There is a need for an electronic health record (EHR) system that is secure, accessible, and user-friendly. Currently, hundred of EHRs exist and different clinical practices may use different systems. If a patient requires an emergency visit to a certain physician, the physician may be unable to access important records and patient information efficiently, requiring extra time and resources that strain the healthcare system. This is especially true for patients traveling abroad where doctors from different countries may be unable to access a centralized healthcare database in another.

In addition, there is a strong potential to utilize the data available for improved analytics. In a clinical consultation, patient description of symptoms may be ambiguous and doctors often want to monitor the patient's symptoms for an extended period. With limited resources, this is impossible outside of an acute care unit in a hospital. As access to the internet is becoming increasingly widespread, patients may be able to self-report certain symptoms through a web portal if such an EHR exists. With a large amount of patient data, artificial intelligence techniques can be used to analyze the similarity of patients to predict certain outcomes before adverse events happen such that intervention can occur timely.

What it does

myHealthTech is a block-chain EHR system that has a user-friendly interface for patients and health care providers to record patient information such as clinical visitation history, lab test results, and self-reporting records from the patient. The system is a web application that is accessible from any end user that is approved by the patient. Thus, doctors in different clinics can access essential information in an efficient manner. With the block-chain architecture compared to traditional databases, patient data is stored securely and anonymously in a decentralized manner such that third parties cannot access the encrypted information.

Artificial intelligence methods are used to analyze patient data for prognostication of adverse events. For instance, a patient's reported mood scores are compared to a database of similar patients that have resulted in self-harm, and myHealthTech will compute a probability that the patient will trend towards a self-harm event. This allows healthcare providers to monitor and intervene if an adverse event is predicted.

How we built it

The block-chain EHR architecture was written in solidity, truffle, testRPC, and remix. The web interface was written in HTML5, CSS3, and JavaScript. The artificial intelligence predictive behavior engine was written in python.

Challenges we ran into

The greatest challenge was integrating the back-end and front-end components. We had challenges linking smart contracts to the web UI and executing the artificial intelligence engine from a web interface. Several of these challenges require compatibility troubleshooting and running a centralized python server, which will be implemented in a consistent environment when this project is developed further.

Accomplishments that we're proud of

We are proud of working with novel architecture and technology, providing a solution to solve common EHR problems in design, functionality, and implementation of data.

What we learned

We learned the value of leveraging the strengths of different team members from design to programming and math in order to advance the technology of EHRs.

What's next for myHealthTech?

Next is the addition of additional self-reporting fields to increase the robustness of the artificial intelligence engine. In the case of depression, there are clinical standards from the Diagnostics and Statistical Manual that identify markers of depression such as mood level, confidence, energy, and feeling of guilt. By monitoring these values for individuals that have recovered, are depressed, or inflict self-harm, the AI engine can predict the behavior of new individuals much stronger by logistically regressing the data and use a deep learning approach.

There is an issue with the inconvenience of reporting symptoms. Hence, a logical next step would be to implement smart home technology, such as an Amazon Echo, for the patient to interact with for self reporting. For instance, when the patient is at home, the Amazon Echo will prompt the patient and ask "What would you rate your mood today? What would you rate your energy today?" and record the data in the patient's self reporting records on myHealthTech.

These improvements would further the capability of myHealthTech of being a highly dynamic EHR with strong analytical capabilitys to understand and predict the outcome of patients to improve treatment options.

Share this project: