Vital signs are essential for assessing a patient's condition. But a definitive cardiovascular diagnosis requires stress testing. Some problems only can be understood by monitoring performance at various levels of physical activity.

Like everyone else, fire-department personnel undergo periods of rest and periods of extreme activity. Unlike everyone else, fire-department personnel constantly are connected to the EKGs of computer-aided dispatch and NFIRS 5 data. We need only tap this valuable data for more definitive insights into performance and staffing issues.

While static performance measurements can identify problems, diagnosing and remedying operational response issues requires an understanding of the interaction between activity and performance. The prescriptive remedies for poor performance during periods of low activity are vastly different from the remedies for poor performance during high levels of activity. Having more-definitive insight into the problem makes the solutions easier to develop and sell to stakeholders.

We begin by defining fire department activity and performance measurements. We then look at methods for evaluating the interaction between activity and performance. Finally, we'll identify prescriptive remedies for monitoring and optimizing performance in tough economic times.

Five Activity Measurements

Fire department incident activity exists in both time and space. There are five ways to quantify fire department activity for any given timeframe and location:

• Number of incidents: the number of distinct events that receive fire-department assistance.
• Number of apparatus responses: the number of apparatus committed to incidents.
• Total staff hours committed.
• Simultaneous incidents department-wide.
• Simultaneous incidents within operational areas such as battalions, station areas or districts.

Each of these activity measurements has its place. The number-of-incidents metric can be used to identify activity patterns based on demand, but it does not distinguish between simple and complex incidents. For instance, a one-ambulance EMS incident does not tax fire-department resources as much as a complex structure-fire incident. If, however, the number of apparatus responses is used to characterize activity, we get a more accurate representation of the resources consumed by incidents.

For a different perspective, staff hours can be used. Staff hours are calculated by multiplying the number of personnel by the duration of the incident. However, there's another activity dynamic that cannot be ignored.

Nothing affects staffing more than simultaneous incident load. Simultaneous incidents require separate staff for each incident. If the rate of overlapping incidents increases, then staffing demands increase. But localized staffing shortages are the most dangerous, as they exhaust staffing resources more quickly. These only can be measured by the number of simultaneous incidents that occur within operational areas smaller than the entire department.

In reality, each of the five activity measurements contributes to understanding the problem. But how can one effectively compare these radically different numbers?

Since we are interested in the rise and fall of activity levels, we don't have to be concerned about the absolute number of incidents vs. the absolute number of simultaneous incidents, for example. Rather, we can create a percentage of activity by determining where activity peaks. So in a given month, if the incident count peaks at 36 incidents per day, then 36 becomes our 100% monthly activity level. If a given day has 18 incidents, then that day can be assigned a 50% activity level, 24 incidents would equate to 67%, and so on.

Figure 1 shows the pattern of each of the activity measurements compared by hour of day. This data was accumulated for a West Coast fire department over a three-year study period. Each of the five activity measurements follows the same general trend. From midnight to 7 a.m., there is minimum activity. This is followed by a steep rise in activity from 7 to 9 a.m., then peaks and plateaus until 6 p.m., when a gradual decline begins. This continues until minimum levels are reached shortly before midnight.

While the timing may differ, most fire departments have these three activity levels:

• Minimal (midnight to 6 a.m.)
• Transitional (7 to 8 a.m. and 6 to 11 p.m.)
• Peak (9 a.m. to 5 p.m.)

These activity levels provide the resting and exertion measurements we'll need for stress-testing departmental performance.

Key Performance Measurements

Like activity, performance can be measured in time and space. Job one is getting emergency personnel to the scene. The amount of time it takes to deliver the needed resources can be expressed in seconds. The location of the incident is space. Space can be defined by station area, district, fire-demand zone and the like.

Many incidents are addressed by the response of a single company. But a significant number of incidents require the assembly of multi-apparatus teams. Both single-company and team response also must be considered when analyzing performance.

There are five key performance indicators that complement activity measurements in time and space:

• Call to arrival: Both first-company arrival and arrival of subsequent units.
• Dispatch to arrival (when call to Arrival is not available).
• Call processing: The time it takes to handle the request for assistance and dispatching of personnel.
• Turnout: The time it takes personnel to assemble on an apparatus and begin "wheels turning" to the scene.
• Travel: The time it takes for the apparatus to go from its location at the moment of dispatch to the emergency scene.

The most-definitive performance measurement is call to arrival. This measurement most closely replicates the experience of those requesting emergency services. Here's how it works.

The fire department establishes a goal for the number of seconds it requires to process a request for assistance, dispatch the correct apparatus, have the apparatus begin traveling and finally arrive at the scene of the emergency. Goals vary by local conditions, but a reasonable goal for a typical single-family structure fire would be to have this activity take place in 380 or fewer seconds — 60 seconds for call-processing, plus 80 seconds for turnout, plus 240 seconds for travel. The timing begins when a request is received and ends when the first apparatus arrives at the scene. If the fire department reacts in 380 or fewer seconds, then the incident passes. If it takes longer than 380 seconds, then the incident fails.

For this single-family building fire, a typical goal for assembling an effective team of apparatus and personnel would add another 240 seconds to allow the last assigned apparatus to arrive at the scene.

Ideally, for every incident that fails, nine or more would pass, i.e., a 90% or better compliance rate. We want to see 90% or better compliance maintained during each of the three activity phases: minimal, transitional and peak.

Figure 2 illustrates activity and performance over time. Each bar indicates the number of incidents by hour. Performance is indicated by the portion of incidents each hour that pass (red) or fail (smaller orange area) given the arrival goal of 380 seconds.

In Figure 3, we shift from time to examining both activity and performance in space. Here we look at pass and fail by district. We observe that this three-district fire department has far fewer incidents in District 2; however, there is a far greater proportion of failed incidents in District 2.

In this simple example, District 1 is a core urbanized area. District 3 also is an urban area but has a smaller geographical area with municipal boundaries on three sides. District 3 can be referred to as a peripheral-core district. District 2 is a peripheral residential area. It covers a larger geographical area with far less population density than District 1 or District 3.

Stress-Test Modeling

Stress-test modeling focuses on incidents that fail to meet performance standards. We want to know three things about these failed incidents — location of failure, activity load at the time of failure and reason(s) for failure.

In this example we'll use "District" as the location of failure, and quantify failed incidents by activity load.

A call-to-arrival goal can fail when call processing, turnout or travel time (by themselves or in combination) fail to meet department standards. In most fire departments, the same call-processing facility is used for all dispatch and communications. It's helpful to analyze this parameter separately. That distills station performance down to three essential components: call to arrival; turnout; and travel.

In Figure 4 (the table below), notice that slower turnout times can be masked by shorter travel times. In other words, time lost in turnout can be recovered in travel. Conversely, a peripheral fire station, with protracted travel times, may have a high failure rate regardless of turnout performance. What makes sense is to evaluate each station by call to arrival, turnout and travel over minimal, transitional and peak activity levels.

Overall, in this example, the fire department has a 380-second call-to-arrival compliance of 88.63%. That's very good. But Figure 4 provides opportunities to improve on even that impressive performance.

For instance, notice that all districts would benefit from improved turnout times. Also, note that District 1 fails the turnout-time test twice as often during minimal activity as in transitional or peak activity.

Travel time is the greatest problem in District 2's area; elsewhere travel time is very good.

Optimizing Performance in Tough Times

Here's a simple step-by-step plan for optimizing performance.

• 1. Make sure your fire department accurately tracks critical times for all apparatus responses
  These times include time of request, dispatch, en route, arrival, patient arrival (for EMS) and clear time. If any of these times are missing, or if your fire department only tracks the first arriving apparatus, make the necessary changes to track these critical times for all apparatus responses.

• 2. Audit call-processing times
  Use a 60-second target for receiving the call and dispatching the appropriate apparatus.

• 3. Monitor turnout time for opportunities to improve performance
  If 60 seconds can be trimmed from call-processing and turnout times, it will have the same effect as moving all fire stations half-mile closer to every emergency.

• 4. Address poor performance in minimal activity periods with training, problem-solving and close monitoring
  If poor performance has become the established routine, break the routine.

• 5. If your fire department regularly experiences a drop in performance during peak hours, define the problem
  Where are the performance drops occurring? Are they driven by localized simultaneous activity or are they part of a department-wide pattern? Slower responses are to be expected in larger districts with lower population densities. But if increased activity regularly causes performance drops in core or peripheral-core station areas, then a serious problem exists. Isolate the nature of the problem and consider redeployment options.

Whether diagnosing a medical or operational condition, the point of stress testing is to identify and correct problems before they debilitate. Stress-testing your department's performance should be a regular part of leading fire operations — especially in tough economic times.

Michael Fay is president of Animated Data, which publishes analytical software. Fay was among the original faculty of the National Fire Academy in Emmitsburg, Md., and is a veteran of the Amherst (Mass.) Fire Department. He can be reached at mike@nfirs5.com.

Related Stories

• Simple Math: 6 Steps to Understanding Performance Statistics

FIGURE 4: Incidents Failing to Meet Goals

		LEVEL OF ACTIVITY
		MINIMAL	TRANSITIONAL	PEAK	TOTAL
DISTRICT 1	Incidents	892	3,411	3,546	7,849
	Call to Arrival Fail	12% (105)	6% (200)	6% (219)	524
	Turnout Fail	89% (792)	81% (2,747)	69% (2,463)	6,002
	Travel Fail	3% (23)	3% (94)	4% (156)	273
DISTRICT 2	Incidents	357	1,267	1,097	2,721
	Call to Arrival Fail	41% (147)	27% (342)	24% (266)	755
	Turnout Fail	88% (313)	77% (970)	61% (670)	1,953
	Travel Fail	23% (81)	21% (271)	21% (234)	586
DISTRICT 3	Incidents	725	2,979	2,932	6,636
	Call to Arrival Fail	18% (130)	9% (279)	9% (269)	678
	Turnout Fail	93% (671)	83% (2466)	71% (2084)	5,221
	Travel Fail	4% (30)	5% (149)	7% (217)	396