Coronary revascularization procedures decreased markedly in California after the introduction of drug-eluting stents and the initiation of public reporting in 2003, resulting in a large number of low-volume heart programs. California hospital discharge data were analyzed from 2006 to 2010 to study the impact of this change. In-hospital mortality and hospital readmission for major adverse events at 90 days and 365 days were determined for patients who underwent isolated coronary artery bypass grafting (CABG) or percutaneous coronary intervention (PCI) either with acute coronary syndrome (ACS) or PCI without acute coronary syndrome (PCI-noACS). Three terciles were chosen by case volume as follows: high-volume (747 ± 336 [SD]/yr total PCI, 210 ± 130 isolated CABG), intermediate volume (362 ± 47 PCI, 106 ± 27 CABG), and low-volume (211 ± 6 PCI, 53 ± 17 CABG) terciles were studied. PCI-noACS procedures decreased 33% and CABG 20%, whereas PCI-ACS procedures increased slightly. Risk-adjusted in-hospital mortality was slightly better in high-volume compared with low-volume terciles for CABG (2.0% vs 2.6%) and PCI-noACS (0.64% vs 0.85%). There was no difference in major adverse events at 90 days or 365 days among volume terciles within procedure groups, and no change in event rates was noted over the 5-year period. Wide variation in outcomes, associated with low volume, contributed to poor statistical discrimination among providers. In conclusion, lower volume hospitals had similar overall outcomes with wider variation. Conservative treatment strategies apparently contributed to decreased procedure volume. Collaboration among hospitals of similar structure and case volume may be the most appropriate performance improvement model to reduce variability among providers.
The initiation of public reporting of coronary artery bypass grafting (CABG) outcomes and the introduction of drug-eluting stents in 2003 coincided with a marked decrease in adverse events after percutaneous coronary intervention (PCI) and CABG procedures in California. This can be explained by a combination of decreased procedure volume, decreased mortality after CABG, and decreased need for reintervention after PCI. Although technical advances and improved medical management have contributed to the need for fewer interventions and better procedural success, the incidence of major complications, such as stroke and MI, has remained unchanged. Decreasing volume has resulted in a large number of “low-volume” cardiovascular programs in California. The present study examines the impact of this change. Variation in outcomes within hospital volume categories was studied by determining hospital readmission rates for adverse events during the first postprocedure year from 2006 to 2010 in California. In-hospital mortality (IHM), major adverse events at 90 days (MAE-90), and major adverse events at 365 days (MAE-365) are compared for isolated CABG, PCI with acute coronary syndrome (PCI-ACS), and all other PCI (PCI without acute coronary syndrome [PCI-noACS]).
Methods
The California Cardiac Surgery and Intervention Project, a program supported by the California Society of Thoracic Surgeons, collects data from the California Office of Statewide Health Planning and Development (OSHPD) hospitalized patient database on all cardiac surgery and PCI procedures performed in the state. Because California requires licensed hospitals to submit data on all discharged patients biannually, a 100% collection rate is assured for patients who underwent either CABG or PCI.
All patients who underwent an isolated CABG or PCI during 2006 to 2010 in California were included. Patients who underwent a concomitant major cardiac procedure (i.e., valve, aneurysm) with coronary revascularization were excluded. Two clearly distinct categories of PCI were identified using the International Classification of Diseases, Ninth Revision (ICD-9-CM), acute myocardial infarction codes 410.xx. Patients who underwent PCI with an admitting diagnosis of any acute MI code were categorized as PCI-ACS. All other PCI procedures were categorized as PCI-noACS. No attempt was made to classify patients with unstable angina as “ACS”; thus, these patients were included in the PCI-noACS group if they did not also have an associated 410.xx admitting diagnosis.
Annual follow-up was complete through 2011. IHM was corrected for “same day” transfers to another health care facility. Documentation of methodology and risk modeling is reported on the California Cardiac Surgery and Intervention Project Web site. Hospital readmission rates for adverse events (death, stroke, acute myocardial infarction and reintervention for another procedure) were determined at 90 days and 1 year postprocedure. The OSHPD patient discharge database did not include information on target vessel reintervention. For this reason, elective readmissions for PCI were excluded and only “unplanned” reinterventions were included in calculation of MAE-90 and MAE-365 days after the index procedure.
The collection of MAE data was accomplished using ICD-9-CM codes. Events were counted as present if they were new discharge diagnoses that occurred within the specified time period. Although it was possible to have a patient sustain more than 1 adverse event within the specified time period, only the first such event was counted for the combined outcome.
The hospitals were divided into terciles by total volume of “open heart” surgery (OHS) or PCI procedures performed over the 5-year period. The high-volume tercile averaged 747 ± 336 (SD) total PCI, 416 ± 224 total OHS, and 210 ± 130 CABG; intermediate volume tercile 362 ± 47 PCI, 168 ± 27 OHS, and 106 ± 27 CABG; and low-volume tercile 211 ± 6 PCI, 82 ± 28 OHS, and 53 ± 17 CABG.
Results
There were 114 cardiac surgery and interventional programs that actively offered both CABG and PCI services in California during 2006 to 2010. The number of procedures performed per year decreased 20% for CABG (from 15,721 in 2006 to 12,511 in 2010), 33% for PCI-noACS (36,299 to 24,906), and increased slightly for PCI-ACS (20,535 to 21,911).
During the 5-year study period, risk-adjusted IHM for CABG was 2.05%, 2.14%, and 2.62% for high, intermediate, and low-volume terciles; IHM for PCI-ACS was 4.02%, 4.07%, and 3.92%; and for PCI-noACS 0.64%, 0.72%, and 0.85%. There was little variation from year to year, and no trend for improving IHM in any volume or procedure category during the 5-year study period.
Annual change in procedural volume that occurred among individual hospitals in the volume terciles for PCI-noACS and CABG is shown in Figure 1 . PCI-noACS procedures decreased more noticeably between 2006 and 2008 among the highest volume sites in each tercile and tended to level off from 2008 to 2010.
There was little variation in risk-adjusted adverse events among volume groups ( Table 1 ). Adverse events averaged 6% to 7% at 90 days for CABG and PCI-noACS and 13% for PCI-ACS. Event rate at 1 year (MAE-365) was 10% to 11% for CABG, 20% for PCI-ACS, and 15% for PCI-noACS. As volume decreased for CABG and PCI-noACS, high-volume hospitals maintained relatively consistent outcomes, whereas low-volume programs showed widely varying results. The mean annual event rate for volume terciles remained constant at approximately 10% for CABG and 15% for PCI-noACS.
| High Volume | Intermediate Volume | Low Volume | ||||
|---|---|---|---|---|---|---|
| MAE-90 (%) | MAE-365 (%) | MAE-90 (%) | MAE-365 (%) | MAE-90 (%) | MAE-365 (%) | |
| CABG | ||||||
| 2006 | 6.2 | 10.0 | 6.0 | 10.3 | 7.2 | 12.0 |
| 2007 | 5.7 | 10.2 | 5.7 | 9.9 | 6.6 | 13.0 |
| 2008 | 6.0 | 10.2 | 5.7 | 9.9 | 7.7 | 12.5 |
| 2009 | 5.3 | 9.6 | 6.2 | 10.4 | 6.0 | 9.5 |
| 2010 | 5.9 | 10.4 | 5.3 | 9.0 | 6.1 | 10.2 |
| 2006–2010 | 5.8 | 10.1 | 5.8 | 9.9 | 6.7 | 11.5 |
| PCI-ACS | ||||||
| 2006 | 13.3 | 20.4 | 14.0 | 20.7 | 12.9 | 19.5 |
| 2007 | 13.6 | 20.9 | 14.2 | 22.1 | 12.3 | 19.1 |
| 2008 | 13.8 | 20.6 | 13.3 | 21.2 | 13.9 | 20.3 |
| 2009 | 13.2 | 20.4 | 12.3 | 19.6 | 12.5 | 20.0 |
| 2010 | 12.9 | 20.1 | 11.6 | 18.8 | 10.9 | 18.5 |
| 2006–2010 | 13.3 | 20.5 | 13.0 | 20.4 | 12.5 | 19.5 |
| PCI-noACS | ||||||
| 2006 | 6.7 | 14.3 | 7.6 | 15.1 | 7.1 | 13.9 |
| 2007 | 7.2 | 14.8 | 7.6 | 15.3 | 7.5 | 15.3 |
| 2008 | 7.3 | 15.3 | 7.4 | 15.2 | 7.5 | 14.8 |
| 2009 | 7.1 | 14.6 | 7.3 | 14.8 | 7.9 | 15.3 |
| 2010 | 7.1 | 14.9 | 6.9 | 13.9 | 7.0 | 14.3 |
| 2006–2010 | 7.1 | 14.8 | 7.4 | 14.9 | 7.4 | 14.7 |
Annual MAE-365 varied between 3 and more than 40% for lower volume hospitals, with less variation at higher volume ( Figure 2 ). This figure also illustrates the impact of volume on statistical significance or outlier status at 95% confidence level. Although the percentage of adverse outcomes may remain the same, the likelihood of having a “better” or “worse” event rate increases with increasing procedure volume. This is further detailed in Table 2 , showing the incidence of statistical significance in any year in the different outcome categories. High-volume sites were more than twice as likely to have better or worse outcomes than low-volume sites in adverse events at 1 year, although the average incidence of events is similar for each group as shown in Table 1 . Low-volume sites were also less likely to have better outcomes than worse outcomes. For all procedure categories and outcome groups, hospitals in the low-volume terciles had 9 instances of better than expected and 59 instances of worse than expected annual outcomes in any year, whereas hospitals in the high-volume terciles had 74 better and 111 worse instances of outlier status.
| CABG | PCI-ACS | PCI-NOACS | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Better | Same | Worse | Better | Same | Worse | Better | Same | Worse | |
| IHM | |||||||||
| High volume | 2 | 184 | 3 | 1 | 182 | 6 | 2 | 181 | 6 |
| Inter volume | 0 | 181 | 8 | 3 | 181 | 5 | 0 | 186 | 3 |
| Low volume | 0 | 182 | 7 | 1 | 185 | 3 | 0 | 182 | 7 |
| MAE-90 | |||||||||
| High volume | 1 | 181 | 7 | 8 | 162 | 19 | 15 | 152 | 22 |
| Inter volume | 3 | 178 | 8 | 1 | 178 | 10 | 7 | 178 | 4 |
| Low volume | 0 | 188 | 1 | 0 | 180 | 9 | 3 | 175 | 11 |
| MAE-365 | |||||||||
| High volume | 11 | 167 | 11 | 11 | 166 | 12 | 23 | 141 | 25 |
| Inter volume | 7 | 175 | 7 | 4 | 175 | 10 | 6 | 177 | 6 |
| Low volume | 0 | 184 | 5 | 3 | 179 | 7 | 2 | 178 | 9 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
