f a l l 2 0 0 6 35

Most local governments in NorthCarolina provide ambulanceand emergency medical services

(EMS) using various arrangements.Growth in population, changes for firstresponders in the post–9/11 era, andchanges in Medicare’s fee schedules forreimbursement for ambulance serviceshave created pressures to manage thecosts and the efficiency of ambulanceservices and EMS in many counties.

This article reports on the efforts ofthree local governments and a county-supported regional medical center in LeeCounty, Alabama, to find an intergovern-mental solution to provision of these vitalservices. In Lee County, pressures to re-duce the appropriations for ambulanceservices and EMS led an ad hoc committeeconsisting of representatives of the fourentities in the county to propose fundingof a comprehensive study using industrial

P O P U L A R G O V E R N M E N T

Searching for Cost-Effectiveness in Emergency Medical ServicesDouglas J. Watson and Floun’say R. Caver

engineering techniques. The results maybe of value to North Carolina officials asthey struggle with similar cost situations.

Watson is director of public affairs programs,and Caver, assistant professor of publicaffairs, at the University of Texas at Dallas.Watson was city manager of Auburn, Ala-bama, when the events reported in this articleoccurred. Contact them at douglas.watson@utdallas.edu and fcaver@utdallas.edu.

36 p o p u l a r g ov e r n m e n t

Comparison of Lee County andNorth Carolina Counties

Lee County and its towns have impor-tant similarities with one out of five NorthCarolina counties. First, Lee County’soverall population of about 123,000 iscomparable to that of twenty North Car-olina counties with populations from90,000 to 170,000.

Second, the challenge of serving oneor two large cities and a more sparselypopulated area of county residents issimilar to that faced by seventeen NorthCarolina counties. Lee County’s popula-tion density is 189 persons per squaremile. North Carolina counties rangingfrom Pasquotank and Wayne in theeast, to Caldwell, Cleveland, and Iredellin the center and the west, have densitiesof 150–225 persons per square mile.

Finally, the overall land area of LeeCounty is more than 600 square miles,much like that in fourteen North Carolinacounties whose land area is 550–700square miles. (For the particular countiesthat are comparable on these variousmeasures, see the sidebar on this page.)

Background

Like many growing southern counties,Lee County has faced escalating costsfor the government services demandedby its increasingly urban population. As a result, local officials are interestedin finding ways to provide services moreefficiently. Lee County consists ofseveral incorporated areas and a vastexpanse of small rural developmentsand farmland. Two contiguous majorcities—Auburn, a jurisdiction of about47,000 people and the home of AuburnUniversity, and Opelika, a jurisdictionof about 24,000 people and the countyseat—make up the urban population.Opelika has stagnated for several decades,but Auburn has doubled its populationin the past twenty-five years. Thecounty and the two cities have workedtogether for two decades to providesome services, such as an airport, jails,visitors’ services, and ambulance services.The cooperation has resulted in costsavings and greater effectiveness.

EMS, however, has been difficult toprovide efficiently and effectively to LeeCounty’s diverse population. Auburn–

Opelika, the urban center, represents onlya small percentage of the land area of thecounty and is relatively easy to serve, butthe county’s sparsely populated ruralarea is spread in all four directions fromthe urban area. Smiths Station, a recentlyincorporated city of a few thousand, ison the eastern border of the county butstill depends heavily on the county gov-ernment to provide urban services.

Before 1980, several private ambulanceservices served the county, primarily thetwo cities. The mayors of the two citiesand Lee County’s probate judge (whoby virtue of that position served as chairof the county commission) consideredthe services inadequate because of theirlack of professionalism, training, and de-pendability. In the early 1980s, Auburn’smayor led an effort to improve ambulanceservices by contracting with the East Ala-

bama Medical Center, the only hospital inthe county, located in Opelika. Throughmonths of skillful negotiation and sup-port building, the mayor developed aconsensus to upgrade ambulance services.

As a result, the cities and the countydeveloped a joint ambulance service toserve the entire county from one centrallocation—the hospital. However, the twocities did not want to give up separateEMS divisions in their fire departments. In-stead, they agreed to cover emergency callsin designated rural sections of the countyat no cost to the county government.

In the late 1980s, Auburn city leadersapproached hospital administrators andOpelika officials with a plan to consoli-date the EMS operations of both citieswith the hospital’s ambulance service.Auburn policy analysts had determinedthat in more than 90 percent of the calls

Population of90,000–170,000

AlamanceCabarrusCatawbaClevelandCravenDavidsonHarnettHendersonIredellJohnstonNew HanoverOnslowOrangePittRandolphRobesonRockinghamRowanUnionWayne

Population Density of 150–225 Persons/Square Mile

BurkeCaldwellClevelandIredellJohnstonLeeLincolnNashOnslowPasquotankPittRandolphRockinghamUnionVanceWayneWilson

Land Area of 550–700Square MilesBertieBuncombeChathamDavidsonGuilfordHarnettHaywoodHydeIredellMoorePittRockinghamUnionWayne

Comparison of Lee County, Alabama, with North Carolina Counties

Source: U.S. Census Bureau, State & County QuickFacts, available at http://quickfacts.census.gov (last visited June 28, 2006).

North Carolina Counties

Lee County, AlabamaPopulation: 123,254Population of municipalities: Auburn, 46,923; Opelika, 23,498Area: 609 square milesPopulation density: 189.1 persons/square mile

f a l l 2 0 0 6 37

in which an ambulance was needed, theemergency response team was not needed.However, both responded to every call.In addition, the fire chief generally sent afire truck, and the police dispatched apatrol car. The result was four emergencyvehicles arriving at the home of a heartattack victim when only an ambulancewas needed. Auburn city leaders ques-tioned the practice of sending fire trucksto every emergency scene.

The fire chief in Opelika convincedthe mayor and other city leaders thatthe consolidated service proposed byAuburn city leaders would lead to alessening of service in his city. Althoughthe chief’s argument was not based onanalysis, the elected officials were notwilling to override him.

As a result, Auburn proceeded aloneto contract with the hospital for a con-

solidated EMS systemthat included both am-bulance services andemergency response.Auburn’s ten em-ployees (all para-medics) were given theoption of joining thenew EMS organization at the hospitalor remaining as firefighters with the fire department. Most of the paramedicsaccepted employment with the hospital.A few opted to remain with Auburn as firefighters. Auburn’s immediatesavings in personnel costs were about$350,000 per year.

Further, Auburn provided space inone of its fire stations for the hospital’sEMS crew so that an ambulance couldbe stationed in the city for the first time.This action eliminated a five-mile com-

mute for the ambu-lance from the hospitalto downtown Auburnand lessened responsetime by 5–7 minuteson most calls.

By the late 1990s,the hospital’s manage-

ment of the combined EMS–ambulanceservice was judged to be a success, withsignificant savings accruing to Auburn.Opelika continued to provide duplicateservice, even though its budget wasstrained.

In 2002, Medicare established a newfee schedule for reimbursement for am-bulance services. They now were to bepaid on the basis of the type of call, witha separate charge for mileage.1 Also, thepayments required by Auburn andOpelika to subsidize the county ambu-

In more than 90 percent of 911 calls, an ambulance andthree other emergency vehicleswere sent when only anambulance was needed.

Figure 2. Probability of Simultaneous Deployment for Emergency and Nonemergency Calls Combined

Source: From Jerry A. Davis et al., Evaluating Emergency Medical Services: Controlling the RisingCost of Saving Lives, 26 JOURNAL OF HEALTH AND HUMAN SERVICES ADMINISTRATION 485 (2004). Reprintedby permission.

Note: The five reserve units are included, for a total of ten.

Figure 1. Demand, in Two-Hour Intervals, for Emergency and Nonemergency Calls Combined

Source: From Jerry A. Davis et al., Evaluating Emergency Medical Services: Controlling the RisingCost of Saving Lives, 26 JOURNAL OF HEALTH AND HUMAN SERVICES ADMINISTRATION 485 (2004). Reprintedby permission.

38 p o p u l a r g ov e r n m e n t

lance service had increased by 60 percentin ten years and were projected to rise20 percent over the next five years.2

As a result, Auburn officials suggesteda comprehensive study of the EMS–ambulance service to see if it could bemade more efficient. At the annualmeeting of the ad hoc oversight commit-tee, Opelika and Lee County represent-atives readily supported the Auburnofficials’ suggestion, as did the hospital

administrators. Each group agreed topay one-fourth of the cost of the study.

In response to a request for proposals,the Industrial and Systems EngineeringDepartment at Auburn University sub-mitted a proposal to conduct the study.A research professor and two graduatestudents would be the principal investi-gators. In the previous year, the samegroup had conducted an in-depth analy-sis of the Auburn Police Department thathad been helpful to the city managerand the city council in responding to thepolice chief’s demands for extra staff.3

After discussing all the proposals, thedecision-making group decided to con-tract with the university.

The Efficiency Study

The study commissioned by the localgovernments was coordinated by the adhoc oversight committee, which in-cluded the mayor, the city manager, andthe public safety director of Auburn; themayor and the fire chief of Opelika; theprobate judge of Lee County; andhospital administrators. Working withthe committee, the Auburn Universityresearchers defined the study’s keyobjectives as follows:

• To observe and document thecurrent EMS delivery method

• To develop key indicators of systemperformance

• To evaluate and interpret the keyindicators

• To identify and quantify expensesand revenue associated with theambulance service

• To identify areas of possibleimprovement and to quantifyassociated cost savings

• To discuss alternative methods ofproviding the ambulance servicesand EMS4

The ad hoc committee and the re-searchers agreed on a research method-ology that would center on proven in-dustrial engineering techniques, such ashistorical data analysis and elementaltime studies. Historical data analysis isused to document and quantify currentdemand patterns. Elemental time studiesare used to study repetitious job duties,breaking them down in order to define

Table 1. Emergency Service Demandand Unit Allocation, by Site

Percentage Percentage

Site of Calls of Units

Opelika 41 40

Auburn 38 40

Smiths Station 21 20

Total 100 100

Table 2. Time Study Results

EMTs Supervisors Billing Office Dispatchers

Productive work 21% 67% 88% 22%

Additional work 7 8 4 3

Voluntary idle [time] 70 25 8 74

Involuntary idle [time] 2 0 0 0

Source: From Jerry A. Davis et al., Evaluating Emergency Medical Services: Controlling the RisingCost of Saving Lives, 26 JOURNAL OF HEALTH AND HUMAN SERVICES ADMINISTRATION 485 (2004). Reprintedby permission.

| | | | | | | | | | | |||

||

||

|

||

||

|| 100.00%

80.00

60.00

40.00

20.00

0.00

0 1 2 3 4 5 6 7 8 9 10

Rel

ativ

ePr

ob

abili

ty

Cu

mu

lati

vePr

ob

abili

ty

Number of Units

| | | | | | | | | | | | |||

||

||

||

||

|15.00%

10.00

5.00

0.00

Perc

enta

ge

100.00%

80.00

60.00

40.00

20.00

0.00

Cu

mu

lati

vePe

rcen

tag

e2 4 6 8 10 12 14 16 18 20 22 24

Hour of the Day

60.00%

50.00

40.00

30.00

20.00

10.00

0.00

f a l l 2 0 0 6 39

key elements and recommend cost-savingchanges. The merging of historical dataand observational data (elemental timestudies) offers a basis for an accuratedescription of a work environment andfor quantifiable, valid, and realistic recom-mendations.

Historical Data AnalysisTo begin the data analysis, the researchersfirst had to identify and choose potentialdata sources. In EMS, important data arecollected in run logs, patient care reports(PCRs), and insurance documents, amongothers. For this study the researchersdecided that the PCRs would be theprimary data source. The PCRs containdata on time of dispatch, time of arrivalon scene, time of departure from scene,time of arrival at destination, and severitytype (emergency versus nonemergency).5

From the PCRs the researchers calcu-lated and analyzed the following metrics:

• Dispatch time—the time elapsedfrom call receipt to dispatch

The results of the analysis of theefficiency of allocations indicated abouta 40- 40-20 split of human resourcesand equipment among Opelika, Auburn,and Smiths Station (see Table 1).7 Thesimilarity between site demand andequipment dispersion suggested that theequipment was efficiently allocated.

The peak-demand analysis captured,in two-hour intervals, the number andthe percentage of emergency and non-emergency calls received, by time of day(see Figure 1). In military time, peak de-mand was between 1000 hours (10 A.M.)and 2000 hours (8 P.M.). This findingresulted in the researchers recommend-ing reduction of coverage at Stations 1and 2 by one unit during the night shift(midnight to 6 A.M.). By doing so, thesystem could realize a savings per stationof $82,600, resulting in a total savingsof more than $160,000 per year.

The researchers conducted similardemand analyses for days of the weekand months of the year. For emergencycalls, the results indicated that Friday

• Service time—the total time fromdispatch until return to service

• Time spent on scene—thedifference between the arrival timeand the departure time

• Call volume and severity type by site

• Time of call per twenty-four-hourday

• Number of units simultaneouslydeployed

These performance metrics allowedthe researchers to analyze service effi-ciency, the efficiency of the allocation ofhuman resources and equipment, andscheduling (peak-demand) efficiency.Findings for service efficiency were amean emergency response time of 8.6 minutes and a mean service time of34.0 minutes. Although the responsetime was slightly higher than Ammons’sfindings on response time for cities withpopulations greater than 100,000, thead hoc committee deemed both metricsacceptable.6

40 p o p u l a r g ov e r n m e n t

and Saturday experienced the highestdemand, with the other days receivingrelatively equal demand. For nonemer-gency calls, Friday again experiencedthe highest demand, and Sunday experi-enced the lowest demand. As for monthsof the year, September and Octoberwere the months of highest demand foremergency and nonemergency calls com-bined.8 (Each fall, more than 90,000 stu-dents return to Auburn University, andeven more people attend the AuburnUniversity football games.)

Bureaucrats and politicians are proneto overestimation or neglect of cuesfrom the environment because ofhuman beings’ inability to process anddecode such cues efficiently.9 The resultis inefficient decisions. In EMS one of

the strongest arguments for increasedstaffing levels centers on the possibilityof simultaneous deployment. Thisargument puts too much weight on thelikelihood of a crisis. Thus an importantanalysis for local EMS and ambulanceservice providers is the probability ofsimultaneous deployment.

The Auburn University researchers’analysis of the relative and cumulativeprobabilities of multiple-unit deploymentin Lee County showed about a 50 per-cent probability that no unit would bedeployed at any one time, about an 80 percent probability that one unit or less would be deployed, and a morethan 90 percent probability that twounits or less would be deployed (seeFigure 2). This finding indicates thatsimultaneous deployment, althoughpossible, should not dominate argu-ments for continuing inefficient staffingarrangements. In the post–9/11 environ-ment, such a metric also should be usedby other emergency services to ensurean informed discussion of emergencyresponse staffing.

Elemental Time StudiesThe ad hoc committee insisted onmoving beyond historical data analysis.The region had been changing quickly,so the committee believed that observa-tional data were needed to make informeddecisions. As a result, the researchersconducted “shift time studies,” in whichthe researchers spent entire work shiftswith the employees, studying major job functions. They performed about405 hours of observation, 325 of whichwere dedicated to observing the work ofemergency medical technicians (EMTs).The time studies focused on identifyingand analyzing the job functions of EMTs,dispatchers, supervisors, and billingpersonnel. The job tasks were identifiedas productive work, additional work,voluntary idle time, and involuntary idletime. “Productive work” was defined aswork directly related to EMS (commu-nications, administration, travel, patientcare, and dispatch); “additional work,”as station and unit cleanup and miscel-laneous tasks; “voluntary idle time,” astime for sleeping, personal hygiene, andsocializing; and “involuntary idle time,”as idle time over which the individualbeing studied had no control.10

EMS staff and equipment were situ-ated throughout Lee County: 2 units atStation I (Opelika), 2 at Station II (Au-burn), and 1 at Station III (Smiths Station).A unit was normally staffed by twoparamedics (EMT– Ps), the highest EMT level and highest pay grade. The 2002 workforce consisted of 3 dis-patchers, 5 supervisors, 2 billing per-sonnel, and 51 EMTs, 84 percent ofwhom were EMT– Ps, 12 percentEMT– II’s (intermediate), and 4 percentEMT–I’s (basic).11

The time studies identified severalareas of opportunity in staffing alloca-tion and job functions. For example,units were normally staffed by EMT– Ps.This practice created units with maxi-mum costs. Thus, one recommendationwas to move to a tiered response systemrelying on units with mixed EMT levelsand pay grades—say, an EMT–P, anEMT– I, and a driver, instead of twoEMT– Ps and a driver. By diversifyingthe units, the system could save morethan $100,000, depending on the costto hire new EMT– I’s or EMT– II’s.

Savings also could come from com-bining the job functions of various officepersonnel. The productive time workedby the billing office, the supervisors, andthe dispatchers suggested an interestingopportunity (see Table 2). The billing office used 88 percent of its time in pro-ductive work. However, the dispatchersspent 77 percent of their time in unpro-ductive work (additional work plusvoluntary idle time), and the supervisorswere involved in voluntary idle tasks 25 percent of the time. Improved effi-ciency could result from the supervisorsand the dispatchers using their voluntaryidle time to perform billing tasks, eliminating or reducing the need for thebilling office. Also, merging the ambu-lance dispatch service with existing local government dispatch operations in Auburn and Opelika might result in savings.12

The IntergovernmentalCooperation

Between 1980 and 2000, Auburn andthe East Alabama Medical Center ex-perienced a successful intergovernmentalrelationship in providing consolidatedEMS and ambulance service. Auburn

f a l l 2 0 0 6 41

realized an immediate savings of about$350,000 per year, and the hospitalreceived additional revenue with just asmall increment in workload (becauseonly about 10 percent of ambulancecalls required rescue equipment).

However, Opelika was reluctant toenter the cooperative, resulting in dupli-cate services and budget inefficiency.Opelika’s hesitance was predicated onoverestimation of worst-case occur-rences, in particular on the judgmentthat a combined operation would resultin reduced service to Opelika.

In light of this, in 2002 the ad hoccommittee acknowledged the continuingneed for the several entities to work to-gether in order to provide a more effi-cient ambulance service. Although mostof the findings of the study supported thecurrent personnel structure and stationlocation, the committee requested thatthe hospital administrators implementseveral recommendations to forestallfuture increases in the annual operatingbudget supplement paid by the localgovernments. Specifically, the committeeasked that the number of personnel bereduced during the early morning hours,when call volume was at its lowest.

However, the ad hoc committee wasreluctant to implement recommendationsthat required true cooperation. Forexample, the study identified potentialsavings from consolidation of the hos-pital’s ambulance dispatch services withAuburn’s and Opelika’s existing dispatchservices. For various reasons related totrust and turf, the committee did notcall for this consolidation. As a result,the potential savings were not realized.

Through their analysis of the Lee Coun-ty ambulance services, the researchersidentified the following enablers andconstraints to intergovernmental coop-eration:

Enablers

• The willingness of all governmentsto cooperate for a workable solu-tion that would benefit all parties

• The use of independent andreputable consultants to obtainobjective recommendations onservice efficiency

• The researchers’ regular reportingto the ad hoc committee on their progress

Constraints

• Perceptions of a power imbalanceamong the parties (representativesof one government sensing that rep-resentatives of another governmentwere having too much influence onthe outcome of the study), in anotherwise competitive environmentamong the governments

• A fear that one party would benefitat the expense of the others

• Personal dislike or distrust basedon previous efforts to cooperate

Conclusion

An interesting element of the study wasits use of industrial engineering tech-niques such as elemental time studiesand historical data analysis to evaluatethe county’s EMS structure. These kindsof techniques are rarely used in thepublic sector, despite constant pressureon local government officials to be more efficient.13 Objective data-collec-tion measures (time studies) provide an accurate representation of the system,and they serve as a solid foundationfrom which to continue generatingrecommendations for improvement.

An unusual outcome of the study was the hospital administrators’ conclusionthat the EMS–ambulance service didnot generate enough money or good will for the hospital. As a result,they encouraged the ad hoc committeeto consider alternative means of pro-viding the service. Since 2003 the com-mittee has met occasionally and studiedvarious options, including privatization.As of this writing, serious discussionsare under way about alternative ar-rangements beyond privatization,including creation of an EMS specialdistrict directly supported by the tax-payers. This development suggests awillingness to consider a long-termarrangement that would institution-alize intergovernmental cooperation in order to provide efficient ambulanceservices. Moreover, it illustratesawareness that competition amongmultiple service providers, public orprivate, might not be efficient and thatan EMS district (a service monopoly)might produce the most cost-effectivestructure.

Intergovernmental cooperation in aLee County EMS will likely result in amore rational and cost-effective ap-proach to the performance of this vitalservice. In a more stable politicalenvironment, the study might haveresulted in immediate implementationand savings. However, in a situationinvolving three local governments, ahospital, and rural interests, developinga long-term solution is more difficult.Regardless of future steps, the detailedstudy of the EMS operations provides asolid basis for decision making.

Notes

1. Section 4531(b)(2) of the BalancedBudget Act of 1997 added Section 1834(1)to the Social Security Act. It mandated anational fee schedule for ambulance servicesperformed under Medicare Part B. The feeschedule went into effect on April 1, 2002.See Medicare Program; Fee Schedule forPayment of Ambulance Services and Re-visions to Physician Certification Require-ments for Coverage of Nonemergency Ambu-lance Services, 67 Fed. Reg. 9100 (2002) (tobe codified at 42 CFR pts. 410, 414).

2. Jerry A. Davis et al., EvaluatingEmergency Medical Services: Controlling theRising Cost of Saving Lives, 26 JOURNAL OF

HEALTH AND HUMAN SERVICES ADMINISTRA-TION 485 (2004); Jerry A. Davis, LEE

COUNTY EMS EFFICIENCY STUDY, PARTS I &II. TECHNICAL REPORT (Auburn, Ala.:Auburn Univ. College of Eng’g, 2002),available at City of Auburn Public Library,749 East Thach Ave., Auburn, AL 36830(Reference 352.9 LEE).

3. Douglas J. Watson et al., Use ofIndustrial Engineering in Measuring PoliceManpower: A Small City Case Study,26 PUBLIC PERFORMANCE & MANAGEMENT

REVIEW 132 (2002). 4. Davis et al., Evaluating Emergency

Medical Services.5. Id.6. DAVID N. AMMONS, MUNICIPAL

BENCHMARKS (2d ed., Thousand Oaks, Cal.:Sage Publications, 2001).

7. Davis et al., Evaluating EmergencyMedical Services.

8. Id.9. See BRYAN D. JONES, POLITICS AND

THE ARCHITECTURE OF CHOICE: BOUNDED

RATIONALITY AND GOVERNANCE (Chicago:Univ. of Chicago Press, 2001).

10. Davis et al., Evaluating EmergencyMedical Services.

11. Id.12. Id. 13. Watson et al., Use of Industrial

Engineering.