients, offers clinical decision support, and distributes outcome-based blockchain healthcare contracts.

There's a shortage of healthcare resources and misalignment of healthcare incentives. Patients have difficulty accessing qualified service. Physicians in developing countries are often underpaid and are incentivized to maximize the volume rather than the quality of care. Echidna Health addresses all these issues.

It consists of 2 independent systems: a health chatbot (prototype available here) and blockchain contracts (conceptual stage).

First, upon needing medical service, a patient would message the chatbot with her or his diseases or symptoms. Using a medical knowledge base and natural language processing, the chatbot would gather relevant details, asking some of the same questions a physician would ask. The info is summarized into a clinical note and distributed to physicians, along with differential diagnoses and/or treatment recommendations.

Along with the clinical note, a blockchain backed outcome based contract would be generated for the patient and providers in his/her locale. For example, for a child presenting with asthma exacerbation status post respiratory infection, the contract would depend on treatment outcomes. In this case, the contract can award to the physician 25% of the patient's payment upfront, 50% in a week contingent on resolution of the asthma exacerbation, and 25% contingent in two months contingent upon good control of underlying asthma. 

Based on the preliminary note, providers can gauge case complexity and put in quotes. Pro bono advice can even be offered to the patient. The patient is then empowered to choose the provider offering reasonable price and with qualified credentials. Payment is made via the blockchain when the patient pays a physical visit, which is much more efficient given the preliminary clinical note compiled by the chatbot.

To enforce the contract, the patient would complete weekly review of systems questionnaires. A small portion of payment can be given back to the patient for questionnaire completion. The results are then used to determine treatment outcome and thus physician payment. 

This platform would change the world.

The AI chatbot directly speeds up patient care by automating much of the information gathering and note writing. It also makes healthcare more accessible; more people have smartphones than have clean water. Complementary decision support can help providers follow evidence based guidelines and consider more relevant diagnoses. The blockchain smart health contracts would compel providers to maximize patient outcome rather than service volume. The healthcare market is highly inefficient because of the asymmetry of leverage and information between counter parties. By aligning provider incentives with the patients', efficiency is vastly improved.

Elevator pitch video: I was having technical difficulties uploading my pitch video in time, but it will get uploaded eventually to YouTube under the title "Echidna Health MIT Solve pitch" You may access it via search. 

We join just a handful of startups applying chatbots to patient interviews. However our chatbot is much more lean, transparent, and scalable in terms of medical knowledge. In fact I prototyped it in only 3 months part-time to compile a medical knowledge base of relationships between diseases, symptoms and metrics on which a chat crawler is run.

In terms of blockchain, there's so much talk on using blockchain to structure health records or track drugs, but we're innovative in proposing its usage for incentivizing outcome based care. An implementation in the field would be a game changer. 

Thanks, we already addressed this issue in previous questions :)

Already accomplished - prototype of patient interview chatbot

Goals

• Add decision support for most primary care conditions to chatbot
• Make chatbot multilingual via NLP
• Streamline chatbot UI, also deploy to WeChat, WhatsApp, SMS
• Field test chatbot 
• Formulate outcome based contract rules for most common conditions
• Prototype blockchain architecture for implementing these contracts 

The chatbot is deployable as a service in just a few months. It'll be free for patients; a small fee maybe charged to physicians who get referred via the chatbot.  I have no costs to recoup and no investors to repay; the operating expenses of the server is minimal.

The blockchain contracts on the other hand require significant capital to validate and scale. I'm here just to propose the architecture.

Our chatbot will be deployed on web, SMS, WeChat, WhatsApp, Facebook. We're using a minimalist, low-data, multilingual messaging interface intuitive and accessible to anyone with a phone. 

None so far. We currently have a chatbot prototype and expect to field test in US and China Q4 2018.

Thanks, please refer to main project description for how patients and physicians may be affected. The chatbot takes only server power to scale and can serve hundreds of millions in a few years. The blockchain proposal is much more capital intensive, and it's too immature to offer an estimate. 

I developed the venture in my personal time and am looking forward to building a team. I'm a Stanford trained electrical engineer and an IU medical student (on leave). I've previously brought a medical hardware (an anti-pollution wearable air filter) project from concept to product, so I'm familiar with the pains and intricacies of breathing life into a novel idea. I wrote the software for the chatbot prototype, and I know clinical workflows as a medical student.

A mix of SaaS and blockchain commission

The chatbot alone offers a time saving service to physicians who can request their patients to use it before the visit. Healthcare facilities can subscribe to it via b2b SaaS. Chatbot can also be sold to EHR companies for integration. 

When chatbot and smart contracts are run as a platform, then we can simply charge a commission on the patient payment. Like "Uber for healthcare," but with transparency of the blockchain and value added from the chatbot. In locales more open to adoption of cryptocurrency (eg EU, Russia, parts of Africa etc), we can bind also bind payment to a tradable token and finance via an ICO.

I built a promising prototype on my own and want to gain exposure, get feedback, and find collaborators through Solve. 

I've little experience in scaling software or services, so Solve will be valuable in helping me partner with those with that expertise.