This case study is my personal work and is not associated with Kinetik.
Context
Kinetik Healthcare Solutions is a SaaS company that automates the scheduling of non-emergency medical trips and billing between health plans and transportation providers with the goal to make preventative care accessible to everyone.
Problem
Due to high volumes of health plan insurance consumers (members) missing scheduled trips, Kinetik’s clients (health insurance companies) may waste millions of dollars on transportation fees alone without members receiving the preventative care they need. This leads to operational inefficiencies for health insurance companies, wasted resources, and negative impacts on member health.
Product
Goals
The main objective is increasing customer satisfaction and experience by creating a product feature that will reduce the number of member no shows and cut unnecessary costs for health plans.
Metrics to improve:
Reduce the number of member no show trips
Decreasing the amount of time users spend scheduling trips
Increase in the number of trips overall
Increase in number of rescheduled trips
How can we increase customer satisfaction and experience by reducing the number of member no shows and cutting unnecessary costs for health insurance companies?
No Show Shield utilizes both proactive and reactive methods to reduce the number of no shows and increase rescheduled trips.
Predictive AI Model & Widget
The No Show Shield would use a predictive AI model to produce a risk score that is based on data from the member's trip request and trip history. The model would take into account the trip location, scheduled day and time, and trip reason. The information would be displayed on a widget on the member's profile for a call center representative— or a trip scheduler (TS) user— to use when scheduling trips over the phone. The widget is expandable so TS users can see the breakdown of data that led to the risk score and take action— flagging the trip or setting custom reminders.
Health Insurance companies can configure a risk score threshold for trips to be automatically flagged (e.g System will flag all trips with a 75% or higher risk score).
To determine what information is important to display, I made a flow chart with some of the most common reasons a member may miss a trip based on the notes TS users will leave about rescheduling or last minute changes. Currently, there is no formal way to collect that information, but occasionally members will call before or after a no show to make last minute changes to a ride or to try to reschedule.
An AI model will be able to predict these reasons based on a member's past behaviors and relay that to a TS user, who can then take actions while booking the trip to lower the risk of no show. Member App and SMS actions can also act as a communication channel and help members remember to make necessary changes to their trips before it's too late.
In addition to the reminders, prescribed by the predictive model, the Kinetik Health app can be improved to reduce member no shows. To collect more data to train the predictive model, I created a new "Missed" status for trips on the Kinetik Health app and a prompt asking why a member missed a trip. The member is then prompted to reschedule the trip either through call or within a reschedule flow in the app. The app saves the member's trip information, except for the date and time so that the member can edit those for their new trip.
With the No Show Shield, the Kinetik Health App will also be able to check a member's location 10 minutes before dispatch (or a configurable amount of time) before a ride and prompt possible rescheduling if the member is too far from the pick up location, so that rides can be can be canceled and rescheduled rather than missed.
Caregivers who book trips for members through the Kinetik Health App will also be able to put their contact information in the app to receive live trip tracking and reminders to also minimize channels of communication and the chance of miscommunication.
Administrators can use the No Show Shield dashboard to manage resources and review flagged trips. They can also view high level projected savings and data trends around no show trips, such as
Geographic heatmap
Member No Shows by Day of the Week
No Show Trip Reasons
Administrators may also manually mark members with a "High Risk No Show Badge" or automate it to be added to a member's profile after a configurable number of no shows.
Conclusion
This project was done through a product intern hackathon during my internship at Kinetik. It was a great opportunity for me to explore product management alongside design, learn how to develop a business case from a product lens, and utilize product metrics in my research. At the conclusion of this project, I was able to pitch my product idea to the entire US branch of the company and it was selected as the winner of the hackathon.
Thank you to the team at Kinetik for supporting me throughout this project and my internship!
