Heart Disease Community
reliability of EKG
About This Community:

This patient support community is for discussions relating to angina, angioplasty, arrhythmia, bypass surgery, cardiomyopathy, coronary artery disease, defibrillator, heart attack, heart disease, mitral valve, pacemaker, PAD, stenosis, and stress tests.

Font Size:
A
A
A
Background:
Blank
Blank
Blank
Blank Blank

reliability of EKG

how reliable is an EKG as a diagnostic tool?  
Related Discussions
40 Comments Post a Comment
Blank
976897_tn?1379171202
It is very reliable with the units used these days. They have built in checks which inform the nurse/doctor if one of the leads isn't attached properly to the patient. This was the biggest issue with EKG reading in the past, loose leads.
Blank
367994_tn?1304957193
An EKG is not very reliable and requires further testing, supporting symptoms, clinical evidence.  If I remember correctly it has a specificity rating of 40%.  For instance, the segment ST interval when elevated can represent many different conclusions, if the interval is depressed there can be different diagnoses.etc. etc.  To procede subsequently from a positive EKG, requires an evaluation by a differential diagnoses with other evidence.

Sensitivity can be a problem due to artifacts meaning a structure or substance not normally present, of no consequence, but it presents an authentic signal, etc.
Blank
Avatar_f_tn
"If I remember correctly it has a specificity rating of 40%"

Can you be more specific with the 'IT' as to which make/model you are refering to? There are portable ekg devices which use a USB interface for connecting to a pc. These have an accuracy of over 77%. Some have 82% and higher accuracy at detecting MI. Technology has come along in leaps and bounds with new algorithms and even my local hospital upgrades EKG machines every year. The first EKG machines simply recorded the electrical signals detected but for years now computer software has made a much more accurate analysis available. This is how the machine is able to tell a nurse when a wire is loose or has a defect before the trace is recorded, the computer knows an anomoly when it sees one. There is a huge variety of EKG manufactures with model ranges with different costs. The more you pay, the higher the accuracy. I really dont believe so many cardiologists would even bother with a 40% accuracy for detemining a patients welfare.
Blank
Avatar_f_tn
Having worked in this field, I can see both sides of how accurate the EKG can be. Actually, it's only as good as the tech running the machine. Artifacts can be caused by several things; any tech with half a brain should be able to tell what is causing the artifact and how to correct it. They shouldn't have to rely on the machine to tell them what the problem is. One of the biggest problem I've seen is the electrodes being placed wrong, which the machine won't tell you. It may tell you if the lead is loose, but as long as it has contact you'll get a reading. ALL EKG leads should be placed AFTER the tech has felt for each of the ribs so the leads are not placed partially or literally over the rib bone. I rarely see any tech these days feeling with their fingers for those intercostal spaces. They don't know how to place an EKG reading on the half or quarter standard, which they need to know how to do to get a good reading. When my daughter's complexes were shooting off of the page, no one knew how to correct that. Areas on her EKG wouldn't even show a complex at all, in other words some areas on her EKG were 'flatlined'. We used to joke about that part of her heart as being 'dead'. The machine had no idea how to diagnose those problems on it's own. This is a tool which is only a foundation for detecting a possible problem. It certainly is incapable of true diagnoses; in fact it can misdiagnose, even through the computer system.    
Blank
367994_tn?1304957193
It is my understanding an EKG unit indicates an improper electrical connection but there is no way to detect without manual manipulation by the tech.  I'm not familiar with the state-of- art for an EKG unit and whether the unit now provides the connection status of each lead, if not, it should, and apparently it now has that ability.

From physiology of the heart's conduction system perspective, it seems if the electrode is not properly connected or adhered to the skin there would be a "flatline", and a tech would be aware.   If the trace is recorded at 25mm/sec, and gain (amplitude of signal) at 10mm/mv, there should be no problem with the output on grid of the graphic paper.

I'm not referring to any specific model regarding reliability of 40%.  It is a general recall from reading on the subject, and I go back a few years so that might not be accurate.  However, the OP asked the reliability of an EKG, and having sourced various tracings of an EKG, I am aware of the limitations. And I gave an example of ST segment depression that can be caused by ischemia, digitalis, rapid heart rate, and temperature or electrolyte abnormality, and causes of ST Elevation: Infarction, Vasospastic angina, Pericarditis and early repolarization

The output of an EKG unit produces a list of possibiliies.   So a differential diagnosis is an appropriate systematic method to identify the unknowns with further testing and clinical evidence as well. This method, essentially a process of elimination, and is used by physicians or other clinicians to diagnose the specific disease in a patient from the abnormal EKG.

Do you have a reference of 77-82% reliability for a specific EKG system?  It seems if there is one out of five possibilities of a particular heart disorder with a given tracing produced from the ekg software, the rating intuitively seems high.


Blank
159619_tn?1318997813
Funny story..... I went in for my annual visit with my cardiologist a few years back and before she came in they did all the normal stuff including an EKG. Well, she came in and was talking to me about how well I was doing with my weight loss and daily exercise while she was looking at my file. When she got to the EKG she stopped dead in her tracks. After an uncomfortable pause, she said "there's an issue with your EKG, there are some major changes from your last one". She then started drilling me with questions like "have you experienced any chest pain in the last six months", "does your chest hurt now" and "how is your exercise tolerence". She was talking to me like I was trying to hide the truth.

After a few minutes of this, she said she was going to have a "more experinced" tech take another EKG. She left the room and I could hear her talking to the tech in the hall when she told her "walk this strip into me right away when you're done, and let the cath lab know we may be on the way down". Wow, not what I wanted to hear!

Long story short, the second EKG was normal and like all my others. She told me it was a matter of a couple leads being placed wrong. It can happen and when it does it can take a few years off your life:)
Blank
Avatar_f_tn
Yes technicians can be a problem due to lack of training in an overstretched budget driven environment. I have seen many nuclear scans with the top or bottom half of the heart missing because the scanner hasn't been aligned correctly. Patients are not pleased to learn they have to lay still for an extra 30 mins due to incompetence. However, the equipment is very good if used correctly.
Blank
Avatar_m_tn
Can you be more specific?
Blank
967168_tn?1343732745
Very interesting post made me think of some questions.  Is it normal for physicians to do the ekg themselves? I wonder if they're more or less accurate than their nurses?

How likely is it for an EKG 3 - 4 years apart to have almost the exact same dx from different techs/drs/offices?

Also, when something significant shows up on an EKG twice in a row on the same day at different levels from different machines; is that proof positive something is wrong?  I'm being really dense here, but why use 2 different machines and why is one more accurate than another?
Blank
Avatar_f_tn
usually a ward or unit is kitted out with the same make/models but the actual cardiac care unit normally has the top of the range models. When a patient is in cardiac care unit, an ecg is normally taken every hour for about 8 hours so a pattern can be seen and obviously any likely errors be ignored.
I would rather a cardiac nurse attach the leads and operate the ecg machine on me rather than a cardiologist. The reasons being, she does it many times a day and is trained on the machine. I've never seen a cardiologist operate an ecg machine, maybe they don't get the training on using them. They concentrate more on the results.
Blank
367994_tn?1304957193
Obviously, there can be operational problems with a tech, that is a given with any employment and that can be corrected with good training and supervision.  I believe the  OP is interested in the reliability of an EKG test to diagnose a heart problem and that is not very good by selectitivity standards for a particular heart disorder.  The only abnormal EKG result that has a 100% rating for selectivity is for tachycardia.

It is the EKG system's software that evaluates the electrical impulses that passes through the heart at different angles, and based on the resistance of the tissue and time for the impulse to conduct at a given angle identifies a possible problematic section.  Other tracings of the EKG may or may not substantiate, and the software prints an output of the tracings and its conclusions.  A qualified cardiologist reads the output and procedes with a differential dx if the result is abnormal.  For instance an abnormal EKG can be a fast heart rate (over 100), and the cardiologist may consider that to be anxiety and not pathological.  Or an abnormal result can be the constellation of serious and non-serious probabilities that the cardiologist will systematically eliminate with further testing and clinical evidence.

Years ago, it required a highly trained professional to read an EKG, and different interpretations were not unusual,  but today with the software involved the results are almost infalliable.

There is the American Registry of Diagnostcians that set standards to be met and tests given to assure competency. If there is incompetency at that level,  I would not trust the general care given by a hospital.  I can only imagine the problems that would exist of the personnel at that level is incompotent.

QUOTE: "Also, when something significant shows up on an EKG twice in a row on the same day at different levels from different machines; is that proof positive something is wrong?  I'm being really dense here, but why use 2 different machines and why is one more accurate than another?"

>>>>There are insignificant events that show up consistently, and some do not...such as an artifact may be EKG unit related.  It is only a tool for further investigation if positive and does not or should not be conclusive.  Sometime the symptoms of an individual require further testing if the results of the EKG is negative as well.

Blank
367994_tn?1304957193
Sorry Cath I don't agree with that assessment, "I would rather a cardiac nurse attach the leads and operate the ecg machine on me rather than a cardiologist. The reasons being, she does it many times a day and is trained on the machine. I've never seen a cardiologist operate an ecg machine, maybe they don't get the training on using them. They concentrate more on the results".
___________________________________________

A doctor knows when an EKG is not within known normal or abnormal parameters, and the output of the tracing that is meaningless and faulty in real time of the testing procedure.  For instance if an electrode lead is misplaced there can be an error in the polarity of the output...when the electrode is placed in the direction of impulse flow, there will be a positive upward slope of the wave form, electrode against the impulse flow will be a negative down sloping wave on EKG monitor.  If the electrode is not conducting, there will be a flatline to base reference. If the amplitude (heigth of waveform) is low the electrode may not be firmly attatched.

During a stress test a doctor is required to be present to supervise the procedure, and during my stress test  the doctor watching the monitor asked the tech to readjust a lead as I remember.  A competent doctor knows exactly where the electrodes should be placed based on knowing the anatomy of the heart and what angles of the heart are associated with each lead....and the knowledge of normal and abnormal wafeforms.  
Blank
Avatar_f_tn
You may disagree but perhaps things are just done differently where you are. In the hospital I work at in the UK, a Doctor certainly doesn't have to be present to ensure a technician does everything correctly. The reason being, the technician is trained to a high level with the equipment and understands it better than anyone. The manufacturer sends in training  staff to teach the technicians over a period of two weeks.
Cardiologists in the hospital know the basics of an ecg machine and will know where to place the leads etc. However, the machines are now so complicated that they have to ask a cardiac nurse to take on the procedure. They can look at the printout and request another procedure if they feel something looks amiss, but they don't actually operate the equipment. Our cardiologists don't have the time to take courses every few months when new equipment is delivered and they personally dont see the point when competent staff are trained. One piece of equipment our cardiologists are trained to use is a portable echocardiogram. They cannot use the large machines, but the portable ones are far simpler. So, perhaps cardiologists where you live are on training courses all the time, whereas in our hospitals they are with the patients.
Blank
967168_tn?1343732745
I just thought a dr doing an EKG with a nurse in the room was a bit weird, but that was in 2007 and the 2nd one I had done.  

When I went to my pcp recently, he sent in 2 seperate nurses to do the different EKG's I had, and both my cardiologist & EP, the nurse did both EKG's but a NP hooked me up for my stress test and stayed there the entire time.

These were all private practice offices, and the hospital I went to the nurses did everything and the physicians only came in to do surgery or talk to me after.
Blank
Avatar_f_tn
Where I worked as a tech, we were the ones who taught the interns about running EKGs and reading EKGs. There is a huge difference between having a doctor in the room during a stress test and having a doctor run an EKG in the office. In the former, he is ONLY in the room in case of an emergency; in the office or on the wards, he doesn't need to run the EKG although I have seen many cardiologists running EKGs, they even ran the machines on my own daughter at times. I think it just depends on the individual doctor.
Blank
976897_tn?1379171202
But getting back to the original question, I would still answer with "very". This is because there is nothing else. Without ecg machines, how else would you measure the electrical activity of the heart. To go from no information to the output from those machines is a huge step in a diagnosis and taking the number of accurate diagnosis into account, it cant be called anything but very accurate in my opinion.
If we had no ecg machines, or scanners then there would be a lot more dead people.
Blank
367994_tn?1304957193
Good point. Yes, the subject of reliability should differentiate arrhythmia from heart dimensions, structure, and diseases, and that may not have been done here. Better evaluation with other testing procedures for heart disorders exclusive of rhythm maladies.

"Healthy people who take the test are at very little risk. It's about the same as if they walk fast or jog up a big hill. Medical professionals should be present in case something unusual happens during the test."  Most people who have a stress test have a condition that may be exacerbated with physical stress, and everything I have ever read and my experience indicates that precaution is observed...healthy individuals notwithstanding.

For details regarding supervision and interpretation of exercise tests, the reader should refer to the ACC/AHA/American College of Physicians’ "Clinical Competence Statement on Stress Testing."Exercise testing should be conducted only by well-trained personnel with a sufficient knowledge of exercise physiology. Only technicians, physiologists, nurses, and physicians familiar with normal and abnormal responses during exercise can recognize or prevent adverse events. Equipment, medications, and personnel trained to provide advanced cardiopulmonary resuscitation (CPR) must be readily available.

I am referring to high end state-of-the art EKG recorders.There are many good recorders designed to capture high-quality ECG data during exercise. Many use microprocessors to generate average waveforms and make ECG measurements. The PHYSICIAN must compare the raw analog data with computer-generated output to validate its accuracy. EKG monitoring by a physician during a stress test is in realtime.  Anything less would be counter productive and put the patient at unnecessary risk..

Computer processing is not completely reliable because of software limitations in handling noise (artifacts) and inadequacy of the available algorithms.  It has been stated in another post that reliability for detecting heart disorders at 70%...no source available.
Blank
976897_tn?1379171202
It has been stated in another post that reliability for detecting heart disorders at 70%

It isn't the machines which are at fault, it is the doctors nearly all the time. A research study was done by NICE in the uk to evaluate 6 doctors by giving them ecg readouts to examine. All the ecg readouts were from patients already selected to have further testing with echocardiography due to possible heart malfunction.
Only TWO out of SIX doctors correctly diagnosed the readouts. There is now to be more intensive training with understanding ecg readings. Of the 90 readouts given to the six doctors, the ecg machine had no errors.
So manufacturers may claim certain accuracies but how can we truly evaluate an accuracy when most doctors can't read the printouts? Feedback could be coming from doctors who have misread the readout or didn't understand it.
Blank
367994_tn?1304957193
Doctors do a differential analysis based the microprosessors's software alogrithms and the reported outcome of the process as many if not most EKGs in the United States now are recorded by digital, automated machines equipped with software that measures EKG intervals and amplitudes, provides a virtually instantaneous interpretation, and often compares the tracing to those recorded earlier by the same system.

I had a stress test and when the ST interval was greater than 1mm, the test was halted by the physician monitoring the EKG output.  Consideration with an  ST elevation can be an  Infarction, Vasospastic angina, Pericarditis and early repolarization (arrythmia problem), and without doing a differential dx to eliminate each possible condtions, I wouldn't be surprised a group of examiners for different opinions for an abnormal ST interval as there are several possibilities.    


Blank
Avatar_f_tn
The most important part of the EKG is the patient's history. My daughter had an EKG that showed a typical WPW complex. She didn't have WPW due to the fact that she had had an open-heart surgery to do the ablation of the pathways and had had several EP Studies which proved she did not have this problem, but the EKG still continued to show the short PR interval and the delta wave as well as the wide QRS. This look was due to the area of the surgery as well as the LVH and a Maheim Fiber in the ventricle. Her doctors (including one of the top pediatric cardiologists in the study of childhood EKGs, in the world) words to us were: "Good luck trying to convince any cardiologist that she doesn't have WPW!"
Blank
976897_tn?1379171202
I don't think you are seeing my point here. If all doctors are trained the same way and reach the same level of knowledge, then their decisions should all result in the same conclusion. An example is, take 10 school children aged 10 and ask them all what is 1+1. Would you expect any of them to say 3? or 0?. No of course not, and if they did, they certainly wouldn't receive the appropriate grade, as a doctor wouldn't.
With the example ecg printouts in the research i mentioned, there was only one possible conclusion which made the test that much easier. It was a simple test, nothing difficult and any doctor would be expected to give the correct analysis. Again, it's like asking these doctors the result of 1+1. Only 2 out of 6 gave the answer as 2 and the others were way off the mark. It had nothing to do with the machine, it gave an accurate readout confirmed by echograms and it had nothing to do with algorithms or software. To make the test easy, there could only be one possible conclusion for the readouts. So to me, looking at it with common sense, it seems like the problem usually lies with the doctor and not the equipment. I have always maintained that it doesn't matter how clever you are, you cannot and will never be able to teach anyone common sense. I believe many doctors just have a huge memory capacity to hold data they've been taught or read, and this is what they mostly call on. Pioneers are the ones who at times push the boundaries of the text books and give medicine new limits. Unfortunately these are very rare and it took me two years to find one.
Blank
367994_tn?1304957193
Grendslori makes a good point that EKG's are "hard wired" to produce a result,  and my point has been that a cardiologist should only accept the result of an EKG after doing further testing to rule out or substantiate all EKG results.

For instance a short PR segment consideration can be Wolff-Parkinson-White syndrome (more than one pathway for an impulse) or hypertrophic obstruction cardiomyopathy, and other possible conditions. Also, a short PR interval may also occur as a normal variant, although it is much more common in the pediatric population and at faster heart rates. In itself it is not dangerous. A short PR interval may be associated with an otherwise normal electrocardiogram or a myriad of bizarre electrocardiographic abnormalities.

If there is only one possible conclusion from an EKG print out, then the expectation would be all doctors or professionals should agree on the result, but I haven't seen any ekg tracing configuration(s) that indicates a single possible condition...exception would be tachycardia.  Or the state-of-art for an ekg's software is to have the ability to diagnose to a single condition (I'm not aware of any),  then it would be unnecesasary for a doctor to do any further analysis it seems to me....maybe the doctors overly depend on the software and  are getting lazy and overlook something on the ekg printout that is relevant. Can you site the reference?
Blank
Post a Comment
To
Top Heart Disease Answerers
976897_tn?1379171202
Blank
ed34
watford, United Kingdom
159619_tn?1318997813
Blank
erijon
Salt Lake City, UT
63984_tn?1385441539
Blank
Flycaster305
97303, OR
212161_tn?1391090750
Blank
heartfluttersflyawayplz
hoschton, GA
Avatar_f_tn
Blank
skydnsr
612551_tn?1247839157
Blank
Jerry_NJ
NJ
MedHelp Health Answers
Blank
BloodPressure Tracker
Monitor Your Blood Pressure
Start Tracking Now
Blank
HeartRhythm Tracker
Track your Heart Condition
Start Tracking Now
Recent Activity
1755089_tn?1376057155
Blank
drewlevas uploaded new photos
11 mins ago
139792_tn?1299416777
Blank
Dalubaba commented on photo
1 hr ago
1236893_tn?1408490528
Blank
National Vaccine Information Center...
2 hrs ago by gymdandee
Heart Disease Community Resources
RSS Expert Activity
242532_tn?1269553979
Blank
Emotional Eating Control: How to St...
Aug 28 by Roger Gould, M.D.Blank
233488_tn?1310696703
Blank
New Cannabis Article from NORTH Mag...
Jul 20 by John C Hagan III, MD, FACS, FAAOBlank
242532_tn?1269553979
Blank
3 Reasons Why You are Still Binge E...
Jul 14 by Roger Gould, M.D.Blank
Top Heart Disease Answerers
976897_tn?1379171202
Blank
ed34
watford, United Kingdom
159619_tn?1318997813
Blank
erijon
Salt Lake City, UT
63984_tn?1385441539
Blank
Flycaster305
97303, OR
212161_tn?1391090750
Blank
heartfluttersflyawayplz
hoschton, GA
Avatar_f_tn
Blank
skydnsr
612551_tn?1247839157
Blank
Jerry_NJ
NJ