coronary artery CT angio vs. observation admission cost

[Hooper]

---

Members don't see this ad.

I'm doing a presentation for my EM elective. Anybody know what the approximate difference in cost is between getting CCTA and being admitted as observation status?

This is for a low to intermediate risk chest pain patient with non-ischemic ECG and negative 1st set of enzymes. The admission would be for chest pain rule out with serial cardiac enzymes and maybe further testing like stress test.

Ball park numbers would be a great start or perhaps a link where I can look this up.

 

[docB]

I think this will be a tough topic.

Sorting out "cost" is difficult because the amount billed vs. actual cost to administer are wildly different. The billing for studies is always a hospital pipe dream that no one ever pays. The insured always get some lower contracted rate, Medicare pays whatever it pays and pays less than cost on a lot of stuff. If you can get a hospital cost it's usually wrong as it doesn't include stuff like staffing the scanner, the read, etc.

I would also suggest comparing the rads exposure and morbidity associated with stress vs. CCTA as an aside for your presentation.

Good luck.

 

[danzman]

I'm doing a presentation for my EM elective. Anybody know what the approximate difference in cost is between getting CCTA and being admitted as observation status?

This is for a low to intermediate risk chest pain patient with non-ischemic ECG and negative 1st set of enzymes. The admission would be for chest pain rule out with serial cardiac enzymes and maybe further testing like stress test.

Ball park numbers would be a great start or perhaps a link where I can look this up.

This is an interesting blog entry about your topic that links to the NEJM article about it.

I think its going to be pretty difficult to actually get any real numbers on all of this. Too many variables.

http://blogs.nejm.org/now/index.php/ct-angiography-for-safe-discharge-of-patients/2012/04/11/

 

[docB]

You might want to talk to one of your hospital's case managers. They might be able to get you a list of the contracted rates for each arm with the various insurers. That's at least some real data. You could tell them you don't need to know who pays what. On second thought that might be proprietary info. They don't want Blue Cross to find out that they are paying more than someone else. Case management is still worth a shot.

 

[Hooper]

To clarify, my presentation is exactly about the article that danzman has linked. I present a low risk chest pain patient and try to answer the question: Can the patient go home straight from the ED with a negative CCTA?

I just wanted to mention the costs in my presentation as an aside to identify if CCTA would save money on healthcare expenditure compared to obs admission. I have absolutely no idea what either costs, but I heard from one ED doc that they are comparable. My guess would be that CCTA costs less but I just wanted to know for sure. I guess the numbers I would be looking for would be amount billed to insurance companies, not what was actually paid. I am only interested in cost difference and I assume amount paid is proportional to amount billed.

docB, that would also be a good question. As far as radiation, I found the following link useful:

http://www.radiologyinfo.org/en/safety/index.cfm?pg=sfty_xray

So CCTA is 12mSV and calcium scoring is 3mSV adding up to 15 mSV which would amount to an equivalent of about 5 years of background radiation. Estimated increased lifetime cancer risk is 1 in 10,000 to 1 in 1,000. As a reference, head CT=2mSV, A/P CT=10mSV, and CXR=0.1mSV. Possible morbidity from CCTA would be contrast reaction (allergy or renal) and beta-blocker induced brady-arrhythmia. I don't really know much about stress testing.

 

[Tiger26]

Obs itself can be incredible expensive, since some hospitals bill it out as outpatient services which may/may not be covered by insurance. I've known of patients with good health insurance who've received bills well above $10,000 for obs protocols.

 

[Arcan57]

To clarify, my presentation is exactly about the article that danzman has linked. I present a low risk chest pain patient and try to answer the question: Can the patient go home straight from the ED with a negative CCTA?

I just wanted to mention the costs in my presentation as an aside to identify if CCTA would save money on healthcare expenditure compared to obs admission. I have absolutely no idea what either costs, but I heard from one ED doc that they are comparable. My guess would be that CCTA costs less but I just wanted to know for sure. I guess the numbers I would be looking for would be amount billed to insurance companies, not what was actually paid. I am only interested in cost difference and I assume amount paid is proportional to amount billed.

docB, that would also be a good question. As far as radiation, I found the following link useful:

http://www.radiologyinfo.org/en/safety/index.cfm?pg=sfty_xray

So CCTA is 12mSV and calcium scoring is 3mSV adding up to 15 mSV which would amount to an equivalent of about 5 years of background radiation. Estimated increased lifetime cancer risk is 1 in 10,000 to 1 in 1,000. As a reference, head CT=2mSV, A/P CT=10mSV, and CXR=0.1mSV. Possible morbidity from CCTA would be contrast reaction (allergy or renal) and beta-blocker induced brady-arrhythmia. I don't really know much about stress testing.

Nuclear stress testing (which is what most hospitals are doing these days) ranges from 9.4 mSV (for technitium) to 40 mSV for thallium. The number you're quoting for the CCTA sound high to me, I've heard 5-10mSV thrown about.

 

[xaelia]

CT-STAT and ROMICAT II compared CCTA-based strategies to "standard care" - usually a nuclear stress - and tracked costs.

CT-STAT tracked ED costs, and they were reduced by 38% compared with "standard care". ROMICAT II reports 19% decrease in ED costs, but then 50% increase in subsequent hospital costs.

The safety aspect, unfortunately, is a bit of self-fulfilling prophecy. Only 2 of the 1357 patients who didn't rule-in for MI in ACRIN-PA at the index visit went on to have an MI within 30 days, which is consistent with the prior literature showing patients with two negative troponins have a ~0.3% risk of MI within 30 days. Basically, when you run all these low- to zero-risk patients through the scanner in these studies, you're not well-demonstrating the incremental prognostic value of the test - and end up with a lot of false positives. This is why you see in CT-STAT the six-fold greater incidence of downstream non-invasive testing, and in ROMICAT II the reporting of 50% increased hospital costs. There's a JAMA article from last year that looks at outpatient Medicare enrollees without known CAD undergoing CCTA - and their costs were ~$12000 in the next six months, compared with ~$8000 for nuclear stress and ~$5000 for stress echo and treadmill. The costs in that study were all CCTA patients undergoing revascularization procedures - and actually, the mortality trend was worse in the CCTA group receiving all these extra interventions and expenditures.

So, in any event, you're asking a very complex question to which the evidence is incomplete and inconclusive. The answer to low-risk chest pain is not in these tests, it's in fixing the broken expectations for the evaluation of low-risk chest pain. These patients should go home after appropriate clinical and biomarker risk-stratification, with the minority receiving outpatient stress testing, something like the recently published strategy from Vancouver in Annals of EM.

 

[WilcoWorld]

The answer to low-risk chest pain is not in these tests, it's in fixing the broken expectations for the evaluation of low-risk chest pain.

Agreed. It feels like Achilles racing the Tortoise when we try to further reduce the miss rate in patients with atypical pain + negative biomarkers + normal or non-diagnostic ECG's. Most of the testing we do is simply to prove to future juries that we met "the standard of care" even though we're probably hurting more patients than we help.

 

[docB]

Agreed. It feels like Achilles racing the Tortoise when we try to further reduce the miss rate in patients with atypical pain + negative biomarkers + normal or non-diagnostic ECG's. Most of the testing we do is simply to prove to future juries that we met "the standard of care" even though we're probably hurting more patients than we help.

The poor resource utilization being driven by this defensiveness is ruining society. We could vaccinate the whole world on what we spend ruling out MI in low risk patients.

 

[Byvi16]

1

 

[Byvi16]

.

 

[Hooper]

Nuclear stress testing (which is what most hospitals are doing these days) ranges from 9.4 mSV (for technitium) to 40 mSV for thallium. The number you're quoting for the CCTA sound high to me, I've heard 5-10mSV thrown about.

I got this number from here: http://www.radiologyinfo.org/en/safety/index.cfm?pg=sfty_xray.

Website developed by ACR and RSNA. It's information for patients, but nonetheless comes from radiologists.

CT-STAT and ROMICAT II compared CCTA-based strategies to "standard care" - usually a nuclear stress - and tracked costs.

CT-STAT tracked ED costs, and they were reduced by 38% compared with "standard care". ROMICAT II reports 19% decrease in ED costs, but then 50% increase in subsequent hospital costs.

The safety aspect, unfortunately, is a bit of self-fulfilling prophecy. Only 2 of the 1357 patients who didn't rule-in for MI in ACRIN-PA at the index visit went on to have an MI within 30 days, which is consistent with the prior literature showing patients with two negative troponins have a ~0.3% risk of MI within 30 days. Basically, when you run all these low- to zero-risk patients through the scanner in these studies, you're not well-demonstrating the incremental prognostic value of the test - and end up with a lot of false positives. This is why you see in CT-STAT the six-fold greater incidence of downstream non-invasive testing, and in ROMICAT II the reporting of 50% increased hospital costs. There's a JAMA article from last year that looks at outpatient Medicare enrollees without known CAD undergoing CCTA - and their costs were ~$12000 in the next six months, compared with ~$8000 for nuclear stress and ~$5000 for stress echo and treadmill. The costs in that study were all CCTA patients undergoing revascularization procedures - and actually, the mortality trend was worse in the CCTA group receiving all these extra interventions and expenditures.

So, in any event, you're asking a very complex question to which the evidence is incomplete and inconclusive. The answer to low-risk chest pain is not in these tests, it's in fixing the broken expectations for the evaluation of low-risk chest pain. These patients should go home after appropriate clinical and biomarker risk-stratification, with the minority receiving outpatient stress testing, something like the recently published strategy from Vancouver in Annals of EM.

I'm gonna definitely look at these studies to be well prepared for my presentation on Friday, so thanks for sharing them.

"Only 2 of the 1357 patients who didn't rule-in for MI in ACRIN-PA at the index visit went on to have an MI within 30 days, which is consistent with the prior literature showing patients with two negative troponins have a ~0.3% risk of MI within 30 days."

I'm a clinically inexperienced medical student and you clearly are more well read on this topic, but I'm not sure I see your point. All the patients in this study were going to be admitted if there was no study. So even though they are low-zero risk, the healthcare climate EP's practice in, leads to chest pain rule out admissions. However, the CCTA group got discharged twice as much directly from the ED and had shorter hospital stays. The fact that there were only two MI's (one in each group) is a good thing, and shows that you can safely discharge these patients from the ED with a negative CCTA. With downstream costs, are you referring to CCTA use in general? Because I cannot speak to that. I'm only talking about CCTA in the isolated setting of low risk cardiac patients with non-ischemic ECG, one set of negative cardiac enzymes, who would otherwise be admitted for two more sets of enzymes and possible stress testing. If negative, they can go home and don't need any further testing, even outpatient. If positive or indeterminate, they can be admitted for further testing which would have happened anyway without a CCTA.

Please correct me if I'm misunderstanding.

 

[xaelia]

In the CCTA studies, all patients had a second set of cardiac enzymes prior to discharge. In ACRIN-PA, it was 90-180 minutes after the first, I believe.

My point is, to establish the negative likelihood ratio for a test, you need to have some reasonable incidence of disease in your patient population. If they're saying a negative CCTA proves you're going to have an event-free period, you need to evaluate CCTA in a group that has lots of events as a whole - but not in the negative CCTA group. In ACRIN-PA, whether you had positive or negative CCTA, if you didn't have an MI on the first visit, you didn't have an MI within 30 days. If you look back at the original ROMICAT, they did the CCTA but blinded the treating physicians to the results. 8 out of 368 in that study had MI, while 23 more were diagnosed with UA at the index visit via conventional testing. In their ~6 month follow-up period, no one had an MI or death from cardiovascular causes. However, because the treating physicians were blinded to the results, they unknowingly were sending home patients with CAD and stenosis on CCTA after ruling them out for MI. And they did fine. Therefore, it's not the CCTA that adds much to the prognostic value after biomarkers and risk stratification.

You should be aware that to rule someone out for MI does not require three sets of enzymes 6-8 hours apart - it's something that can be done with two sets 2-4 hours apart in the ED. When we're admitting for a r/o, we're really asking the inpatient team to determine if this fleeting chest pain is unstable angina - which is entirely different.

Also, it should make a difference whether you trust these studies when you look at:
http://www.medical.siemens.com/siemens/en_US/gg_ct_FBAs/files/brochures/ArticleSession24.pdf
and consider that Hoffman, Litt, and Raff are lead authors or heavily involved in the centers responsible for ROMICAT, ACRIN-PA, and CT-STAT, respectively.

 

[Arcan57]

I got this number from here: http://www.radiologyinfo.org/en/safety/index.cfm?pg=sfty_xray.

Website developed by ACR and RSNA. It's information for patients, but nonetheless comes from radiologists.

With new scanners and dose-saving techniques, some of the scanners are in the 1.3-3.5 mSV range. 3 years ago the standard was 10-20mSV, but with the advent of 256 (or more) slice scanners and dose saving techniques the amount of radiation has dropped significantly. It may be worth trying to find out what scanner your institution has, since if they are using a 64 slice scanner under old protocols the numbers you were quoting are probably as accurate as anyone can get when estimating rad dosing.

In terms of how much CCTA helps with risk stratifying low-risk patients, we honestly need to not do any testing on low-risk patients once they've ruled out for MI. When I admit a low-risk chest pain patient to our cardiologists, if they are young and healthy they get a couple more sets of cardiac markers. It's vanishingly rare that they get any further in-patient testing, and most don't get a stress test as an outpatient either. CAD exists in young patients with no risk factors, but chasing after it is so low-yield that the expense and morbidity outweighs the benefits. I don't think CCTA changes the equation in TIMI 0-1 patients.

I think CCTA could be all kinds of useful in the patient that's intermediate risk or in patients that constantly present with chronic angina (ie the drug-seekers).

 

[Greenbbs]

"I think CCTA could be all kinds of useful in the patient that's intermediate risk or in patients that constantly present with chronic angina (ie the drug-seekers)."

The problem with this is that in the medicolegal climate we are currently in, people who constantly present may constantly get CCTA testing, especially when people don't do appropriate chart reviews, etc. Then you get into increased rads exposure, contrast reactions, etc, and the NNH may equal the NNT

The other thing that could prove interesting is seeing what the incidence of post CCTA contrast induced nephropathy would be. Xaelia had an interesting article on his blog the other day about CTA for PE that showed a higher incidence of nephropathy tha positive PE findings. One would have to wonder if this was true for the CCTA groups as well

 

[Arcan57]

"I think CCTA could be all kinds of useful in the patient that's intermediate risk or in patients that constantly present with chronic angina (ie the drug-seekers)."

The problem with this is that in the medicolegal climate we are currently in, people who constantly present may constantly get CCTA testing, especially when people don't do appropriate chart reviews, etc. Then you get into increased rads exposure, contrast reactions, etc, and the NNH may equal the NNT

The other thing that could prove interesting is seeing what the incidence of post CCTA contrast induced nephropathy would be. Xaelia had an interesting article on his blog the other day about CTA for PE that showed a higher incidence of nephropathy tha positive PE findings. One would have to wonder if this was true for the CCTA groups as well

The CCTA has less morbidity than a cath, which is currently the recommended test of choice for refractory chest pain. The current standard of care in my area is to get x number of stress tests until eventually the cardiologist gets sick of stressing them, at which point they move on to another cardiologist/hospital system to repeat the cycle.

I don't think CCTA makes the current situation worse (caveat: I don't have access to it at my current shop), and the NNT for someone presenting who actually has chronic, non-ischemic chest pain approaches infinity. Since they don't have the disease, we need to focus on minimizing resource utilization and minimizing harm to them. If you've got the medical records showing recent negative caths, rule them out and send them home. However, this patient population bounces from hospital to hospital (and sometimes city to city) and the first time they present you have to take them seriously. I honestly wonder what the negative predictive value of "this feels like when I had my last heart attack but worse" and having no records at your facility is.

In regards to nephropathy, the current accepted research definition for CIN is an increase in Cr of 25% from baseline or an absolute increase of 0.5 mg/dl after a contrast exposure. However, in most patients the bump is transient and doesn't require anything more than supportive care. This absolutely matters if you have someone with multiple co-moribidities and concommitant serious illness, but in healthy patients without pre-existing renal disease that you'd discharge after a negative test it's usually without significance.

 

[Hooper]

In the CCTA studies, all patients had a second set of cardiac enzymes prior to discharge. In ACRIN-PA, it was 90-180 minutes after the first, I believe.

My point is, to establish the negative likelihood ratio for a test, you need to have some reasonable incidence of disease in your patient population. If they're saying a negative CCTA proves you're going to have an event-free period, you need to evaluate CCTA in a group that has lots of events as a whole - but not in the negative CCTA group. In ACRIN-PA, whether you had positive or negative CCTA, if you didn't have an MI on the first visit, you didn't have an MI within 30 days. If you look back at the original ROMICAT, they did the CCTA but blinded the treating physicians to the results. 8 out of 368 in that study had MI, while 23 more were diagnosed with UA at the index visit via conventional testing. In their ~6 month follow-up period, no one had an MI or death from cardiovascular causes. However, because the treating physicians were blinded to the results, they unknowingly were sending home patients with CAD and stenosis on CCTA after ruling them out for MI. And they did fine. Therefore, it's not the CCTA that adds much to the prognostic value after biomarkers and risk stratification.

You should be aware that to rule someone out for MI does not require three sets of enzymes 6-8 hours apart - it's something that can be done with two sets 2-4 hours apart in the ED. When we're admitting for a r/o, we're really asking the inpatient team to determine if this fleeting chest pain is unstable angina - which is entirely different.

Also, it should make a difference whether you trust these studies when you look at:
http://www.medical.siemens.com/siemens/en_US/gg_ct_FBAs/files/brochures/ArticleSession24.pdf
and consider that Hoffman, Litt, and Raff are lead authors or heavily involved in the centers responsible for ROMICAT, ACRIN-PA, and CT-STAT, respectively.

If you want to establish a negative likelihood ratio for a test, I think you only need to study the population in which the test will be used. In this case it's the low risk patients with low prevalence of disease that are being admitted unnecessarily. In the ACRIN-PA trial, 14% percent of CCTA group got stress tested and 4% got cathed. In the traditional group, 58% got stress tested, 4% got cathed, 6% got CCTA. All of these stress tests were done on inpatients. So with CCTA you can significantly cut down on admissions and inpatient stress tests.

I see what you are saying about the 2 sets of enzymes 2-4 hours apart for MI. So all these people are being admitted to rule out unstable angina. CCTA can rule out unstable angina (by ruling out CAD, right?)

I wish you never showed me that link to Siemens because it pretty much shatters my whole presentation. I've done most of my research and am presenting tomorrow so I'm not finding a new topic. I'll just disclose it at the end and hopefully get a chuckle out of it.

 

[Hooper]

To their credit, the authors do cite their conflicts of interest in the paper.

 

[xaelia]

I mostly mean to say the CI for the negative likelihood ratio is really wide. The authors themselves admit that you need 40,000 patients to truly show a difference between CCTA and another similar clinical strategy due to the paucity of events.

Having <50% stenosis on CCTA doesn't rule out MI, because the rupture of mixed atherosclerotic/fibrofatty lesions is a common etiology for cardiac ischemia. I'm sure increasingly absent CAD observed on CCTA has a correlation with ruling out for MI, but troponins are much cheaper.

Regarding conflict of interest - in the current academic culture, it's essentially an ethics violation to fail to report your COI - so, no, the authors don't get any credit for compliance. Now, what the COI actually means is another matter entirely - literature describing the effects of COI on both reporters and readers shows some complexity to the compensation mechanisms, but, suffice to say, the less COI the better.

But, yes, when I presented this study for our journal club a couple months back, I finished with that PDF featuring Litt, Hoffman, and Raff as salespersons for Siemens. You have to be careful about ad hominem attacks, but it can be funny if presented in a non-accusatory fashion.

 

[Arcan57]

I mostly mean to say the CI for the negative likelihood ratio is really wide. The authors themselves admit that you need 40,000 patients to truly show a difference between CCTA and another similar clinical strategy due to the paucity of events.

Having <50% stenosis on CCTA doesn't rule out MI, because the rupture of mixed atherosclerotic/fibrofatty lesions is a common etiology for cardiac ischemia. I'm sure increasingly absent CAD observed on CCTA has a correlation with ruling out for MI, but troponins are much cheaper.

Regarding conflict of interest - in the current academic culture, it's essentially an ethics violation to fail to report your COI - so, no, the authors don't get any credit for compliance. Now, what the COI actually means is another matter entirely - literature describing the effects of COI on both reporters and readers shows some complexity to the compensation mechanisms, but, suffice to say, the less COI the better.

But, yes, when I presented this study for our journal club a couple months back, I finished with that PDF featuring Litt, Hoffman, and Raff as salespersons for Siemens. You have to be careful about ad hominem attacks, but it can be funny if presented in a non-accusatory fashion.

And actually disclosing a COI makes the audience trust you more than they otherwise would, despite what common sense would say.