Being in favor of delivering ambulance service in well-equipped units that usually arrive on scene within a few minutes of an emergency is a bit like saying you want the best heart surgeon in the land performing your triple-bypass. What’s there to argue about? However, beyond the basic “motherhood and apple pie” premise, almost everything about response times can be debated. Standards, scientific evidence demonstrating efficacy, measurement methods and how EMS organizations can improve response times are all complex issues.
Because it has the appearance of objectivity, response time is the single key operational measure used to assess system performance from the citizen’s perspective. Yet this seemingly straightforward measure is frequently misunderstood, calculated in ways that make benchmarking difficult and often not independently validated.
This article debunks the myths about response time and discusses the important issues your service should be concerned with when recording and using response times.
Response time standards ƒ an urban myth?
No universally accepted response-time system requirement exists. However, in urban areas, the most widely used ambulance response-time standard is eight minutes and 59 seconds (8:59), with 90% compliance reliability ƒ measured on a fractile, not average, basis. (Fractile response time measurement includes a reliability factor and measures all time intervals between the time the service received enough information to initiate a response and the time a properly equipped and staffed ambulance arrives on scene.)
This 8:59 target is consistent with the response time recommended by the National Fire Protection Association when adjusted to include call-processing time. (NFPA 1710, _18.104.22.168.2, requires communities to “provide for the arrival of an ALS company within an eight minute response time to 90% of incidents Ú” The NFPA standard excludes call-processing time intervals.) In addition, a recent survey of the 200 most populous cities in America found that more than three-quarters (77%) of the survey respondents report a target of 8:59 or less using the fractile measurement method (February 2005 JEMS).
Response times are confusing. I recently had to explain the 8:59 benchmark to a group of Chinese fire officials discussing options to reduce response time. One observed, “8:59 is very tricky. It’s not really eight minutes. It’s only one second from nine minutes.” I agree and believe we should begin referring to 8:59 as “<9/90” (meaning nine minutes and zero seconds, with 90% reliability). We delude ourselves that 8:59 is truly an eight-minute benchmark.
The “standard” was derived while cardiac arrest survival was being studied in the early 1970s by Alvarez and Cobb in Seattle. They noted dramatic increases in survival, which they attributed to at least three factors: reduced response times of less than eight minutes, first responders performing CPR and citizen CPR training. Later, Mickey Eisenberg, MD, et al catapulted Seattle’s results to national attention by publishing several landmark studies. The 59 seconds were added as an accommodation for punch clocks before computers were widely used in communications centers. These studies resulted in countless systems being designed during the past four decades to deliver response times to all emergency cases within nine minutes.
But cardiac arrests represent a small portion of EMS system responses. What about other conditions, such as drownings, traumas and strokes? Little research has been published that formally demonstrates that fast response times save lives in those circumstances.
It doesn’t take a PhD to discern that, despite the lack of definitive clinical evidence, outcomes are often positively influenced by EMS systems that consistently deliver a predictable response time. In the simplest of terms, when you’re thirsty, you don’t need to know that the scientific composition of water includes organic and inorganic materials, dissolved gases, nutrients, inorganic salts, trace elements, heavy metals and potentially toxic chemicals. You simply turn on the tap and expect a response. The same is true with EMS. Detailed clinical research isn’t needed to know what 9Ï1Ï1 callers expect.
EMS customers universally perceive value in a rapid intervention for loved ones. Long response times often negatively affect scene interactions and perceptions of care. And they are frequently the cause of public complaints. In the absence of a regulatory or defined system standard, court proceedings provide ultimate guidance. In litigation, experts often cite 8:59 as the defined community standard of care.
One size does not fit all
As a designer of EMS systems, Fitch & Associates [the author’s firm] recommends urban transport ambulance response time benchmarks ranging from 8:59 to 12:59, depending on a variety of carefully considered system factors. These include use of a protocol-driven medical priority dispatch system (MPDS) consistent with International Academies of Emergency Dispatch standards, active medical supervision and vigilant quality control, the availability of ALS first response, and other demographic and economic factors unique to the local community.
In suburban and rural areas and for non-life-threatening calls it’s common to set a slightly longer target response time on the basis of geographic and call-density patterns.
With response times, one size clearly does not fit all. Well-meaning personnel have expanded the benchmark for life-threatening emergencies set at 8:59/90% (e.g., calls prioritized as Echo and Delta level calls under MPDS) to include provision of that same level of response to all 9Ï1Ï1 requests. Some would cynically argue that this has been done primarily to increase call volumes. Use of a single priority response to every 9Ï1Ï1 request is like saying that all hospital patients need a CT scan or an enema on admission, regardless of condition. That doesn’t make sense. Neither does sending a maximum-level response to a person choking and to a patient who suffered a ground-level fall without any loss of consciousness or bleeding ƒ just because they called 9Ï1Ï1. For less critical calls, a lights-and-siren response isn’t necessary. An unnecessary hot response by first responders and ambulances is dangerous, and often deadly, for both responders and civilians.
The reason for standards
Why are response time standards a good idea? The core reason for establishing a standard or target is to measure performance. Ideally, it gives decision makers an objective mechanism to know when additional resources are required. It also allows departments to meaningfully benchmark performance with similar agencies.
Research is demonstrating the importance of also determining the time it takes EMS to physically reach the patient’s side. Most systems measure only up to wheel stop on scene. However, vertical response times (up steps and elevators) are a significant issue in urban areas. According to data published in the Canadian Journal of Emergency Medicine, “The vertical response interval represents a significant component of the total EMS response time and is significantly longer for calls originating in apartment buildings.” This same finding was documented in a New York study, which reported, “The scene-to-patient response time is substantial in a large urban area, particularly in office or residential buildings.”
So, what is the optimal response time? Recent research by Blackwell and Pons indicates that to change clinical outcomes paramedic response times would have to be consistently reduced to less than five minutes. Economically, this standard is simply not feasible in most communities.
To improve clinical performance with the resources at hand, EMS systems need to identify a response-time target, objectively measure that benchmark and improve specific response components.
The method by which the response time is measured is as significant as the actual standard or target. EMS has been hampered by the lack of a standard way to measure response times. Historically, averages were used. Reporting response times as an average means taking a designated period, adding all the response times and dividing by the number of responses. However, the use of an average response time doesn’t represent an accurate picture of response time. In contrast, the fractile method uses a specific response goal for a type of call and a percentage that represents the level of compliance in meeting that goal.
The only correct way to measure a response time is from the patient’s perspective. Many private and public agencies report only the times under their direct control as their response time, often excluding dispatch and mobilization time. Others report only travel time. A volunteer agency recently reviewed tried to exclude both the three-minute call-taking/dispatch time and the seven-minute time responding to the station from their overall “response time”.
First responder and transport ambulance times should be maintained and reported separately. The time to patient should be captured, even if not currently reported as an element of the system response time. This information aids in system evaluation and improvement.
Determining what should be included and excluded from response time calculations also has a significant effect on compliance to standard. More than 25 specific factors must be considered to determine if response time reports between systems are comparable, and failing to consider these factors can skew reported performance.
For example, are call-taking and “out of chute” times counted? Calls cancelled en route excluded? Multiple units responding to the same incident counted? Do changes in call priority restart or otherwise affect the response time (e.g., upgrading a call from a sick person to a cardiac arrest, downgrading a reported school bus wreck with multiple victims to a minor fender bender without reported injuries)? Do BLS or ALS first responders stop the clock? When the system experiences unusual demand or overload, are specific response times excluded?
5 fundamental strategies
Although many approaches and methods can be suggested to improve response times, the following five are essential.
1. Match supply and demand: Historical analysis of where and when calls occur can assist in predicting call demand. Adding resources that match demand patterns rather than in full 24-hour shifts is more effective. Best practice in busy, high call volume systems is to plan resources, support services and training around demand rather than using a static shift pattern. These systems also focus on resource deployment and tailor re-deployment strategies to improve coverage as calls ebb and flow throughout a given shift without abusing staff.
2. Manage component times and lost unit hours: The importance of setting internal component targets (see Figure 2, p. 50) and measuring performance against those benchmarks cannot be over-emphasized. Like the dials and gauges on a car’s dashboard, component times help leaders and workers monitor current performance, speed and direction for the system.
Among these internal component targets, call-taking and dispatch times, out-of-the-chute and off-load times at hospitals are typically problematic areas and need routine monitoring. In Canada and some parts of the United States, hospital delays have reached epidemic proportions. For example, for the Region of Peel, just north of Toronto, hospital delays increased 210% during the six-year period from 1996Ï2002. The net effect was reducing available coverage hours by two full ambulances each day.
Clearly, lost unit hours rob your system of coverage. Examples include units without a full crew, non-roadworthy ambulances, out-of-service meetings, on-duty training and restocking or other equipment changes that remove the unit from the coverage plan.
Monitoring subprocess component times and externally validating response times are critical to improving service. Routinely reporting response time exceptions, the reasons for those times and any corrective actions taken focuses energy on improving performance. Using an independent auditor to routinely proof a random sample of response times via computer records and 9Ï1Ï1 tapes is also recommended.
3. Don’t run hot when it’s not required: Having a single level of response, such as in Hong Kong (see sidebar, p. 52), is neither an effective nor a safe strategy. Using a defensible priority dispatching system and medically controlling when first responders and ambulances respond with lights and siren are reasonable ways to optimize outcomes and follow the Hippocratic Oath to first do no harm. If you remain unconvinced, review recent emergency vehicle death statistics.
4. Harness technology and innovations wisely: True progress requires working with people, refining work processes and fully using technology. Available and emerging technology is mind blowing. Geographic Information Systems (GIS) and automatic vehicle location (AVL) should be required in most areas to facilitate sending the right vehicle to the right call and in the right response mode.
Managing coverage is complex, and dispatching in busy systems can become counterintuitive. Software that shows relationships between calls, available resources and the best routing at different times of the day is inexpensive compared with paying for additional unit hours.
Getting information recorded by medics into a format that meets the needs of receiving hospitals, quality improvement and research staff, billing departments and myriad other users who need this information is an ongoing problem. Current methods include entering handwritten reports into a database, scanning input of paper forms, dictating reports to a central call documentation operator and using laptop computers in ambulances. Rapid processing, data interaction with separate billing systems and paperless medical-record retrieval are moving agencies toward new scanning and laptop solutions.
5. Be fully accountable and transparent: To be sustainable over time, emergency service organizations need to be held accountable for performance, be fully transparent and have functioning external oversight.
Historically, the term transparency has rarely been used in the context of EMS organizations. In other disciplines, such as architecture, engineering and technology, transparency is commonly understood and used. In current management literature, various definitions exist for organizational transparency, one of which really makes the point: a condition opposite of secrecy. Transparency is the deliberate attempt to move from a secretive or opaque organization to one that encourages open access to information, participation and decision making, which ultimately creates a higher level of trust among stakeholders.
The importance of EMS response times is no myth. It’s common sense. 9Ï1Ï1 callers expect a response as reliable (and almost as quick) as opening the tap for water. The actual time benchmark used by systems is not as important as consistently and accurately measuring performance. Matching supply and demand, improving component times, responding appropriately, harnessing technology and ensuring transparency are all essential to improving response time performance.
In many parts of the world that don’t have access to modern EMS systems, response times are typically measured in hours or days, not minutes. For countries that do have modern systems, specific response benchmarks vary widely.
North America: For life-threatening emer-gencies, providing a transport-capable unit within 8:59 with 90% reliability is the most common urban benchmark. Common rural and wilderness benchmarks are within 15/90% and 30/90%, respectively. Response time targets have been shaped by position papers published by the American Heart Association, the American College of Emergency Physicians, the National Association of EMS Physicians, the National Fire Protection Association and the Commission on the Accreditation of Ambulance Services. Reporter and former paramedic Bob Davis authored a hard-hitting 2004 series in USA Today that told the truth about emergency transports and caught many EMS leaders and elected officials by surprise. According to Davis, “Los Angeles is one of many cities that routinely lie to themselves about their true response times to medical emergencies. The result is needless deaths.” His 18-month investigation revealed that more than 1,000 “savable” lives are lost needlessly each year in the nation’s biggest cities because of inefficiencies in the cities’ emergency medical systems. USA Today’s investigation found that the root of these inefficiencies, in many cases, is the simple matter of counting seconds. “Cities that rely on these imprecise response times leave precious minutes unaccounted for.”
United Kingdom: A national standard (including both urban and rural areas) was established in 2001. The most critical emergency calls, referred to as “Category A” calls, have a response time requirement of eight minutes and zero seconds, with a 75% compliance requirement, and the additional stipulation that 95% of these calls must be reached within 14 minutes in urban areas and 19 minutes in rural areas. (Category A in the U.K. approximates American Echo- and Delta-level calls under MPDS. Non-life-threatening calls in the U.K. have an established target of 14 [urban] and 19 [rural] minutes without response time compliance percentage requirements.) Since 1999, the U.K. has invested more than $90 million U.S. into improving ambulance services. A recent investigation by the BBC indicated that although 26 of the 31 ambulance services reported meeting the targets, wide discrepancies remain as to when the clock was started. One official was quoted as saying, “We are cheating the system. We need to be open and honest about how long it takes.” U.K. Health Minister Lord Warner denied that the situation was as dire as the BBC investigation suggested. “In any human system, there will be bits of frailty, but all ambulance services have improved in the last few years, and that has saved lives.”
Australia: Response times are typically measured at 5-, 10- and 15-minute intervals. Statistics from several large services indicate that compliance to the 10-minute standard is approximately 50%. In other words, response times average 10 minutes for both urban and rural areas. In urban areas, times are slightly faster. For example, in Sydney, response-time compliance to the 10-minute standard is reported at 55%.7 Compared with the 90% North American fractile benchmark, response times in Australia would likely range between 14 and 15 minutes. Herald Sun reporter Kate Jones wrote May 30, 2005, that sick and dying Victorians are waiting well beyond scheduled response times ƒ even up to 47 minutes for ambulances to answer emergency calls. “Lagging ambulance response times are putting lives at risk and eaving ill patients in limbo. Freedom of Information documents show that more than 2,000 Victorians waited for at least 16 minutes for an ambulance, including 88 who waited for half an hour or longer.”
Hong Kong: Response times are described as a “performance pledge.” The current pledge is to arrive within 12 minutes for all types of requests with 92% reliability. Heart attacks and toothaches get the same response commitment. Hong Kong’s 7 million residents reside in an area of about 500 square miles. The system responds to almost a half million calls annually. It’s strained by the increased demand for service, and vertical response times (currently not measured) are lengthy. Press reports that the Hong Kong Fire Services Department had not met its performance pledge led to a recent review of dispatch processes, changes in scheduling and the study of demand management principles.
- Williams DM. “2004 JEMS 200-city survey: A snapshot of facts & trends to allow you to make benchmarks for your service.”. JEMS 2005;30:42-60.
- Page JO. The Paramedics: An Illustrated History of Paramedics in Their First Decade in the U.S.A..„www.jems.com/paramedics/ch1e.html. Accessed July 2005
- Silverman RA, Galea S, Blaney S. “Vertical response time: The hidden component of ambulance response time.” Academic Emergency Medicine.„www.aemj.org/cgi/content/abstract/12/5_suppl_1/142. Accessed July 2005
- Blackwell TH, Kaufman JS. “Response time effectiveness: Comparison of response time and survival in an urban emergency medical services system.”. Academic Emergency Medicine 2002;9:288-295.
- Pons PT, Haukoos JS, Bludworth W. “Paramedic response time: Does it affect patient survival?”. Academic Emergency Medicine 2005;12:594-600.
- BBC News. “Ambulance call figures questioned.”.„http://news.bbc.co.uk/1/hi/health/4113624.stm. Accessed July 2005
- Ambulance Service of New South Wales. Year in Review. (Figure 6).„www.asnsw.health.nsw.gov.au/docs/workload.pdf. Accessed July 2005„
Jay Fitch, PhD, is the founder of Fitch & Associates, a leading resource for EMS/public safety organizations (www.fitchassoc.com). The firm provides strategic support, and independent review and compliance services for EMS organizations.