The Deployment Plan

What the plan specifies#

1 · The schedule: when the unit hours exist.
The unit hours the procuring entity chose to buy are laid out across the week and packaged into shifts. Shift length is a design lever: 24-hour, 12-hour, 10-hour, and 8-hour shifts each trade coverage shape against fatigue, cost, and recruitment. Shift length is bounded by the fatigue and performance evidence, not chosen on cost alone. The schedule expresses the chosen model as a staffing pattern that produces the intended readiness across the week. It does not impose a demand curve on the crews.

2 · The posting plan: where units sit and how coverage is held.
Units post by clinical level (BLS/ALS) so that coverage matches the geography and timing the plan protects. Posting matters because a unit hour in the wrong place does not protect the area the plan covers. There are two kinds of deployment and within the second there are two move-up strategies.

• Static (fixed) deployment. There is no dynamic plan. Units have fixed homes and return to them and committing one unit does not move the others. Fire apparatus typically work this way: a crew heads home when its mission is complete, often without waiting for an assignment.
• Dynamic deployment. Committing one unit triggers a planned repositioning of others to preserve coverage. Coverage recovers faster and crews are worked harder. Two move-up strategies exist within it:
  • Direct assignment. Units may leapfrog past each other to fill a post. When the southernmost post opens, the north unit may be sent all the way to it.
  • Cascading assignment. Units shuffle down the line without crossing. The south unit moves to the southernmost post, the central unit fills south, and the north unit fills central.

The choice between direct and cascading assignment is a design decision rather than a dispatch habit and it interacts with the compliance goal: cascade when the system is below the coverage level that meets its compliance target and assign direct at or above it. Both are dynamic behaviors.

Neither static nor dynamic deployment is wrong. They buy different readiness at different cost and the cost benefit analysis says whether the one a system wants is affordable. What the plan requires is that the choice is explicit and that its readiness implications are modeled.

3 · Response assignment: which unit answers a given call.
Separate from where units sit is which unit is committed to a given call. The fastest capable unit is the common default but not the only valid method. "Fastest" is deliberate: the closest unit is not always the fastest and time is what is wanted.

• Fastest. Always send the capable unit that will arrive first.
• Territorial. Units answer within defined areas.
• Acuity-differentiated. With good dispatch triage (EMD, MPDS), assignment is matched to the call's acuity and time sensitivity. Time-critical, life-threatening calls get the fastest unit, while lower-acuity calls are routed to preserve the readiness of strategically placed units for the next life threat.

The acuity-differentiated case is where the assignment logic itself produces readiness and it fits an integrated, multi-producer system. In a blended fire and private system, a suspected life threat always draws the fastest unit, often a strategically posted fire company, while a low-acuity call is run by a private producer's move-and-cover units, leaving that fire company in place and ready for the next life threat. The two producers work under one deployment plan, which conserves readiness deliberately.

Diversion. Diversion is pulling a unit off a lower-priority call to send it to a higher-priority one. ODS treats diversion as a feature. Some systems cap the number of diversions and some treat a call as non-divertible after one diversion, on the reasoning that the low-acuity patient would otherwise wait too long. ODS does not, because that reasoning inverts the priority. A patient with a time-sensitive emergency needs the unit now and that patient's physiology is indifferent to whether the low-acuity patient has already been bumped. ODS sends the resource to the patient who needs it most, however many times the lower-priority call has already been diverted from, with one condition: if the deferred patient's condition changes, that is a re-triage event and the call is re-evaluated. A system that diverts chronically, repeatedly bumping the same low-acuity calls, is receiving a signal that other services have a role: mobile integrated health, community paramedicine, telemedicine, or nurse triage.

Fatigue management#

Crew fatigue is a design constraint that can override coverage. A plan that produces coverage by overworking crews produces readiness that degrades, which readiness measurement exists to expose. This is the one place the deployment plan is not purely a matter of community preference: a schedule may not be adopted if it degrades safety or performance.

The fatigue science is well established. Sleep loss degrades essentially every aspect of human performance (Rosekind et al., NASA Technical Memorandum 2001-211385, 2001). Alertness follows a two-process model (Borbély, Human Neurobiology, 1982): homeostatic sleep pressure that builds with hours awake, plus a circadian rhythm with a circadian low overnight, when errors and accidents spike. Risk rises with hours on shift and with successive shifts (Folkard and Tucker, Occupational Medicine, 2003). Sleep debt accumulates and does not recover quickly (Van Dongen et al., Sleep, 2003; Belenky et al., Journal of Sleep Research, 2003). Self-assessment of fatigue is unreliable, because fatigue impairs the judgment needed to assess it. Schedules are benchmarked against the evidence, not against how rested crews feel.

Apparent efficiency can mislead. The same 24 hours of coverage can be produced with roughly 6 FTEs on a 24-on/48-off rotation, or about 8 FTEs on rotating 12-hour shifts, which makes the 24-hour shift look cheaper per hour of coverage. The arithmetic hides the fatigue cost that evidence describes, so the "efficient" option is not the same product. ODS surfaces the dollar cost and the fatigue consequence together.

System workload and human workload are different measurements. System workload is what the system can absorb: concurrency, read on the curve of the marginal utility of unit hours (Financial Sustainability). Human workload is how much a crew can do before fatigue degrades them.

Fatigue enters the plan at four levels:

• Shift design. Length, consecutive shifts, start times, rotation.
• In-shift workload. Time utilized against time activated, plus movement time. Concurrency is not a measure of a crew's workload. It is a property of the system and needs many units to exist; a single unit has none. Airline pilots are not measured by how many aircraft are in the sky, which is an air traffic control measure. A pilot's workload is time on duty, door closed to door open. A crew's in-shift workload is read the same way.
• Recovery. Measured separately: rest between duty periods and the staffing depth that makes that rest real.
• Deployment posture. Whether the model keeps crews perpetually in motion or allows necessary recovery during the shift.

In-shift workload is measured as per-resource utilization: time utilized against time activated, for a specific unit. This is a workload measure on a crew, not a performance surrogate or unit hour utilization (UHU) under a new label. Every producer runs the method on its own system: analyze each resource individually (its time activated, its time utilized, and its safety record) to locate the shift length at which safety and performance hold, avoiding both over-utilization and under-utilization. The indicators that identify optimal shift length are accidents, injuries, and near misses. They are measurable, already reported, and rise with hours on shift and successive shifts (Folkard and Tucker, Occupational Medicine, 2003).

ODS publishes no threshold shift length and no utilization figure, because the answer is system-specific. Within the safety and performance bounds a system finds, any shift configuration the workforce finds beneficial is supported: the evidence sets the guardrails and the workforce's preferences fill the space inside them.

Who owns the plan#

Unit-hour procurement (quantity), the deployment plan (location), response-time accountability, and billing travel together.

The party that holds all four is the procuring entity, defined with the regulator and the producer in System Structure. Whoever holds the plan buys the unit hours and takes on the revenue, so the producer of unit hours does not carry the transport-revenue risk. A producer controls whether its committed unit is staffed, serviceable, and out the door quickly. It does not control the posting plan, the call classification, or the zone definitions that jointly determine when a unit arrives. A producer cannot be held to a number that someone else's decisions largely determine (see Measurement & the Score).

Under public procurement (the alliance model), a fire department can serve in communications, first response, and transport, hold unit-hour procurement and the deployment plan, and place the order a private producer fills, while the regulator remains purely regulatory.