Nobody likes to wait for anything, especially for a 911 call taker during an emergency. 911 Center Public Safety Answer Points (PSAP) Average Call Answer Times have been under scrutiny lately in several states. But before you throw the baby out with the bath water and criticize 911 call taker efficiency, you need to validate the data.

Before you measure something, you need to define where to start measuring. For example, when you run a race, whether it’s 50 yards or a 5K, there is a distinct start- and finish point to the race. The clock starts when you cross the start line, and stops when you cross the finish. Looking at statistics for emergency calls is no different, as long as you understand where the clock started, and why.

While recently discussing average 911 answer times with a colleague, a point came out in the conversation that created a significant amount of confusion around this very topic. In an emergency, seconds matter and some emergency dispatchers are being penalized for not meeting state-mandated answer times. Other agencies have been accused of “fudging the numbers” to make their statistics fall within acceptable guidelines. While looking at the data, it became apparent that the REAL problem was possibly that no one was paying attention to when the clock started… only when it ended.

Imagine running a 5-minute mile, only to find out later that your time was actually 10 minutes, because the clock started while you began tying your shoes. Not too fair, right? Well, the same goes with 911 calls.

Unless you’ve listened to the trunk side of a 911 call, you’d likely be astonished at the archaic, analog nature of getting a call from point A to point B. About two years ago, I was fortunate enough to receive an audio clip from a 911 call that I quite often use for training purposes, as it highlights several points that otherwise aren’t very obvious.

911 call preamble: Getting ready to get ready
911 Centralized Automatic Message Accounting (or CAMA) trunks that connect the Public Safety Answer Points to the 911 tandem central office are specialized analog circuits similar to Centrex lines. When a call is presented to them from the 911 network, this signaling mechanism is not ringing voltage, as found on a normal telephone line. The central office will “wink” toward the PSAP by applying reverse battery on the circuit. The PBX will then “wink” back towards the central office confirming its readiness to accept a call. The central office will then “wink” back at the PBX confirming that the response was received, and digits will be coming down the line.

When you look at audio as a sound wave, these “winks” are clearly noticeable as a sharp spike in the audio file and can even be heard as a loud click on the line.

In this example you can clearly see the three winks at the very beginning of the call, and if you are measuring answer time from the central office side, this would be a likely spot to start counting from zero.

Audible "winks" on a 911 call.
Audible “winks” on a 911 call.

At this point in time, the audio path is now open between the central office and the 911 PSAP call taking equipment. The central office then signals to the Customer Premise Equipment (or CPE) in band information using multi-frequency tones for digits and specialized signaling characters to indicate the Psuedo ANI (pANI) of the inbound 911 call. Depending on the area, and the carrier, the ANI that is received could be 7 (NNX-XXXX), 8 (I-NNX-XXXX), or 10 digits (NPA-NNX-XXXX)in length. Once again, looking at our example audio, the MF tones are clearly discernible in the audio wave.

ANI tones on a 911 call.
ANI tones on a 911 call.

You will also notice that there is another audio spike, which is the PBX signaling a “wink” back to the central office acknowledging receipt and acceptance of the ANI information. It also serves as a go-ahead signal for the central office to open up the audio channel between the original caller and the PSAP.

At this point, based on the audio in this example, the PBX applies ringing to the line, and you can see the abrupt change in audio as the callers audio is now also patched through.

As an interesting side note, what has happened up until this point is fairly critical in processing and delivering the 911 call to the PSAP. I have seen cases in the past where adjunct equipment has been installed on the CAMA trunks to capture the ANI information and send it over to the CPE 911 equipment for processing. But, because they are signaling back to the central office was not in proper sequence, they returned answer supervision too early to the central office and the caller’s audio actually corrupted the receipt of the MF tones. In fact, as it turns out, a woman screaming can mimic an MF tone, causing the system to process garbage data and potentially make the call fail.

Another interesting thing happens at this point. The CPE equipment is now aware of the call, has the information required to process it, and typically generates a Call Detail Reporting (CDR) start record. Once again, another potential starting point for the call. The only problem here is that this starting point is 3 seconds out of sync with the central office starting point.

The next step in the sequence would be for the CPE or PBX to process the call, and deliver it to a 911 call taker. After analyzing and listing closely to this sample recording, it appears that the 911 call taker answered the line immediately after the first ring (and remember, a ring cycle is to second on, followed by 4 seconds of silence). Since we cannot see or hear the second ring, we can assume that the call was picked up almost immediately after the first ring, and in fact you can see a small blip of audio when the line is connected, which is immediately followed by the dispatcher saying “911. What is the location of your emergency?”

The 911 call begins.
The 911 call begins.

At this point on the timeline, 9 seconds have passed from when the central office initiated the call. Deciding where the starting point is can significantly skew the data, and the dispatcher could actually be penalized for a 9-second delay, when in fact, they answered the call within 2 seconds of it being presented to them.

Keeping it fair for everyone
We certainly want to make sure that our nation’s public safety operators are doing their job, and are performing within accepted national specifications. What we have to be careful of though, is making sure we are not penalizing them by looking at bad data.