CCS Medical, a chronic care management company, is taking a personalized approach to chronic care treatment adherence through its new AI tool, PropheSee. Built in partnership with Accenture, the model pools patient data to generate personalized messages to keep patients on track with their diabetes treatment.
CCS told Fierce Healthcare a one-size-fits-all patient outreach model doesn’t work for all patients. Notifications are most effective, they say, when they are customized and delivered at the right time.
The company is using a version of an attrition model, which has been used in other industries like telecommunications, financial services and entertainment, Chinmoy Barua, managing director at Accenture, said in an interview.
Though the type of model is not new, Richard Mackey, chief technology officer of CCS, said the company believes this is the first time the model has been deployed in the chronic care industry, where there is a persistent need for resupplying medical equipment.
CCS’ goal in launching PropheSee is to hyperpersonalize the patient care experience. Taking into account a patient’s data profile, the model can push out a message to a patient through the patient’s preferred means of communication when it thinks the message will be most effective. The timing and format of the notification will differ between patients.
“We think that is meaningful, differentiated and innovative in this space. So it's a combination of taking tried and true approaches and methods from other industries that are more mature, but being able to apply them in a new and novel way here,” Mackey said.
Accenture and CCS have made use of a host of large data sets including claims data, socioeconomic data and outcomes data to build fleshed-out profiles of CCS’ customers that could help the model predict whether a patient is at risk for discontinuing their treatment.
“We can leverage both the first-party and third-party data and kind of mash the data together to create a holistic picture for our customers,” Barua said.
Barua contends that the model is 90% accurate with a 90-day forecast of a patient’s likelihood to falter with treatment. The model could account for economic hardships, like a patient losing their job, and prompt the patient with a personalized message reminding them of the importance of adhering to their therapy based on patient personas and in line with the patient's communication channel preferences.
CCS is focused on collecting social determinants of health data that patients may not feel comfortable disclosing to a provider through alternative means to provide the patient with appropriate support.
Barua hopes the accuracy of the prediction will increase with additional data sources.
“As (patients) go through a care journey, the way that they interact with us today could be very different in two or three months … We built this framework so that as they change and go through their life's journey and have bends in the roads and ups and downs in the journey, we can respond and bring them different kinds of messages,” Mackey explained.
Before the messages are sent to customers, they are tested on a control group through Accenture’s Learn, Test and Validate procedure. CCS aims to prove the efficacy of the messages so that patients do not receive unnecessary or unhelpful communications. Patients may also opt out of communications altogether.
CCS shared the following examples with Fierce Healthcare.
The first example, seen below, is a message that a caregiver with diabetes would receive. The message, sent via text, emphasizes wellness.

In a second example, CCS contacts a patient through email and reminds them to reorder diabetes supplies.

CCS and Accenture are also collaborating on a generative AI framework that will guide CCS’ future endeavors in AI, the companies said.