MediRoutes launches predictive ETAs for NEMT dispatchers

MediRoutes has launched Predictive ETAs for non-emergency medical transportation providers. The feature is available in the latest version of its dispatch software.

The tool uses machine learning and GPS trip history to estimate how long a route is likely to take, giving dispatchers a more detailed basis for scheduling than standard map-based travel times. It combines historical trip data with live traffic conditions and updates estimates as a journey progresses.

The estimates are based on GPS readings collected from provider vehicles about every five seconds. The model also accounts for variables including previous travel on the same route, time of day, and day of the week.

The approach is designed to reflect how non-emergency medical transportation journeys differ from standard driving routes. In addition to time on the road, these trips can include loading and unloading, wheelchair or stretcher securement, and waiting at facility entrances.

According to MediRoutes, it has supported more than 100 million NEMT trips since 2008, and the model becomes more accurate as more data is added.

The launch comes as transport brokers apply greater scrutiny to GPS reporting and operators face growing pressure to back on-time performance records with documented trip data. Against that backdrop, MediRoutes is positioning historical GPS information as both an operational record and a source for route planning.

Dispatchers can use the estimates when building routes and deciding whether to accept broker trips. A more precise view of expected travel times could influence whether a provider decides an additional job will fit within an existing schedule.

The effect could extend beyond dispatch teams. Clinics, care facilities, and riders waiting for collection can receive ETAs that reflect the full trip rather than only the drive time shown by consumer mapping applications.

MediRoutes is integrated with several large brokers, including ModivCare, MTM, and MAS, as well as more than a dozen regional brokers. Those integrations already rely on GPS reporting, and the new feature uses the same data set for scheduling estimates.

David Reinkensmeyer, Chief Executive Officer of MediRoutes, said route planning is only as strong as the time estimates used at the outset.

"A planned route is only as good as the travel times behind it," Reinkensmeyer said.

"We've been collecting granular GPS data on NEMT trips for nearly two decades. AI lets us put all of that to work so dispatchers can build a day they can actually deliver on," he added.

Operational Detail

John Bierman, Chief Technology Officer of MediRoutes, said the frequency of GPS readings allows the model to identify differences in route timing across the week.

"A route doesn't take the same time on a Tuesday morning as it does on a Friday afternoon, and readings every few seconds give the model enough detail to see the difference," Bierman said.

He added that the model is designed to learn from completed journeys rather than idealised route assumptions.

"It learns from how those trips truly went, so the numbers a dispatcher builds the day on reflect how each route really runs," he said.

The latest version of the software is being rolled out to customers ahead of a wider launch of the updated dispatch platform later this month.