By Mark Reddish
Apple recently announced that devices running iOS 13 will enable location fixes quickly enough to be used for routing wireless 9-1-1 calls. This is a significant development because routing 9-1-1 calls based on the device’s estimated location, rather than relying on the location of the cell sector serving the caller, will allow more calls to be delivered to the correct emergency communications center (ECC) on the first try. (Google has reportedly enabled a comparable capability for Android devices.) This will make it possible to decrease emergency response times, and thereby save lives.
Background on “Misrouted” Wireless 9-1-1 Calls
When someone calls 9-1-1 from a cellphone, the call is sometimes delivered to the wrong ECC. These “misroutes,” are not necessarily due to a technical error in the routing process. They can occur because the 9-1-1 system is configured to route wireless calls to an ECC based on the cell tower sector that handles the call, which may not lead to the appropriate ECC for the caller’s actual location. Misroutes occur more frequently near ECC service boundaries and in areas where multiple jurisdictions are in close proximity, such as the Washington, DC, region.
In March 2018, the Federal Communications Commission (FCC), in a Notice of Inquiry, sought public comments on how frequently misrouting occurs for wireless 9-1-1 calls, how it impacts emergency response, and how to prevent it. Quantifying the scope of the problem is difficult because ECCs may have different policies and capabilities for transferring as well as tracking misrouted calls. As part of the FCC’s public comment process, APCO made clear that the delays in response can mean the difference between life and death. Further, the problems created by misrouted 9-1-1 calls are made worse by the fact that – despite the sophistication of modern communications technology – even basic call transfers (regardless of the reason to transfer a call) are not universally possible. There are still instances in which ECCs have to manually call each other to convey the information about an emergency. Even when ECCs are able to transfer the voice portion of a 9-1-1 call, they too often cannot include ANI and ALI, let alone data contained in the computer-aided dispatch system.
Eliminating misroutes is challenging, in part, because the network routes a wireless 9-1-1 call within a matter of seconds of the caller pushing “send,” but estimating the location of a caller historically required 20-30 seconds. That’s changing. Modern location technologies are making it possible to estimate a device’s location within seconds, at least with enough accuracy for routing purposes. According to Apple’s announcement, iOS 13 offers wireless carriers the option to enable location-based routing (LBR) and to fall-back to the cell sector routing mechanism if a device-based location estimate cannot be provided quickly enough.
A Rough Estimate of the Impact of Misrouting
To estimate the impact of misrouting, we can modify an equation used by the FCC to estimate the impact of improving the accuracy of location information delivered with 9-1-1 calls. The FCC looked at how many 9-1-1 calls for ambulances resulted in deaths that could have been prevented by a one-minute reduction in response time, the percentage of wireless 9-1-1 calls that would experience a one-minute reduction in response time with improved location accuracy, and the percentage of 9-1-1 calls that come from wireless devices. Based on this, the FCC estimated more than 10,000 lives a year could be saved by improving 9-1-1 location accuracy.
For estimating the impact of misrouted 9-1-1 calls, the inputs for the equation are slightly different. We’ll still use 80% as the percentage of 9-1-1 calls from wireless phones, but we insert new numbers for the percentage of misrouted calls and the number of lives that could be saved by eliminating misroutes. A study by the Alliance for Telecommunication Industry Solutions estimated that an average of 12% of wireless calls are misrouted nationwide. According to a study cited by the FCC, transferring a call between ECCs could add an average of 40 seconds to the call time. Thus, compared to the FCC’s estimate of lives saved by improving 9-1-1 location accuracy, a larger percentage of calls are at issue, but the reduction in response time per call is smaller.
|The FCC ‘s Math for 9-1-1 Location Accuracy||The Math for Location-based Routing|
|0.8||% 9-1-1 calls from wireless||0.8||% 9-1-1 calls from wireless|
|0.05||% of wireless 9-1-1 calls that would see a one-minute reduction in response from improved location info||0.12||ATIS estimate of % of misrouted wireless 9-1-1 calls|
|x 253,000||# of deaths that could have been prevented by a one-minute reduction in response time||x 168,575||# of deaths that could have been prevented by a 40 second reduction in response time|
|10,120||estimated lives saved by reducing response times by one minute by improving location accuracy||16,183||estimated lives saved by reducing response times by eliminating wireless misroutes|
This equation estimates that more than 16,000 lives could be saved by eliminating misroutes altogether, but no LBR solution will be perfect. Therefore, we still need to factor in an estimate of how many misroutes LBR solutions would prevent. The ATIS study provided estimates of how many misrouted calls could be prevented by various degrees of accuracy in a location-based routing decision. For a horizontal uncertainty of 300 meters, which is what ATIS recommended for providers of positioning technologies to be used in LBR solutions to strive for, the estimated reduction would be 50%. Thus, according to a modified version of the FCC’s equation for calculating the impact of 9-1-1 location accuracy improvements, achieving the ATIS recommendation for LBR would save more than 8,000 lives a year.
Of course, this is a rough estimate. There’s no way to know how many lives could be saved by eliminating “misrouted” 9-1-1 calls, but if we can rely on the FCC’s equation for a reasonable estimate, there should be no shortage of motivation to adopt LBR solutions.
What Happens Next?
More needs to happen before ECCs start seeing the benefits of LBR. For one thing, wireless carriers need to update their networks to make routing decisions with the device-based location when it’s available, rather than routing based on the cell sector. Devices and operating systems will need to be modified, similar to Apple’s, to make a quick location available for LBR, and the entire industry should work to improve the accuracy of quick location fixes as well as ALI. This issue will ultimately return to the FCC for further consideration of creating rules to ensure 9-1-1 calls are routed to the ECC that serves the caller’s location.
 Note, the ATIS report uses the term “sub-optimally routed” to distinguish a network error from the situation in which the network performs how it’s designed but due to inaccurate location information delivers the call to the wrong ECC. For simplicity and consistency with the FCC proceeding, we’ll use the term “misroute.”
 This number was derived from the FCC’s assumptions, based on emergency medical response research, showing that: for 73,706 observed incidents, a one-minute delay in response time caused a mortality increase of 746 deaths (from 4,386 to 5,132 deaths, a 17% increase); the relationship between response time and mortality is linear, meaning that if a one-minute delay causes a 17% increase in mortality, a one-minute decrease could cause a 17% decrease in mortality; and that there are 25 million calls for an ambulance each year. If the relationship between response time and mortality is linear, a 40 second reduction in response time would save 2/3 the number of lives that a 60 second reduction would: 497 lives per 73,706 incidents (instead of 746 lives). For 25 million calls for an ambulance, that translates to 168,575 lives that could be saved by a 40 second reduction in response time.
 ATIS’s estimates of the percentage of misrouted calls that could be mitigated using an LBR solution were based on findings that if, for example, 81.1% of misrouted calls were within 1,000 meters of the routed ECC boundary, an LBR solution with position estimates with horizontal uncertainty values of 1,000 meters could potentially resolve 19% of misrouted calls. 50 meters of accuracy corresponded to 15.4% of misrouted calls (which means approximately 85% preventable by an LBR solution with 50 meters uncertainty). It isn’t clear how accurate the quick location from Apple devices will be, but the device-based hybrid location technology used by Apple devices is capable of providing location estimates for an ALI fix that are much more accurate than 300 meters, even for indoor environments in urban areas.
 For example, a study by RapidSOS provided estimates that differed from those the FCC relied on in the location accuracy proceeding – a larger percentage of 9-1-1 calls that would be affected by better indoor location technology, but a more conservative estimate of how many lives could be saved by an overall reduction of one minute in response time. It’s also worth considering that the FCC’s equation only takes into account the potential lives saved for EMS calls, but reducing response times would certainly save lives for other types of 9-1-1 calls.