Healthcare often falls over during the last mile - at the customer touchpoint. While the technology and communication industries have marched along with real-time, instant-reach models, healthcare is stuck with faxes and phone calls. Beyond the obvious inefficiencies, these methodologies have the dubious distinction of being extremely expensive as well. Incorporating these new advances in communication, especially in personal communication using messengers with smart AI driven agents can drastically optimize the last mile problem in healthcare. Additionally, it serves to extend the reach of healthcare to far-flung, geographically distributed areas thereby truly achieving the tenet of global health.
What it does
Conversational Health is an AI agent that manifests as a chatbot on any messaging medium including Facebook Messenger, SMS etc. This system ties into standardized repositories of data like NIH databases and EMRs and enables patients and care providers to instantly call up information on demand.
This information could be routine updates that need to be sent to the patient such as lab results, medication reminders, appointment confirmations etc. or care pathways that need to be sent to case-workers remotely to administer an uncommon treatment.
Beyond the push mechanism, consumers can use this agent and the backing intelligence to self-triage their conditions and get recommendations and appointments with clinics and hospitals optimizing for cost and coverage. Finally, this system could tie into the social fabric and communities surrounding the patient helping them with their social needs and wellness coaches to help them stay on their care regimen.
How I built it
- redoxengine to tie into EMRs
- IBM Watson Conversation for entity resolution and conversational pathways
- Symcat, MeSH for access to symptoms/conditions/drug
- Node-red for orchestrating the services
What's next for Conversational Health
I envision bringing this conversational health model to health systems and augment their existing mechanisms for patient contact and outreach - like the Stanford myHealth app and to integrate with services like zocdoc to ensure that patients make the correct decision in terms of the kind of care they seek.