A
Panacea - Intelligent Preventive Care
The purpose of this application was to design a physician On-demand house call service. Our group was approached by a team of Harvard Graduates who wanted to create a Premium on-demand house calls service in Lagos, Nigeria.
OVERVIEW
Panacea is an Uber for medical house-calls.
While experiencing unexpected, non-critical symptoms of illness or injury, users can “hail” a doctor to
come to their homes at short notice and provide medical services.
Interacting with the service involves two primary software components; an user-facing and a doctor-facing smartphone app.
PROBLEM
1. How to use contextual information to probe for pertinent information beyond a standard form?
​
2. Apps are short lived. People download applications when they need them but forget after a while. We had to create a solution that could engage users passively for preventive care
CONCEPTUALIZATION
Intelligent User Interface
The design leverages an intelligent user interface to aid doctors in collecting relevant diagnostic information and offering suggestions based on contextual information that might not be apparent to the doctor, or might not come up in conversation with the patient. Rather than replacing the doctors' expertise, this system supplants it. Doctors can spend their time probing for additional relevant details and engaging with patients rather than collecting a wide breadth of basic information.
​
User-Facing App
While experiencing an ailment, users report their symptoms using the app on their smartphones. The app uses contextual information such as epidemiological trends, location data, and user history to probe for any additional information that might aid diagnosis. A doctor is then dispatched. Additionally, the app can occasionally notify users of any public health concerns and
schedule follow-up services if needed.
​
Doctor-Facing App
Doctors in the field receive notifications on their app when a patient orders medical services. The app displays the patient’s location, medical history, reported symptoms, and any diagnostic information that might have been gleaned from contextual data collected by the app. If specific tests or equipment might be required to treat the patient based on this information, the app will alert the doctor to ensure he or she is properly equipped.
SOFTWARE ARCHITECTURE
FEATURES
Scenario 1: Provide an approximate diagnosis based on location-based trends and the symptoms
mentioned by the patient.
​
How we tackled this:
A user requests a doctor complaining about flu-like symptoms. Given these symptoms and information concerning current public health crises, the app asks the user if she has recently been to South America. Since the early stages of a Zika viral infection are similar to more common infections, the app is able to gauge whether the doctor on call should be dispatched prepared to treat a routine illness or whether the patient should seek more urgent care without
needing to expose the practitioner.
​
User Facing App - low fidelity wireframes
Doctor Facing App - low fidelity wireframes
Scenario 2: Inform users to about virus outbreak in their region and remind them to seek preventive therapies such as flu shots etc.
​
How we tackled this:
After not needing to use the service after several healthy months, the app notifies the user that it is flu season and gives an approximate number of reported cases in the user’s area as well as a quick actionable link to order a flu shot. This interaction leaves the user more aware of his potential exposure to the virus as well as reminded of the service, whether he orders a vaccine or not.
User Facing App - low fidelity wireframes
Doctor Facing App - low fidelity wireframes
Scenario 3: The intelligent feature of the application notifies the doctors of the medical condition in the surrounding areas and preps the doctors with all necessary medication or medical kits.
​
How we tackled this:
Several users living in the same apartment building have recently been diagnosed with strep throat. A new user calls for a doctor complaining of flu-like symptoms and a developing sore throat. The app uses this information to dispatch a doctor with a strep test, which the doctor may not have automatically tested for without knowing that the patient was at an increased risk.
Doctor Facing App - low fidelity prototype
Scenario 4: The smart reminders for any follow-ups required
​
How we tackled this:
After using the service to request a visit that resulted in stitches for a minor laceration, the user
opens the app and sees an option to quickly schedule a follow-up visit to remove the stitches.
After checking out, a doctor in the field receives a notification detailing the service requested as
well as a record detailing the previous visit.
User Facing App - low fidelity prototype
Doctor Facing App - low fidelity prototype
​
The primary downside to this approach is the potential high development cost, as there is a need to pull information from a wide range of sources and develop highly predictive algorithms. Additionally, benefiting from longitudinal patient data might only be possible after the service has been established for some time. Implementing an intelligent system likely makes sense as a longer term goal. After the basic infrastructure necessary for an "Uber for healthcare" is established, these elements can be rolled out over time as resources allow.