Zone dispatching for NEMT: splitting your service area so vans stay close to work

What zone dispatching is

Zone dispatching divides your service area into a few smaller areas and gives each one its own drivers. A van based in the north zone takes north zone pickups, so its next trip usually starts close to its last drop-off. The alternative is dispatching the whole map as one area, where any van can take any trip.

Transit agencies running ADA paratransit, a demand-response service much like NEMT, use both approaches. A 2013 final report for the Transportation Research Board’s Transit IDEA program (Project 73, Texas A&M Transportation Institute) summed up the trade-off:

• What zones buy. Smaller, independent zones are easier and cheaper to manage, give riders better on-time performance, and tend to leave call centers and drivers more satisfied, since drivers work a smaller, familiar range. Agencies in very large, sprawled cities adopted zones for those reasons.
• What zones cost. Under strict zoning, a van may not pick up riders outside its zone, which blocks some ride sharing and adds empty miles. Riders whose pickup and drop-off sit in different zones made up as much as 30 percent of daily demand in Los Angeles County.
• The benchmark. The report treats the whole-map approach, which Houston still used, as the one that minimizes operating cost and maximizes service, with drawbacks that grow with the size of the area.

The same idea already shapes broker work. Georgia runs NEMT in five regions (North, Atlanta, Central, East, and Southwest), and Verida has served all five since April 1, 2026. Texas managed care defines a service area as the counties in each state-defined area. Brokers also assign trips by the areas providers say they serve, which is covered below. A company zone is the same concept at the scale of one fleet. For the empty-mile side of the math, see how to reduce deadhead miles.

Draw zones from your own pickups

Zones drawn from city limits or a map’s grid rarely match where your trips are. Draw them from your records.

1. Pull four to eight weeks of completed trips. Plot pickups and drop-offs separately. Morning pickups cluster around homes, and drop-offs cluster around a handful of facilities.
2. Build from ZIP codes or counties. They are easy to explain to dispatchers and drivers, and they match how brokers describe service areas. New York’s MAS manual refers to every ZIP code for which a provider asks to be assigned trips, and Texas service areas are built from counties.
3. Mark the anchors. Dialysis centers, hospitals, and day programs generate the steadiest flows. Put each large anchor inside a zone, not on a boundary.
4. Balance vehicle hours, not trip counts. A zone with many wheelchair and stretcher riders needs more van time per trip than an ambulatory zone of the same size. The pickup and drop-off time guide covers measuring stop time by level of service.
5. Follow real barriers. A river with two bridges or a highway with few crossings makes a natural line, because trips across it are slow anyway.
6. Share the border ZIP codes. Let a ZIP code on a boundary belong to both neighboring zones, so a pickup there goes to whichever van is closer that day.
7. Check the vehicle mix in each zone. Every zone with wheelchair riders needs a wheelchair van on each shift, or a clear rule for borrowing one.

Here is a hypothetical result for a 12-van fleet, to show the output. The figures are examples.

Name the zones plainly, and give every dispatcher the same map.

Anchor drivers to their zones

A zone works only if its drivers start and stay there most of the day.

• Start inside the zone. A driver who begins at a garage or at home inside the zone reaches the first pickup with a short empty drive. If that means vans go home with drivers, read up on insurance and policy first in the take-home vehicle guide.
• Keep recurring riders with the same driver. For standing order riders, MTM Health’s Virginia handbook (for fee-for-service trips from October 1, 2026) expects a best-faith try at keeping one familiar driver, with a note on file whenever someone else fills in. A driver who stays in one zone makes that easy.
• Learn the zone. Drivers who work the same area learn the side entrances, gate codes, and which facilities have residents ready. The IDEA report credits a limited, familiar range with better driver job satisfaction.
• Keep one floater. A van with no zone of its own absorbs a breakdown, a surge of will-call returns, or a long cross-zone trip without pulling a zone below its minimum.

Build shifts around zone demand the same way you would for the fleet as a whole. The driver shifts guide covers turning a demand curve into start times.

Morning and afternoon flows

Zone demand moves through the day, and the flows are predictable enough to plan around.

• Early morning, inbound. Riders travel from homes toward facilities, and appointment times fix the drop-offs. Texas schedules arrivals between 15 and 60 minutes ahead of the appointment. Louisiana’s Medi Trans manual will not let a rider with an 8:00 a.m. appointment be picked up before 6:00 a.m. Most early vans run from the edges of a zone toward its anchors.
• Late morning, the first turnover. Under NIDDK’s description, in-center hemodialysis patients keep a set slot, most often three days a week, with sessions of roughly four hours. A rider in a 6:00 a.m. chair, for example, is ready to go home around 10:00, just as the next shift arrives, so the zone around a dialysis center gets a second peak before noon.
• Midday, will-call returns. Clinic riders call when ready, and the broker’s response window starts. The Virginia and Rhode Island handbooks from MTM Health give 45 minutes, and Georgia’s manual gives one hour. Keep a van near each zone’s clinic cluster during those hours.
• Afternoon, outbound. Riders travel from facilities back to the edges. The van that dropped a rider at the hospital in the morning is often not the one nearest when that rider is ready, and a zone rule decides who goes.
• Day of week. Dialysis riders follow Monday, Wednesday, Friday or Tuesday, Thursday, Saturday patterns, so a zone’s load can differ sharply between those days. Rebalance vans by weekday, not once for the whole week.

Facility hours set the edges of the day. On return rides, CareOregon wants drivers to aim for a pickup within 15 minutes of the facility’s closing time. In Texas, a rider whose appointment is the provider’s last of the day gets a pickup scheduled within 15 minutes of closing time. The last van in a zone should be near its latest-closing anchor.

Trips that cross zones

Every zoned fleet has trips that start in one zone and end in another. Decide the rule once, write it down, and apply it the same way every day.

The IDEA report tested the third rule. In simulations built on real paratransit demand from Houston, Los Angeles, and Boston, letting providers pick up out-of-zone riders who needed a return trip into their own zone cut empty trip miles by up to 23 percent, without lowering service. In the best cases it also used 6.6 percent fewer vehicles, cut total mileage by 9 percent, and raised passenger trips per revenue hour by 7.8 percent. The report also notes that some transit systems send cross-zone riders to taxis or carriers that mostly serve cross-zone rides, which is the fourth rule.

Whatever the rule, the maximum ride time still applies. A cross-zone van that picks up a second rider on its way home adds minutes to the first rider’s clock.

Measure whether the zones are working

Track a few numbers by zone each month and compare them with the months before you zoned.

• Deadhead share. Empty miles as a share of all miles, per zone. This is the number zones exist to lower.
• Trips per hour. Completed legs per vehicle hour. A zone that runs well above the others may need another van, and one well below may need to merge.
• On-time performance by zone. A zone with late pickups clustered at the same hours is short a van at those hours.
• Cross-zone share. The share of trips that start and end in different zones. A rising share means the boundaries no longer match where riders go.
• Will-call response time. Minutes from the rider’s call to pickup, per zone, against the broker’s window.

The NEMT KPIs guide covers how to calculate and trend each of these.

When zones stop helping

Zones are a simplification. They help when they match the work and hurt when they do not.

• Too few trips. A zone with one van has no backup for a breakdown or a burst of will-call returns.
• Too many crossings. When a large share of riders cross boundaries, as in the Los Angeles County figure in the IDEA report, the boundaries cost more empty miles than they save. Redraw first, and drop zones if redrawing does not help.
• Heavy same-day work. In the IDEA report’s data, roughly 15 to 20 percent of demand arrived during the day of service. The more trips arrive late, the more often the nearest van sits across a line.
• Rural service. Long trips cross any line you draw. The rural NEMT guide covers planning around a few regional hubs instead.
• Whole-day planning. The IDEA report’s benchmark was the whole-map approach, which it calls the lowest-cost option, with drawbacks tied to the size of the area. A planning tool that weighs every van against every trip works on the whole map directly, and zones then become a default starting point rather than a hard rule.

The IDEA report notes that some systems run hybrid policies. A workable hybrid for a fleet: drivers start in home zones, but dispatch allows a cross-zone assignment whenever a nearby van is idle and the rider’s times still work. The route optimization guide covers whole-day planning in more detail.

Planning the whole map in HealthRide

Ryder Go plans and assigns your whole day in one click, matching each trip to a driver and van using pickup times, appointments, where each driver is, and who needs a wheelchair van. By default, you review the plan before it applies, and the dispatch board still takes drag and drop. The live map shows every vehicle at once, so a van far from its usual area stands out. See Ryder Go for how the planning works.