Health Harmony
Key Considerations
Notable Risks and Potential Business Implications
There are two very notable risks with this application: Data Security and Cost
Because this application is intended to house Personal Health Information (PHI), it will be extremely important to ensure the safekeeping of data both on the device and within a cloud environment. This product would also be subject to the same laws governing other health institutions, specifically HIPAA requirements. These data security concerns cannot be overlooked and must be built into the foundation of the application.
These data security needs directly tie into the second risk of this application, Cost. This could potentially be a very expensive application to build and maintain. There would be the need for potential investment by third-party players interested in the type of application. While private investment could certainly get the application started, for the long-term future of the application, we would ideally need partnerships with other major players in healthcare, such as Medicare/Medicaid, Insurance Companies, or Hospital Systems. If a partnership were established, we could potentially expand the application to do a two-way data integration to pull in more health record information from a hospital source while also pushing back data regarding symptom management and medication compliance. There is significant potential here, but the upfront investment is also significant.
Integrating AI into the Platform and How It Will Help the User Make More Useful Decisions
There are so many uses for Artificial Intelligence (AI) and it is perhaps the most talked about new technology in the market. A little background on AI - AI is a field that houses many different kinds of technology sort of how Physics is a field that houses many subtypes.
For this product, I'm envisioning the use of both Machine Learning and a trained AI as well as a Deep Learning Model and Large Language Models. My current idea is to train a model on pharmacological data, flag potential complications or allergies, and warn the user that they should talk to their pharmacist or doctor. This could potentially prevent accidental negative side effects due to drugs that interact poorly or prevent accidental allergic reactions. This model-trained AI could also be used to uncover health metric trends that should be raised to a doctor's attention.
​
There are other uses that I have in mind. One such use would be mimicking the user to log into third-party applications such as hospital portals. Once logged in, the platform could then ingest any updated health information and then the AI could categorize appropriately into the Health Timeline. This would prevent the user from having the manage multiple portals and give them a more comprehensive vision of their health.
​
Below is a mocked-up AI workflow that outlines at a high-level how the AI would work both during the first time set up and post-connection.
