Problems withg "upgrade" from Zoll E-Series to X-Series

[A204]

I manage an ambulance company with about 70 ALS vehicles doing primary 9-1-1 work. We recently upgraded our cardiac monitors from the Zoll E-Series to the X-Series and are having some significant problems. Our issues mostly center around the new 12-lead interpretion algorithm. The X-Series requires a MUCH cleaner 12-lead before it will do the interpretation. As a result we are seeing about a third of our 12-leads not being interpretted due to data quality limitations. We never had this problem with the E-Series monitors and nothing has changed with our process other than the monitors. Also, the E-Series gace a clear *** Acute MI*** readout that was used for cath lab activation. The X-Series provides ***STEMI*** which does not correlate to the E-Series. That part is hard for me to explain, but if you're familiar with both monitors, you know what I mean. For those of you with X-Series monitors, what are you using for cath lab activation triggers? I'm sure some of you are going to say that we should just train our medics to read the 12-lead, but there is no support for that in our community. In fact, a review of more than 100 12-leads interpretted by our old E-Series monitors showed the monitor interprettation to be more reliable than that of most ER docs. Enough of my ramblings... please use this as a general thread to discuss issues with the transition from E-Series to X-Series monitors.

 

[Aidey]

Why are you having so man problems with data quality? That is what you need to look at and fix.

What do you mean about the Acute MI and STEMI not correlating? Because they use different terms? Or because the X-Series doesn't agree with the E-Series about what is and isn't an MI? I don't think there is anything unclear about the machine printing out STEMI rather than Acute MI.

Who did the 12 lead review and determined that it was "more reliable" than most ER docs?

 

[xrsm002]

Yup I'm glad we did some 12 lead training in my medic school. Not in depth but I have the basics of it down

 

[usalsfyre]

Remember the X is based off of a ProPaq, and doesn't share architecture with previous Zolls, meaning they may have different filtering algorithms.

The first place I would look is your electrodes, a fair number of problems can be traced there. Make sure they're not sitting out open for >24 hours and spend a bit of extra money and buy something better than bargain basement.

As for "STEMI" vs "Acute MI". First make sure it's not changeable. Second tell the cardiology/EM group or cath lab that's causing the issue if they refuse to activate because if a minor difference in terminology your 70 something trucks will be taking all the STEMIs they encounter elsewhere until they learn to play well with others.

 

[STXmedic]

Who did the 12 lead review and determined that it was "more reliable" than most ER docs?

This. I've had those monitors tell me a blatant NSR was an Acute MI. I highly doubt that those algorithms are better than an EM Physician.

As Aidey said, figure out why you're getting poor quality.

And although you said it's not what you want to hear, training will be your medics' best friend. They shouldn't need any algorithm to tell them when to initiate a Heart/STEMI Alert.

 

[A204]

Thanks for the replies. Yes, we went through the expected troubleshooting of evaulating electrodes, retraining employees on lead placement and proper technique, and we still have problems. ALl of our paramedics are trained for basic 12-lead interpretation, but our medical control chose to trust the machine. In fact, the review of the E-Series interpretation v. ED doc was done by the QI guy from one of our lecoal ED physican groups to validate the machine's interpretation, noth the other way around. The difference between ***Acute MI*** and ***STEMI*** is far more than terminology. It is the underlying triggers in the algorithm. It is hard for me to explain the difference in a message thread, but you absolutely know what I'm talking about if you are familiar with the details of both monitors. I've talked to a couple other ambulance companies who are having similar experiences with the transition from E-Series to X-Series. Maybe you can't fully appreciate the shallenges unless you've been through it. Anyone else out there have an E-Series to X-Series transition story?

 

[Christopher]

Thanks for the replies. Yes, we went through the expected troubleshooting of evaulating electrodes, retraining employees on lead placement and proper technique, and we still have problems. ALl of our paramedics are trained for basic 12-lead interpretation, but our medical control chose to trust the machine.

The E-Series uses the GE Marquette 12SL algorithm which the LifePak 12 also uses. The Zoll X-Series uses the Inovise algorithm which they bought. This is a newer algorithm and thus it will have some quirks.

That being said the E-Series and LifePak 12 often were set to a 40 Hz low pass filter while the X-Series by default runs a 150 Hz low pass filter. This significantly changes the amount of artifact seen on the ECG. You can switch the X-Series to use a 40 Hz low-pass and that may help you out.

In fact, the review of the E-Series interpretation v. ED doc was done by the QI guy from one of our lecoal ED physican groups to validate the machine's interpretation, noth the other way around.

Validation sets for both the GE Marquette and Inovise were done using in hospital acquired ECGs. I cannot stress this enough.

They are sensitive to artifact and are both horrible in the face of tachycardias (>110), atrial flutter, and some bundle branch blocks.

The difference between ***Acute MI*** and ***STEMI*** is far more than terminology. It is the underlying triggers in the algorithm. It is hard for me to explain the difference in a message thread, but you absolutely know what I'm talking about if you are familiar with the details of both monitors.

This does not make any sense. I'm familiar with both monitors and use 4 different models of the 3 major brands between the services I work for. None of them operate differently when a STEMI message is received.

That being said, they both arrive at those messages in different ways so maybe that is what you meant. Either way, they are not sufficiently different in that respect. No algorithm does a great job arriving at the message so this complaint is moot.

I've talked to a couple other ambulance companies who are having similar experiences with the transition from E-Series to X-Series. Maybe you can't fully appreciate the shallenges unless you've been through it. Anyone else out there have an E-Series to X-Series transition story?

We went from the LP12 to the X-Series and are running into issues related to the 12-Lead as well. However, it has nothing to do with their interpretation algorithm. Your system is setup for failure if you rely on the box to tell you what it is.

The current legitimate problems with the X-Series and 12-Leads include:

• Cropping of high amplitude complexes on the 12-leads
• Not saving 12-Leads to memory when the log is full
• Printed copies are very difficult to scan/copy due to the mode of printing
• The 4x3 view is not continuous in time w.r.t. left-to-right reading of the paper which is confusing when compared to traditional print-outs

I have no financial incentive to stand up for the X-Series, but it certainly is a step up from the E-Series.

As for your 12-Lead "interpretation woes"...bad news: none of the current generation of EMS cardiac monitors uses the GE Marquette 12SL algorithm! Zoll is Inovise, LifePak is Glasgow, and Philips is their own DXL.

You'll never get the same messages again.

Here is a link to the various interpretation guides for the monitors:

• Zoll X-Series Inovise 12L Interpretive Algorithm Physician's Guide
• Philips MRx Philips DXL 12-Lead Algorithm Guide for Physicians
• Physio-Control LP15 Glasgow 12-lead ECG Analysis Program

 

[A204]

Christopher,

Thank you for your very thoughtful, well-articulated response. You're obviously a smart guy. Clearly you were willing to put more effort in your detailed response than I was in my question.

As for the low pass filter, our units were already set to 40 Hz (filtered diagnostic) as part of our troubleshooting process.

I don't think I'm being clear regarding my concern that there are differences between ***Acute MI*** and ***STEMI***. For the X-Series, findings for the acute portion of the 12L Interpretive Algorithm's MI algorithm are divided into five categories. We were able to determine ***STEMI*** from the X-Series correlated to Category 1, but even the Zoll engineers we talked to were unable to confirm that meant the same thing as ***Acute MI*** for the E-Series. I have a conference call with Zoll and another ambulance company in less than an hour to try and resolve this.

For reference, here's a summary of five categories for the acute portion of the 12L Interpretive Algorithm's MI algorithm for the X-Series:

Acute MI, Category I: ST Elevation Acute MI. This finding is reported when no
ST-T confounding conditions are present, and when ST levels measured at J+20 ms meet the ACC (American College of Cardiology) guideline definition for STE Acute MI.

Acute MI, Category 2: Probable ST Elevation Acute MI. This finding is reported when significant ST-T abnormalities specific for Acute MI are present but do not meet the ACC guideline definitions for STE Acute MI; or when ST-T confounding conditions are present.

Acute MI: Category 3: Acute MI. This finding is reported when a pattern of localizing, abnormal ST deviation in two or more contiguous leads is present. Levels of ST deviation are less than that of Category 1 or Category 2 acute MI

Acute MI: Category 4: MI, probably acute. This finding is reported when prominent ST-T abnormalities specific for Acute MI are present, but are accompanied by QRS abnormalities that indicate the infarction is evolving.

Acute MI: Category 5: MI, possibly acute. This finding is reported when ST-T abnormalities specific for Acute MI are present, but are accompanied by significant QRS abnormalities consistent with evolved, or healed MI.


Thankfully we don't have all of the problems you referred to in your post. Our ePCR system was written by our own software developers. The X-Series has been fantastic when it comes to data integration. Through the SDK we're able to continually pull the data across in the background whenever both the monitor and tablet computer are within range of the ambulance's wifi. We seldom print from the monitor since it is cleanly shown in our ePCR system and on the printed PCR report. At least that's one major improvement with the X-Series that has worked as promised.

 

[Christopher]

Christopher,

Thank you for your very thoughtful, well-articulated response. You're obviously a smart guy. Clearly you were willing to put more effort in your detailed response than I was in my question.

For reference, our area utilizes Paramedic interpretation for activation and many services have at one point or another turned off device interpretation to discourage relying on it.

As for the low pass filter, our units were already set to 40 Hz (filtered diagnostic) as part of our troubleshooting process.

I don't think I'm being clear regarding my concern that there are differences between ***Acute MI*** and ***STEMI***. For the X-Series, findings for the acute portion of the 12L Interpretive Algorithm's MI algorithm are divided into five categories. We were able to determine ***STEMI*** from the X-Series correlated to Category 1, but even the Zoll engineers we talked to were unable to confirm that meant the same thing as ***Acute MI*** for the E-Series. I have a conference call with Zoll and another ambulance company in less than an hour to try and resolve this.

Well, the *** ACUTE MI SUSPECTED *** message was really narrow in its definition. Good specificity but horrible insensitive.

Figuring out how to best use the Zoll messages means you should evaluate them in context to their applicability to Rokos' definitions of MI's and the ECG. I'll walk thru them below:

For reference, here's a summary of five categories for the acute portion of the 12L Interpretive Algorithm's MI algorithm for the X-Series:

Acute MI, Category I: ST Elevation Acute MI. This finding is reported when no
ST-T confounding conditions are present, and when ST levels measured at J+20 ms meet the ACC (American College of Cardiology) guideline definition for STE Acute MI.

This level will be the closest to *** ACUTE MI SUSPECTED ***. It is probably roughly as sensitive and specific, ignoring differences in their "exclusions".

Acute MI, Category 2: Probable ST Elevation Acute MI. This finding is reported when significant ST-T abnormalities specific for Acute MI are present but do not meet the ACC guideline definitions for STE Acute MI; or when ST-T confounding conditions are present.

This is a great addition to the algorithm, but alas it means it doesn't meet the arbitrary guideline of 1mm+ in 2+ leads. This category is for ECG's which have ST-e and ST-d. Likely if a paramedic saw this ECG they would activate or would run serial ECG's and activate at some point. The *** message may trigger on this ECG as well.

Acute MI: Category 3: Acute MI. This finding is reported when a pattern of localizing, abnormal ST deviation in two or more contiguous leads is present. Levels of ST deviation are less than that of Category 1 or Category 2 acute MI

This category, by definition, would probably not be equivalent to the Marquette definition. A shrewd paramedic may activate.

Acute MI: Category 4: MI, probably acute. This finding is reported when prominent ST-T abnormalities specific for Acute MI are present, but are accompanied by QRS abnormalities that indicate the infarction is evolving.

This category is tricky. When you read their definition in full you see that the ECG actually meets the ACC definition, yet other ECG features make it not as solid of a call. I would need to see sample ECG's from this category to form an opinion, but it seems reasonable to activate.

Acute MI: Category 5: MI, possibly acute. This finding is reported when ST-T abnormalities specific for Acute MI are present, but are accompanied by significant QRS abnormalities consistent with evolved, or healed MI.

Probably not a *** style ECG.

Thankfully we don't have all of the problems you referred to in your post. Our ePCR system was written by our own software developers. The X-Series has been fantastic when it comes to data integration. Through the SDK we're able to continually pull the data across in the background whenever both the monitor and tablet computer are within range of the ambulance's wifi. We seldom print from the monitor since it is cleanly shown in our ePCR system and on the printed PCR report. At least that's one major improvement with the X-Series that has worked as promised.

We're working on that stuff, we use emsCharts but their mobile software is trash. I'm a software engineer so we may end up doing this on our own. The problems I described w/ printing etc are actually due to how the Inovise Algorithm obtains 12-Leads, so simply using the SDK will still have the problem. (Maybe not the cropping bug)

 

[recruiter1]

I am having the same issue. Shoot me an email/pm and lets talk about this, perhaps we can find a way to resolve it.

 

[A204]

Please send me your contact information and I'll contact you. We've made considerable progress with the Zoll engineers. It may be mutually beneficial if we spoke directly. I'll be in the office every day next week except Tuesday the 1st.