Dialing Down the Delta
2016; Elsevier BV; Volume: 68; Issue: 6 Linguagem: Inglês
10.1016/j.annemergmed.2016.10.004
ISSN1097-6760
Autores Tópico(s)Cardiac Arrest and Resuscitation
ResumoWhen a few taps on a smartphone can bring a pizza, a ride, or a map displaying happy hours within a walkable radius, some telephone users are wondering why they can't summon emergency medical services (EMS) just as conveniently. The information technology revolution has not bypassed the 911 system—far from it, say national experts in emergency response communications—but to date it has failed to make 911 as reliable as many citizens unwittingly assume it is. The gap between potential and observable performance in geolocation technology used for emergency dispatching is substantial; it varies widely among local systems and is difficult to measure precisely in minutes, yet it is known to affect clinical outcomes, including mortality. Any “disruptive technology,” as coined and popularized by Harvard business scholar Clayton M. Christensen, MPhil, MBA,1Christensen C.M. The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail. Harvard Business Review Press, Brighton, MA1997Google Scholar can be bad news not only for users and vendors of the products it outperforms and replaces but also for their supporting infrastructure. Such processes have rarely been orderly. In the case of mobile telephones, now used by more than 90% of the US adult population and described by a Pew Research Center commentator as “the most quickly adopted consumer technology in the history of the world,”2Rainie L. Cell phone ownership hits 91% of adults. 2013. Available at: http://www.pewresearch.org/fact-tank/2013/06/06/cell-phone-ownership-hits-91-of-adults/. Accessed September 27, 2016.Google Scholar one casualty of this transition has been the precision of emergency response systems in pinpointing the source of a 911 call. For EMS and for emergency departments, the mismatch between contemporary telephones and 911 location methods generates disturbing delays during the out-of-hospital period. Robert E. O'Connor, MD, MPH, chair of the board of the American College of Emergency Physicians, recognizes that automated location of a cell telephone is a problem. “With the landline phones, it was always easy because you could search an address to the phones, so when the caller ID came up, the dispatcher would know exactly where the call was coming from. With mobile phones, it's complicated because the person could be anywhere…. From what I know, there is no foolproof system” for determining the position of a mobile telephone. Global positioning system (GPS) hardware is standard on contemporary smartphones, a useful advance for those using such telephones, Dr. O'Connor noted, provided the user has activated it, which is not always the case. (Curiously, he observed, GPS deactivation rates are high in Silicon Valley, although he did not speculate on what this implies about technology-industry workers.) GPS allows tracking by satellite, augmenting or replacing triangulation according to cellular towers' signal strength, a process that can be wildly inaccurate when high call volume causes signals to be rerouted to more distant towers. Yet GPS is no magic bullet: positioning of a call placed from indoors, or from many dense urban sites, is difficult for GPS and similar global navigation satellite systems using microwaves, which are scattered by building components, including walls, roofs, and metal-coated glass. More recent approaches use WiFi access points and Bluetooth beacons, which can provide detailed indoor locational information but raise other issues involving security and privacy. Other technologies offer the possibility of indoor positioning but have encountered resistance from large telephone carriers. Location may be the most important aspect of EMS dispatch, Dr. O’Connor said, but it is just part of a full spectrum of activity. In time-sensitive clinical situations such as possible cardiac arrest, stroke, or blood vessel laceration, he noted, fast automated location helps a dispatcher obtain further information for first responders and in some cases guide patients or bystanders in performing prearrival measures such as cardiopulmonary resuscitation. Anecdotally, he reported, dispatchers told him they can locate a telephone within 10 to 20 m2 more than half the time. That proportion clearly needs to increase for arrival times to decrease. “The technology is there,” Dr. O’Connor summarized. “I think we have to learn how to harness it.” “911 Is the most important number in the United States of America,” said Stephen H. Souder, director of the Department of 9-1-1/Public Safety Communications in Fairfax County, VA, and chair of the Federal Communications Commission’s (FCC’s) Task Force on Optimal Public Safety Answering Point Architecture. “In my mind, it is the most fundamental social contract that local government has with its residents.” Having worked for 62 years in public safety communications, including the entire 48-year-long era of 911 service and the current development of next-generation 911 service, Souder is an experienced and outspoken advocate of upgrading the system to eliminate “a widespread delta, meaning gap, in where people call from and where the 911 center is provided information through the system and where they really are…. The whole issue of location accuracy has been around virtually since the dawn of cellular 911 capability.” Federal regulatory attention to the problem dates back to the FCC's first Notice of Proposed Rulemaking (1994) and initial report and order establishing enhanced 911 (E911) location-accuracy rules (1996),3Federal Communications Commission. Revision of the commission’s rules to ensure compatibility with enhanced 911 emergency calling systems, CC docket No. 94-102, report and order and further notice of proposed rulemaking, 11 FCC Rcd 18676 (1996).Google Scholar recurrently amended with exceptions and extensions over the years. “Here we are 22 years later,” Souder continued, “and the problem is still a very significant one. So it's regrettable that the location-technology part of mobile communications has not in any way been given the time and attention by the carriers, because it's their job to do that.” The national 911 system was launched in an era when all telephones were easy to find: they were hardwired lines with fixed addresses, known to the telephone company (singular) from the time each line was activated. Telephony was a drastically different industry in nearly every respect: the American Telephone and Telegraph (AT&T) monopoly had not yet been broken up, rotary telephones and pulse dialing were still standard, and the earliest 911 calls involved mechanical switching (an improvement over the previous system, manual routing by switchboard operators). “The FCC's original location-accuracy rules,” said Trey Forgety, JD, director of government affairs for the National Emergency Number Association, “were created for an era when a mobile phone was really a car phone, had a 3-watt transmitter, weighed 7 pounds, sat in the floor of your car, and was only ever used outside, because of course if you were at home, you'd use your land line.” The 21st-century communications environment has little in common with the conditions under which AT&T and the FCC established the original 911 system, rolling out the service from 1968 onward to reach approximately 50% of the US population by 1987 and 93% by the end of the 20th century.4National Emergency Number Association. 9-1-1 origin and history. Available at: http://www.nena.org/?page=911overviewfacts. Accessed September 27, 2016.Google Scholar Nearly everything about telephones is now faster, more capable, and more flexible, but as commentators from the emergency response profession to mainstream journalists have increasingly observed, the 911 system can no longer do what it once did, and what Google Maps, ride services such as Uber and Lyft, and delivery services can currently do: locate a caller with reliability, speed, and accuracy appropriate to the purpose. “Today's 911 systems really are stuck in the telephone era,” Forgety commented. “They rely on last-century technology that isn't terribly intelligent [and] can't take advantage of a lot of the capabilities that modern smartphones have inside.” The 911 system works best when callers are outdoors and able to provide a location, the first essential piece of information needed by a dispatcher at one of the nation's 6,100 Public Safety Answering Points (PSAPs). Because many callers cannot, however—vocally incapacitated by their medical condition, unaware of their position on a distant road, menaced into silence by an intruder, flustered by the conditions of the emergency, or otherwise incommunicado—the E911 system's automated location capability is an essential tool allowing dispatchers to direct first responders to the site regardless of what the caller can say. Today's PSAPs use 2 primary systems for estimating locations, beginning with “a hardware and software system that accepts 911 calls and uses information delivered along with the call to look up the caller's location in the Automatic Location Identification database,” Forgety said. “In the wire line case, that database is populated at the time a caller establishes wire line service; it's their address. In the wireless case, however, that has to be done at the time of the call by the mobile network.” A second system, computer-aided dispatching, lets dispatchers combine data from the call-taking system and the caller's voice to plan a response. “Unfortunately,” Forgety continued, “in some cases computer-aided dispatching systems…have been designed in a way that they essentially overwrite the automatically retrieved data if the call taker uses something called the rebid function, [which] allows the call taker to request that a mobile network operator update their location estimate for the caller. That's very important because often the first estimate that a network develops is not the most accurate estimate that's possible for that user's circumstances.” Measurements of GPS signals from satellites “thousands of miles away, moving thousands of miles an hour, and putting out only about as much power as a 20-watt light bulb” are imprecise; this is why a Google map on a smartphone includes a light blue circle indicating uncertainty in the estimate of the telephone's position, shown by a smaller central blue dot, with the circle's radius narrowing over time as uncertainty diminishes. Souder boiled down the typical scenario: “Generally speaking, in the 911 world, we have about 1 minute on the phone with a caller to find out where is the emergency, what is the emergency, and to acquire the phone number that the emergency is being reported from, so that if we need to call back, we can do that. The carriers tell us that we should wait about 20 to 30 seconds into that one-minute time and do this rebid to get a better location than they provided to us initially. We don't have that luxury, you know?” Forgety cited another cause of imprecision: call routing. “Location-based routing is a capability that we don't actually have today. 911 Calls are routed on the basis of which cell tower and cell-tower sector a given call comes into. That can result in sometimes minutes-long delays while a call is transferred to the correct 911 center if it was made near a jurisdictional boundary.” Mobile-telephone calls placed from indoors present an additional problem, as do calls from above the ground floor of a building, which add the vertical elevation (aka z axis) to the 2-dimensional horizontal positioning problem. Responders in urban settings who are given only a dispatchable street address (eg, 123 Park Place, Atlantic City) or x/y map coordinates (eg, 40.7394826°N, 73.9914821°W for New York's Flatiron Building) without the z dimension can consume valuable minutes searching for the floor or room where the emergency is occurring. GPS, Forgety noted, is “not very good at developing a z axis estimate…quickly, in particular, because the geometry of the satellites and the receiver [is] very bad. All of the satellites are at roughly the same altitude, and they're all way overhead, and you don't have a transmitter down below your feet to make it possible to get a vertical range.” The FCC estimates that approximately 70% of 911 calls now come from wireless telephones, and 64% of those wireless calls originate indoors. Approximately 2 of 5 households are now wireless only. With the populace swiftly adopting a technology the dispatching system wasn't designed for, the result is that a 911 call from a mobile telephone inside a building—as in nearly half of all 911 calls—has a highly variable chance of yielding an accurate location for the dispatcher and first responders. Although the FCC does not yet collect nationwide data on location accuracy, and state or local data can be spotty, one national newspaper investigation found that “your chance of 911 getting a quick fix on location ranges from as low as 10% to as high as 95%.”5Kelly J, Keefe B. 911's deadly flaw: lack of location data. 2015. Available at: http://www.usatoday.com/story/news/2015/02/22/cellphone-911-lack-location-data/23570499/. Accessed September 12, 2016Google Scholar Many citizens, especially those who have never had to make a 911 call themselves, were taken by surprise in May 2016 to see the national 911 system highlighted in a 16-minute “deep dive” segment on the news/satire program Last Week Tonight With John Oliver.6Last Week Tonight. 911: Last Week Tonight with John Oliver (HBO). 2016. Available at: https://www.youtube.com/watch?v=A-XlyB_QQYs. Accessed September 26, 2016.Google Scholar Souder appeared on Oliver's program and recommended it as a pungent, succinct introduction to the problem. Oliver pounces on the persistent gap in accuracy, some states' diversion of targeted 911 financial support to general funds, and the contrast with non-911 geolocation applications (a pizza deliveryman makes an inevitable appearance) to underscore the urgency of improving performance. He quoted the FCC's estimate that “the location accuracy improvements…could save approximately 10,120 lives annually.” Unlike a better-known pop-culture critique—Public Enemy's 1990 song “911 Is a Joke,” a gallows-humor protest over slow service to minority neighborhoods (“they come when they wanna/so get the morgue truck and embalm the goner”)—Oliver's spotlight acknowledges an important systemic upgrade. The FCC adopted an order in January 2015 establishing a timeline for improving indoor accuracy7Federal Communications Commission. In the matter of wireless E911 location accuracy requirements: fourth report and order. 2015. Available at: https://apps.fcc.gov/edocs_public/attachmatch/FCC-15-9A1.pdf. Accessed September 21, 2016.Google Scholar: within 2, 3, 5, and 6 years of adoption, cell telephone providers must provide dispatchable location or x/y axis location within 50 m for 40%, 50%, 70%, and 80%, respectively, of all wireless 911 calls. This order incorporates some elements of the “Roadmap” agreed on by the 4 national wireless carriers (AT&T, Sprint, T-Mobile, and Verizon), the National Emergency Number Association, and the Association of Public Safety Communications Officials in November 2014, along with the similar “Parallel Path” of the regional and rural Competitive Carriers Association. It also addresses vertical location, mandating that carriers “use an independently administered and transparent test bed process to develop a proposed z-axis accuracy metric” within 3 years of adoption and achieve that metric to cover 80% of the top 25 Cellular Market Areas within 6 years. Even with the staged incremental improvements outlined in this order, however, the goal of accuracy within 50 m for 80% of all wireless 911 calls will not be met until that 6-year point, January 2021. Souder notes that “even 5 years out there will still be a delta of about 20% given to the carriers.” As Oliver pointed out, “People making emergency calls are on the very short list of things we expect to be found 100% of the time.” The 2015 order represents a negotiated compromise, not an ideal. “Those of us that are on the front line,” Souder said, “practitioners, the 911 center directors, and most importantly our 911 call takers, viewed it as just that, a compromise. It was nowhere near as robust and demanding [as] we had hoped it would be…. Even 4 years out now, there's still going to be a delta of about 20%. So we've got some work to do.” One reason the standard is not as rigorous as its predecessor proposal, a 2014 Notice of Proposed Rulemaking,8Federal Communications Commission. In the matter of wireless E911 location accuracy requirements: third further notice of proposed rulemaking. 2014. Available at: https://apps.fcc.gov/edocs_public/attachmatch/FCC-14-13A1.pdf. Accessed September 30, 2016.Google Scholar he added, is that the FCC was reluctant to “arbitrarily issue a directive; they wanted it very much to be industry driven. There's always a danger in that because then the industry is going to be looking out for its best interest.” One report on a private initiative to address the persistent gap through a dedicated smartphone application, RapidSOS,9Rifkin G. A lifesaving smartphone app inspired by a brush with tragedy. 2015; (Available at:) (Accessed September 18, 2016)http://www.nytimes.com/2015/10/01/business/smallbusiness/a-lifesaving-smartphone-app-inspired-by-a-brush-with-tragedy.htmlGoogle Scholar described the 2015 order in passing as a special-interest bureaucratic victory: “The communications industry, saying the deadline was unrealistic, has succeeded in getting the regulations modified.” The Center for Public Integrity takes a similar position, more aggressively and in more detail, charging the wireless carriers with lobbying to weaken the FCC's accuracy-proportion benchmarks, slow its timeline, and substitute WiFi and Bluetooth for more expensive but potentially resilient solutions to the indoor and urban location problems, proposed by the FCC in 2014 but replaced by the 2015 rules.10Holmes A. Cell phone lobby win means “more people will die.” 2015. Available at: https://www.publicintegrity.org/2015/09/29/17935/cell-phone-lobby-win-means-more-people-will-die. Accessed September 30, 2016.Google Scholar (A key drawback of reliance on WiFi, this article noted, is that it would underserve lower-income households, which are less likely to have smartphones and routers.) This analysis contends that organizations favoring the faster timetable and more stringent rules of the 2014 proposal (particularly distinct indoor and outdoor metrics; these are combined in the 2015 order) were crowded out of the final deliberations by the better-connected, deeper-pocketed carriers. Competing business interests also underlie the controversy. A firm developing network-based location technologies, TruePosition, has funded a coalition called Find Me 911 that includes the American Academy of Emergency Medicine and multiple first-responder organizations.11Find Me 911. About the Find Me 911 coalition. Available at: https://ecfsapi.fcc.gov/file/7520964296.pdf. Accessed September 30, 2016.Google Scholar Find Me 911 advocates faster FCC action and technology neutrality on indoor location-accuracy requirements and accuses the large carriers of using their influence to bend the rules toward their favored technologies; rhetorical battles by the coalition and its opponents have been heated.12Heaton B. 911 Location accuracy group calls out wireless industry. 2013. Available at: http://www.govtech.com/public-safety/911-Location-Accuracy-Group-Calls-Out-Wireless-Industry.html. Accessed September 30, 2016.Google Scholar, 13Goldstein P. Verizon blasts Find Me 911 Coalition report on 911 location data. 2014. Available at: http://www.fiercewireless.com/wireless/verizon-blasts-find-me-911-coalition-report-911-location-data. Accessed September 30, 2016.Google Scholar (Perhaps coincidentally, or perhaps not, the coalition's entire Web site [http://www.findme911.org] remained inaccessible through a large cable-based Internet provider—not one of the 4 major cell telephone carriers—throughout the research and drafting period for this article, although many materials related to Find Me 911 were available elsewhere, including its Facebook14Facebook. Find Me 911 Available at: https://www.facebook.com/FindMe911.org/. Accessed September 30, 2016.Google Scholar and YouTube15YouTube. Find Me 911. Available at: https://www.youtube.com/watch?v=VI7OTTHGVCA. Accessed September 30, 2016.Google Scholar sites.) In Forgety's view, the critique of the lobbying resulting in the 2015 rules “loses one key important nuance, which is, how long could the industry have successfully abated updates in the rules if they had really wanted to?… Historically, whenever the FCC has updated the rules, at least 1 of the carriers, if not several, have sued and have been successful in delaying the implementation of the rules, if not preventing them outright, for some period of time. Now, that costs everybody a lot of money, it costs a lot of time, and from the standpoint of our enterprise, it costs lives.” To remove conflict between unilateral action and obstructive litigation, he continued, “[W]e have spent the last 2 years moving out with the carriers as partners rather than as adversaries to get this work done.” John Marinho, MBA, vice president of technology and cybersecurity for the trade organization CTIA—The Wireless Association (previously the Cellular Telephone Industries Association), pointed out that the new 50-m accuracy metric, although still broad by urban spatial standards (roughly half the length of a soccer field), is “certainly an improvement from the previous standard, which was about 150 meters.” The progress in indoor penetration using WiFi and Bluetooth information, he observed, can augment other locational technologies being tested or applied—those discussed in the 2015 order include observed time difference of arrival, uplink time difference of arrival, terrestrial beacons, radiofrequency fingerprinting, barometric sensors, and alternative global navigation satellite systems such as the Russian GLONASS, Chinese BeiDou, and European Galileo systems—to create a highly capable hybrid system. “As you have more and more constellations that can be used for purposes of the kind of function that you see with GPS in terms of location on the surface of the planet, those satellite constellations provide you with improved accuracy,” Marinho said. The 2015 order's key observation from the medical perspective may appear in paragraph 160 (the source of Oliver's mortality reduction estimate) and the subsequent discussion, drawing on a study conducted in 2001 in Salt Lake City16Wilde E.T. Do emergency medical system response times matter for health outcomes?.Health Econ. 2013; 22: 790-806Crossref PubMed Scopus (70) Google Scholar and a 2002 cardiac emergency study in Pennsylvania17Athey S. Stern S. The impact of information technology on emergency health care outcomes.RAND J Econ. 2002; 33: 399-432Crossref PubMed Scopus (107) Google Scholar to bolster the intuitive correlations among location accuracy, response time, and medical outcomes. Quantitative data on location accuracy and its consequences are otherwise surprisingly scarce. “Unfortunately, there are no nationwide statistics or even in many cases statewide statistics kept about the prevalence of this problem,” Forgety reported. “And efforts to collect data at individual 911 centers have been somewhat frustrated in the past because of the limitations of hardware and software in legacy 911 systems.” In the absence of such performance metrics, journalistic coverage has typically presented narratives involving deaths or other poor outcomes when location inaccuracy has delayed responders.18Knutson R. Cellphones leave gaps for emergency services. 2016. Available at: http://www.wsj.com/articles/SB10001424052702304579404579231913503559556. Accessed September 20, 2016.Google Scholar, 19Fung B. A survey of 911 dispatchers reveals the horrible, human cost of bad technology. 2015. Available at: https://www.washingtonpost.com/news/the-switch/wp/2014/04/24/a-survey-of-911-dispatchers-reveals-the-horrible-human-cost-of-bad-technology/. Accessed September 20, 2016.Google Scholar, 20Fitts A.S. When 911 operators can't find their callers. 2015; (Available at:) (Accessed September 20, 2016)http://www.theatlantic.com/technology/archive/2015/11/when-911-operators-cant-find-their-callers/416730/Google Scholar “A lot of the reports that I've seen,” added Marinho, are “based upon anecdotal evidence, as opposed to having real hard data. In fact, that's what we're involved in doing now with the work that we're doing in response to the FCC rules…to generate that kind of data that can tell us exactly how the existing technologies are improving, as well as how we can improve on those existing technologies.” (The 2015 order's provisions for data collection, beginning in 2017, may help resolve the disputes between the major carriers and the Find Me 911 participants over which indoor location technologies perform well enough to justify their expense.) Rather than focus on unflattering comparisons between nonemergency geolocation systems and E911—the meme in question, circulated by Oliver and others, being some version of “if pizzerias, Facebook, and Uber can find me quickly, 911 should be able to as well”—Marinho stressed nuances, stringent FCC requirements, and continual improvement. “There's a big difference between those systems that don't have to operate with a level of reliability and consistency and throughput that you're gonna see with 911 in the priority that it’s given and how it's handled compared to Google Maps or to other forms of location-based solutions because 911 has to operate at a much higher level of reliability and has to rely upon using a variety of different technologies…. The fact is they operate differently. A 911 call gets a much different level of priority through the network; it's a much different level of throughput in terms of how it's routed to a dispatch center. It then has the ability to determine who to dispatch: do you dispatch EMS? Do you dispatch fire? Do you dispatch the police?” Forgety concurred that E911 and nonemergency systems are fundamentally different. “In 911, we need to find you precisely more than we need to find you quickly,” he said. “Google or Uber wants to show you a position estimate the moment you open the application, so it's OK if there's a big blue circle around your location because in that case you're looking at the map, and you can drag the pin around or enter an address to get it exactly where you want it. Most of the time, you're not paying attention to the fact that the location estimate is actually very uncertain…. [I]n 911, though, we don't have that kind of user interactivity, number 1. And number 2, we need a much more precise location estimate to be actionable.” Consequences of mislocations are severe: in situations in which police may be kicking in doors, “you don't want to kick down the wrong door. That can have tragic consequences for everybody involved.” The location-accuracy problem calls for expanded information in 2 realms: more extensive performance assessments, as well as more channels for information usable by PSAPs. One farsighted aspect of the 2015 order, Marinho added, is its expansion of positional databases: in accordance with the order, CTIA is planning a national emergency address database that will include dispatchable addresses for 25% of the population in the top 25 Cellular Market Areas, plus locations of WiFi hotspots and Bluetooth beacons to allow visibility and scanability of mobile devices, regardless of whether they are registered with these access points. An even broader advance under development largely at the state level, Next Generation 911 (NG911),21National Highway Traffic Safety Administration National 911 Program. 2015 National 911 progress report. 2016. Available at: http://www.911.gov/pdf/National-911-Program-2015-ProfileDatabaseProgressReport-021716.pdf. Accessed September 30, 2016.Google Scholar is building an environment in which callers can transmit requests through Internet protocol and send more sophisticated data to PSAPs, including text, images, and video, as in commercial multimedia applications but with appropriate security for emergency settings. The National Emergency Number Association has been planning a future path for NG911 since 2000, has developed the i3 standard for the architecture of a dedicated emergency services Internet protocol network,22National Emergency Number Association. NENA i3 solution—stage 3. Available at: https://www.nena.org/?page=i3_Stage3. Accessed September 30, 2016.Google Scholar and supports federal legislation providing matching grants to states for system upgrades. NG911 systems will face substantial financial and operational challenges but are fully or partially operational in several states. Souder's FCC task force has marshaled 50 experts with some 800 years of collective experience to guide the overhaul of the system. “Five years from now,” he projected, “the only thing about 911 that you and I know about it that's going to be the same is going to be the number, 911.” As the national emergency address database and NG911 are developed, their utility will depend on universal integration, particularly because private sector efforts to solve the location problem take multiple forms. “A lot of these systems are trying to solve that problem by making an end around the existing 911 systems in some cases,” Forgety said, “and we do have concerns that those nonintegrated systems will result in very detrimental fragmentation. Now, there have also been efforts like that by RapidSOS that have worked very hard to integrate with the existing 911
Referência(s)