SEARCH

SEARCH BY CITATION

Keywords:

  • chest pain;
  • multidetector computed tomography (MDCT);
  • cost–effectiveness;
  • computed tomography of the coronary arteries;
  • echocardiography

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References
  9. Supporting Information

Objectives:  The aim was to use a computer model to estimate the cost–effectiveness of 64-slice multidetector computed tomography (MDCT) of the coronary arteries in the emergency department (ED) compared to an observation unit (OU) stay plus stress electrocardiogram (ECG) or stress echocardiography for the evaluation of low-risk chest pain patients presenting to the ED.

Methods:  A decision analytic model was developed to compare health outcomes and costs that result from three different risk stratification strategies for low-risk chest pain patients in the ED: stress ECG testing after OU care, stress echocardiography after OU care, and MDCT with no OU care. Three patient populations were modeled with the prevalence of symptomatic coronary artery disease (CAD) being very low risk, 2%; low risk, 6% (base case); and moderate risk, 10%. Outcomes were measured as quality-adjusted life years (QALYs). Incremental cost–effectiveness ratios (ICERs), the ratio of change in costs of one test over another to the change in QALY, were calculated for comparisons between each strategy. Sensitivity analyses were conducted to test the robustness of the results to assumptions regarding the characteristics of the risk stratification strategies, costs, utility weights, and likelihood of events.

Results:  In the base case, the mean (±standard deviation [SD]) costs and QALYs for each risk stratification strategy were MDCT arm $2,684 (±$1,773 to $4,418) and 24.69 (±24.54 to 24.76) QALYs, stress echocardiography arm $3,265 (±$2,383 to $4,836) and 24.63 (±24.28 to 24.74) QALYs, and stress ECG arm $3,461 (±$2,533 to $4,996) and 24.59 (±24.21 to 24.75) QALYs. The MDCT dominated (less costly and more effective) both OU plus stress echocardiography and OU plus stress ECG. This resulted in an ICER where the MDCT arm dominated the stress echocardiography arm (95% confidence interval [CI] = dominant to $29,738) and where MDCT dominated the ECG arm (95% CI = dominant to $7,332). The MDCT risk stratification arm also dominated stress echocardiography and stress ECG in the 2 and 10% prevalence scenarios, which demonstrated the same ICER trends as the 6% prevalence CAD base case. The thresholds where the MDCT arm remained a cost-saving strategy compared to the other risk stratification strategies were cost of MDCT, <$2,097; cost of OU care, >$1,092; prevalence of CAD, <70%; MDCT specificity, >65%; and a MDCT indeterminate rate, <30%.

Conclusions:  In this computer-based model analysis, the MDCT risk stratification strategy is less costly and more effective than both OU-based stress echocardiography and stress ECG risk stratification strategies in chest pain patients presenting to the ED with low to moderate prevalence of CAD.

ACADEMIC EMERGENCY MEDICINE 2008; 15:1–10 © 2008 by the Society for Academic Emergency Medicine

Chest pain is the most common emergency department (ED) diagnosis in patients 50 years or older.1 These patients have a broad spectrum of disease likelihood, and ultimately, well over 50% will not be diagnosed with coronary artery disease (CAD).2–5 The identification of ED chest pain patients with significant CAD remains a challenge.

One strategy is the use of chest pain observation units (OUs) in the ED. Several EDs have developed OUs to efficiently and safely manage low-risk chest pain patients with serial cardiac enzymes and subsequent cardiac stress testing.6–10 Although these units are an improvement over simply admitting or discharging patients with chest pain, there is still a significant cost and time investment involved.11–14

Recently, the 64-slice multidetector computed tomography (MDCT) of the coronary arteries has emerged as a modality for evaluation of CAD.15–24 Some have advocated use of MDCT in the ED for low-risk chest pain patients as an alternative to stress testing.25,26 Although currently not the standard of care in the ED setting, in the future, MDCT may become a first-line screening instrument for detecting significant CAD in low-risk patients presenting to the ED with chest pain. It is unclear whether the increased cost of the MDCT test itself could be associated with better patient outcomes and if these higher costs with improved outcomes could potentially represent a good use of resources.

The aim of this study was to estimate the cost–effectiveness of MDCT in the ED compared to an OU stay plus stress electrocardiogram (ECG) or stress echocardiography for the evaluation of low risk chest pain patients presenting to the ED.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References
  9. Supporting Information

Study Design

A decision analytic model was developed to compare the health outcomes and costs that result from different risk stratification strategies for ED patients with low-risk chest pain. Computerized decision analysis is a technique intended to aid in decision-making under uncertain conditions that incorporates a variety of inputs and multiple outcomes to provide a structured approach in addressing the question of interest. Ethics approval was not required for this study because it did not involve the use of human subjects or medical records.

Study Setting and Population

The characteristics of the population included in the analysis were similar to those of a 54-year-old male, which reflects the average age and most prevalent gender in our OU low-risk chest pain population. Another reason for the specific age and gender used in the model is to allow for determination of life expectancy.27 Three estimates of the prevalence of clinically significant CAD were evaluated with the model very low risk (2%), low risk (6%; base case), and moderate risk (10%). Analyses were run separately for each of these groups. CAD was defined as stenosis of at least 50% in the left arterial descending artery or 70% in any other coronary artery as measured by angiography.28

The structure of the decision tree is shown in Figure 1. The tree begins with a patient with chest pain who is considered low risk for symptomatic CAD presenting to an ED. Upon presentation, each patient is evaluated under one of the following scenarios: 1) stress ECG testing after OU care, 2) stress echocardiography after OU care, or 3) MDCT with no OU care. With the exception of accounting for MDCT-indeterminate results, the structure of the remainder of the decision tree is identical for each of the three arms following the risk stratification scenario.

image

Figure 1.  Tree structure (mathematical model) used in the decision model. The three therapies evaluated are presented after the decision node, indicated by a square. Given the size of the tree, only two of the trees is shown; however, stress ECG is identical to stress echocardiogram and begins at the noted [+]. At the chance nodes, indicated by a circle, the patient has either the presence or the absence of significant CAD, then gets a positive or negative test and continues from left to right. The outcomes and costs are entered at the payoff node, indicated by the triangle at the end of the model. The total cost for that simulated patient is added together and reported to one of the three risk stratification managements for which the patient underwent. CABG = coronary artery bypass graft; CAD = coronary artery disease; CP = chest pain, ECG = electrocardiogram; MDCT = multidetector computed tomography; MI = myocardial infarction; OU = observation unit; PCI = percutaneous coronary intervention.

Download figure to PowerPoint

For the two stress test arms, patients get OU care, which includes cardiac enzymes and cardiac monitoring. This is followed by a stress echocardiography or stress ECG test, and a disposition is made from these results. For the MDCT arm, patients get the MDCT in the ED, and depending on the result, a disposition will be made. Test performance characteristics were abstracted from the literature and inputted into the model to determine the rate of true-positive and false-negative tests for those with CAD and the false-positive and true-negative tests for those without CAD. Those with CAD and a positive test are assumed to then have angiography with the outcome of dying from the test or surviving the test and then receiving percutaneous coronary intervention (PCI), coronary artery bypass grafting (CABG), or medical management. The endpoint of these three management strategies is one of the following health state outcomes: alive in good health, alive with a myocardial infarction (MI), or death. Patients with a false-negative test do not receive any interventions and have one of the previously described outcomes. Those without CAD and a false-positive test have angiography with the outcomes of alive in full health or dying from the test. Finally, a negative test result leads to no further testing. The analysis was conducted from a hospital perspective. Direct costs of patient care and outcomes were estimated for 30 days after presentation.

Input Parameters

We estimated prevalence of CAD, sensitivities, specificities, morbidity, mortality, costs, and utilities based on the published literature5,13,27,29–43 with the exception of OU hospital costs and cost of MDCT, which came from the our institution. The input parameters for the model are listed in Table 1, and further details regarding these input parameters can be found in Data Supplement S1 (available online at http://www.blackwell-synergy.com/doi/suppl/10.1111/j.1553-2712.2008.00161.x/suppl_file/acem_161_sm_DataSupplementS1.pdf). All input parameters are reported with 95% confidence intervals (CI) if available; otherwise, ranges are reported.

Table 1.   Model Inputs and Ranges Used in Sensitivity Analysis
Inputted ParameterBase Case AnalysisRange Used in Sensitivity AnalysisReference
  1. ECG = electrocardiogram; MDCT = multidetector computed tomography; ICER = incremental cost–effectiveness ratios; QALY = quality-adjusted life-year; CAD = coronary artery disease; MI = myocardial infarction; OU = observation unit; PCI = percutaneous coronary intervention; CABG = coronary artery bypass graft

Sensitivity/specificity
 MDCT (%)
  Sensitivity9980–100Vanhoenacker et al.34
  Specificity8460–100Vanhoenacker et al.34
 Stress echocardiography (%)
  Sensitivity8560–90Kertai et al.36
  Specificity7060–90Kertai et al.36
 Stress ECG (%)
  Sensitivity7460–90Kertai et al.36
  Specificity6965–90Kertai et al.36
Probabilities (%)
 CAD61–70Goodacre and Calvert13 
 Missed CAD
  Death7.72–75Pope et al.5
  MI4410–75Estimated
  Health4910–75Estimated
 Death after angiography
  CAD-positive0.10.05–0.5Lozner et al.37
  CAD-negative0.020.018–0.02Lozner et al.37
  CABG101–50Yusuf et al.38,39
  Medical management101–50Yusuf et al.38,39
  PCI8050–99Yusuf et al.38 Anonymous39
  CABG and death3.01–10Shroyer et al.40
  CABG and MI3.31–10Mercado et al.41
  Medical management and death21–5Anonymous39 Lenzen et al.42
  Medical management and MI21–5Lozner et al.37 Lenzen et al.42
  Death after PCI2.51–10de Winter et al.43
  MI after PCI8.51–50de Winter et al.43
  Indeterminate MDCT3.80–30Sun and Jiang35
Cost ($)
 MDCT1,500$750–$3,000Institution
 Stress echocardiography277$188–$750Medicare 2007
 Stress ECG105$78–$312Medicare 2007
 Angiography2,278$1,282–$5,126Medicare 2007
 PCI12,228$8,273–$33,092Medicare 2007
 CABG35,723$23,240–$92,958Medicare 2007
 OU physician and hospital1,712$856–$3,424Institution
 Missed CAD and Death58,745$29,373–$117,490Estimated and Shaw33
 Missed CAD and MI15,549$7,776–$31,098Mahoney et al.32
Utility
 Alive
  Health1 Tsevat et al.29
  MI0.880.5–0.95Tsevat et al.29
  Death0 Tsevat et al.29
Life expectancy (yr)
 After health24.77 Social Security Administration27
 After MI11.2 Peeters et al.30 Mark et al.31

Utilities

Utility weights, which can range from 0 (death) to 1 (perfect health), were assigned to each outcome. Each utility was multiplied by the probability of being in a health state (for example, MI) to estimate quality-adjusted life-years (QALYs). Therefore, years of life with symptoms of disease receive less “credit” in calculating QALYs. Estimates of quality-of-life reduction after an MI were from previously published literature.29 Utilities were assigned to patients depending on correct or incorrect diagnosis of their CAD. Patients with false-negative tests, who were discharged home with clinically significant CAD, and who survived and had no MI are assumed to have eventually received treatment and survived with a utility of 1. Overall utility of outcomes used for death, MI, and health were 0, 0.88, and 1, respectively.29

Life Expectancy.  Estimates of life expectancy were based on longitudinal data from the literature.27 The life expectancy reduction after a nonfatal MI was obtained from previous published analyses.30,31 In this model we used a 54-year-old male, which is the average age and most prevalent gender in our OU and another studied OU.44,45

Health Care Costs.  Because we were interested in the incremental difference between the three risk stratification strategies, we measured only components of costs that differ among the alternative risk stratification strategies. For this reason, costs start after initial ED workup is finished. The costs after this workup and subsequent management contribute to differences in the incremental costs of each testing strategy. We used our institution-specific costs for MDCT, OU care, and cost of missed MI and death, whereas costs of stress echocardiography, stress ECG, coronary angiography, PCI, and CABG were from Medicare payments based on national average relative value units for 2006.46 Costs, such as missed CAD with MI, were taken from the literature.32 The cost of missed CAD with death was taken from our institution’s cost and compared with the literature33 (see Table 1 for description and sources of all health care costs used in the model). All costs were adjusted to 2007 U.S. dollars using the Medical Care component of the Consumer Price Index.47 No discounting was necessary, because costs and cost–effectiveness were examined for a 30-day period after adjustment to 2007 dollars.

Data Analysis

For each of the arms of the tree, we estimated the average costs and outcomes associated with each of the diagnostic strategies. The incremental cost–effectiveness ratios (ICERs) were calculated for comparisons among each strategy. TreeAge Pro 2006 Suite (TreeAge Software Inc., Williamstown, MA) was used to calculate costs and outcomes.

One-way Sensitivity Analysis/Threshold Sensitivity Analysis.  Several sensitivity analyses were conducted to test the robustness of the results to changes in the assumptions and estimates used in the model. We compiled these analyses into tornado diagrams. One-way sensitivity analyses were performed by individually varying the sensitivities/specificities of MDCT, stress echocardiography, stress ECG; probability of missed MI and death, missed MI, and health; probability of death after angiography, with and without CAD; probability of getting CABG, PCI, or medical management only; probability of death and AMI after PCI; probability of death and AMI after CABG; costs of MDCT, stress echocardiography, stress ECG, angiography, PCI, CABG, and OU; and the utility of patients who survived MI with health. Similarly, a threshold sensitivity analysis was done to determine at which point these input parameters resulted in a substantial impact on cost, effectiveness, or cost–effectiveness of each modality.

The ranges used in the sensitivity analysis are shown in Table 1. Each parameter was varied through a credible range of values to determine how much this changed the findings of the analysis. For costs, each base case analysis was ranged from 0.5 to 2.0 times the dollar amount of the parameter used for the base case.

Probabilistic Sensitivity Analysis.  We conducted a probabilistic sensitivity (Monte Carlo) analysis to evaluate uncertainty by varying all of the input model variables simultaneously to assess the overall variability of the model.48–50 Each scenario was simulated 10,000 times using Monte Carlo simulation, by drawing with replacement from beta distributions associated with each of the input parameters, to estimate the variance associated with each of the screening arms. The outcome of each individual was recorded, and all of the outcomes were averaged. This method accounts for variability among individuals and tests, which more closely resembles reality. The 95% CIs of the ICERs were determined from the simulations by taking the 2.5th percentile and the 97.5th percentile of the simulations.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References
  9. Supporting Information

Costs

In the base case analysis (Table 2), total costs of care after ED workups for low-risk patients with chest pain for each risk stratification were as follows: MDCT arm, $2,684 (95% CI = $1,773 to $4,418); stress echocardiography arm, $3,265 (95% CI = $2,383 to $4,836); and stress ECG arm, $3,461 (95% CI = $2,533 to $4,996). This cost relationship was constant throughout the prevalence of CAD studied.

Table 2.   Model Results of Cost, Quality Adjusted Life Years, and Incremental Cost-effectiveness Ratios Comparing Three Cardiac Risk Stratification Managements
RiskStress ECGStress EchoMDCTICERs
Echo vs. ECG (95% CI)MDCT vs. ECG (95% CI)MDCT vs. Echo (95% CI)
  1. CAD = coronary artery disease; CI = confidence interval; ECG = electrocardiogram; echo = echocardiogram; ICER = incremental cost–effectiveness ratios; MDCT = multidetector computed tomography; QALY = quality-adjusted life-year.

Very low, 2% CAD
Cost$2,844$2,631$2,051Dominant (dominant, $382,152/QALY)Dominant (dominant, 19,796/QALY)Dominant (dominant, $68,272/QALY)
QALYs24.7124.7224.74
Low (base case), 6% CAD
Cost$3,461$3,265$2,684Dominant (dominant, $123,467/QALY)Dominant (dominant, $7,332/QALY)Dominant (dominant, $29,738/QALY)
QALYs24.5924.6324.69
Moderate, 10% CAD
Cost$4,086$3,907 $3,288Dominant (dominant, $45,127/QALY)Dominant (dominant, $4,868/QALY)Dominant (dominant, $20,598/QALY)
QALYs24.4624.5224.63

QALYs

Table 2 presents the QALYs at each risk stratification arm of the model while changing the prevalence of CAD. The QALYs are presented to two decimal places to make apparent the small differences that result from using different test strategies, not because the estimates are as precise or clinically meaningful as this number of decimal places would imply. In the base case, patients whose prevalence of CAD was 6%, who underwent the MDCT arm, had a QALY of 24.69 years (95% CI = 24.54 to 24.76 years), whereas patients who underwent stress echocardiography and stress ECG arms had a QALY of 24.63 years (95% CI = 24.28 to 24.74 years) and 24.59 years (95% CI = 24.21 to 24.75 years), respectively. In lower-risk patients, the MDCT arm also had the highest QALYs at 24.74. This was followed by QALYs of 24.72 for stress echocardiography and 24.71 for stress ECG arms. Patients with a 10% prevalence of CAD had the same QALY trends as the base case.

Life Expectancy

Life expectancy varied little compared to QALYs. In the base case, patients whose prevalence of CAD was 6% and who underwent the MDCT arm had a life expectancy of 24.70 years, whereas patients who underwent stress echocardiography and stress ECG arms had life expectancies of 24.61 and 24.59 years, respectively. The higher- and lower-risk patients had the same trends as the base case.

The 30-day mortality in the 6% CAD prevalence was calculated per 10,000 patients for each risk stratification arm. For stress ECG, stress echocardiography, and MDCT arms, there were 27 deaths per 10,000 patients (95% CI = 12 to 46), 24 deaths per 10,000 patients (95% CI = 11 to 42), and 17 deaths per 10,000 patients (95% CI = 8 to 30), respectively. Comparing the mortality of the stress ECG arm to the MDCT arm, the MDCT arm had a statistically significant decrease in mortality with 10 fewer deaths per 10,000 patients (95% CI = 4 to 20). Comparing the mortality of the stress echocardiography arm to the MDCT arm, the MDCT arm had a statistically significant decrease in mortality with seven fewer deaths per 10,000 patients (95% CI = 3 to 10).

The total MIs in the base case were calculated per 10,000 patients for each arm. For stress ECG, stress echocardiography, and MDCT arms, there were 124 MIs per 10,000 patients (95% CI = 56 to 219), 103 MIs per 10,000 patients (95% CI = 46 to 182), and 42 MIs per 10,000 patients (95% CI = 19 to 74), respectively. There was a statistically significant decrease in MIs in the MDCT arm compared to the stress ECG arm, with 82 fewer MIs per 10,000 patients (95% CI = 40 to 140). There were fewer MIs in the MDCT arm compared to the stress echocardiography arm, with 61 fewer MIs per 10,000 patients (95% CI = 27 to 110).

ICER

Table 2 presents the effects of each testing strategy for a 54-year-old male on costs, QALYs, and ICERs. In the base case, those with a 6% prevalence of CAD, the MDCT arm dominated (lower cost, more effective) the stress ECG arm with the 95% CI ranged from dominant to $7,332 per QALY gained. When comparing the MDCT arm to the stress echocardiography arm, the MDCT arm was dominant and the 95% CI ranged from dominant to $26,738 per QALY gained. This result is displayed graphically in the cost–effectiveness plane (Figure 2). Each of the points in Figure 2 represents 10,000 estimated incremental cost and effectiveness pairs derived using Monte Carlo simulation. The majority of the points are located in the bottom right quadrant of the graph, indicating lower costs and higher effectiveness (negative ICER or dominant). A few of the points are located in the top right quadrant of the graph, indicating that the MDCT has higher costs but higher effectiveness versus the stress echocardiography. These results suggest that the strategy using MDCT is the dominant strategy in the base case.

image

Figure 2.  Cost–effectiveness plane. Each dot represents an estimate of the incremental effectiveness versus the cost of using a strategy with MDCT versus OU + stress echocardiography for all 10,000 values tested in our sensitivity analysis. The predominance of values in the bottom right quadrant represents the MDCT strategy as more effective and less costly than stress echocardiography combined with OU stay. MDCT = multidetector computed tomography; OU = observation unit.

Download figure to PowerPoint

In chest pain patients with a 2% prevalence of CAD, the MDCT risk stratification arm dominated the stress ECG arm, and the 95% CI ranged from MDCT being dominant to an ICER of $19,796 per QALY gained. We found that the MDCT arm dominated the stress echocardiography arm in 2% prevalence CAD patients where the 95% CI ranged from MDCT being dominant to an ICER of $68,272 per QALY gained. In moderate-risk chest pain patients, those with a 10% prevalence of CAD, the MDCT arm dominated the stress ECG arm (95% CI = dominant to $4,868/QALY) and the MDCT arm dominated the stress echocardiography arm (95% CI = dominant to $20,598/QALY).

The stress echocardiography risk stratification arm was also compared to the stress ECG arm with probabilistic sensitivity analysis. The stress echocardiography arm dominated the stress ECG arm in each of the three CAD prevalence groups; however, the upper bound of the 95% CI varied depending on the prevalence of CAD. In very-low-risk chest pain patients (2% CAD prevalence), the 95% CI ranged from the stress echocardiography arm being dominant to an ICER of $382,152/QALY gained. As prevalence of CAD increased, the lower bound of the 95% CI remained in the quadrant where the stress echocardiography arm dominated the stress ECG arm, but the upper bound of the 95% CI for the ICER decreased (low-risk patients, 6% CAD prevalence, upper bound = $123,467/QALY gained; moderate-risk patients, 10% CAD prevalence, upper bound = $45,127/QALY gained).

Although the results of different prevalence of CAD are not identical to the base case, the MDCT risk stratification arm consistently dominated the stress echocardiography and stress ECG arms; that is, the MDCT arm was less costly and had superior outcomes compared to the other two alternatives.

Sensitivity Analysis

The MDCT risk stratification strategy remained dominant compared to the other strategies because it was associated with better outcomes and lower costs. At our base case prevalence of 6% CAD, there were five input parameters that resulted in the MDCT risk stratification strategy having higher costs than the other alternatives. The point at which the MDCT arm was associated with higher costs was cost of MDCT >$2,097, cost of OU care <$1,092, prevalence of CAD >70%, specificity MDCT <65%, and indeterminate rate of MDCT >30%. Under each of these scenarios, the MDCT risk stratification strategy was associated with higher costs; however, it remained the more effective alternative, and under all of these scenarios, the ICER was well below $10,000 per QALY gained. The following variables were not influential within reasonable ranges: costs of stress echocardiography, stress ECG, angiography, and PCI and sensitivities of MDCT, stress echocardiography, and stress ECG. A tornado diagram of one-way sensitivity analyses for five of the input parameters is shown in Figure 3. Tornado diagrams graphically depict the effect of the selected input variables over a range described to demonstrate the greatest contributors to the outcome variable (ICER). For example, when comparing stress echocardiography to MDCT in Figure 2, the third bar from the top, “Cost of MDCT,” which ranges from $750 to $3,000, gives an ICER of $–13,000 per QALY gained to $10,000 per QALY gained. In this model, if MDCT were $750, the hospital would be a gaining $13,000 per QALY compared to stress echocardiography, whereas if the MDCT cost $3,000, it would cost the hospital $10,000 per QALY gained compared to the stress echocardiography. The dotted line illustrates the break-even point; everything to the right of this dotted line is where MDCT is more costly than the stress echocardiography.

image

Figure 3.  Tornado diagram made by ranges of sensitivity analyses comparing stress echocardiogram versus MDCT. Vertical dotted line represents reference point of incremental cost–effectiveness, where everything to the right of this makes MDCT more costly than stress echocardiogram. Note that only five variables touch or cross this line; however, all are still within $10,000 incremental cost–effectiveness. CAD = coronary artery disease; MDCT = multidetector computed tomography; OU = observation unit.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References
  9. Supporting Information

In this analysis we found that MDCT in the ED is more cost-effective than observation and monitoring in an OU followed by a stress echocardiography or stress ECG in low-risk patients presenting to the ED with chest pain. Furthermore, the MDCT risk stratification strategy, across most of the analyses, was a cost-saving approach of evaluating the low-risk chest pain population.

Only one prospective study evaluating cost of 64-slice MDCT versus standard of care has been done so far. Goldstein et al.26 evaluated the cost of MDCT compared to nuclear stress testing in low-risk patients presenting to the ED with chest pain. They reported an overall lower cost for the MDCT arm of about $300. Of note, their cost of MDCT was reported as $507 versus nuclear stress testing of $538. Overall, they showed MDCT was cost beneficial in low-risk chest pain patients arriving to the ED, mostly due to time saving and earlier discharge. This study is an example of why MDCT done early in the workup is less costly relative to standard care in this low-risk population. However, due to MDCT being a relatively new technology and this being a prospective study, only a total of 197 patients were enrolled, with 9 (4.6%) patients having clinically significant disease. The computer analytic decision model used in our work can lead to a better understanding of the generalizability of these findings. In our model, we were able to determine whether the disease prevalence, cost of tests, sensitivity, specificity, MDCT-indeterminate rate, and other variables alter the conclusions regarding the cost–effectiveness of the MDCT risk stratification strategy. We found that the results were robust to nearly all of the assumptions that were made in developing the model and the MDCT strategy may be a cost-effective alternative for low-risk chest pain in the ED when tested prospectively.

For MDCT to be cost-saving, the cost of MDCT must be less than $2,097, the cost of OU care must be greater than $1,092, the prevalence of CAD in patients should be less than 70%, the MDCT specificity must be greater than 65%, and the MDCT-indeterminate rate should be less than 30%. However, assuming that MDCT becomes standard-of-care for patients presenting with low- to moderate-risk chest pain, and a societal willingness to pay per QALY gained of $50,000, MDCT performed in the ED may be cost-effective compared to the other two alternatives mentioned previously in all of the ranges used in the sensitivity analysis given in Table 1.

The MDCT’s effectiveness does hinge on its high sensitivity and specificity. At a sensitivity of 99% versus a sensitivity of 74% for stress echocardiography, there is little question of why the MDCT is more effective than stress echocardiography. MDCT is more accurate in identifying patients who require intervention and those who can be safely discharged from the ED. Failure to detect disease can diminish life expectancy and the number of QALYs. Thus, test sensitivity is a major cause of variation in health outcomes. Furthermore, the significant upfront cost of the MDCT radiologic test in the ED prevents admissions to the OU and therefore foregoes OU cost. This ultimately makes the MDCT arm less costly than the other options.

Limitations

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References
  9. Supporting Information

Our analysis relies heavily on published assessments of test performance that met a set of predefined quality criteria. Because of this, there is potential to have a significant publication bias; researchers who studied diagnostic test performance who found lower sensitivity and specificity may be less likely to submit or have their work accepted for publication. This is particularly true for sensitivity and specificity of MDCT, which is a relatively new technology that has only been described in a few small studies. For this reason, a wide range was used in the sensitivity and specificity while performing the probabilistic analysis.

Our performance characteristics of MDCT came from the literature;34,35 however, due to the relatively few studies using 64-slice MDCT, comparing this result to the criterion standard, angiography, we use sensitivity and specificity of the MDCT from a group of studies with a CAD prevalence of 64%.34 Although this is different from our low-risk population, we believe that the sensitivity and specificity of the MDCT may be as good, if not greater, in this low-risk population due to less calcification of the coronary arteries, which has been cited as a contributor to poor image quality and indeterminate MDCT tests.35

Mortality rates of PCI are based on literature that studied patients with non-ST segment elevation MIs (NSTEMIs). Our patient populations are neither NSTEMI patients nor are they stable CAD patients, given their complaint of chest pain. Thus, the probability of death and AMI after PCI may be overestimated. To address this, a wide range was used in the sensitivity analysis and results were consistent across this range.

Along with sensitivity, specificity, and mortality rates, all input parameters taken from the literature may be over- or understated in the low-risk OU population. This is secondary to multiple factors, including differences in CAD prevalence, patient population, expertise of the physician interpreting the image or test result, test accuracy, and cost differences in the local market. As mentioned, there is only one randomized controlled trial of MDCT versus OU care, which limits the quality of the data available for determining precise point estimates for MDCT diagnostic performance for this decision analysis. For this reason, a wide range was used in the probabilistic analysis including utility scores and costs. In addition, we incorporated both meta-analyses of MDCT in general CAD populations and multiple observational studies for assumptions that were not constructed on randomized controlled trial data. As better estimates for the input parameters become available, it will be important to reevaluate the results of this analysis.

This cost–effectiveness analysis compares MDCT to two modalities of stress testing. Although some of the patients in the OU are admitted to the hospital for cardiac evaluation before stress testing due to worsening chest pain, development of an abnormal cardiac enzyme, or having ECG changes, we believe that this is a low number in most institutions. It is less than 2% in our OU.44

Our estimate of cost–effectiveness assumes ED access to timely MDCT, which may not be possible at all institutions. In cases where MDCT may not be available, OU stay costs may need to be incorporated into future cost–effectiveness analyses.

Debate is still ongoing regarding approaches to minimize soft plaques causing nondetectable lesions on MDCT,51 as well as radiation dosage issues.52,53 A recent study shows the radiation from a 64-slice scanner evaluating the coronary arteries has a lifetime attributed risk of cancer, and it is not negligible.54 The increased risk of cancer may influence the long-term cost–effectiveness, although we only evaluated the 30-day outcomes. Importantly, this may change the patients’ willingness to undergo MDCT. Both of these factors ultimately may impact the ICERs. Further studies need to be done to understand the actual risks.

Patients who have significant CAD on MDCT need to get another catheterization and most likely PCI. This requires another dye load. Because most of these lesions are stable in this low-risk population, intravenous fluids and other modalities to protect renal function can be employed for 24 hours until a second dye load is administered. There is an extremely small potential of renal failure due to the double dye load; however, this is not addressed in our model.

Conclusions

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References
  9. Supporting Information

With newer, more expensive technologies, it is imperative that a cost–effectiveness analysis be done to justify their use. A large randomized controlled trial must be done to compare MDCT to current standard-of-care to definitively change practice. Computer-simulated cost–effectiveness analysis must also be done to evaluate these risk stratification modalities. In this computer-based model analysis, the MDCT risk stratification strategy is less costly and more effective than both OU-based stress echocardiography and stress ECG risk stratification strategies in chest pain patients presenting to the ED with low to moderate prevalence of CAD. This is largely due to the diagnostic test performance of MDCT and the avoidance of OU costs.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References
  9. Supporting Information
  • 1
    McCaig L, Nawar E. National Hospital Ambulatory Medical Care Survey: 2004 emergency department summary. Adv Data. 2006; 372:129.
  • 2
    Lee TH, Rouan GW, Weisberg MC, et al. Clinical characteristics and natural history of patients with acute myocardial infarction sent home from the emergency room. Am J Cardiol. 1987; 60:21924.
  • 3
    McCarthy BD, Beshansky JR, D’Agostino RB, et al. Missed diagnoses of acute myocardial infarction in the emergency department: results from a multicenter study. Ann Emerg Med. 1993; 22:57982.
  • 4
    Mehta RH, Eagle KA. Missed diagnoses of acute coronary syndromes in the emergency room--continuing challenges. N Engl J Med. 2000; 342:120710.
  • 5
    Pope JH, Aufderheide TP, Ruthazer R, et al. Missed diagnoses of acute cardiac ischemia in the emergency department. N Engl J Med. 2000; 342: 116370.
  • 6
    Gibler WB, Runyon JP, Levy RC, et al. A rapid diagnostic and treatment center for patients with chest pain in the emergency department. Ann Emerg Med. 1995; 25:18.
  • 7
    Nichol G, Walls R, Goldman L, et al. A critical pathway for management of patients with acute chest pain who are at low risk for myocardial ischemia: recommendations and potential impact. Ann Intern Med. 1997; 127:9961005.
  • 8
    Polanczyk CA, Johnson PA, Hartley LH, Walls RM, Shaykevich S, Lee TH. Clinical correlates and prognostic significance of early negative exercise tolerance test in patients with acute chest pain seen in the hospital emergency department. Am J Cardiol. 1998; 81:28892.
  • 9
    Storrow AB, Gibler WB. Chest pain centers: diagnosis of acute coronary syndromes. Ann Emerg Med. May 2000; 35:44961.
  • 10
    Zalenski RJ, Rydman RJ, Ting S, Kampe L, Selker HP. A national survey of emergency department chest pain centers in the United States. Am J Cardiol. 1998; 81:13059.
  • 11
    Polanczyk CA, Kuntz KM, Sacks DB, Johnson PA, Lee TH. Emergency department triage strategies for acute chest pain using creatine kinase-MB and troponin I assays: a cost-effectiveness analysis. Ann Intern Med. 1999; 131:90918.
  • 12
    Roberts RR, Zalenski RJ, Mensah EK, et al. Costs of an emergency department-based accelerated diagnostic protocol vs hospitalization in patients with chest pain: a randomized controlled trial. JAMA. 1997; 278:16706.
  • 13
    Goodacre S, Calvert N. Cost effectiveness of diagnostic strategies for patients with acute, undifferentiated chest pain. Emerg Med J. 2003; 20:42933.
  • 14
    Nucifora G, Badano LP, Sarraf-Zadegan N, et al. Comparison of early dobutamine stress echocardiography and exercise electrocardiographic testing for management of patients presenting to the emergency department with chest pain. Am J Cardiol. 2007; 100:106873.
  • 15
    Leschka S, Alkadhi H, Plass A, et al. Accuracy of MSCT coronary angiography with 64-slice technology: first experience. Eur Heart J. 2005; 26:14827.
  • 16
    Mollet NR, Cademartiri F, van Mieghem CA, et al. High-resolution spiral computed tomography coronary angiography in patients referred for diagnostic conventional coronary angiography. Circulation. 2005; 112:231823.
  • 17
    Nikolaou K, Knez A, Rist C, et al. Accuracy of 64-MDCT in the diagnosis of ischemic heart disease. AJR Am J Roentgenol. 2006; 187:1117.
  • 18
    Raff GL, Gallagher MJ, O’Neill WW, Goldstein JA. Diagnostic accuracy of noninvasive coronary angiography using 64-slice spiral computed tomography. J Am Coll Cardiol. 2005; 46:5527.
  • 19
    Pugliese F, Mollet NR, Runza G, et al. Diagnostic accuracy of non-invasive 64-slice CT coronary angiography in patients with stable angina pectoris. Eur Radiol. 2006; 16:57582.
  • 20
    Ropers D, Rixe J, Anders K, et al. Usefulness of multidetector row spiral computed tomography with 64- x 0.6-mm collimation and 330-ms rotation for the noninvasive detection of significant coronary artery stenoses. Am J Cardiol. 2006; 97:3438.
  • 21
    Meijboom WB, Mollet NR, Van Mieghem CA, et al. 64-slice computed tomography coronary angiography in patients with non-ST elevation acute coronary syndrome. Heart. 2007; 93:13256.
  • 22
    Muhlenbruch G, Seyfarth T, Soo CS, Pregalathan N, Mahnken AH. Diagnostic value of 64-slice multi-detector row cardiac CTA in symptomatic patients. Eur Radiol. 2007; 17:6039.
  • 23
    Oncel D, Oncel G, Tastan A, Tamci B. Detection of significant coronary artery stenosis with 64-section MDCT angiography. Eur J Radiol. 2007; 62:394405.
  • 24
    Hoffmann U, Nagurney JT, Moselewski F, et al. Coronary multidetector computed tomography in the assessment of patients with acute chest pain. Circulation. 2006; 114:225160.
  • 25
    Gallagher MJ, Ross MA, Raff GL, Goldstein JA, O’Neill WW, O’Neil B. The diagnostic accuracy of 64-slice computed tomography coronary angiography compared with stress nuclear imaging in emergency department low-risk chest pain patients. Ann Emerg Med. 2007; 49:12536.
  • 26
    Goldstein JA, Gallagher MJ, O’Neill WW, Ross MA, O’Neil BJ, Raff GL. A randomized controlled trial of multi-slice coronary computed tomography for evaluation of acute chest pain. J Am Coll Cardiol. 2007; 49:86371.
  • 27
    Social Security Administration. Actuarial Publications: Period Life Table, 2003. Social Security Online. July 9, 2007. Available at: http://www.ssa.gov/OACT/STATS/table4c6.html. Accessed Feb 7, 2008.
  • 28
    Williams SV, Fihn SD, Gibbons RJ, et al. Guidelines for the management of patients with chronic stable angina: diagnosis and risk stratification. Ann Intern Med. 2001; 135:53047.
  • 29
    Tsevat J, Goldman L, Soukup JR, et al. Stability of time-tradeoff utilities in survivors of myocardial infarction. Med Decis Making. 1993; 13:1615.
  • 30
    Peeters A, Mamun AA, Willekens F, Bonneux L. A cardiovascular life history. A life course analysis of the original Framingham Heart Study cohort. Eur Heart J. 2002; 23:45866.
  • 31
    Mark DB, Harrington RA, Lincoff AM, et al. Cost-effectiveness of platelet glycoprotein IIb/IIIa inhibition with eptifibatide in patients with non-ST-elevation acute coronary syndromes. Circulation. 2000; 101:36671.
  • 32
    Mahoney EM, Jurkovitz CT, Chu H, et al. Cost and cost-effectiveness of an early invasive vs conservative strategy for the treatment of unstable angina and non-ST-segment elevation myocardial infarction. JAMA. 2002; 288:18518.
  • 33
    Shaw LJ. Myocardial perfusion imaging in the evaluation of chest pain in the acute care setting: Clinical and economic outcomes. J Nuclear Cardiol. 2007; 14(3 Suppl):S1338.
  • 34
    Vanhoenacker PK, Heijenbrok-Kal MH, Van Heste R, et al. Diagnostic performance of multidetector CT angiography for assessment of coronary artery disease: meta-analysis. Radiology. 2007; 244:41928.
  • 35
    Sun Z, Jiang W. Diagnostic value of multislice computed tomography angiography in coronary artery disease: a meta-analysis. Eur J Radiol. 2006; 60: 27986.
  • 36
    Kertai MD, Boersma E, Bax JJ, et al. A meta-analysis comparing the prognostic accuracy of six diagnostic tests for predicting perioperative cardiac risk in patients undergoing major vascular surgery. Heart. 2003; 89:132734.
  • 37
    Lozner EC, Johnson LW, Johnson S, et al. Coronary arteriography 1984-1987: a report of the Registry of the Society for Cardiac Angiography and Interventions. II. An analysis of 218 deaths related to coronary arteriography. Catheterization and cardiovascular diagnosis. 1989; 17:114.
  • 38
    Yusuf S, Zucker D, Peduzzi P, et al. Effect of coronary artery bypass graft surgery on survival: overview of 10-year results from randomised trials by the Coronary Artery Bypass Graft Surgery Trialists Collaboration. Lancet. 1994; 344:56370.
  • 39
    Anonymous. Coronary artery surgery study (CASS): a randomized trial of coronary artery bypass surgery. Survival data. Circulation. 1983; 68(5):93950.
  • 40
    Shroyer AL, Coombs LP, Peterson ED, et al. The Society of Thoracic Surgeons: 30-day operative mortality and morbidity risk models. Ann Thorac Surg. 2003; 75:185664.
  • 41
    Mercado N, Wijns W, Serruys PW, et al. One-year outcomes of coronary artery bypass graft surgery versus percutaneous coronary intervention with multiple stenting for multisystem disease: a meta-analysis of individual patient data from randomized clinical trials. J Thorac Cardiovasc Surg. 2005; 130:5129.
  • 42
    Lenzen M, Boersma E, Bertrand M, et al. Management and outcome of patients with established coronary artery disease: the Euro Heart Survey on coronary revascularization. Eur Heart J. 2005; 26:116979.
  • 43
    de Winter RJ, Windhausen F, Cornel JH, et al. Early invasive versus selectively invasive management for acute coronary syndromes. N Engl J Med. 2005; 353:1095104.
  • 44
    Khare R, Powell E, Fernandez E, et al. Prevalence of non-diagnostic stress tests in observation unit patients: potential implications for CT coronary angiography [abstract]. Acad Emerg Med. 2007; 14(5 Suppl 1):S198.
  • 45
    Goodacre SW, Morris FM, Campbell S, Arnold J, Angelini K. A prospective, observational study of a chest pain observation unit in a British hospital. Emerg Med J. 2002; 19:11721.
  • 46
    Gold MR, Siegel J, Russell L, et al. Cost-Effectiveness in Health and Medicine. New York, NY: Oxford University Press, 1996.
  • 47
    U.S. Department of Labor BoLS. Consumer Price Index for Medical Care. Available at: http://www.bls.gov/cpi/home.htm. Accessed Apr 26, 2008.
  • 48
    Briggs AH. Handling uncertainty in cost-effectiveness models. Pharmacoeconomics. 2000; 17:479500.
  • 49
    Critchfield GC, Willard KE, Connelly DP. Probabilistic sensitivity analysis methods for general decision models. Comput Biomed Res. 1986; 19:25465.
  • 50
    Doubilet P, Begg CB, Weinstein MC, Braun P, McNeil BJ. Probabilistic sensitivity analysis using Monte Carlo simulation. A practical approach. Med Decis Making. 1985; 5:15777.
  • 51
    Horiguchi J, Fujioka C, Kiguchi M, et al. Soft and intermediate plaques in coronary arteries: how accurately can we measure CT attenuation using 64-MDCT? AJR Am J Roentgenol. 2007; 189:9818.
  • 52
    Hausleiter J, Meyer T, Hadamitzky M, et al. Radiation dose estimates from cardiac multislice computed tomography in daily practice: impact of different scanning protocols on effective dose estimates. Circulation. 2006; 113:130510.
  • 53
    Morin R, Gerber T, McCollough C. Physics and dosimetry in computed tomography. Cardiol Clin. 2003; 21:51520.
  • 54
    Einstein A, Henzlova M, Rajagopalan S. Estimating risk of cancer associated with radiation exposure from 64-slice computed tomography coronary angiography. JAMA. 2007; 298:31723.

Supporting Information

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References
  9. Supporting Information

Data Supplement S1. 64-slice CT of the coronary arteries data supplement (PDF file)

Please note: Blackwell Publishing are not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
ACEM_161_sm_DataSupplementS1.pdf26KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.