Reduction of in-hospital mortality among California hospitals meeting Leapfrog evidence-based standards for abdominal aortic aneurysm repair
Article Outline
Background
The Leapfrog Group established evidence-based standards for abdominal aortic aneurysm (AAA) repair, including targets for case volume and perioperative β-blocker usage. The purpose of this study was to determine whether meeting these benchmarks correlated with improved patient outcomes over time.
Methods
We studied California hospitals that responded to consecutive Leapfrog Group Hospital Quality and Safety Surveys between 2000 and 2005. Survey results of compliance with Leapfrog standards were linked to patient outcomes for AAA repair using the California state discharge database for the corresponding years. A random-effects Poisson regression analysis was performed to measure the effect of meeting β-blocker and case volume standards on hospital mortality and average length of stay after elective open and endovascular AAA repair (EVAR) during the early (2000-2002) and later (2003-2005) phase of Leapfrog implementation.
Results
Among 140 hospitals that performed open AAA repair, 25 (17.4%) met the Leapfrog case volume standard, 32 (22.2%) were compliant with routine perioperative β-blocker use, 5 hospitals (3.5%) met both criteria, and 78 control hospitals failed to meet either standard. After controlling for temporal differences in hospital and patient characteristics, hospitals that implemented a policy for perioperative β-blocker usage were found to have an estimated 51% reduction of in-hospital mortality (relative risk, 0.49; 95% confidence interval, 0.24-0.99; P < .05) after open AAA repair cases compared with control hospitals over time. There was no improvement in mortality outcomes over time, however, after open AAA repair in hospitals meeting case volume standards. Among 111 California hospitals in which EVAR was performed, there was an estimated 61% reduction of in-hospital mortality over time (relative risk, 0.39; 95% confidence interval, 0.07-1.80) among hospitals meeting Leapfrog case volume standards compared with control hospitals, although these results did not reach statistical significance. Finally, there was no reduction in length of hospital stay over time after either EVAR or open AAA repair for hospitals meeting Leapfrog standards compared with control hospitals.
Conclusions
This population-based study supports the effectiveness of meeting Leapfrog AAA repair standards towards improving mortality outcomes over time and suggests that their impact depends upon procedure type. Further studies are needed to help promote the standardization of evidence-based measures that may improve vascular surgery outcomes.
High-quality clinical research is published every year that pertains to the diagnosis, treatment, and management of patients with vascular disease. The practice of evidence-based medicine (EBM) attempts to identify those interventions with the best outcomes and implement them into clinical practice.1 This translation of evidence from well-designed observational studies and randomized controlled trials (RCTs) into clinical practice and health care policy constitutes an important source of quality improvement for vascular surgery.
One of the largest efforts to standardize EBM in the United States is being led by the Leapfrog Group, an alliance of large public and private health care purchasers representing >37 million individuals across the United States.2 This health care consortium was founded in 2000 with the aim to exert its combined leverage towards improving nationwide standards of health care quality, optimizing patient outcomes, and ultimately lowering health care costs.3, 4, 5 The Leapfrog Group's strategy to achieve these goals is through providing patient referral, financial incentives, and public recognition for hospitals that practice or implement EBM health care standards. These include hospital use of computerized physician order entry systems, compliance with 24-hour intensive care unit (ICU) physician staffing, evaluation using a 30-point composite Leapfrog Safe Practices Score, and evidence-based hospital referral standards for five high-risk operations.
Elective AAA repair is one of the high-risk operations targeted by the Leapfrog Group, with the goal to improve perioperative surgical care and optimize patient outcomes by promoting referral to hospitals compliant with specific EBM standards for this procedure. To date, two Leapfrog standards have been developed for hospitals in which elective AAA repair is performed and are based on collective evidence obtained from well-designed randomized trials and observational studies:
Hospitals and surgeons are increasingly being evaluated by their compliance with meeting Leapfrog standards for elective AAA repair, although the effectiveness of these measures over time has not been evaluated on a population level for each type of AAA repair procedure. To address this question, we performed a retrospective longitudinal analysis of in-hospital mortality outcomes after elective EVAR and open AAA repair among California hospitals according to whether they met Leapfrog standards. California, with a large number of hospitals that perform elective AAA repair, was among the first states targeted by the Leapfrog initiative. We hypothesized that hospitals in this state that met Leapfrog criteria for AAA repair have benefited from improvements in postoperative mortality and length of stay over time compared with hospitals that did not meet these EBM standards.
Methods
Hospital and patient data
We reviewed response data obtained from the Leapfrog Group Hospital Quality and Safety Surveys sent to California hospitals annually between 2001 and 2005. Surveys were sent to all acute-care hospitals in urban areas, and data collected represented self-reported information on hospital demographics and annual compliance with each of the Leapfrog Group safety standards. Of a total of 337 hospitals in California targeted by this initiative, 140 hospitals were identified in which AAA repair procedures were performed every year and responded to consecutive Leapfrog Group Hospital Quality and Safety Surveys during the time period. Outcomes among patients who underwent elective open or endovascular AAA repair procedures at these hospitals were identified using the California Office of Statewide Health Planning and Development (OSHPD) database for the corresponding years (2000-2005) and were linked by OSHPD identification number to the Leapfrog Group survey results.
International Classification of Diseases, 9th Revision (ICD-9) procedure codes were used to identify open AAA repair (38.34, 38.36, 38.44, 38.64, 39.25, 39.52) and EVAR (39.71) procedures from the OSHPD database, and were cross-referenced with ICD-9 diagnosis codes for nonruptured AAA (441, 441.0, 441.4, 441.9). The EVAR procedure code was introduced in October 2000; therefore, data on this procedure are available from this point on. The ICD-9 codes used to designate emergency or ruptured AAA repairs were excluded from analysis. The combination of diagnosis and procedure codes ensured that only outcomes from elective AAA repair cases were analyzed. An Institutional Review Board approved a protocol for this study.
Study design
We compared hospital characteristics, in-hospital mortality, and hospital length of stay (LOS) for two consecutive time periods:
Given that in-hospital outcomes after elective AAA repair are known to differ by whether the procedure is performed using EVAR or open techniques, all analyses were performed after stratifying by the type of procedure.11, 12 Hospitals were stratified into one control group and three comparison groups according to their fulfillment of Leapfrog AAA repair volume and perioperative β-blocker usage standards:
Hospital level covariates that were analyzed included the number of total annual admissions, number of floor admissions, number of ICU admissions, number of staffed floor beds, number of licensed floor beds, number of staffed ICU beds, number of licensed ICU beds, hospital membership status in health organization, presence of an Accreditation Council for Graduate Medical Education (ACGME) accredited general surgery or vascular surgery residency training program, the total number of elective AAA repairs reported by survey, and the proportion of EVAR to open AAA repair cases performed within individual hospitals over time. Patient level variables that were analyzed included age, sex, race, ethnicity, insurance status, and Charlson Comorbidity Index score. The main outcome measures were in-hospital mortality and median hospital LOS after elective AAA repair.
Statistical analysis
We compared hospital characteristics, in-hospital mortality, and LOS before and after Leapfrog AAA standards between each of the defined groups using χ2 tests for categoric variables and analysis of variance (ANOVA) or Student paired t tests for continuous variables that were normally distributed. The Wilcoxon signed rank test was used to compare nonparametric data between time periods. A Poisson regression model, with a random intercept for each hospital, was used to estimate the effect of meeting Leapfrog β-blocker and case volume standards on in-hospital mortality with and without adjusting for hospital and patient-level confounders.
To estimate the independent effect of these two EBM standards, data were pooled from hospitals that met both Leapfrog standards to create two hospital subgroups that met either β-blocker or case volume standards, respectively. The effect of Leapfrog standards on in-hospital mortality after AAA repair was calculated as the mortality rate ratio for hospitals that met each Leapfrog standard category (2 subgroups) vs hospitals that did not meet any Leapfrog standards (control group) for each time period.
A ratio of mortality rate ratios was calculated separately for the comparison of each Leapfrog subgroup with the control group over time. The effect of Leapfrog standards on hospital LOS was analyzed using a linear regression model and fit using a random intercept for each hospital and a log-normal distribution. Values of P less than α = 0.05 (two-sided) were considered to be significant for all statistical tests and models. Stata 9.0 statistical software (StataCorp, College Station, Texas) was used for all analyses.
Results
Hospital and patient characteristics
Baseline characteristics of the California hospitals analyzed in this study are reported in Table I, with the four groups defined by hospital compliance with Leapfrog standards after stratifying by AAA repair procedure type. Of 140 hospitals in which open AAA repair procedures were performed, 25 (17.4%) met only the Leapfrog case volume standard during either time period, 32 (22.2%) reported compliance with only the perioperative β-blocker usage standard beginning in 2003, 5 (3.5%) reported meeting both of these Leapfrog AAA repair standards during the defined time period, and 78 (56%) reported not meeting either Leapfrog AAA standard during the entire study period and were defined as controls. In addition, 111 California hospitals responded to consecutive Leapfrog surveys and performed EVAR cases during the study time period, including 23 (21%) that met case volume criteria alone, 25 (23%) that implemented the β-blocker standard alone, 5 (5%) that met both AAA standards, and 58 (52%) that did not meet either standard.
Table I. Characteristics of California hospitals responding to consecutive Leapfrog surveys between 2000 and 2005 in which elective open and endovascular abdominal aortic aneurysm repair was performeda
| Hospital characteristics | Open abdominal aortic aneurysm repair | Endovascular aneurysm repair | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Control | β-Blocker | Volume | Both | Pb | Control | β-Blocker | Volume | Both | Pb | |
| No. of hospitals | 78 | 32 | 25 | 5 | 58 | 25 | 23 | 5 | ||
| Total admissions, mean (SD)c | 12.8 (6.5) | 12.4(5.0) | 19.8(6.8) | 26.5(11.1) | <.05 | 13.2(5.6) | 13.4(4.8) | 19.6(6.2) | 26.5(11.1) | <.05 |
| ICU admissions, mean (SD)c | 1.6(1.7) | 1.5(0.8) | 2.1(1.3) | 4.4(2.8) | <.05 | 1.8(2.7) | 1.5(0.7) | 2.2(1.2) | 4.4(2.8) | <.05 |
| Staffed floor beds, mean (SD) | 196.6(93.2) | 187.8(81.8) | 289.2(109.7) | 449.0(228.1) | <.05 | 210.8(77.0) | 210.8(93.8) | 310.1(104.1) | 449.0(228.1) | <.05 |
| Staffed ICU beds, mean (SD) | 21.5(10.6) | 21.1(9.9) | 39.7(21.5) | 81.0(40.8) | <.05 | 22.6(9.7) | 22.7(10.5) | 40.6(19.8) | 81.0(40.8) | <.05 |
| Annual AAA volume, mean (SD) | 15.4(10.1) | 15.7(9.1) | 53.1(31.4) | 62.1(22.7) | <.05 | 17.2(10.7) | 19.3(9.6) | 60.5(36.4) | 65.7(25.7) | <.05 |
| Member of health system, No. (%) | 78(79) | 27(84) | 22(88) | 4(80) | 0.75 | 48(83) | 20(80) | 20(87) | 4(80) | 0.93 |
| ACGME Surgical Training, No. (%) | 8(10) | 2(6) | 4(16) | 3(60) | <.05 | 6(10) | 1(4) | 4(17) | 3(60) | <.05 |
aHospital strata based on meeting Leapfrog AAA standards. Control: No Leapfrog standard met; β-blocker: β-blocker standard met; Volume: Case volume standard met; Both: Both β-blocker and case volume standards met. |
bOne-way analysis of variance used for comparison of means, and Chi-square test used to compare proportions between groups. |
cExpressed as mean number of total hospital admissions per year in units of thousands. |
For hospitals in which both EVAR and open AAA repair cases were performed, those that met case volume or both Leapfrog standards had a significantly higher mean annual number of total and ICU admissions (P < .05) and higher number of floor and ICU beds (P < .05) compared with control hospitals that met none of the Leapfrog standard or hospitals that met only the β-blocker standard. These larger hospitals were also more likely to contain an ACGME-accredited general surgery or vascular surgery training program, although no difference existed between hospitals' Leapfrog status and their association with an organized health care system (Table I). Finally, although hospitals meeting case volume or both Leapfrog standards averaged a higher number of elective endovascular and open AAA cases per year, there was no significant difference in annual case volume between hospitals complying with the β-blocker standard and those hospitals failing to meet any Leapfrog standard.
The demographic makeup of patients' that underwent open and endovascular AAA repair in California hospitals responding to Leapfrog Group Hospital Quality and Safety Surveys between 2000 and 2005 is reported in Table II. During the study period, 8570 patients underwent open AAA repair, including 1497 (17%) at hospitals that implemented routine perioperative β-blocker use after 2003, 2,996 (35%) at hospitals that met the case volume standard, 670 (8%) at hospitals that met both standards, and 3407 (40%) at control hospitals. During the same time period, EVAR was performed on 4323 patients, including 704 (16%) at hospitals that met the perioperative β-blockade standard, 1783 (41%) at hospitals at met the minimum AAA volume standard, 499 (12%) at hospitals that met both Leapfrog standards, and 1337 elective cases (31%) at hospitals that failed to meet any Leapfrog standard.
Table II. Characteristics of patient population undergoing elective open and endovascular abdominal aortic aneurysm repair in California hospitals stratified by compliance with Leapfrog standardsa
| Demographics | Open abdominal aortic aneurysm repair | Endovascular aneurysm repair | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Control | β-Blocker | Volume | Both | Pb | Control | β-Blocker | Volume | Both | Pb | |
| Patient total | 3407 | 1497 | 2996 | 670 | 1337 | 704 | 1783 | 499 | ||
| Men, No. (%) | 2,702(21) | 1,169(22) | 2,366(21) | 533(20) | .84 | 1,123(84) | 603(86) | 1,524(86) | 436(87) | .49 |
| Age, mean (SD) y | 73.5(7.8) | 73.4(7.6) | 73.1(7.6) | 73.0(8.5) | 76.2(6.9) | 75.6(7.0) | 75.3(7.6) | 76.4(8.1) | ||
| Age, No. (%) y | .18 | .70 | ||||||||
| ≤34 | 0(0) | 0(0) | 0(0) | 0(0) | 1(0) | 1(0) | 0(0) | |||
| 35-64 | 573(17) | 224(15) | 520(17) | 109(16) | 140(10) | 70(10) | 202(11) | 58(12) | ||
| ≥65 | 2,833(83) | 1,272(85) | 2,476(83) | 559(83) | 1,197(90) | 634(90) | 1,581(89) | 441(88) | ||
| Race, No. (%) | <.05 | <.05 | ||||||||
| White | 2,942(86) | 1,318(88) | 2,766(92) | 569(85) | 1,187(89) | 619(88) | 1,596(89) | 398(80) | ||
| Black | 108(3) | 38(3) | 27(1) | 35(5) | 20(2) | 20(3) | 30(2) | 41(8) | ||
| Latino | 112(3) | 66(4) | 65(2) | 18(3) | 32(2) | 31(4) | 51(3) | 10(2) | ||
| Asian | 152(4) | 58(4) | 91(3) | 44(7) | 69(5) | 22(3) | 74(4) | 44(9) | ||
| Other | 92(3) | 18(1) | 47(2) | 4(1) | 29(2) | 12(2) | 32(2) | 6(1) | ||
| Insurance, No. (%) | .15 | .28 | ||||||||
| Medicare | 2,550(75) | 1,175(79) | 2,264(76) | 516(77) | 1,094(82) | 586(83) | 1,459(82) | 414(83) | ||
| Medicaid | 78(2) | 21(1) | 60(2) | 24(4) | 22(2) | 11(2) | 28(2) | 8(2) | ||
| Private | 709(21) | 276(18) | 635(21) | 125(19) | 216(16) | 101(14) | 254(14) | 74(15) | ||
| Self-pay, other | 31(1) | 13(1) | 10(0) | 2(0) | 5(0) | 6(1) | 42(2) | 3(0) | ||
| Charlson Index, No. (%) | .25 | .17 | ||||||||
| 1 | 1,369(40) | 582(39) | 1,225(41) | 294(43) | 573(43) | 320(45) | 827(46) | 199(40) | ||
| 2 | 1,141(33) | 506(34) | 1,009(34) | 211(32) | 464(35) | 237(34) | 586(33) | 181(36) | ||
| 3 | 472(14) | 206(14) | 428(14) | 96(14) | 175(13) | 89(13) | 239(13) | 67(13) | ||
| ≥4 | 425(12) | 203(14) | 334(11) | 69(11) | 125(9) | 58(8) | 131(7) | 52(10) | ||
aHospital strata based on meeting Leapfrog abdominal aortic aneurysm standards. Control: No Leapfrog standard met; β-blocker: β-blocker standard met; Volume: Case volume standard met; Both: Both β-blocker and case volume standards met. |
bOne-way analysis of variance used for comparison of means, and χ2 test used to compare proportions between groups. |
For patients undergoing both types of AAA repair procedures, there was no significant difference in the mean age, sex, insurance status, or Charlson Comorbidity Index score among the patient populations treated in hospitals that belonged to the different Leapfrog strata (Table II). However, hospitals that met both AAA standards performed a higher proportion of both EVAR and open cases on African American and Asian patients compared with other hospitals in the analysis.
In-hospital mortality
The average in-hospital mortality rates after elective open and endovascular AAA repair among California hospitals in different Leapfrog strata during both time periods are summarized in Table III. Among open AAA repair cases, a trend towards increasing mortality rates over time was observed among control hospitals as well as hospitals that met case volume and both Leapfrog standards. In comparison, in-hospital mortality after open AAA repair trended lower after low-volume hospitals implemented the β-blocker standard by the later time period (4.93% vs 3.25%; P = .12, Table III). When both low and high-volume hospitals that implemented a hospital policy for routine perioperative β-blocker use after 2003 were analyzed together, this mortality benefit persisted (Fig 1).
Table III. Clinical outcomes after endovascular and open abdominal aortic aneurysm repair in California hospitals responding to Leapfrog surveys during early (2000-2002) and later (2003-2005) phase of implementationa
| Leapfrog standard(s) | Elective cases, No. | In-hospital mortality, No. (%) | Length of hospital stay, mean (SD) days | |||||
|---|---|---|---|---|---|---|---|---|
| 2000-02 | 2003-05 | 2000-2002 | 2003-2005 | Pb | 2000-2002 | 2003-2005 | Pc | |
| Control (no standard) | ||||||||
| Open AAA (n = 78) | 1923 | 1484 | 74(3.85) | 75(5.05) | .09 | 7.09(1.7) | 7.48(3.0) | .33 |
| EVAR (n = 58) | 397 | 940 | 8(2.02) | 18(1.91) | .99 | 4.10(2.9) | 3.44(2.4) | .21 |
| β-Blocker standard | ||||||||
| Open AAA (n = 32) | 913 | 584 | 45(4.93) | 19(3.25) | .12 | 7.23(2.3) | 7.19(1.6) | .94 |
| EVAR (n = 25) | 261 | 443 | 3(1.15) | 7(1.58) | .75 | 5.70(5.5) | 4.21(5.7) | .39 |
| Case volume standard | ||||||||
| Open AAA (n = 25) | 1743 | 1253 | 69(3.96) | 55(4.39) | .56 | 6.92(1.2) | 6.94(1.0) | .95 |
| EVAR (n = 23) | 517 | 1266 | 6(1.16) | 8(0.63) | .25 | 3.02(2.2) | 2.30(1.4) | .19 |
| Both standards | ||||||||
| Open AAA (n = 5) | 406 | 264 | 11(2.71) | 11(4.17) | .30 | 6.30(1.9) | 6.20(1.5) | .93 |
| EVAR (n = 5) | 172 | 327 | 2(1.16) | 1(0.31) | .08 | 3.50(1.3) | 2.30(0.4) | .09 |
aHospitals stratified into four groups based on compliance with individual Leapfrog standards for AAA repair. |
bP values calculated for overall in-hospital mortality rates between time periods using χ2 and Fisher exact test. |
cP values calculated for mean length of hospital stay using Wilcoxon signed rank test. |
Fig 1.
Comparison of mean death rates for elective open abdominal aortic aneurysm (AAA) and endovascular AAA repair (EVAR) in California hospitals meeting Leapfrog standards for the 3-year periods during early (2000-2002) and later (2003-2005) phases of initiative. A significant decrease in mean death rates after open AAA repair was observed in the later time period for hospitals that implemented a policy for routine perioperative β-blocker use. In comparison, mortality rates after open AAA repair for control hospitals and those meeting Leapfrog case volume standards trended higher over consecutive time periods. Average in-hospital death rates after EVAR cases trended lower for all hospital groups over time, although the greatest reduction in mortality was observed among hospitals that met AAA case volume. Error bars represent the standard error of the mean.
A random-effects regression model was used to control for changes in patient and hospital variables over time, and an estimated 51% reduction in mortality (relative risk [RR], 0.49; 95% confidence interval [CI], 0.24-0.99; P < .05) after open AAA cases was found among all California hospitals that implemented a policy for routine β-blocker use compared with control hospitals. In comparison, adjusted regression models found no significant difference in open AAA in-hospital mortality during the consecutive time periods among all hospitals that met Leapfrog case volume standards (Table IV).
Table IV. Mortality rate ratio estimates for open abdominal aortic aneurysm repair at California hospitals that met Leapfrog β-blocker and case volume standards compared with hospitals that met neither standard during early (2000-2002) and later (2003-2005) phase of implementationa
| Leapfrog standard | Rate ratio Leapfrog standard vs control | Ratio of rate ratios (95% CI) | P | |
|---|---|---|---|---|
| 2000-2002 (95% CI) | 2003-2005 (95% CI) | |||
| Control, no standard (n = 78) | 1.00(Ref) | 1.00(Ref) | 1.00(Ref) | |
| β-blockade standard (n = 37) | ||||
| Unadjusted | 1.12(0.75-1.67) | 0.70(0.45-1.08) | 0.63(0.36-1.09) | .097 |
| Adjustedb | 1.56(0.97-2.49) | 0.69(0.39-1.22) | 0.49(0.24-0.99) | .048 |
| Case volume standard (n = 30) | ||||
| Unadjusted | 1.01(0.71-1.44) | 0.86(0.62-1.20) | 0.86(0.54-1.37) | .53 |
| Adjustedc | 1.32(0.84-2.07) | 1.02(0.66-1.59) | 0.80(0.44-1.45) | .47 |
aRandom effects Poisson regression models used with intercept for each hospital to estimate difference in mortality rate between time periods. |
bPoisson models adjusted for race, insurance type, sex, age, Charlson comorbidity index score, total hospital and intensive care unit admissions, case-mix, and hospital abdominal aortic aneurysm volume. |
cPoisson models adjusted for race, insurance type, sex, age, Charlson comorbidity index score, total hospital and ICU admissions, case-mix, and hospital compliance with β-blocker standard. |
In-hospital mortality rates after EVAR cases were not observed to significantly change during the consecutive time periods among control hospitals and low-volume hospitals that met the β-blocker standard (Table III). In comparison, a trend towards lower mortality rates between time periods after EVAR cases was found among hospitals that met only case volume (1.16% vs 0.63%; P = .25) and both Leapfrog standards (1.16% vs 0.31%; P = .08).
When all California hospitals in the study that met case volume criteria were analyzed together, compliance with this Leapfrog standard was found to reduce in-hospital mortality after EVAR during the consecutive time periods to a larger extent than control hospitals or those that met β-blocker standards (Fig 1). The adjusted in-hospital mortality rate after EVAR cases was estimated to have been reduced by 61% over time (RR, 0.39; 95% CI, 0.07-1.80; P = .26) among hospitals that met case volume standards compared with control hospitals, although this result did not reach statistical significance (Table V).
Table V. Mortality rate ratio estimates for endovascular abdominal aortic aneurysm repair at California hospitals that met Leapfrog β-blocker and case volume standards compared with hospitals that met neither standard during early (2000-2002) and later (2003-2005) phase of implementationa
| Leapfrog standard | Rate ratio Leapfrog standard vs control | Ratio of rate ratios, (95% CI) | P | |
|---|---|---|---|---|
| 2000-2002 (95% CI) | 2003-2005 (95% CI) | |||
| Control, no standard (n = 58) | 1.00(Ref) | 1.00(Ref) | 1.00(Ref) | |
| β-blockade standard (n = 30) | ||||
| Unadjusted | 0.57(0.19-1.75) | 0.54(0.24-1.25) | 0.95(0.23-3.81) | .94 |
| Adjustedb | 0.78(0.14-4.36) | 0.90(0.34-6.73) | 1.14(0.16 −8.33) | .90 |
| Case volume standard (n = 28) | ||||
| Unadjusted | 0.61(0.20-1.88) | 0.29(0.13-0.66) | 0.54(0.15 −1.98) | .35 |
| Adjustedc | 1.25(0.28-4.72) | 0.62(0.22-1.70) | 0.39(0.07-1.80) | .26 |
aRandom effects Poisson regression models used with intercept for each hospital to estimate difference in mortality rate between time periods. |
bPoisson models adjusted for race, insurance type, sex, age, Charlson comorbidity index score, total hospital and intensive care unit admissions, case-mix, and hospital AAA case volume. |
cPoisson models adjusted for race, insurance type, sex, age, Charlson comorbidity index score, total hospital and intensive care unit admissions, case-mix, and whether hospital reported compliance with β-blocker standard. |
Hospital length of stay
The mean ± standard deviation hospital LOS among patients undergoing elective open and endovascular AAA repair at California hospitals stratified by Leapfrog standard compliance for the consecutive time periods is reported in Table III. No significant change in LOS after open AAA repair was found among control hospitals and those meeting β-blocker and case volume standards over time (Fig 2). In comparison, a trend towards lower hospital LOS after EVAR cases over time was observed among both control hospitals and those meeting Leapfrog standards (Table III and Fig 2). When hospitals meeting Leapfrog β-blocker and case volume standards were compared with control hospitals in adjusted log-normal regression models, however, no significant difference was found in LOS after EVAR cases during the consecutive time periods.
Fig 2.
Comparison of mean hospital length of stay for abdominal aortic aneurysms (AAA) repaired electively by open and endovascular (EVAR) techniques in California hospitals meeting Leapfrog standards for early (2000-2002) and later (2003-2005) phases of the initiative. No decrease in hospital length of stay over time after open AAA repair was observed among control hospitals and those meeting β-blocker and case volume standards. In comparison, hospital length of stay after EVAR cases decreased over consecutive time periods to a similar extent among both control hospitals and those meeting Leapfrog standards. Error bars represent the standard error of the mean.
Discussion
Abdominal aortic aneurysm repair is among five high-risk operations targeted by the Leapfrog Group for evidence-based hospital referral standards. The specific criteria for these standards is based on evidence that elective AAA repair performed at hospitals meeting minimum procedure volume and complying with perioperative β-blockade should benefit from a significant reduction in postoperative morbidity and mortality.5, 7, 8, 9, 10 The generalizability of applying these EBM standards toward improving AAA postoperative outcomes remains controversial, however, given that their effectiveness has not been demonstrated using data from participating hospitals. To our knowledge, this is the first population-based study to directly measure the temporal impact of meeting Leapfrog standards for AAA repair on clinical outcomes.
Our data indicate that patients who had open AAA repair in California hospitals that implemented the Leapfrog standard for perioperative β-blockade experienced an estimated 50% reduction in postoperative mortality over time compared with patients who had operations in control hospitals that did not meet any quality standards. Patients who underwent EVAR in California hospitals that met Leapfrog case volume standards also appeared to benefit from a reduction in mortality over time, although this finding did not reach statistical significance. These results suggest that hospital compliance with Leapfrog standards for elective AAA repair is an effective means to help improve in-hospital mortality outcomes over time and support further efforts aimed at standardizing patient referral to hospitals that comply with other EBM standards.
There has been considerable controversy regarding referral patterns and the development of centers of excellence for high-risk operations such as AAA repair.13, 14, 15, 16, 17 Much of the debate has centered on deciding what specific measures should be used to assess hospital and surgical quality. Hospital procedure volume emerged as an important metric early in this process and was used as the sole standard for AAA repair when the first Leapfrog surveys were released in 2001. The use of a volume threshold alone to define hospital quality has been challenged, however,15, 16, 18 including an inability to arrive at a consensus for appropriate volume cutoffs, the exclusion of low-volume hospitals that have excellent outcomes, and the argument that hospital volume serves as merely a proxy for many other process measures that ultimately determine clinical outcomes.
Furthermore, although AAA volume standards were originally developed as a metric for open AAA cases, our data suggest that this measure has the greatest impact on improving EVAR outcomes over time. Our study was underpowered to achieve statistical significance; nevertheless, these findings are consistent with the results of a recent large nationwide cohort study that showed that outcomes for EVAR steadily improved from 2000 to 2003 as volume increased, whereas open AAA repair outcomes remained fairly constant.19 Given that hospitals in our study that met Leapfrog case volume criteria performed most of the elective AAA repairs after 2003 using EVAR techniques (Table III), perhaps it should not come as a surprise that this procedure benefited the most within hospitals meeting this standard.
Despite the benefits of case volume on AAA outcomes, most hospitals across the country do not have the referral pattern or the infrastructure to meet this standard.15, 16 The addition of hospital compliance with perioperative β-blockade in 2003 as a second criteria for elective AAA repair helped address this limitation by allowing hospitals of all sizes to implement a process-based measure.5 As our data show, this standard was put into practice primarily by lower-volume hospitals in California that still do most elective AAA cases with traditional open repair (Table I, Table III).
An interesting finding of this study is that the dramatic reduction of in-hospital mortality for the time period after implementation of the β-blocker standard was achieved through reduction of in-hospital mortality after open AAA cases, but not EVAR procedures (Table III). These data suggest that process measures such as perioperative β-blocker usage are most critical to implement during procedures such as open AAA repair that require longer hospital LOS. Together, both Leapfrog standards provide quality benchmarks for AAA repair that large and small hospitals in which endovascular and open AAA repair is performed should aim to achieve.
The study has several important limitations. First, the validity of this study relies on controlling for factors that may have changed differentially over time between control hospitals and those meeting Leapfrog standards. These include temporal changes in risk stratification, device improvements, and technical issues related to the operative repair. Our model adjusts for some patient factors that influence operative risk, but the use of an administrative database precludes the ability to control for all of these other important variables that may have changed differentially over time.
Second, the assignment of Leapfrog compliance for β-blocker and case volume by individual hospitals was made using self-reported survey data, and clinical outcomes were assessed retrospectively. Only urban California hospitals that returned surveys during consecutive periods were included in the analysis, but it is possible that hospitals not responding to Leapfrog surveys might have worse clinical outcomes and thus increase the overall treatment benefit for hospitals meeting Leapfrog standards for AAA repair.
Third, this analysis is limited to 140 hospitals in a single state, although the diverse demographic makeup of California provides a reasonable population from which to extrapolate the effect of Leapfrog AAA repair standards over time to the rest of the nation.
Fourth, the survey data used for this analysis only allowed us to document whether there was a hospital-level policy for perioperative β-blocker use, and we did not have patient level data showing medication usage or hemodynamic variables to confirm this. Nevertheless, our analysis was designed to demonstrate the effectiveness of implementing a hospital policy for β-blocker use on a population level, not to re-evaluate their efficacy at the patient level.
Finally, the analysis of in-hospital dependent variables for this study does not tell us whether long-term outcomes were different among hospitals that met different Leapfrog criteria. Indeed, it appears that a longer follow-up period or a larger population sample size will be necessary to show a statistically significant effect of meeting Leapfrog case volume standards on mortality outcomes after EVAR cases.
Conclusions
This population-level analysis among California hospitals provides evidence supporting patient referral to medical centers meeting Leapfrog case volume and β-blocker standards for EVAR and open AAA repair, respectively. It will be important to confirm these findings among larger national data sets. Continuing to apply EBM standards toward improving patient care and clinical outcomes for elective AAA repair, as well as other high-risk vascular procedures, will ensure that patients receive the highest standard of care.
Author contributions
We are grateful to J. Dennis Bush (Thomson Healthcare) and Sarah Collins (LeapFrog Group) for their help obtaining LeapFrog Group survey data, and Aidan McDermott for assistance with executing the statistical analysis of this study.
References
- . What the surgeon of tomorrow needs to know about evidence-based surgery. Arch Surg. 2006;141:317–323
- . Fact sheet. http://www.leapfroggroup.orgAssessed: Dec 12, 2007
- . Has the Leapfrog Group had an impact on the health care market?. Health Aff (Millwood). 2005;24:228–233
- . Shaping the future of surgery: the role of private regulation in determining quality standards. Ann Surg. 2006;243:304–312
- . Potential benefits of the new Leapfrog standards: effect of process and outcomes measures. Surgery. 2004;135:569–575
- . Results of elective abdominal aortic aneurysm repair in the 1990s: a population-based analysis of 2335 cases. J Vasc Surg. 1999;30:985–995
- Hospital volume and surgical mortality in the United States. N Engl J Med. 2002;346:1128–1137
- . Volume standards for high-risk surgical procedures: potential benefits of the Leapfrog initiative. Surgery. 2001;130:415–422
- The effect of bisoprolol on perioperative mortality and myocardial infarction in high-risk patients undergoing vascular surgery (Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography Study Group). N Engl J Med. 1999;341:1789–1794
- . Beta-blocker therapy in noncardiac surgery. N Engl J Med. 2005;353:412–414
- . Update on open repair of abdominal aortic aneurysms: the challenges for endovascular repair. J Am Coll Surg. 2004;199:946–960
- . Guidelines for the treatment of abdominal aortic aneurysms (Report of a subcommittee of the Joint Council of the American Association for Vascular Surgery and Society for Vascular Surgery). J Vasc Surg. 2003;37:1106–1117
- . The Leapfrog standards: ready to jump from marketplace to courtroom?. Health Aff (Millwood). 2003;22:46–59
- . Using volume criteria: do California hospitals measure up?. J Surg Res. 2003;113:96–101
- . The case against volume as a measure of quality of surgical care. World J Surg. 2005;29:1222–1229
- . Leapfrog patient-safety standards are a stretch for most hospitals. Issue Brief Cent Stud Health Syst Change. 2004;1–6
- . Strategies for improving surgical quality–should payers reward excellence or effort?. N Engl J Med. 2006;354:864–870
- . Estimating the potential impact of regionalizing health care delivery based on volume standards versus risk-adjusted mortality rate. Int J Qual Health Care. 2007;19:195–202
- . Improving aneurysm-related outcomes: nationwide benefits of endovascular repair. J Vasc Surg. 2006;43:446–451discussion 451-2
Competition of interest: none.
PII: S0741-5214(08)00075-X
doi:10.1016/j.jvs.2008.01.021
© 2008 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
