Author + information
- Received September 23, 2016
- Revision received November 30, 2016
- Accepted December 8, 2016
- Published online March 6, 2017.
- Edward L. Hannan, PhDa,∗ (, )
- Zaza Samadashvili, MDa,
- Kimberly Cozzens, MAa,
- Foster Gesten, MDb,
- Alda Osinaga, MDc,
- Douglas G. Fish, MDc,
- Constance L. Donahue, AuDc,
- Ronald J. Bass, BAc,
- Gary Walford, MDd,
- Alice K. Jacobs, MDe,
- Ferdinand J. Venditti, MDf,
- Nicholas J. Stamato, MDg,
- Peter B. Berger, MDh,
- Samin Sharma, MDi and
- Spencer B. King III, MDj
- aDepartment of Health Policy, Management and Behavior, University at Albany, State University of New York, Albany, New York
- bOffice of Quality and Patient Safety, New York State Department of Health, Albany, New York
- cOffice of Health Insurance Programs, New York State Department of Health, Albany, New York
- dHeart and Vascular Institute, Johns Hopkins University, Baltimore, Maryland
- eDepartment of Cardiology, Boston Medical Center, Boston, Massachusetts
- fDepartment of Cardiology, Albany Medical Center, Albany, New York
- gDepartment of Cardiology, Campbell County Memorial Hospital, Gillette, Wyoming
- hDepartment of Cardiology, Northwell Health, Great Neck, New York
- iClinical Cardiology, Mt. Sinai Medical Center, New York, New York
- jAcademic Affairs, St. Joseph’s Health System, Atlanta, Georgia
- ↵∗Address for correspondence:
Dr. Edward L. Hannan, School of Public Health, State University of New York, University at Albany, One University Place, Rensselaer, New York 12144-3456.
Background Recent studies have demonstrated relatively high rates of percutaneous coronary interventions (PCIs) classified as “inappropriate.” The New York State Department of Health shared rates with hospitals and announced the intention of withholding reimbursement pending demonstration of clinical rationale for Medicaid patients with inappropriate PCIs.
Objectives The objective was to examine changes over time in the number and rate of inappropriate PCIs.
Methods Appropriate use criteria were applied to PCIs performed in New York in patients without acute coronary syndromes or previous coronary artery bypass graft surgery in periods before (2010 through 2011) and after (2012 through 2014) efforts were made to decrease inappropriateness rates. Changes in the number of appropriate PCIs were also assessed.
Results The percentage of inappropriate PCIs for all patients dropped from 18.2% in 2010 to 10.6% in 2014 (from 15.3% to 6.8% for Medicaid patients, and from 18.6% to 11.2% for other patients). The total number of PCIs in patients with no acute coronary syndrome/no prior coronary artery bypass graft surgery that were rated as inappropriate decreased from 2,956 patients in 2010 to 911 patients in 2014, a reduction of 69%. For Medicaid patients, the decrease was from 340 patients to 84 patients, a decrease of 75%. For a select set of higher-risk scenarios, there were higher numbers of appropriate PCIs per year in the period from 2012 to 2014.
Conclusions The inappropriateness rate for PCIs and the use of PCI for elective procedures in New York has decreased substantially between 2010 and 2014. This decrease has occurred for a large proportion of PCI hospitals.
- Percutaneous Coronary Interventions Reporting System
- stable coronary artery disease
In 2009, the American College of Cardiology (ACC) and the American Heart Association (AHA), in conjunction with the Society for Cardiovascular Angiography and Interventions, the Society of Thoracic Surgeons, the American Association for Thoracic Surgery, and the American Society of Nuclear Cardiology, released appropriate use criteria (AUC) for coronary revascularization to serve as a supplement to the earlier ACC/AHA guideline documents (1). The criteria were updated in 2012 (2). Although nearly all patients with acute coronary syndromes (ACS) were judged to be appropriate for revascularization, many types of elective patients were rated as “inappropriate” or “uncertain.” This terminology was changed in 2013 to “appropriate care,” “may be appropriate care,” and “rarely appropriate care.” However, the older terminology is used here because that was in place when the 2012 AUC were issued.
Three recent studies applied these criteria to examine the extent of inappropriateness of revascularization procedures from 2009 to 2010 (3–5). The studies found percutaneous coronary intervention (PCI) inappropriateness rates for patients without acute conditions that ranged from 12% to 17% (3–5). The studies also found that many PCI patients could not be rated for appropriateness, with the unrated percentage among patients without acute conditions varying from 28% to 50% (3–5). Two other studies tracked changes in appropriateness, 1 from Washington state and another from the National Cardiovascular Data Registry (NCDR), in conjunction with efforts to inform hospitals about their appropriateness data (6,7).
In an effort to reduce the rate of inappropriate PCI, the New York State Department of Health (the Department) presented information to hospitals about their own inappropriateness rates, published data on the topic in a peer-reviewed journal (8), and recommended denial of reimbursement for inappropriate PCIs received by Medicaid patients. The purpose of this study was to examine whether any such denials have occurred in New York since shortly after the release of the AUC and: 1) the rate of inappropriateness of PCI procedures for patients without acute conditions; 2) the percentage of cases that could not be rated because of missing information; and 3) the total number of patients undergoing PCI and elective PCI.
The database used in the study is the state of New York’s Percutaneous Coronary Interventions Reporting System, which was developed in 1991 to collect information on all New York patients undergoing PCI in nonfederal hospitals in the state. This PCI reporting system contains information on demographics, comorbidities, left ventricular function, hemodynamic state, vessels diseased, hospital and operator identifiers, and in-hospital adverse outcomes. As of July 2009, it has also contained information on the extent of anti-ischemic medical therapy used by patients and noninvasive test findings, both of which are needed in the scenarios used in the AUC. The stress test information includes standard exercise stress tests, stress echocardiogram, stress testing with single photo emission computed tomography and testing with cardiac magnetic resonance. Also, low-, intermediate-, and high-risk findings are defined and recorded. Data are audited and cross-checked against the Department’s acute care hospital discharge database (the Statewide Planning and Research Cooperative System), to ensure accuracy and completeness. There is also an intensive auditing process that involves medical record reviews in samples of hospitals each year.
The revascularization AUC were subdivided into 3 broad categories: 1) ACS; 2) no ACS/no prior coronary artery bypass graft (CABG) surgery (henceforth referred to as stable PCI patients); and 3) no ACS/prior CABG surgery. Patients were then separated into scenarios that were intersections of clinical characteristics required to determine the appropriate intervention. This study concentrated on the stable PCI patients because they are the ones most at risk for procedures considered inappropriate and because they were the subject of an earlier study in New York documenting procedures rated as inappropriate (3).
Scenarios for stable PCI patients were defined by the number of diseased vessels and presence of proximal left anterior descending coronary artery (PLAD) disease, findings on noninvasive testing, amount of anti-ischemic medical therapy received, and Canadian Cardiovascular Society (CCS) angina class. In keeping with the RAND methodology, patients were rated as appropriate, uncertain, or inappropriate for revascularization. Cases with insufficient information to be rated were also noted. We used the updated 2012 AUC criteria for our study so that an equitable comparison could be made across years (2).
All 58 nonfederal New York State hospitals in which PCI was performed during the entire study period were included in the study. Our analyses included all stable PCI patients who underwent PCI in these hospitals between January 1, 2010, and December 31, 2014.
For each year between 2010 and 2014, the number of patients undergoing PCI was identified by primary payer. “Medicaid Primary” included individuals without Medicare or private coverage and was determined based on matching data from New York State’s Medicaid Eligibility System, while the “other” category encompassed all other patients, including those with Medicaid Plus Medicare insurance.
Also, the number of stable PCI patients were identified by year and primary payer. The number and percentage of patients who could not be rated because of missing information and the number and percentage of stable PCI patients rated as “inappropriate” for revascularization were then obtained by year and payer status. Chi-square tests for trend were used to examine the significance of trends in the volumes of inappropriate PCI, the patients who could not be rated, and the number of elective patients undergoing PCI. An interrupted time series analysis with adjustment for autoregression and generalized estimating equations to account for clustering of patients within hospitals were used to examine the nature of reductions in inappropriateness rates over time.
These analyses were repeated after eliminating patients whose inappropriateness rating could have changed between 2010 through 2011 and 2012 through 2014 because of a change in the definition of CCS class. Specifically, in 2012, New York changed the definition from “typical CCS class” to “worst in last 2 weeks” to align with NCDR definitions; therefore, patients seen between 2012 and 2014 who had a typical CCS class of I or II could have been assigned a CCS class of III or IV if they had only occasional class III or IV angina. Unfortunately, neither the 2010-to-2011 nor the 2012-to-2014 databases had both “typical” and “worst” classes coded; thus, there was no way to make the data more consistent across time periods. Consequently, patients in scenario 14 (1- or 2-vessel CAD without PLAD involvement, with low-risk stress test findings and less than maximal medical therapy) were eliminated after reviewing AUC scenarios segregated by CCS class. This was the only scenario for which a shift from CCS class I or II to class III or IV could have resulted in a shift from inappropriate to not inappropriate (uncertain). Additionally, since ACS unstable angina (UA) patients (patients with a code of UA and CCS class III or IV) are in the AUC’s Table 1 but not in the document’s Table 2 group (no ACS/no prior CABG surgery; this study’s stable PCI patients), but patients with a code of UA who have CCS class I or II angina are in AUC Table 2; an upgraded CCS class could affect appropriateness ratings by moving a patient from AUC Table 2 to AUC Table 1. Thus, patients with a code of UA were eliminated in our repeat analyses, regardless of their CCS class.
Other analyses were conducted to determine the range across hospitals in inappropriateness ratings. Tertiles of hospital inappropriateness rates were computed for the period before (2010 to 2011) and after (2012 to 2014) sharing data with hospitals. The period before informing hospitals that appropriateness could be assessed and providing hospitals with their own results (2010 to 2011) was regarded the baseline. Also, tertiles of hospital-level changes in inappropriateness rates were computed to determine the widespread nature of any reduction in the rate of inappropriateness.
We also examined the possible unintended consequence of reducing appropriate PCIs while attempting to reduce inappropriate PCIs. Because of changes in the definition of CCS classes during the study period, and because of the marked increase in the number and percent of PCIs that could be rated in later years, we were limited to examining a subset of all scenarios that were rated as appropriate. The change in definition of CCS classes (from typical level to the worst experienced in the past 2 weeks) necessitated limiting the scenarios examined to those for which patients with classes I or II, as well as classes III or IV, were rated as appropriate (2). Improvements in reporting the severity of positive stress test findings during the study period could also have biased the results in favor of later years. Because it was possible to rate more patients, a higher number of patients could be rated as “appropriate” for revascularization. We minimized this potential bias by further limiting the scenarios we examined to ones that did not require severity of stress test findings. For example, there were 3 scenarios for patients with 3-vessel/PLAD disease with maximal medical therapy depending on whether the stress test results were low, medium, or high risk. However, because all 3 of these scenarios were appropriate for CCS classes I and II, and classes III and IV, we were able to include patients in earlier years as appropriate despite missing severity. Consequently, we examined 22 scenarios consisting of symptomatic patients (CCS classes I through IV) with either stable or unstable angina who had: 1) 1- or 2-vessel PLAD disease or 3-vessel disease with maximal medical therapy; 2) 3-vessel disease with abnormal left ventricular systolic function; or 3) left main disease (AUC scenarios 31, 33, 35, 37, 39, 41, 43, 45, 47, 48, and 49 for CCS classes I and II, and for CCS classes III and IV).
All tests for statistical significance were 2-sided, with a p < 0.05 significance level. All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, North Carolina).
Table 1 presents the total number of PCI patients in New York by year, broken out by payer. The total decreased by 6,982 (12.9%) between 2010 and 2014, although there was a slight increase between 2012 and 2014. Overall, Medicaid volume decreased from 7,142 to 7,070 (1.0%) across the entire study period, although there were fluctuations, such as a 6% increase from 2012 to 2014.
The total number and percentage of PCI patients defined as stable decreased considerably during the study period (Table 2). The number of stable PCI patients decreased by 54.0% from 20,142 in 2010 to 9,261 in 2014, and the percentage of all PCIs that were performed in stable patients decreased from 37% to 20%. For Medicaid Primary patients, there was a 51.5% decrease from 2,729 in 2010 to 1,324 in 2014, and the percentage of all PCIs done in stable patients decreased from 38% to 19%. For other patients (not Medicaid Primary), there was a 54.4% decrease from 17,413 in 2010 to 7,937 in 2014, and the percentage of all PCI patients who were stable decreased from 37% to 20% (chi-square test for trend p < 0.001 for all groups).
The number and percentage of stable PCI patients (i.e., no ACS, no prior CABG surgery) who were rated as inappropriate for revascularization according to the AUC are reported by payer and year in Table 3. As indicated, the percentage of patients rated as inappropriate decreased from 18.2% in 2010 to 10.6% in 2014 for all PCI patients, from 15.3% to 6.8% for Medicaid Primary patients, and from 18.6% to 11.2% for other patients (chi-square test for trend p < 0.001 for all groups). When the relatively small group of no ACS/prior CABG patients was added to our analyses, the results were similar: reduction from 17.5% inappropriate in 2010 to 10.2% in 2014 for all patients, and from 14.7% to 6.6% for Medicaid Primary patients. There was no significant difference in the reduction of the inappropriateness rate between Medicaid Primary stable PCI patients and other stable PCI patients (p = 0.46). Figure 1 shows the inappropriateness rate for Medicaid and other patients for each 6-month period between 2010 and 2014. An interrupted time series analysis demonstrated that there was no significant trend in the inappropriateness rate for the entire study period (p = 0.67) or during 2012 through 2014 (p = 0.99), but there was a significantly lower rate in 2012 to 2014 than in 2010 to 2011 (odds ratio: 0.60; 95% confidence interval: 0.44 to 0.83). Similarly, for Medicaid patients, there was no significant trend for the entire study period (p = 0.93) or during 2012 to 2014 (p = 0.58), but the rate was significantly lower in 2012 to 2014 than in 2010 to 2011 (odds ratio: 0.48; 95% confidence interval: 0.31 to 0.73).
The total number of stable PCI patients rated as inappropriate fell from 2,956 in 2010 to 911 in 2014, a decrease of 69.2% (Table 3). For Medicaid Primary patients, the decrease was from 340 to 84 patients, a decrease of 75.3%, and for other patients, the decrease was from 2,616 patients to 827, a 68.4% drop. These decreases in inappropriateness volumes were much larger than the decreases in inappropriateness rates because of the large decline in patients with nonacute conditions going to PCI.
After eliminating 12,440 (21%) patients whose inappropriateness ratings could have changed because of definitional changes in CCS class in 2012, the percentage of patients rated as inappropriate decreased from 17.1% in 2010 to 9.0% in 2014 for all PCI patients, from 14.2% to 5.9% for Medicaid Primary patients, and from 17.6% to 9.6% for other patients (chi-square test for trend p < 0.001 for all groups). The total number of PCIs rated as inappropriate diminished from 1,984 in 2010 to 676 in 2014, a decrease of 65.9%, and the number of Medicaid Primary patients rated as inappropriate dropped from 229 to 60, a decrease of 73.8%.
For the 50 hospitals with an annual volume of at least 20 patients without ACS or prior CABG in the 2010-to-2011 and 2012-to-2014 time periods, the percentage of stable PCI patients rated as inappropriate ranged from 1.4% to 55.5% in 2010 to 2011 and from 0.0% to 29.9% in 2012 to 2014. In 2010 to 2011, 23 of 50 hospitals had inappropriateness rates higher than 20% for this subgroup, whereas only 8 hospitals had inappropriateness rates higher than 20% in the 2012-to-2014 time period (Figure 2).
With regard to individual hospital changes in inappropriateness rates, more than two-thirds of the hospitals decreased their inappropriateness rates by at least 2.9% and one-half of the hospitals decreased their rates by at least 6.3% between 2010 to 2011 and 2012 to 2014 (Figure 3). However, the interquartile range in the inappropriateness rate remained large (10% to 25% in 2010 and 4% to 16% in 2014).
For stable PCI patients who could not be rated by the AUC because of missing information, the percentage of cases that could not be rated decreased from 19.3% in 2010 to 7.0% in 2014 for all patients, from 18.8% in 2010 to 6.6% in 2014 for Medicaid Primary patients, and from 19.4% to 7.1% for other patients (chi-square test for trend p < 0.001 for all groups) (Table 4). The total number of patients who could not be rated decreased from 3,895 in 2010 to 652 in 2014, a decrease of 83.3%. The corresponding decrease in unrated cases for Medicaid Primary patients was from 513 to 88, a decrease of 82.8%, and for other patients, the numbers dropped from 3,382 to 564, a decrease of 83.3%. All but 4 hospitals reduced their rate of stable PCI patients who could not be rated (Figure 3).
Table 5 presents the numbers of appropriate PCIs in AUC scenarios containing symptomatic patients (CCS classes I through IV) with stable or unstable angina who had either 1- or 2-vessel PLAD disease or 3-vessel disease with maximal medical therapy, 3-vessel disease with abnormal left ventricular systolic function, or left main disease. The total number of PCIs for these groups dropped somewhat between 2010 and 2011, but was higher in each of the 3 years following dissemination of inappropriateness rates to hospitals.
Between 2010 and 2014, the percentage of inappropriate PCIs in New York state for patients without ACS or prior CABG surgery dropped from 18.2% to 10.6% for all patients, from 15.3% to 6.8% for Medicaid Primary patients, and from 18.6% to 11.2% for other patients (Central Illustration). The decreases in inappropriateness rates for Medicaid Primary and other patients were not statistically different. Most of this decrease occurred after hospitals were provided their own inappropriateness rates and additional information in early 2012. With regard to the change in appropriate PCIs performed during the study period, for a select set of higher-risk scenarios, there were more appropriate PCIs in each of the years after the release of inappropriateness rates to hospitals than there were in each of the years before the release. Thus, at least for these scenarios, there was no evidence of a decrease in appropriate PCIs being performed subsequent to initiatives undertaken to reduce inappropriate PCIs.
We also found that although the range of hospital inappropriateness rates remained large, more than two-thirds of all hospitals decreased their inappropriateness rates by at least 2.9%, and one-half of the hospitals decreased their rates by more than 6.3%. Thus, although it appears that most hospitals have successfully reduced their inappropriateness rates, there is room for improvement among many hospitals.
There was also a substantial decrease in the number of stable CAD patients undergoing PCI. This number decreased by 54% between 2010 and 2014 for all patients and by 52% for Medicaid patients. Also, the percentage of all PCI patients who had stable CAD decreased from 37% to 20% during the 2010-to-2014 time period.
These reductions in rates of inappropriate PCI in stable CAD patients might have been influenced in part by initiatives undertaken by the New York Department of Health. In November 2011, the Department presented a seminar to hospitals describing the AUC and the aggregate New York state findings for 2010. In February 2012, the Department shared detailed information on appropriateness of 2010 PCIs with each hospital in which PCI was performed. Also, in November 2011, the New York State Medicaid Redesign Team recommended that the Department deny reimbursement for PCIs performed on individuals with no ACS/no prior CABG surgery (i.e., stable PCI patients) that were rated as inappropriate. This recommendation was approved by the Department; the policy was informally made public in late 2011 and formally announced in the June 2013 New York State Medicaid Update (8). Because of lags in obtaining and cleaning data, and improvements in PCI rates over time, the denial process was not implemented, and no denials have occurred. In May 2012, a study documenting cases rated as inappropriate was published in this journal (3).
However, it is very important to note that there was intense national scrutiny in the period following the 2009 release of the AUC, including national studies of the inappropriateness rates in the NCDR and feeding back data to hospitals in the NCDR. Thus, the changes reported here might largely reflect national changes. In comparison to the national findings using the NCDR, the total percentage decrease in inappropriateness rates in NCDR from 2010 to 2014 (50%) is higher than the percentage decrease in New York in that time period (42%). However, in the 2011-to-2014 period (with 2011 used because it was the last year before New York hospitals were given their inappropriateness rates), the decrease was 38% for NCDR and 39% for New York. With regard to changes in the volume of PCIs for patients with stable CAD, the decrease was 34% for NCDR between 2010 and 2014 versus 54% for New York. Thus, the decrease in the inappropriateness rate in New York mirrored the NCDR rate, but the decrease in the volume of PCIs for stable CAD was higher in New York than in the NCDR.
A caveat of the study is that the definitions of Canadian Cardiovascular Society class changed in 2012 in the New York registry. As noted previously, coding changes in CCS class could have biased results in favor of later years because of this definitional change. We attempted to eliminate this bias by removing the 32% of patients who could have been affected by this bias in sensitivity analyses. When this was done, the results were essentially the same, suggesting that the findings were robust despite this major limitation. However, it is possible that if these patient subgroups could have been included without any possible bias, the results would have been different.
Another caveat is that providers were incentivized to decrease their inappropriateness rates by upcoding patient severity of illness rather than improving patient selection. Although the department audits registry data for accuracy, data on CCS class in particular were not based on objective clinical measurements and were consequently difficult to ascertain.
The rate of inappropriate PCIs and the use of PCI for elective procedures decreased substantially between 2010 and 2014. These decreases were due, at least in part, to national publicity and national quality improvement efforts resulting from the publication of the AUC and other studies related to the appropriateness of revascularization. Future quality improvement efforts will continue to inform hospitals about their rates and about decreases in the overall statewide rate that may incentivize hospitals to improve further.
A consequence of this utilization decrease has been the reduction in valuable resources spent by health care payers and the U.S. health care system, as well as a reduction in occasional complications that would have occurred with additional PCI procedures. It remains to be seen whether payers will begin to deny payment on the basis of inappropriateness, but it appears to be unnecessary given that large national and New York decreases in “inappropriate” PCIs and in PCI use, in general, for stable CAD patients have occurred following national publications and confidential sharing of hospital and operator rates to providers.
The success in using AUC to inform evidence-based medicine by professional societies and policy makers can serve as a model for improving the cost effectiveness of other costly health care services. In general, this approach can serve as an aid to providers in an era when payment will depend on health care outcomes rather than on the services provided. It can also serve as a model for payers and governmental agencies in their goal of assuring that procedures and other services will be as cost effective as possible.
COMPETENCY IN PRACTICE-BASED LEARNING: Dissemination of inappropriateness rates to hospitals by the New York State Department of Health was followed by a substantial reduction in the proportion of inappropriate percutaneous coronary interventions, as classified according to the ACC/AHA appropriate use criteria.
TRANSLATIONAL OUTLOOK: Further surveillance should assess appropriate use in other areas of cardiovascular care and the durability of the impact of hospital-based reporting over time.
The authors thank New York State’s Cardiac Advisory Committee (CAC) members for their encouragement and support of this study; and Cynthia Johnson and the cardiac catheterization laboratories of the participating hospitals for their tireless efforts to ensure the timeliness, completeness, and accuracy of the data submitted.
Dr. Jacobs is the site principal investigator on research grants from Abbott Vascular, Boston Scientific Corporation, Bristol-Myers Squibb/Sanofi, Cordis Corporation, Eli Lilly/Daiichi-Sankyo, and Medtronic Vascular. Dr. Sharma is a member of the speakers bureaus of Abbott, Boston Scientific Corporation, Cardiovascular System Incorporated, Abiomed, Trireme, and The Medicines Company. Dr. King is a member of the data safety monitoring boards for Harvard Clinical Research Institute, Duke University, Capicor, Inc., Merck & Company, and Stentys. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Frederick A. Masoudi, MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- acute coronary syndrome
- appropriate use criteria
- coronary artery bypass graft
- Canadian Cardiovascular Class
- National Cardiovascular Data Registry
- percutaneous coronary intervention
- proximal left anterior descending coronary artery
- unstable angina
- Received September 23, 2016.
- Revision received November 30, 2016.
- Accepted December 8, 2016.
- American College of Cardiology Foundation
- Patel M.R.,
- Dehmer G.J.,
- Hirshfeld J.W.,
- Smith P.K.,
- Spertus J.A.
- Patel M.R.,
- Dehmer G.J.,
- Hirshfeld J.W.,
- Smith P.K.,
- Spertus J.A.
- Hannan E.L.,
- Cozzens K.,
- Samadashvili Z.,
- et al.
- Bradley S.M.,
- Maynard C.,
- Bryson C.L.
- Bradley S.M.,
- Bohn C.M.,
- Malenka D.J.,
- et al.
- New York State Department of Health