Original Article: Are you asking the right questions about prehospital response times?

15 Sep Original Article: Are you asking the right questions about prehospital response times?

© 2014 MedicHealth.com

This past week a podcast by EMS Chief Chris Cebollero and critical care paramedic Kelly Grayson included a discussion on whether or not “response time goals really matter.” [1] The link to the podcast was well distributed in social media.

This isn’t the first time an EMS leader has raised questions about response time goals. Last year, the Chief Medical Officer of American Medical Response, Ed Racht, M.D. and AMR manager Lee Turpin made an evidence-based case to the National Association of EMS Officials and the year before Matt Zavadsky, Director of Public Affairs at MedStar Mobile Healthcare in Ft. Worth, made a similar case in a presentation at EMS World Expo and in an article in EMS Insider. [2] [3] [4]

Racht, Turpin and Zavadsky’s presentations are very well cited and make a compelling argument that EMS response time goals do not matter, but are we asking the right question when it comes to response times? Much of the debate centers on the response time goal of 8 minutes 59 seconds as the evidence continues to grow that the there is no causation between that goal and clinical outcomes outside of cardiac arrest. Missing from the conversation is the continued importance of patient access and reliability.

History of 8 minutes 59 seconds and 90th compliance

Much angst is aimed at the response time standard of 8 minutes and 59 seconds at the 90th, but few appreciate how the “standard” came about.

As the story goes (I welcome anyone to add/clarify the facts in the comments), in the late 1970s EMS consultant Jack Stout was designing the first Public Utility Model system in Tulsa, OK, USA. The system was to involve performance based contracting and Stout wanted to include clinical outcomes as one of the standards. EMS was in its infancy and there was very little peer-reviewed evidence related to prehospital intervention and associated outcomes.

In 1979, Mickey Eisenberg, MD published a paper in the Journal of the American Medical Association (JAMA) that concluded: if CPR was initiated within 4 minutes and definitive care within 8 minutes, cardiac arrest survival could achieve 43%. [5] With this single research paper, Stout designed an ambulance response time with the assumption that, if time mattered for saving the most critical patients, it would have a trickle down affect on other patients. The National Fire Protection Association (NFPA) standard 1710 for EMS response times today sets a less than 4 minute goal for first response and a less than 8 minute goal for the paramedic ambulance. [6]

Why 8 minutes and 59 seconds versus “within 8 minutes” as Eisenberg concluded in his paper? Eisenberg’s paper is clear to say “less than 8 minutes”, which would be 7 minutes and 59 seconds or less. The “59 seconds” is a dated accommodation allowed for EMS systems calculating their response times from manual dispatch cards that couldn’t capture minutes and seconds like automated systems. [7]

Finally, Eisenberg’s clock started with patient collapse and ended with “definitive care”. Eisenberg is not clear to define whether the clock stops when the care provider reached the patient or the actual care (e.g., defibrillation) was administered. This is important because, if Eisenberg’s time definition included the whole cycle time (collapse to care), then Stout’s original definition would have never achieved the aim.

Operational Definitions

When discussing response time measurement, it’s important to appreciate the definition used to collect and measure the data. A study of a 120 providers in Florida published in the EMS Management Journal in 2004 revealed 9 different definitions of response times. [8]

The response time definition widely adopted in American EMS systems begins when the call is received at the medical dispatch center and ends when the ambulance arrives at the call location. Stout originally defined the time initiation as when the call taker had the call back number, location, and chief complaint and ending when the ambulance arrived at the call location. [7] Stout described this as the only definition to capture the patient’s point of view. I would predict that if we asked patient’s or taxpayers when the clock starts and stops they would argue a different definition with the clock starting at 911 pickup and ending when help that can relieve their suffering arrives at their side. Stout also applied this response time goal standard to “life threatening” emergencies only. Response time goal standards for all other calls vary widely and there’s no clear expectation.

One variation frequently allowed for in most EMS systems is to have the clock begin when the EMS system receives the call, which eliminates the 911 call taker processing time that happens before transferring the call to the secondary EMS PSAP. Collecting the time stamps from 911 call pick up to receipt of the call at the EMS call center can be elusive and out of the EMS system’s control. That said, it is still part of the patient’s experience of the time.

Brief Literature Review

To this point, we’ve covered the history and definition of EMS response time, but does all of this matter. The quick answer is very much yes and very much no. Let’s begin by talking about the evidence that supports the no argument.

Several studies have looked at the relationship between response times and outcomes. The most frequently cited are two by Blackwell et al. (2002 & 2009) in Charlotte, NC, USA and one by Pons et al. in Denver, CO, USA (2005). [9] [10] [11]. In these studies the investigators reviewed the outcomes of patients with varying response times and concluded that a response time of less than 8 minutes (Pons) or less than 10 minutes 59 seconds (Blackwell) did not have an affect on outcomes. Two of the studies did note an improvement if the response time was less than 4 or less than 5 minutes.

A couple of interesting notes about these systems.

Charlotte, NC, USA had a response time goal of 10 minutes and 59 seconds that was instituted following a system redesign in the mid-1990s. The studies cited were an effort to determine if it was clinically warranted to pursue the 8 minute 59 second goal adopted by peer system’s nation-wide. Recognizing the economic investment required to reduce response times by 2 minutes, the studies helped them make an evidence-based decision. The results of their research supported remaining at their longer goal.

At the same time the Pons study was published, Denver Health Paramedics were engaged in a Lean Quality Improvement initiative that included focus on improving response time reliability. [12] I sat in on a session at the Institute for Healthcare Improvement National Forum where the work was presented and asked the Lean consultant if the Denver system adapted their response time goal based on Blackwell and Pons findings and she reported they had not.

A very common criticism of the 8 minute 59 second goal is that it doesn’t make sense to build a system around a single patient group (cardiac arrest) that is such a small fraction of the total call volume. Using Medical Priority Dispatch System (MPDS) data from four cities as an example, the 9-card (sudden cardiac arrest) averaged just 1.5% of calls. Being critical of the standard being applied to all calls is a fair argument, however out-of hospital cardiac arrest survival remains highly variable across communities and it’s estimated that annually 350,000 people die of out-of-hospital sudden cardiac arrest. [13] If EMS can achieve the 43% survival Eisenberg suggested in 1979, that would result in 150,500 lives saved per year!

We have covered some of the leading evidence that questions the validity of the 8 minute 59 second goal for EMS response times and now let’s consider why response time reliability matters. This takes us back to considering if we are asking the right questions. As the evidence above shows, in some time sensitive instances, time matters and, in the remainder of cases, quick response may not be critical. This does not however mean that response reliability doesn’t matter. Note I said “response reliability” not “response time goal.”

From the patient’s perspective, one of the key arguments for response time reliability is a commitment to reliable and timely access. The Institute of Medicine (IOM) Report EMS at the Crossroads describes EMS systems through the lens of the Six Quality Aims resulting from another IOM report called Crossing the Quality Chasm and highlights response reliability and the need to pursue patient-centered, timely, and equitable considerations.[14] [15] This means speed matters when conditions benefit, but in most other cases the critical factors are reliability and predictability.

Finally, it’s important to appreciate the public expectation. If I call for help, can I reliably expect it to arrive within a predictable time? This is where response reliability remains very relevant for EMS systems today.

Why has the 8 minutes 59 seconds response time goals stuck?

There are likely many reasons. Initially (and largely still today), it was the result of legacy consultants building the goal into competitive procurement processes for ambulance service and recommending them in their consulting reports. The private sector of the profession became an early adopter to both improve their competitive advantage but also as a symbol of quality. Later, it made its way into standards like the Commission on the Accreditation of Ambulance Services and NFPA 1710. In spite of the emerging evidence, updating the response time standards has been slow and mostly at the local level. For example, the EMSA system in Tulsa/Oklahoma, OK, USA recently extended its response time contractual goal from 8 minutes 59 seconds to 10 minutes 59 seconds becoming one of the first large EMS systems to do so. [16]

Where do we go from here with Response Times?

  • The profession needs to not just point out the problem of antiquated response time standards. EMS leaders need to follow through to lobby and make local changes too.
  • Recognize there are time sensitive conditions (cardiac arrest, respiratory arrest, traumatic injuries, hemorrhage) where time sensitive interventions are life saving and specific sub systems need to be developed to intervene quickly.
  • EMS systems require a measurement definition that truly captures the patient perspective. Ideally, that is from 911 call pick up until help that can relieve suffering is standing by the patient’s side.
  • Response time goals should be set that are evidence-based but also reasonable to the lay public. This will require public education and community input.
  • There are papers tying MPDS dispatch determinants to patient condition and the findings should help guide response plans and system design. Research should be continued to better use this data for planning.
  • Response time reliability for all patients is important and should remain part of any EMS system design.
  • We need to evolve from measuring response times to the 90th percentile and use the more sophisticated statistical process control methods (coming soon…see below)


Coming Soon.

In 1987, Jack L. Stout wrote that measuring response times, as an average, was the leading response time goal scam. He went on to make the case for measuring compliance at the 90th. [7] Stay tuned for an original article on why measuring response times using compliance at the 90th is so 1980s and how response time measurement needs to evolve.

David M. Williams, Ph.D. is a researcher, consultant, & chief executive of the international consulting firm Medic Health. He is also an improvement advisor and faculty for the Institute for Healthcare Improvement. Contact him at www.medichealth.com.