Author + information
- Received July 23, 2012
- Revision received September 28, 2012
- Accepted November 9, 2012
- Published online March 1, 2013.
- ↵⁎Reprint requests and correspondence:
Dr. Howard J. Willens, Miami VAMC (D-26), 1201 Northwest 16th Street, Miami, Florida 33125
Objectives This study evaluated utilization of stress echocardiography (SE) at our institution, the impact of the updated 2011 appropriate use criteria (AUC) on appropriateness ratings, correlation of AUC to radiology benefits managers' (RBM) pre-certification guidelines and the effect of temporal trends and an AUC-based educational project on appropriateness.
Background The AUC for SE have been developed to improve efficiency of utilization and promote optimal patient care.
Methods We classified the appropriateness of 209 SEs from 2008 using the original and updated AUC. We also performed pre-authorization determinations on these SEs using the guidelines of 2 RBMs. We then classified and compared the appropriateness of 209 SEs from 2011 using the updated criteria to that of the 2008 cohort. Finally, we rated and compared 111 SEs requested by cardiologists after an educational project to 111 SEs referred before the intervention.
Results Overall, nearly one-third of SEs were requested for inappropriate indications. Using 2011 AUC, the original ratings of 52 (25%) studies by AUC 2008 were changed and the number of unclassified SE decreased from 20 (9.6%) to 2 (1%). Correlation between RBM pre-authorization determination and AUC ratings was substantial for the first RBM (κ = 0.625) and fair for the second (κ = 0.358). However, 12.9% and 41.9% of studies classified as appropriate or uncertain by the AUC would not have received pre-authorization according to the guidelines of the first and second RBMs, respectively. Referrals of inappropriate SE did not decrease over time or with an educational intervention.
Conclusions The revisions in the updated AUC improve their clinical application by encompassing nearly all indications for SE. The limited correlation between AUC ratings and RBM determinations suggests a need for greater consistency. The large number of SE requested for inappropriate indications at our institution did not decrease with time or education.
In response to the dramatic growth rate of cardiac imaging and concerns about excessive utilization, health plans adopted a variety of strategies to manage utilization, including pre-authorization programs administered by radiology benefits managers (RBMs). Concurrently, the American College of Cardiology Foundation developed and published appropriate use criteria (AUC) for several imaging modalities (1,2) to eliminate unnecessary testing and promote optimal care. Updated AUC for transthoracic, transesophageal, and stress echocardiography (SE) (3) were published in 2011 with the goal of reducing the gaps of coverage in the original criteria noted in application studies (4–8) and to reflect changes in the literature. We previously reported that the revised AUC for echocardiography improved their clinical application for transthoracic echocardiography (TTE) by substantially reducing the number of unclassified studies (9).
The basis for an RBM's decision regarding authorization of an echocardiogram or other procedure varies widely, and transparency is often lacking. Although internal policies of RBMs used to determine pre-authorization of echocardiograms often cite the AUC, the extent to which these pre-certification algorithms are consistent with AUC is largely unknown. We previously demonstrated poor correlation of the pre-authorization guidelines for TTE of 2 of the largest RBMs with the AUC and between themselves (9).
Compared to TTE, there are few reports on the application of AUC for SE. The impact of the recent revisions on their clinical application and the correlation of RBMs' pre-certification algorithms for SE with AUC have not been well studied. Therefore, we undertook an investigation of the application of the 2008 and 2011 AUC for SE and the appropriateness of utilization at our institution. We sought to evaluate the impact of the updated AUC on the ratings of SEs previously classified using the 2008 AUC and the correlation of the 2008 AUC with the pre-certification algorithms for SEs of 2 RBMs. Additionally, we wanted to examine temporal trends in utilization of SE from 2008 to 2011 and the effect of an educational initiative for cardiologists on their SE referral patterns.
This study was a retrospective survey using the electronic medical records of the University of Miami Health System. The study population consisted of 3 cohorts of patients who had SE at an academically affiliated outpatient laboratory during 3 different time periods in 2008 and 2011 (Fig. 1). The first cohort consisted of 209 consecutive patients undergoing SE examination between August and September of 2008. The appropriateness of this cohort's SEs was determined using the original and revised AUC to assess the impact of the updates on appropriateness ratings. The SEs of this cohort were also used to analyze the correlation of AUC with RBM pre-authorization determinations and as the baseline studies for the evaluation of temporal trends. The second cohort consisted of 209 consecutive patients referred for SE between July and September of 2011, 4 months after the publication of the revised AUC. Their SEs were classified using the updated AUC, and the ratings were compared to those of the 2008 cohort by the same criteria to assess temporal trends in utilization. The third cohort consisted of 111 consecutive patients referred for SE by cardiologists between the third week in October and the final week of December 2011, 2 weeks after an educational initiative. To determine the impact of this intervention, the appropriateness of their SE was compared to the ratings of all studies from the second cohort ordered by our division's cardiologists (n = 111) during the 3-month period before the educational project. Patients referred for SE as part of research protocols were excluded from each of the 3 cohorts. The institutional review board of the University of Miami approved the protocol and waived patient consent.
Determination of appropriateness
A common methodology for classification of appropriateness was used for the SEs from all 3 time frames in this study. One experienced echocardiographer (H.J.W.) reviewed the requisitions and pertinent electronic records of the University of Miami Health System dating back to 1999 of patients undergoing SE from the 3 time periods. Data extracted from the medical records to classify appropriateness included demographics, cardiac history, clinical and laboratory assessment of coronary artery disease (CAD) and coronary heart disease risk, and prior imaging and invasive procedures. Pre-test probability of CAD was estimated using the method of Pryor et al. (10). We categorized dyspnea as atypical angina to be consistent with prior reports (8). Coronary heart disease risk was estimated according to the Adult Treatment Panel III of the National Cholesterol Education Program and Framingham Risk Score using the blood pressure and fasting lipid level obtained at a time point closest to the performance of SE regardless of medication use (11). Referring provider type was also recorded.
Using clinical data contained in our electronic records and indications on the requisitions, SEs were rated according to the 2008 and 2011 AUC. Requests for which an indication in the criteria could not be assigned were considered unclassified. The remaining studies were matched to indications for SE in the 2008 and 2011 AUC and classified as appropriate, inappropriate, or uncertain. In cases in which multiple indications could be equally applied, the indication with the most favorable level of appropriateness was selected.
Correlation of AUC and RBM guidelines for stress echocardiography
We searched the Internet for RBMs that had developed detailed pre-authorization guidelines for SE and referenced the 2008 AUC as a resource for creating their algorithms. For our analysis of the correlation of AUC and RBM algorithms, we were interested in evaluating RBM guidelines that were currently being applied to members of health plans referred for SE. Therefore, we selected the 2010 utilization management guidelines of 2 large RBMs that had contracted with health plans to screen SE according to a payer policies update distributed by the American Society of Echocardiography (12). The same investigator (H.J.W.) who performed the appropriateness classification determined whether the SE of the 2008 cohort would have received pre-authorization according to the guidelines for SE of these 2 RBMs blinded to the AUC ratings. Those studies not meeting RBMs' criteria for pre-certification were considered deferred.
Our primary analysis compared these pre-authorization determinations to the appropriateness ratings of the 189 studies from the 2008 cohort that were classified using the original 2008 AUC. As a secondary analysis, we also assessed the correlation between RBMs' pre-authorization determination and the appropriateness ratings of the 207 studies of the cohort that could be classified using the 2011 AUC. Because the AUC state that SEs with indications rated uncertain should be considered reimbursable (3), we considered SEs requested for both appropriate and uncertain indications suitable for receiving pre-authorization and pooled them in the analysis of level of agreement between RBMs' algorithms and AUC.
Effect of an AUC-based educational intervention
An AUC-based educational intervention targeting referring cardiologists was performed in the first week of October 2011 in an attempt to improve utilization of SE at our institution. This consisted of a cardiology grand rounds lecture regarding the rationale for the creation and contents of AUC for echocardiography. The lecture included an extensive discussion of the 5 most common inappropriate indications for TTE and SE identified at our institution during prior analyses and clinical scenarios where stress tests without imaging would be more appropriate (6,9). The lecture was attended by 15 (54%) of the 28 members of the division. In addition, printed materials describing the common inappropriate indications for SE were distributed at the time of the lecture, and a follow-up e-mail with the same list was sent to all of the division's cardiologists 1 week later. The appropriateness of 111 consecutive SEs referred by division cardiologists beginning 2 weeks after the lecture was assessed and compared to that of the 111 consecutive SEs referred by cardiologists from the 2011 cohort.
Continuous variables are reported as mean ± SD and categorical variables as number and proportions. Continuous variable were compared using analysis of variance and categorical variables by chi-square test. Age was dichotomized to <65 years and >65 years and treated as a categorical variable. For the evaluation of impact of revised AUC on appropriateness classification, we compared the ratings of the SE from 2008 by the 2011 AUC to their ratings by the 2008 AUC using McNemar's test. The relationship between the appropriateness ratings by the 2008 AUC and pre-approval determination by RBM algorithms was evaluated with Cohen's kappa statistic (k) and also reported as the percent agreement between a favorable pre-authorization determination and pooled appropriate or uncertain ratings by the criteria. Appropriateness rating and indication according to AUC 2011 and pre-authorization determinations according to both RBMs were repeated in 20 randomly selected cases by the original reviewer (H.J.W.) 6 months after the completion of the initial evaluation to assess intraobserver variability. Interobserver variability for selecting appropriateness rating and indication was assessed in 20 randomly selected cases 4 months after the initial analysis. Intraobserver and interobserver variability are reported as kappa statistic (k) and percent agreement. A 2-sided p value <0.05 was considered statistically significant. All analyses were performed with SPSS Version 19 for Windows (SPSS Inc., Chicago, Illinois).
Utilization of stress echocardiography at our institution
Overall, 529 SEs were classified using the AUC 2011. Table 1 shows the age, sex, referring providers, and appropriateness ratings for the entire study group and for each time frame. For the entire study population, patients were 56.6 ± 14.0 years old, and 252 (47.6%) were men. Of the 529 SEs, 230 (43.5%) were requested for appropriate indications, 121 (22.9%) for uncertain indications, and 163 (30.8%) for inappropriate indications according to the 2011 AUC. A total of 15 (2.8%) could not be classified using the updated AUC. The percentage of studies requested for inappropriate indications in the different cohorts was similar. The SEs for women and patients <65 years old were significantly more likely to be requested for an inappropriate indication compared with those of men and patients >65 years old (35.7% vs. 25%, p = 0.020, and 34.7% vs. 20.3%, p < 0.001, respectively). Appropriateness rating was not significantly related to referring provider type (p = 0.140).
Table 2 shows the 5 most common appropriate, inappropriate, and uncertain indications according to the revised AUC for the entire study group and the 3 individual cohorts. The evaluation of ischemic equivalent with low pre-test probability of CAD, interpretable electrocardiogram (ECG), and able to exercise was the most common inappropriate indication, accounting for 66 (40.5%) inappropriate studies and 12.5% of all referrals. The 5 most common inappropriate indications listed in Table 2 accounted for 110 (67.5%) of the 163 inappropriate SEs and 20.8% of overall referrals. Of the 15 unclassifiable studies, 5 (33.0%) were requested for evaluation of inducible outflow tract gradient in hypertrophic cardiomyopathy.
Impact of AUC updates on appropriateness ratings
Figure 2 compares the appropriateness ratings of the 2008 SE by the original AUC (Fig. 2A) and the revised AUC (Fig. 2B). Table 3 shows the reassignment of the 2008-based appropriateness ratings by the updated 2011 criteria. Overall, the appropriateness classification of 52 (25%) SEs was changed by the updated AUC (p < 0.001), including 18 (90%) of the previously unclassified studies, 19 (18%) of those previously classified as appropriate, 11 (18%) of those previously classified as inappropriate, and 4 (17%) classified uncertain by AUC 2008. The number of unclassifiable SE was markedly reduced from 20 (9.6%) to 2 (1%), and more SEs were classified as uncertain (44 [21.1%] vs. 23 [11.0%]) on the basis of the new criteria. The number of SEs classified appropriate and inappropriate by the original and updated AUC was similar. The evaluation of asymptomatic patients with frequent premature ventricular complexes, exercise-induced ventricular tachycardia, or nonsustained ventricular tachycardia and patients with new or worsening symptoms with normal coronary angiography or normal prior stress imaging study were the most common indications used to reclassify SEs that were reassigned.
Correlation of AUC and RMB algorithms
According to the algorithm of RBM #1, 132 (63%) studies from the 2008 cohort would have been pre-authorized, and 77 (37%) would not have received pre-certification. For RBM #2, 94 (45%) of studies would have been pre-authorized, and 115 (55%) would have been deferred. Table 4 shows the relationship between these RBM pre-authorization determinations and appropriateness ratings according to the 2008 and 2011 AUC. For the 189 studies that could be classified by AUC 2008, there was substantial agreement with RBM #1 (84% agreement, k = 0.633) and fair agreement with RBM #2 (65% agreement, k = 0.330). Of the 127 SEs rated appropriate or uncertain using the 2008 AUC, 17 (13.4%) according to the algorithm of RBM #1 and 55 (43.3%) according to that of RBM #2 would not have been authorized (Fig. 3). Using the 2011 criteria, the agreement between pre-authorization determinations and appropriateness ratings for the 207 classified studies was also substantial for RBM #1 (83% agreement, k = 0.616) and fair for RBM #2 (61% agreement, k = 0.246). Of the 144 SEs rated appropriate or uncertain using the 2011 AUC, 25 (17.4%) according to the algorithm of RMB #1 and 67 (46.5%) according to that of RBM #2 would not have been authorized. The large majority of SEs with a discordant pre-authorization determination were rated appropriate by both versions of the AUC as opposed to uncertain (Table 4). For both RBMs, evaluation of anginal equivalent, intermediate pre-test probability of CAD, interpretable ECG, and able to exercise (indication 3 in 2008 AUC) was the common indication with discordance between pre-authorization determination and 2008 AUC appropriateness rating, accounting for 25.8% of the discordant determinations for RBM #1 and 59.1% for RBM #2. Applying the 2011 AUC, indication 116 (the equivalent of indication 3 in 2008 AUC) and detection of CAD in patients with ventricular arrhythmias without angina (indication 130 in 2011 AUC) were common appropriate indications with a discordant authorization determination for both RBMs, accounting for 16.7% and 13.9% of the discordant determinations for RBM #1 and 44.4% and 6.2% for RBM #2. Indication 152, evaluation of new or worsening symptoms in a patient with prior normal angiography or stress imaging, was also a common clinical scenario with disagreement between RBM #2 determination and AUC 2011 rating, accounting for 9.9% of discordant cases. Agreement between the pre-authorization determinations by the 2 RBMs was only fair (67.5% agreement, k = 0.366). Indication 3 of the 2008 AUC was also the most common clinical scenario with disagreement between RBM, accounting for 31 (54.2%) of 57 discordant determinations.
Temporal trends in utilization and effectiveness of an educational intervention
Figure 4 shows the appropriateness ratings of SEs from 2011 using the updated AUC. Although fewer SEs were ordered for appropriate indications in 2011, overall there was no significant change in appropriateness ratings of SEs requested in 2011 compared to SEs of the 2008 cohort (p = 0.497).
Figure 5 compares the appropriateness ratings of SEs referred by cardiologists before the educational initiative (Fig. 5A) and after it (Fig. 5B). More SEs were requested for appropriate indications after the education project, but the number of inappropriate SEs did not decrease (31.5% vs. 32.4%). Overall, there was no significant difference in appropriateness ratings before and after the initiative (p = 0.339).
Intraobserver reproducibility for determining appropriateness rating and prior authorization according to RBM #2 guidelines was good (90% agreement, k = 0.831 for appropriateness rating and 90% agreement, k = 0.800 for RBM #2 pre-authorization determination). Intraobserver reproducibility was substantial for determining the indication for SE and the prior authorization according to RBM #1 guidelines (70% agreement, k = 0.670 for indications and 90% agreement, k = 0.792 for pre-authorization by RBM #1). Interobserver reproducibility for appropriateness rating was substantial (80% agreement, k = 0.701) and moderate for choosing the indication (60% agreement, k = 0.593).
This is the second study to evaluate the application of the updated 2011 AUC to SEs and the first to compare RBM guidelines for SEs to the AUC. Additionally, we described temporal trends in utilization of SE after the publication of the revised AUC and the impact of an educational initiative on referrals. We observed that SE was often requested for inappropriate indications according to both sets of AUC. We also observed that the revised criteria addressed more indications for SE than the original criteria. We noted that fair to substantial correlation between AUC and the SE pre-certification guidelines of 2 major RBMs was present. Additionally, we report that the number of SEs requested for inappropriate indications did not change over time despite enhanced recognition of AUC. Unfortunately, the educational initiative directed toward cardiologists also did not result in a reduction in the number of SEs requested for inappropriate indications.
Diagnostic cardiac imaging has come under scrutiny and become a common topic in the national dialogue about health care. Therefore, an improved understanding of the utilization of echocardiography has become increasingly important to the point that the Intersocietal Accreditation Commission/Echocardiography has incorporated appropriateness into laboratory accreditation requirements for 2012 (13). The Commission's current document on adult echocardiography focuses on monitoring, reporting, and educational feedback, but quality will likely be assessed based on a standard of appropriateness. The current paper provides additional information regarding appropriate use of SE.
Three prior publications reported that SEs are often requested for inappropriate indications, with the rate of inappropriate referrals ranging from 19% to 27% (6,7,14). The inappropriate indication rate was higher at our institution. Our population was younger and had more women than the other 3 reports, which would predispose our site to having more inappropriate ratings (7). The inclusion of inpatients in the report by Mansour et al. (7) may have contributed to a lower prevalence of inappropriate referrals in their study compared to ours. However, these observations may also reflect a more liberal use of imaging with stress testing in lower risk groups that has been reported for our geographic locale (15).
Overall, a relatively small number of inappropriate indications accounted for the majority of inappropriate SE in all published reports (7,8,14). Although the most common inappropriate indications varied, SE requested for 1) evaluation of ischemic equivalent with low pre-test probability of CAD, an interpretable ECG, and able to exercise (indication 114); 2) pre-operative risk for intermediate risk surgery with no clinical risk factors (indication 156); and 3) asymptomatic patients with intermediate global CAD risk and an interpretable ECG (indication 125) accounted for many of the inappropriate studies and would, therefore, be appropriate targets for future educational interventions.
Studies of AUC for SE using the original 2008 criteria reported that 12% and 19% of SEs could not be classified by the original AUC (7,8). Italian investigators at a single center recently reported that the updated AUC addressed the indications of all the pharmacological SEs requested in their practice between 2001 and 2007 (14). We observed that the updated criteria captured more of the indications for exercise and pharmacological SEs than the original AUC, nearly eliminating SEs that could not be classified. We and others have reported similar improvement in the clinical application of the revised AUC to TTE (9,16–18). These observations suggest that with publication of the revised 2011 AUC for echocardiography, the AUC have become a reliable tool for assessing utilization that can be used by insurance companies and health plans.
Many health plans have introduced prior authorization programs for high-cost diagnostic imaging procedures. Although often citing AUC, the extent to which the algorithms used by these companies to determine pre-certification are literature-based is largely unknown. Studies of pre-authorization for computed tomography and magnetic resonance imaging have shown reductions of utilization. However, whether patients with legitimate clinical indications for these procedures were either not referred or were denied the exam was not analyzed in these reports (19–21). We previously reported that only modest agreement was present between the RBMs pre-authorization determination and AUC appropriateness classification for TTE, and that a substantial numbers of TTE studies receiving an appropriate rating would have been deferred (9). Compared to our findings in that report on TTE, the correlation of AUC and RBM guidelines is somewhat better for SE. Still, >40% of SE requested for indications rated either appropriate or uncertain by the AUC would not have been authorized by a RBM. This finding implies a potential for under-utilization of SE in some health plans using pre-certification programs. Indication 3 from the 2008 AUC was the most common clinical scenario with disagreement between AUC and RBM guidelines because the recommendation for stress testing without imaging is quite broad in the algorithm of RBM #2. The discordance between AUC and RBMs for TTE and SE suggests a need for greater collaboration between medical societies and health plan RBMs.
Although 1 of the goals of AUC is to provide guidance to providers on how to select and use diagnostic imaging efficiently, their impact on utilization of imaging remains largely unknown. There have been no prior studies on temporal trends in appropriateness of utilization of SE. Data on the temporal changes for the other diagnostic cardiovascular imaging modalities after publication of original and revised AUC are also sparse (22,23). Our observations about lack of improvement in the appropriateness of referrals of SE over a 3-year period and 4 months after publication of the revised AUC are consistent with the few reports on trends in utilization for other types of diagnostic cardiovascular imaging. Rahimi et al. (22) reported no change in the appropriateness ratings of TTE of patients referred between 2000 and 2008 despite interim publication of the initial AUC for TTE. A report examining the utilization of stress single-photon computed tomography myocardial perfusion imaging (SPECT-MPI) over time reported an initial improvement after publication of AUC for SPECT-MPI. However, this favorable trend was not sustained at later follow-up (23,24). Data about the effectiveness of provider education on improving utilization of diagnostic imaging are also limited. A multicenter study of SPECT-MPI did find substantial improvement at 1 of several participating sites but not at the other imaging centers (25). Another study reported no benefit of an educational project on appropriateness of SPECT-MPI (24). Studies involving other clinical activities suggest that casual passive educational projects such as ours do not significantly modify providers' behavior (26). It has been recommended that more intensive interventions may be needed to actuate real change. Potentially useful approaches include repetitive longitudinal multifaceted education, interactive case-based learning, personalized feedback to practitioners with comparisons to benchmarks, and provision of clinical decision aids that can be used at point of care (27).
The size of each of the cohorts we studied was modest. The study was performed at an academic outpatient laboratory, and results may not be generalizable to other settings. Classification of appropriateness rating and pre-authorization status remains subjective, requiring interpretation of the AUC, RBM algorithms, and electronic medical record data. Interobserver reproducibility for choosing indications was 62% in our study, and was broadly similar to the range of 56% to 72% reported by other studies (4,8). Retrospective review of electronic medical records may not capture all the details and unique circumstances of an individual case and may be a source of classification error. The RBM algorithms in this study cited the 2008 AUC and were developed before publication of the updated AUC. Because of this, we considered the correlation between the algorithms and the original 2008 criteria as most relevant and used this in our primary analysis. Finally, changes in clinical practice from 2008 to 2011 unrelated to AUC may effect utilization and skew the assessment of temporal trends.
The revisions in the updated AUC have improved their clinical application by addressing nearly all indications for SE. Correlation of AUC with the algorithms of 2 RBMs used by health plans to manage utilization of SE is substantial for the first but only fair for the second. Stress echocardiography continues to be commonly requested for inappropriate indications despite the publication of the second AUC, and an educational intervention directed at cardiologists did not improve their referral patterns. These observations suggest a need for greater dissemination of the AUC among providers and greater collaboration between RBMs providing services to health plans and the professional societies.
The authors have reported they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations And acronyms
- appropriate use criteria
- coronary artery disease
- radiology benefits manager
- stress echocardiography
- single-photon emission computed tomography myocardial perfusion imaging
- transthoracic echocardiography
- Received July 23, 2012.
- Revision received September 28, 2012.
- Accepted November 9, 2012.
- American College of Cardiology Foundation
- Douglas P.S.,
- Khandheria B.,
- Stainback R.F.,
- et al.
- Douglas P.S.,
- Khandheria B.,
- Stainback R.F.,
- et al.
- Douglas P.S.,
- Garcia M.J.,
- Haines D.E.,
- et al.
- Ward R.P.,
- Mansour I.N.,
- Lemieux N.,
- et al.
- Willens H.J.,
- Gomez-Marin O.,
- Heldman A.,
- et al.
- McCully R.B.,
- Pellikka P.A.,
- Hodge D.O.,
- Araoz P.A.,
- Miller T.D.,
- Gibbons R.J.
- National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III)
- ↵ASE Update Payer Policies: April 2011. www.asecho.org/files/Preauthsumm201001.pdf. Accessed July 11, 2011.
- ↵Intersocietal Accreditation Commission (IAC) Echocardiography. www.intersocietal.org/echo/. Accessed July 15, 2011.
- Wennberg D.E.,
- Birkmeyer J.D.,
- et al.
- Cortigiani L.,
- Bigi R.,
- Bovenzi F.,
- Molinaro S.,
- Picano E.,
- Sicari R.
- Parikh P.B.,
- Asheld J.,
- Kort S.
- Ballo P.,
- Bandini F.,
- Capecchi I.,
- et al.
- Mitchell J.M.,
- LaGalia R.
- Gibbons R.J.,
- Askew J.W.,
- Hodge D.,
- Kaping B.,
- Carryer D.J.,
- Miller T.
- Hendel R.C.,
- Cerqueira M.,
- Douglas P.S.,
- et al.
- Sclafani J.J.,
- My J.,
- Zacher L.L.,
- Eckart R.E.
- Arrighi J.A.