Route Optimization for NEMT
Customized route optimization for non-emergency medical transport (NEMT) — pickup-and-delivery pairs, appointment windows, vehicle-type matching, and driver work-time limits calibrated to your fleet.
NEMT operations carry constraints most routing tools handle piecemeal: patient round-trips that must stay on the same route in the right order, medically time-sensitive appointment windows, wheelchair or stretcher requirements that gate vehicle choice, and driver hours that need to be respected across long routes. BOSOPT handles these together — pickup-and-delivery pairing, appointment-window discipline, vehicle-type matching, and daily work-time budgets — calibrated to your operation.
What BOSOPT accounts for in NEMT
NEMT operations combine standing orders for recurring appointments (dialysis, oncology, adult day care), on-demand rides (hospital discharges, doctor visits), and will-call return trips. Each patient trip is a pickup-and-delivery pair with a medically time-sensitive appointment window. Wheelchair and stretcher vehicles have hard requirements — no substitution. BOSOPT handles pickup-and-delivery pairing, appointment-window discipline, vehicle-type matching, and daily work-time budgets — calibrated to your fleet.
Typical constraints we model in NEMT operations — your operation may have some or all of these:
- Patient appointment windows
Patient pickup, appointment start, and return-trip windows — modeled as hard constraints in the route plan. - Pickup-and-delivery pairing
A patient's home-to-facility trip and facility-to-home return stay on the same route in the correct order, with both pickup and delivery times respected. - Vehicle-type matching
Wheelchair vans, stretcher vans, sedans, and buses matched to patient mobility requirements — a wheelchair patient doesn't get assigned to a sedan. - Heterogeneous fleet
Wheelchair-accessible vans, stretcher vans, ambulatory sedans, and larger passenger vehicles — each with capacity constraints for passengers and equipment. - Standing orders and on-demand mix
Recurring appointment slots (dialysis 3x per week, monthly checkups) planned alongside same-day on-demand rides. - Driver work-time limits
Daily work-time budgets and mandatory meal breaks built into the plan before dispatch, not flagged after.
What changes when you optimize NEMT
- Patient appointment windows held.
Routes respect patient pickup and appointment-start windows. Late arrivals to time-sensitive treatments surface before dispatch, not in scheduling calls from the clinic. - Round-trip pairs stay together.
Home-to-facility and facility-to-home rides ride the same route in the correct order. No second dispatch to retrieve a stranded patient at the clinic. - Right vehicle for the patient.
Wheelchair-accessible vans go to wheelchair patients. Stretcher vans go to stretcher patients. Vehicle-type matching is part of the solve, not a manual check after. - Standing orders and on-demand planned together.
Recurring appointment slots and same-day on-demand rides planned as one problem. Standing orders don't get displaced when a same-day request comes in — capacity is accounted for. - Re-optimize before the next wave.
Will-call return trips, driver call-outs, or same-day appointment changes are batched into the next wave's plan. Re-optimization runs in minutes at wave boundaries — not rebuilt by hand. - Driver work-time limits respected.
Daily work-time limits and meal breaks built into the optimization. Routes that would push a driver past their daily budget get re-planned before dispatch. - Transparent capacity limits.
When trip demand exceeds the day's fleet, the system surfaces unassigned trips with locations and reasons — no silent drops, no surprise calls from the clinic.
How this looks in practice
NEMT operations typically see 10–20% miles reduction, 25–40% fewer missed appointment windows, and improved vehicle-type utilization in pilot evaluations — driven by pickup-and-delivery pairing, appointment-window discipline, and vehicle-type matching. Every engagement starts with a baseline comparison on your data.
Typical improvement
| Metric | Range |
|---|---|
| Total miles | 10–25% reduction |
| Total route time | 8–15% reduction |
| Late deliveries | 30–50% reduction |
| Vehicle utilization | 5–15% improvement |
Ranges are based on comparisons against operational data from 80+ organizations. Every engagement starts with a baseline comparison on your data.
Common questions from NEMT teams
We have patients who need wheelchair vans specifically. Can the optimizer respect that?
Yes — vehicle-type matching is modeled as a hard constraint. A wheelchair patient is only assigned to a wheelchair-accessible van. Stretcher patients go to stretcher vans. The optimizer respects mobility requirements in the solve, not as a manual check after.
Our patients have round-trip rides — home to clinic, then clinic back home. Does the optimizer keep those on the same route?
Yes — pickup-and-delivery pairs are modeled directly. The home-to-facility pickup and the facility-to-home return stay on the same route in the correct order, with both trip times respected. Patients don't get stranded at the clinic because the return trip landed on a different route.
We mix standing orders (weekly dialysis, monthly checkups) with same-day on-demand rides. Can the optimizer handle both?
Yes — standing orders and on-demand rides are planned together as one problem. Recurring appointment slots hold their capacity; same-day requests fit into what's left. When a same-day request can't fit, the system surfaces it as unassigned rather than silently displacing a standing patient.
Our patient appointment windows are strict — arriving late to dialysis means the treatment slot is gone. How does the optimizer handle that?
Appointment windows are modeled as hard constraints. Routes that would arrive outside the window surface as infeasible before dispatch, not at the treatment center. If demand exceeds what the fleet can serve within-window, the unroutable trips are flagged with reasons.
What happens if a driver decides to change their route order, or add an unplanned pickup?
The optimizer produces the recommended sequence, but drivers retain operational autonomy. Sequence changes and ad-hoc pickups don't break anything — execution data feeds back into the next wave's optimization, and recurring deviations get investigated as signals the engine should learn from.
Our drivers have shift preferences and vehicle-type comfort. Does the optimizer respect those?
Yes — shift times, vehicle-type preferences, and work areas are honored in the plan. The optimizer respects them rather than fighting them. Driver familiarity with specific vehicle types and patient routines stays where it belongs.
Continuous calibration
The optimizer plans the day; operations executes it. We capture how the day actually ran — sequences, stops, miles, times — and compare against the plan. When drivers consistently change a sequence or add stops without saving miles or time, there’s usually a real-world constraint the optimizer doesn’t see yet. The engine gets recalibrated to reflect what the data shows. Every change is reviewed and applied by our engineers — calibration with judgment, not autonomous drift. The longer the system runs in your operation, the sharper the plan.
See better routes on your data
Send us 5 to 10 days of your delivery data — stops, time windows, vehicle constraints. We’ll run it through BOSOPT and show you a side-by-side comparison against your current plans — and a realistic monthly savings estimate. No commitment.