Author + information
- Received March 24, 2017
- Revision received May 26, 2017
- Accepted July 2, 2017
- Published online November 20, 2017.
- Javier A. Valle, MD, MSCSa,
- Lisa A. Kaltenbach, MSb,
- Steven M. Bradley, MD, MPHc,
- Robert W. Yeh, MD, MScd,
- Sunil V. Rao, MDe,
- Hitinder S. Gurm, MDf,
- Ehrin J. Armstrong, MD, MSca,g,
- John C. Messenger, MDa and
- Stephen W. Waldo, MDa,g,∗ ()
- aDivision of Cardiology, University of Colorado School of Medicine, Aurora, Colorado
- bDuke Cardiovascular Research Institute, Durham, North Carolina
- cMinneapolis Heart Institute, Minneapolis, Minnesota
- dSmith Center for Outcomes Research in Cardiology, Beth Israel Deaconess Medical Center, Boston, Massachusetts
- eDivision of Cardiology, Duke University School of Medicine, Durham, North Carolina
- fDivision of Cardiovascular Medicine, University of Michigan School of Medicine, Ann Arbor, Michigan
- gSection of Cardiology, Veterans Affairs Eastern Colorado Health Care System, Denver, Colorado
- ↵∗Address for correspondence:
Dr. Stephen W. Waldo, Section of Cardiology, Veterans Affairs Eastern Colorado Health Care System, 1055 Clermont Drive, Denver, Colorado 80220.
Objectives The study sought to define patient, operator, and institutional factors associated with transradial access (TRA) in ST-segment elevation myocardial infarction (STEMI) percutaneous coronary intervention (PCI), the variation in use across operators and institutions, and the relationship with mortality and bleeding.
Background TRA for PCI in STEMI is underutilized. Factors associated with TRA are not well described, nor is there variation across operators and institutions or their relationship with outcomes.
Methods The authors used hierarchical logistic regression to identify patient, operator, and institutional characteristics associated with TRA use as well as determine the variation in TRA for STEMI PCI from 2009 to 2015. They also described the relationship between operator- and institution-level use and risk-adjusted bleeding and mortality.
Results Among 692,433 patients undergoing STEMI PCI, 12% (n = 82,618) utilized TRA. TRA increased from 2% to 23% from 2009 to 2015, but with significant geographic variation. Age, sex, cardiogenic shock, cardiac arrest, operators entering practice before 2012, and nonacademically affiliated institutions were associated with lower rates of TRA. There was significant operator and institutional variation, wherein identical patients would have >8-fold difference in odds of TRA for STEMI PCI by changing operators (median odds ratio: 8.7), and >5-fold difference by changing institutions (median odds ratio: 5.1). Greater TRA use across operators was associated with reduced bleeding (rho = −0.053), whereas TRA use across institutions was associated with reduced mortality (rho = −0.077).
Conclusions Transradial access for STEMI PCI is increasing, but remains underutilized with significant geographic, operator, and institutional variation. These findings suggest an ongoing opportunity to standardize STEMI care.
Substantial evidence supports the use of transradial access (TRA) for percutaneous coronary intervention (PCI). Several randomized trials have demonstrated a reduction in bleeding and access site complications among patients treated with radial access, compared with those treated with the femoral approach (1–5). Additional studies and meta-analysis suggest an even greater benefit among those with ST-segment elevation myocardial infarction (STEMI), with a reduction in observed mortality (3,4,6). Despite these data, use of transradial techniques remains limited in the United States when compared with European use, particularly among patients with STEMI (7–11).
Factors associated with the limited use of TRA for percutaneous treatment of myocardial infarction have been understudied. Previous research has identified patient-level characteristics associated with limited use of a transradial approach, including advanced age or the concomitant presence of peripheral artery disease (7,12). Factors beyond the patient may also contribute to its limited adoption, especially as the emergence of door-to-balloon time as a publicly reported quality metric has led to operator and institutional concerns over delays in reperfusion with TRA (13). Preliminary work has supported the possibility of operator and institutional barriers to the use of transradial techniques (14). Despite this, there is a paucity of data exploring potential operator- and institutional-level characteristics associated with the limited adoption of TRA in STEMI. Furthermore, there has not been a complete evaluation of the association between operator and facility use of contemporary transradial techniques and clinical outcomes in a “real-world” setting.
Accordingly, we sought to evaluate the contemporary prevalence of TRA for STEMI in the United States, the operator and institutional characteristics associated with use of TRA, and the variability in its use across operators and institutions. Finally, we assessed the relationship between the use of TRA at the operator- and facility-levels with periprocedural bleeding and inpatient mortality among patients undergoing PCI for STEMI.
The study sample was derived from the National Cardiovascular Data Registry (NCDR) CathPCI Registry, which has been described previously (15,16). Briefly, this registry collects data on patients undergoing PCI at over 1,400 institutions within the United States. The data elements included in this registry were prospectively defined by a committee of the American College of Cardiology and are available on the NCDR website. Each procedure is linked to an operator using a National Provider Index (NPI) number as well as to the institution where the procedure is performed. The entirety of the data is entered at each participating site using a standardized interface and then exported to a central repository. The data subsequently undergo a comprehensive and rigorous auditing program to ensure optimal data integrity (17).
The present analysis focused on all patients undergoing PCI for STEMI between July 1, 2009, and December 31, 2015. As the analysis stratified the population by primary site of arterial access, patients that did not have access site recorded in the registry were necessarily excluded. Additionally, patients requiring mechanical support devices like intra-aortic balloon pumps were excluded, as access site choice would likely be significantly influenced by the requirement of such devices. Subsequent analyses also focused on operator characteristics, and so patients treated by operators with invalid NPI numbers were also excluded from the cohort. To provide an accurate comparison of radial and femoral access approaches among operators and institutions with reasonable procedural experience, patients treated by operators (<5 cases annually) and institutions (<25 cases) with low clinical volumes were excluded. Finally, although data from 2009 were used to demonstrate temporal trends in TRA use, subsequent analyses were performed using a contemporary cohort of patients undergoing PCI for STEMI from 2012 to 2015. This population was selected to more accurately represent current practices, and contemporaneous operator and institutional characteristics associated with TRA use.
All patient characteristics and attributes were derived from the data fields within the NCDR CathPCI Registry. Operator characteristics included the timing of initial registration of their NPI number within the registry, stratified as before or after 2012. Institutional characteristics included teaching status, annual interventional procedural volume, geographic region, and hospital referral region (HRR) as defined by the Dartmouth Atlas project (18). HRRs are defined based on referral patterns for tertiary care using Medicare claims data, and have been used to approximate local practice culture and evaluate geographic patterns and variation in care (19–21). Clinical outcomes including in-hospital mortality and bleeding events were derived from the dataset. The definition of inpatient post-PCI bleeding within the NCDR CathPCI registry has been previously described (22). Briefly, bleeding was defined as any bleeding occurring within 72 h after PCI or before hospital discharge. Bleeding events included hematoma formation, retroperitoneal, gastrointestinal, genitourinary bleeding, intracranial hemorrhage, cardiac tamponade, hemoglobin decrease of >3 g/dl, or transfusion not related to bypass surgery.
The initial analysis sought to measure the temporal trends and geographic variation in TRA for STEMI. We first determined the overall proportion of patients treated with transradial techniques stratified by quarter-year, as well as by HRR, defining the prevalence of TRA across operators and institutions. To estimate percentiles and variance of TRA use at the operator and institutional levels, we created 2 unadjusted hierarchical logistic regression models using random intercepts for institution and operator. A sensitivity analysis was performed on a contemporary cohort treated between 2012 and 2015 to assess for more recent changes in treatment patterns.
We then identified patient, operator, and institutional characteristics associated with TRA in STEMI within a contemporary subgroup of patients (2012 to 2015) to more accurately represent current practice patterns. Using this subpopulation, we determined the characteristics associated with access site choice for STEMI. Differences between categorical variables were assessed using chi-square or Fisher exact testing, whereas the median of continuous variables was compared using the Kruskal-Wallis test. We then assessed the factors associated with the transradial approach accounting for operator and institutional clustering. To do so, we created 2 hierarchical regression models with random intercepts for operator and institution, respectively. The intraclass correlation coefficient (ICC) was then calculated to determine the proportion of total variance in TRA attributable to a given operator or institution. We also calculated a median odds ratio (MOR) for further ease of interpretation of the variability associated with the between-operator and -institution rates of TRA. The MOR is a method of presenting the difference in odds of 2 randomly selected identical patients with STEMI receiving PCI via TRA if treated by different operators or at different institutions. A MOR of 1 demonstrates no variability (i.e., equivalent odds of undergoing PCI via TRA across operators or institutions), with increasing values supporting increasing levels of variability.
Finally, we assessed the association between operator and institutional rates of TRA with risk-adjusted post-PCI bleeding and risk-adjusted in-hospital mortality. Risk-adjusted bleeding rates were defined as the ratio of observed to expected events (bleeding) multiplied by the observed bleeding rate, where the observed bleeding rate is defined as the total number of in-hospital bleeds divided by the total number of eligible PCI admissions. Expected bleeding rates for individual operators and institutions were calculated using separate models, but with identical adjustment variables. Risk-adjusted mortality rates were calculated using analogous methods. We then calculated the Spearman correlation coefficient for both percent TRA use and risk-adjusted bleeding rate, and percent TRA use and risk-adjusted mortality rate to determine the relationships between TRA for STEMI and in-hospital bleeding and mortality for operators and institutions within the CathPCI Registry. All statistical analyses were performed with SAS version 9.4 (SAS Institute, Cary, North Carolina). A p value <0.05 was considered statistically significant.
We identified 718,770 patients who underwent percutaneous revascularization for STEMI between July 2009 and December 2015. After excluding patients without an access site recorded (n = 2,027) or with mechanical support placed at the start of the procedure (n = 5,996), we identified 710,746 patients. Additional patients were excluded if undergoing PCI at low-volume institutions (n = 12,379) or if an operator NPI number was invalid (n = 5,935), resulting in a final cohort of 692,433 STEMI interventions performed by 9,257 operators across 1,381 institutions (Figure 1).
Temporal trends and prevalence of TRA for STEMI PCI
Among the study cohort, 82,618 (12%) of the percutaneous coronary interventions were performed via a transradial approach. The proportion of patients undergoing transradial intervention increased with time, from 2% in 2009 to over 23% in 2015 (p value for trend <0.0001). Average operator and institutional use of transradial intervention similarly rose over time as shown in Figure 2. The median observed proportion of TRA used was 0% (interquartile range [IQR]: 0% to 9%) across operators and 5% (IQR: 1% to 17%) across institutions. A sensitivity analysis excluding sites or operators with <10 procedures resulted in an increase in the median across operators to 1% (IQR: 0% to 11%) and a similar value of 5% (IQR: 1% to 17%) across institutions (Online Table 1). After accounting for clustering within operators and institutions, the estimated median rate of TRA was 3% (IQR: 1% to 33%) across operators and 5% (IQR: 2% to 15%) across institutions, with the distribution of use for both operators and institutions demonstrated in Figures 3A and 3B.
Geographic variation in use of TRA for STEMI PCI
Geographic patterns in the use of a transradial approach for STEMI are graphically depicted in the Figure 4. As shown, referral regions in the Northeast and Midwest appear to have the highest rates of TRA for STEMI, but without a demonstrable geographic pattern for use, and significant variability. HRRs with the highest rates of TRA use lie within similar geographic areas as both high- and low-rate HRRs.
Patient, operator, and institutional characteristics associated with use of TRA for STEMI PCI
When evaluating a more contemporary subpopulation of patients undergoing percutaneous revascularization for STEMI between 2012 and 2015, we identified 453,769 interventions performed by 7,745 operators across 1,376 institutions. Similar to the overall cohort, a minority of contemporary procedures (n = 73,064, 16%) were performed using TRA. Patients undergoing intervention via transradial approach were younger and more likely to be men when compared with those undergoing the procedure via femoral access, as shown in Table 1. The medical comorbidities between the 2 groups were also disparate, with lower rates of prior myocardial infarction, heart failure, peripheral vascular disease, cerebrovascular disease, and coronary artery bypass grafting among those treated with TRA. Procedural characteristics associated with TRA include reduced door-to-device time, decreased total contrast use, increased rates of procedural success (successful lesion dilation), and reduced rates of nonsystems delays for door-to-balloon time. These factors are demonstrated in Table 2. The operator and institutional characteristics associated with the transradial approach are summarized in Table 3. As shown, TRA use was more frequently employed at academic institutions with higher annual interventional volume and more inpatient beds. Patient, procedural, operator, and institutional factors associated with increasing tertiles of rates of use at the operator and institutional level are demonstrated in Online Tables 2 and 3.
After adjustment and accounting for operator and institutional clustering, characteristics independently associated with the use of TRA for STEMI are depicted in Figures 5A and 5B. As shown, the presence of concomitant cardiogenic shock or cardiac arrest was associated with decreased use of the transradial approach. Operators entering practice prior to 2012 were also less likely to perform the procedure via TRA. Finally, academic institutions and facilities with increasing size were more likely to perform TRA for STEMI.
Operator and institutional variability
Measures unadjusted for patient and institutional factors demonstrated significant intraclass correlation across both operators and institutions for the use of TRA, with an ICC of 0.60 across operators and 0.47 across institutions. The MOR was 8.28 across operators and 5.15 across institutions. After adjustment, these measures remained significant. ICC for TRA use across operators was 0.61, with a MOR of 8.69. ICC across institutions was 0.47, with a MOR of 5.02. In other words, 2 identical patients would have a nearly 9-fold difference in their odds of undergoing TRA for STEMI PCI just by changing operators. Similarly, 2 identical patients would have over a 5-fold difference in the odds of undergoing TRA for STEMI PCI by presenting at a different institution. These relationships are depicted in the Central Illustration.
The risk-adjusted median institutional rate of in-hospital bleeding was 8.29% (IQR: 5.99% to 10.47%). The risk-standardized median operator rate of in-hospital bleeding was 8.14% (IQR: 5.01% to 11.86%). There was a correlation between operator rates of TRA use and risk-standardized operator rates of in-hospital bleeding, with increasing use of TRA for STEMI PCI associated with decreased in-hospital bleeding (rho = −0.053; p < 0.0001) (Online Figure 1A). The association between institutional rates of TRA use and risk-standardized rates of in-hospital bleeding did not reach statistical significance (rho = −0.042; p = 0.12) (Online Figure 1B).
The risk-adjusted median institutional rate of in-hospital mortality was 5.98% (IQR: 4.72% to 7.17%). The risk-standardized median operator rate of in-hospital mortality was 5.58% (IQR: 3.11% to 8.12%). There was no correlation between operator rates of TRA use and risk-standardized operator rates of in-hospital mortality (rho = −0.012; p = 0.30) (Online Figure 1C). Institutional rates of TRA use and risk-standardized rates of in-hospital mortality did demonstrate a significant association, with increasing use of TRA for STEMI PCI associated with decreased in-hospital mortality (rho = −0.077; p = 0.004) (Online Figure 1D).
The present study sought to evaluate the operator and institutional factors associated with use of TRA for STEMI PCI. As demonstrated, the use of the transradial approach has increased over time. Although a number of patient characteristics are associated with limited adoption of transradial techniques, there are also operator and institutional characteristics that independently influence access site. Furthermore, the operator and institutional variation account for up to a nearly 9-fold difference in the use of TRA for STEMI, independent of patient characteristics. Across this substantial variability, increased rates of TRA use were associated with decreased inpatient bleeding at the operator level and with decreased mortality at the institution level. These results have important implications for ensuring the optimal treatment of patients presenting with STEMI.
Recent research has reinforced the potential benefits of a transradial approach to percutaneous revascularization, with demonstration of lower bleeding and vascular complications (1,2,4,5). Further work has suggested lower rates of major adverse cardiac events across randomized trials (2,4,5) and meta-analyses (3,23), with a reduction in mortality demonstrated among patients with STEMI. The present analysis suggests that this growing body of published work may have increased adoption of the transradial approach, with a 10-fold increase in the proportion of STEMI patients that are treated in this manner over the last 6 years. These data are consistent with trends in the use of TRA for PCI overall, with a doubling in the utilization of radial techniques from 2011 to 2014 (24). However, even this increased rate of use still remains far below that in other countries, which report TRA in over 50% of their cases (10,11,25). Further analysis of our data suggests that there is significant geographic variation in the adoption of the transradial approach. Tightly knit HRRs in the Northeast and Midwest have higher rates of adoption, whereas regions in the South and West are significantly more likely to use the transfemoral approach. This geographic variation suggests that operator and institutional characteristics play a role in the adoption of TRA for STEMI.
The present study builds on prior research defining factors associated with the use of TRA in STEMI PCI. Prior research has identified patient characteristics associated with the use of TRA, demonstrating that patients with lower medical acuity and fewer comorbidities were more likely to receive the procedure (7,26). Our findings were supportive of this “risk-treatment paradox,” again finding that increasing medical comorbidities and heightened acuity of presentation were associated with a lower likelihood of the use of TRA. The present study adds to this information by identifying novel operator and institutional characteristics that are also associated with the limited use of transradial techniques. For example, operator experience is significantly associated with access site choice, wherein operators entering practice after 2012 were far more likely to pursue radial access. This may reflect a shift in training for interventional cardiology fellows toward an increased familiarity with this approach. Further, patients treated at institutions affiliated with medical training programs were more likely to undergo a procedure with TRA, again suggesting the effects of an evolving emphasis on TRA in interventional cardiology training. Alternatively, these differences may reflect the effect of institutional goals or explicit foci of physician groups to develop specific practice models (27). The geographic and HRR variability noted in our analysis may also support the hypothesis of institutional- or provider group–driven efforts to locally transition to transradial approaches. Interestingly, and in contrast with prior studies, longer door-to-balloon times were present among patients undergoing femoral access, although these findings are highly likely to be affected by patient selection. For example, the increased rate of non–systems-related delays for door-to-balloon time in transfemoral access patients suggests operator or institutional bias toward a transfemoral approach in patients where door-to-balloon time may be prolonged. In aggregate, these operator and institutional factors result in significant heterogeneity in access site choice for patients. Two identical patients would have a nearly 9-fold difference in the likelihood of undergoing percutaneous revascularization when treated by different operators and a nearly 6-fold difference at different institutions independent of patient characteristics. These findings suggest an opportunity to further explore and reduce the variation in the application of the transradial approach, with the hope of improving patient outcomes.
The most compelling reasons for the adoption of a transradial approach in STEMI have been the reduced bleeding and mortality demonstrated in several randomized trials, as compared with a transfemoral approach (1–5). Observational studies have mirrored these findings in real-world settings, although residual confounding may influence these results when performed at the patient level (7,10,11). The present analysis is distinct from other published data by evaluating the association between the transradial approach and outcomes of bleeding or mortality at the operator and institutional level, thereby minimizing effects of patient selection and case mix. This study demonstrated associations similar to that found in the prior published data for bleeding, with a modest but statistically significant association between operator TRA use and decreased bleeding events, and an association between institutional TRA use and decreased bleeding that did not reach significance. It is possible that the outcomes for a transradial approach in STEMI PCI at the operator and institutional level may be influenced by the “learning curve” in nascent transradial STEMI programs (28), with differences in operator experience confounding the relationships demonstrated. This study’s findings for inpatient mortality were conflicting, with increased TRA use not associated with inpatient mortality at the operator level, but associated with a small reduction in mortality at the institutional level. At the operator level, it is possible that a reduction in bleeding may be more evident earlier in transradial adoption than effects on mortality, affecting the association. At the institutional level, increased use of TRA for STEMI PCI may signify a broader effort to promote quality cardiovascular care, with an observed reduction in inpatient mortality. Ultimately, these findings demonstrate the challenge in study of outcomes of an emerging but still nascent transradial approach to STEMI in the United States, where the prevalence of TRA for STEMI PCI likely remains too low and variance too high across both operators and institutions to reliably draw conclusions on the impact of its use. Ongoing analyses in other datasets could be designed to address this question more fully after further maturation of its implementation and more uniformity in operator experience.
First, this analysis was performed using observational data, raising the possibility of residual confounding. Although robust statistical methods for adjustment were employed, the conclusions of this analysis should be viewed as hypothesis generating and associative. Second, although our analysis is able to demonstrate significant heterogeneity in the use of TRA for STEMI PCI in the study population, evaluation of the reasons for this heterogeneity is limited, as data capture for operator and institutional characteristics is less robust than that for patient-level factors. Third, the facility- and operator-level relationships between TRA and inpatient mortality may not reflect individual patient-level relationships between TRA and outcome. However, any such relationships are likely to be heavily influenced by patient selection, as has been previously demonstrated (7,26). By performing our analysis at the operator and facility levels and performing risk standardization, the relationship driven by practice patterns and overall care should be represented, and the unmeasured effects of patient selection should be minimized. Fourth, our analysis cannot account for the “learning curve” for operators and institutions in the implementation of a transradial program for STEMI PCI (28,29). It is possible that our findings may still be influenced by the initial experience of operator adoption of TRA in emergent PCI. As TRA use continues to grow in the United States, it is possible that the operator-level relationships demonstrated in our analyses may change and become congruent with that seen at the facility level. Last, data from the NCDR may not be reflective of all patients undergoing PCI for STEMI in the United States; however, the NCDR captures data from a large number of participating institutions, representing a majority of patients undergoing PCI in the United States.
The transradial approach for percutaneous revascularization is increasing in the United States, albeit with significant geographic variation and ongoing opportunities for adoption. Operator and institutional characteristics are associated with a significant proportion of the variability of transradial adoption, independent of patient factors. This heterogeneity may serve as a potential target for overall quality improvement in the management of STEMI.
WHAT IS KNOWN? TRA for STEMI PCI remains underutilized in the United States. Factors influencing the use of TRA remain understudied, especially at the operator and institutional levels.
WHAT IS NEW? The present study identified patient, operator, and facility characteristics associated with the use of TRA in STEMI PCI. Furthermore, the analysis demonstrated that operator and institutional variation play a significant role in the use of TRA. Within this cohort, identical patients experienced 8-fold differences in the odds of receiving TRA for STEMI PCI by changing operators or institutions, respectively.
WHAT IS NEXT? Further research is needed to understand the rationale for the variation across operators and institutions in the use of TRA for STEMI PCI. In doing so, there may be an opportunity to standardize the care that our patients receive to improve overall clinical outcomes.
The views expressed in this paper are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the U.S. government. Dr. Rao has served as a consultant for Medtronic. Dr. Gurm has served as a consultant for Osprey Medical; and has received research funding from the National Institutes of Health. Dr. Armstrong has served as a consultant or on the advisory board for Abbott Vascular, Boston Scientific, Cardiovascular Systems, and Spectranetics. Dr. Waldo has received research grants from Abiomed, Cardiovascular Systems, and Merck Pharmaceuticals to the Denver Research Institute. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- hospital referral region
- intraclass correlation coefficient
- interquartile range
- median odds ratio
- National Cardiovascular Data Registry
- National Provider Index
- percutaneous coronary intervention
- ST-segment elevation myocardial infarction
- transradial access
- Received March 24, 2017.
- Revision received May 26, 2017.
- Accepted July 2, 2017.
- Bernat I.,
- Horak D.,
- Stasek J.,
- et al.
- Karrowni W.,
- Vyas A.,
- Giacomino B.,
- et al.
- Romagnoli E.,
- Biondi-Zoccai G.,
- Sciahbasi A.,
- et al.
- Mehta S.R.,
- Jolly S.S.,
- Cairns J.,
- et al.
- Baklanov D.V.,
- Kaltenbach L.A.,
- Marso S.P.,
- et al.
- Bradley S.M.,
- Rao S.V.,
- Curtis J.P.,
- et al.
- Rao S.V.,
- Ou F.S.,
- Wang T.Y.,
- et al.
- Johnman C.,
- Pell J.P.,
- Mackay D.F.,
- et al.
- Ratib K.,
- Mamas M.A.,
- Anderson S.G.,
- et al.
- Feldman D.N.,
- Swaminathan R.V.,
- Kaltenbach L.A.,
- et al.
- Brindis R.G.,
- Fitzgerald S.,
- Anderson H.V.,
- Shaw R.E.,
- Weintraub W.S.,
- Williams J.F.
- Moussa I.,
- Hermann A.,
- Messenger J.C.,
- et al.
- Messenger J.C.,
- Ho K.K.,
- Young C.H.,
- et al.
- ↵(2017) The Dartmouth Atlas of Health Care (Trustees of Dartmouth College, Lebanon, NH).
- Rao S.V.,
- McCoy L.A.,
- Spertus J.A.,
- et al.
- Masoudi F.A.,
- Ponirakis A.,
- de Lemos J.A.,
- et al.
- Hess C.N.,
- Peterson E.D.,
- Neely M.L.,
- et al.