Abstract
Introduction
Black patients who experience acute myocardial infarction and receive care in high minority-serving hospitals have higher readmission rates. This study explores how hospital system affiliation (centralized versus decentralized/independent) impacts 30-day readmissions after acute myocardial infarction in black men.
Methods
In 2018, the Healthcare Cost and Utilization Project State Inpatient Database (2009– 2013) was used to observe 30-day readmission for acute myocardial infarction by race, and data from the American Hospital Association Annual Survey of Hospitals (2009–2013) to determine hospital system affiliation for the states Arizona, California, North Carolina, and Wisconsin. A series of hierarchic logistic regressions were conducted to determine if hospital system affiliation mediates the relationship between race and 30-day readmission.
Results
Of 63,743 hospitalizations for acute myocardial infarction among men between 2009 and 2013, black men accounted for 7.1% of hospitalizations and 8.0% of readmissions. In both models, race significantly predicted 30-day readmission (unadjusted OR=1.25, 95% CI=1.14, 1.37, p<0.001, AOR=1.13, 95% CI=1.03, 1.25, p=0.046). After controlling for system type, black men were more likely to be readmitted after acute myocardial infarction than white men in both models (unadjusted OR=1.25, 95% CI=1.14, 1.38, p<0.001, AOR=1.14, 95% CI=1.03, 1.25). There was no difference in odds of being readmitted by race and hospital system type (unadjusted OR=0.88, 95% CI=0.25, 3.07, p=0.84, AOR=1.02, 95% CI=0.21, 5.10, p=0.98).
Conclusions
Black men appear to be more likely to be readmitted after acute myocardial infarction. Centralization does not appear to mediate the relationship between race and 30-day readmissions for acute myocardial infarction.
INTRODUCTION
Readmissions for acute myocardial infarction (AMI) persist despite the availability of preventive measures, such peer advisors post-discharge1 and disease management programs.2 Black patients are at higher risk for readmission after an AMI3 than white patients. Blacks are also at higher risk for mortality after AMI than whites.4
Blacks typically receive care in hospitals that have higher morbidity and mortality rates5; this may account for the some of the disparities in readmissions after an AMI. Hospitals that serve a high volume of black patients are typically large, located in the southern U.S., urban teaching hospitals,6 not-for-profit, and see more Medicaid patients.5 Hospitals that treat a high volume of black patients also have slightly lower performance on AMI quality measures including aspirin given upon arrival and at discharge, and use of angiotensin-converting enzyme–inhibiting drugs or angiotensin receptor blockers.5 In a study by Jha et al.5 this difference in AMI performance disappeared after adjusting for hospital referral region, suggesting that hospitals treating high volumes of black patients are located in geographic areas with lower overall quality of care. Risk-adjusted mortality after AMI is significantly higher in hospitals treating a high volume of black patients, even after accounting for income, hospital ownership, hospital volume, Census region, urban status, and the intensity of hospital surgical treatment.7
Fluctuations in the structure of health systems in recent decades have implications for the availability and quality of health care received by racial or ethnic groups. The past 20 years have seen remarkable consolidation in the hospital industry, as hospitals have joined into larger health systems. Currently, 65% of U.S. community hospitals are part of a health system.8 Health systems are groups of hospitals and other healthcare providers operating under common ownership. However, within health systems there is variation in organizational structure and management practices. In particular, systems can differ at the organizational level at which decisions are made. Among centralized health systems, decisions about service provision and physician arrangements are typically made at the system level. This differs from fully independent hospitals, which have no formal relationships to other hospitals, and decentralized hospitals, which are part of multihospital systems but have greater autonomy in decision making.9 This difference in the degree to which a hospital’s decision making is centralized can impact aspects of hospital operations including operating costs,10,11 adoption of patient safety measures,12 and outcomes for AMI patients.13 It is thought that centralization promotes standardization of processes and improvements in communication between providers.14,15 Relative to fully independent hospitals and decentralized system hospitals, centralized hospitals may have particular advantages that facilitate efforts to reduce readmissions, because improved communications and standardized discharge processes are two key elements of programs that have been successful in readmission reduction.16,17 Moreover, membership in a centralized system is thought to encourage the diffusion of best practices15 and financial resources across system hospitals.18
Given that black men face unique barriers in communicating with providers and implicit bias,19–23 communication and standardization improvements that encourage centralization could help reduce disparities in readmissions. The transfer of process knowledge and financial resources that can occur within a centralized hospital system may be particularly valuable to hospitals that treat high volumes of minority patients. Hospital characteristics may play an important role in mediating the relationship between race and the likelihood of readmission. Hospitals that are more likely to treat black patients have higher rates of readmissions and differ in important structural characteristics like ownership, teaching status, and geographic region.24 The degree of centralization within a hospital system may affect disparities in readmissions both directly, by enabling practices that reduce readmissions, and indirectly, if black patients are more likely to be seen by hospitals in independent systems.
Prior studies of AMI readmissions and hospital affiliation focused on hospital characteristics and used Medicare claims data to explore the effects of hospital size, hospital teaching status, financial stress, and public ownership.24 This study seeks to explore the impact of hospital system affiliation (centralized versus decentralized/independent hospitals) on 30-day readmissions after an AMI in black men. The hypotheses for this study are that black men will have higher 30-day readmissions for AMI than men of other racial groups; black men are more likely to be seen in decentralized/independent hospitals; and the relationship between race and the likelihood of readmission will be weaker after controlling for the admitting hospital’s system type (i.e., centralized or decentralized/independent).
METHODS
Study Population
The hypotheses were tested using data from two sources: the Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases (SIDs) and the American Hospital Association (AHA). The HCUP SID database is sponsored by the Agency for Healthcare Research and Quality and includes hospital inpatient information from across the U.S.25 HCUP data are collected through state data organizations, hospital associations, and private data organizations. The states of Arkansas (2009–2013), California (2010–2011), North Carolina (2010), and Wisconsin (2013) were selected and data were gathered on hospital readmissions for AMI and patient race. The states and years included were selected for analysis based on the availability of revisit variables in the SID and the diversity of the population within the state. Four states in different geographic regions were included to increase the generalizability of results. State and year were included in the regression analysis to account for differences between states and years of admission. The AHA database contains national survey data collected through the AHA Annual Survey of Hospitals and contains data on ≅6,500 hospitals and >400 health systems.26 This observational study was considered exempt from IRB oversight at the University of Alabama at Birmingham.
Measures
To identify the index AMIs in this analysis, hospitalizations were first included with primary diagnosis code (ICD-9 codes 410–410.91 and excluded ICD-9 codes 410.X2 for subsequent episodes of care). AMI hospitalizations within the observation period for each state were included. Observations that occurred during the first or last month in the state dataset were excluded in order to ensure that the observation was not a readmission or an admission for which a readmission could not be counted. Only men and live discharges were included in this analysis. Exclusion criteria for this analysis were observations in which days to event were missing, and observations where the hospital for index hospitalization was not found in the AHA survey.
Thirty-day readmission rate was defined as number of AMI admissions in which there was at least one subsequent hospital admission within 30 days, divided by total number of AMI admissions. Each qualifying hospital stay with AMI as the principal diagnosis was counted as a separate index admission. Readmissions could include hospitalization during the 30-day time period at the same hospital or a different hospital. The index hospitalization for the patient’s first AMI was used in analysis. Patients undergoing elective revascularization during their readmission were considered to be not readmitted (n=753, 7.6% of readmissions).
Patient comorbidity was defined using the Agency for Healthcare Research and Quality Elixhauser27,28 Comorbidity Index for readmissions. The Elixhauser comorbidity adjustment was developed in 1998 for use with hospital administrative discharge data and includes a set of 30 clinical conditions that exist prior to hospitalization, unrelated to the principal diagnosis, and likely to influence mortality and resource utilization within the hospital.27 The Elixhauser Index, a single score that incorporates weights for each comorbidity, has been tested using HCUP databases and effectively incorporates the influence of comorbid conditions on the risk of readmission.28
The AHA database contains information about whether a hospital is a member of a multihospital system and, if so, the degree of centralization in that system. System membership is captured by the dichotomous variable named MSHMEMB, which is equal to 1 for all hospitals owned by multihospital systems and 0 for all hospitals not owned by multihospital systems. For hospitals that are part of systems, the AHA data assigns one of six different system types that reflect the degree of centralization within the system. These six levels include (1) centralized, (2) centralized health system, (3) moderately centralized health system, (4) decentralized health system, (5) independent health system, and (6) blank. The sixth, blank, category is reserved for hospitals that do not report sufficient data to be categorized. For analysis, a two-level centralization variable was employed, similar to that used by Moore and colleagues29 in which centralized was defined as health systems that centrally organize hospital service delivery, physician arrangements, and insurance product development (levels 1, 2, and 3 of the centralization variable), and decentralized/independent was defined as having decentralized or no system affiliation (MSHMEMB=0; levels 4 and 5 of the centralization variable). Hospitals for whom system affiliation could not be determined (level 6) were excluded from the analysis (n=1,119, 1.34% of all patients, 1.97% of patients in “other” non-independent hospitals). This analysis differed from Moore and colleagues29 in that decentralized hospitals were included with the independent hospital classification.
Statistical Analysis
Because of the Center for Healthy African American Men through Partnerships focus on African American men’s health, women were excluded from the analysis. Of the 81,954 men discharged alive after a primary hospitalization for AMI, 489 (0.6%) were excluded because they could not be linked to the AHA survey data, 1,129 were excluded because they resided in a different state than the hospital to which they were admitted, and the two states did not border one another. Patients who were transferred in from another hospital (n=13,450, 17.0%), those admitted to hospitals that were system members but whose centralization status was not known (n=1,119, 1.4%), and those who were missing values for race or age (n=2,021, 3.1%) were all excluded. The final cohort number is 63,743. All models predicting readmission adjusted for random hospital effect and other variables as fixed effects. Models predicting decentralized/independent hospital status used data aggregated to the hospital level (n=584).
This analysis, conducted in 2018, explored the characteristics of patients with AMI between 2009 and 2013 in Arkansas, California, North Carolina, and Wisconsin using chi-square and t-tests to compare characteristics of patients with and without a 30-day readmission for AMI. Next, a hierarchic logistic regression model with hospital-level random effects and patient-level fixed effects was used to compare the odds of readmission by racial group, controlling for patient characteristics. Finally, the potential mediating role of hospital system affiliation (centralized versus decentralized/independent hospitals) on the relationship between race (the risk factor of interest) and 30-day readmission (the primary outcome of interest) was examined. Using a four-step mediational approach as first described by Baron and Kenny30 where steps three and four were combined,31 regression analysis was used to predict readmission within 30 days with race as the only predictor. Next, regression predicted system affiliation, the putative mediator, with race as the predictor. Finally, regression was used to predict readmission within 30 days with race and system affiliation as the predictors. All tests were conducted at the p<0.05 significance level. Statistical analysis was carried out in SAS, version 9.3.
RESULTS
Within this dataset, there were 63,743 hospitalizations for AMI among men during years 2009–2013 in Arkansas, California, North Carolina, and Wisconsin (Table 1). Of those, 2,324 were readmissions for AMI (3.6%). The in-hospital mortality was 8%–9% and was consistent across races. Men aged ≥60 years had the highest proportions of readmissions. Black men accounted for 7.1% of hospitalizations during the observation period and 8.0% of the readmissions. The majority of hospitals in this sample (69.5%) were decentralized/independent, and not part of a hospital system. The largest proportion of hospitalizations and readmissions in the sample were in California (65.1% and 68.4%, respectively). When comparing readmissions by race in centralized and in decentralized/independent hospitals (Figure 1), black men had significantly higher readmission than men of other races; there was no significant difference in readmission by race in centralized hospitals.
Table 1.
Characteristics of Patients With Acute Myocardial Infarction (AMI) Treated in Centralized and Independent Hospitals Between 2009–2013 in Arkansas, California, North Carolina, and Wisconsin
| Characteristics | Total AMI hospitalization | Centralized hospitals | Independent/decentralized hospitals | ||||
|---|---|---|---|---|---|---|---|
|
| |||||||
| 30 day readmissions | No 30 day readmissions | p-valueb | 30 day readmissions | No 30 day readmissions | p-value | ||
| Age, years | <0.001 | <0.001 | |||||
| <50 | 8,179 (12.8) | 218 (9.4) | 2,378 (13.9) | 464 (9.3) | 5,119 (13.0) | ||
| 50–59 | 14,906 (23.4) | 412 (17.7) | 4,133 (24.2) | 843 (16.9) | 9,518 (24.2) | ||
| 60–69 | 16,540 (25.9) | 561 (24.1) | 4,522 (26.4) | 1,182 (23.6) | 10,275 (26.1) | ||
| 70–79 | 12,733 (19.9) | 549 (23.6) | 3,294 (19.3) | 1,204 (24.1) | 7,686 (19.6) | ||
| ≥80 | 11,385 (17.9) | 584 (25.1) | 2,779 (16.2) | 1,308 (26.2) | 6,714 (17.1) | ||
| Racea | 0.094 | <0.001 | |||||
| White | 45,176 (70.9) | 1,639 (70.5) | 12,417 (72.6) | 3,369 (67.4) | 27,751 (70.6) | ||
| Black | 4,502 (7.1) | 185 (8.0) | 1,216 (7.1) | 437 (8.7) | 2,664 (6.8) | ||
| Other | 14,065 (22.1) | 500 (21.5) | 3,473 (20.3) | 1,195 (23.9) | 8,897 (22.6) | ||
| Insurance | <0.001 | <0.001 | |||||
| Medicare | 31,682 (49.7) | 1,426 (61.4) | 7,916 (46.3) | 3,149 (63.0) | 19,191 (48.8) | ||
| Medicaid | 4,457 (6.9) | 201 (8.6) | 1,222 (7.1) | 480 (9.6) | 2,554 (6.5) | ||
| Private | 19,154 (30.1) | 472 (20.3) | 5,602 (32.7) | 938 (18.8) | 12,142 (30.9) | ||
| Self-pay | 4,613 (7.2) | 90 (3.9) | 1,176 (6.9) | 254 (5.1) | 3,093 (7.9) | ||
| Other/Unknown | 3,837 (6.0) | 135 (5.8) | 1,190 (7.0) | 180 (3.6) | 2,332 (5.9) | ||
| Year | 0.03 | 0.051 | |||||
| 2009 | 2,695 (4.2) | 51 (2.2) | 462 (2.7) | 220 (4.4) | 1,962 (5.0) | ||
| 2010 | 28,466 (44.7) | 1,111 (47.8) | 8,401 (49.1) | 2,144 (42.9) | 16,810 (42.8) | ||
| 2011 | 23,671 (37.1) | 945 (40.7) | 6,624 (38.7) | 1,886 (37.7) | 14,216 (36.2) | ||
| 2012 | 2,928 (4.6) | 38 (1.6) | 301 (1.8) | 280 (5.6) | 2,309 (5.9) | ||
| 2013 | 5,983 (9.4) | 179 (7.7) | 1,318 (7.7) | 471 (9.4) | 4,015 (10.2) | ||
| State | 0.39 | 0.004 | |||||
| Arkansas | 14,228 (22.3) | 290 (12.5) | 2,209 (12.9) | 1,326 (26.5) | 10,403 (26.5) | ||
| California | 41,482 (65.1) | 1,589 (68.4) | 11,398 (66.6) | 3,282 (65.6) | 25,213 (64.1) | ||
| North Carolina | 4,786 (7.5) | 333 (14.3) | 2,611 (15.3) | 169 (3.4) | 1,673 (4.3) | ||
| Wisconsin | 5,983 (9.4) | 112 (4.8) | 888 (5.2) | 224 (4.5) | 2,023 (5.1) | ||
| Bed size | 0.61 | 0.21 | |||||
| 6–99 beds | 3,406 (5.3) | 95 (4.1) | 621 (3.6) | 336 (6.7) | 2,354 (6.0) | ||
| 100–299 beds | 28,937 (45.4) | 778 (33.5) | 5,874 (34.3) | 2,478 (49.6) | 19,807 (50.4) | ||
| 300–499 beds | 23,681 (37.2) | 864 (37.2) | 6,379 (37.3) | 1,861 (37.2) | 14,577 (37.1) | ||
| ≥500 beds | 3,247 (5.1) | 587 (25.3) | 4,232 (24.7) | 326 (6.5) | 2,574 (6.5) | ||
| Length of stay (days), | 63,743/4.0 (5.0) | 2,324/ | 17,106/4.4 | <0.001 | 5,001/5.9 | 39,312/3.9 | <0.001 |
| n/mean (SD) | 6.5(7.5) | (4.8) | (6.3) | (4.6) | |||
| Volume AMI | 63,743/173 | 2,324/183.3 | 17,106/187.5 | 0.079 | 5,001/162.8 | 39,312/168.2 | <0.001 |
| patients/year, n/mean (SD) | (120.5) | (106.7) | (109.2) | (124.7) | (124.7) | ||
| Elixhauser index, | 63,743/11 | 2,324/16.2 | 17,106/9.9 | <0.001 | 5,001/16.6 | 39,312/10.5 | <0.001 |
| n/mean (SD) | (12.5) | (13.4) | (11.9) | (13.7) | (12.3) | ||
Note: Data presented as n (%) unless otherwise noted. Boldface indicates statistical significance (p<0.05).
Missing=2,564 total hospitalizations, 3.19% of the total sample.
t-test p-value
Figure 1.
Thirty day readmissions by race in centralized and independent hospitals in Arkansas, California, North Carolina, and Wisconsin, 2009–2013.
In the series of hierarchic logistic regression models, the unadjusted OR with 95% CI (Table 2) demonstrates that black race was associated with significantly higher odds of 30-day readmission after AMI than white race, OR=1.25 (95% CI=1.14, 1.37, p<0.001). Controlling for patient characteristics (age, state, insurance type, comorbidities, and length of stay) had an impact on the magnitude of the association and it remained statistically significant, OR=1.14 (95% CI=1.03, 1.25, p=0.034).
Table 2.
Associations of Race With 30-day Readmission After AMI, Adjusting for Patient Characteristicsa
| Variable | 30 day readmission
|
|
|---|---|---|
| OR (95% CI) | p-value | |
| Unadjusted | ||
| Race | ||
| White | ref | |
| Black | 1.25 (1.14, 1.37) | <0.001 |
| Other | 1.06 (0.99, 1.13) | <0.001 |
| Adjusted | ||
| Race | ||
| White | ref | 0.03 |
| Black | 1.14 (1.03, 1.25) | |
| Other | 1.02 (0.95, 1.09) | |
| Age, years | 1.01 (1.008, 1.013) | <0.001 |
| State | ||
| North Carolina | ref | <0.001 |
| Arkansas | 1.21 (1.03, 1.42) | |
| California | 0.97 (0.85, 1.10) | |
| Wisconsin | 0.97 (0.77, 1.22) | |
| Insurance | ||
| Private | ref | <0.001 |
| Medicare | 1.37 (1.27, 1.48) | |
| Medicaid | 1.74 (1.57, 1.93) | |
| Self | 1.03 (0.91, 1.17) | |
| Other/Unknown | 1.03 (0.90, 1.17) | |
| Elixhauser Index Score | 1.04 (1.04, 1.05) | <0.001 |
| Length of stay days | 1.04 (1.036, 1.045) | <0.001 |
| Year | ||
| 2009 | ref | <0.001 |
| 2010 | 1.26 (1.08, 1.47) | |
| 2011 | 1.29 (1.11, 1.51) | |
| 2012 | 1.13 (0.95, 1.34) | |
| 2013 | 1.19 (1.00, 1.42) | |
Note: Boldface indicates statistical significance (p<0.05)
Patient characteristics include age, state, insurance type, comorbidities, and length of stay.
AMI, acute myocardial infarction
Table 3 reports the results of the mediation analysis before and after controlling for patient and hospital-level characteristics. In the unadjusted model, race significantly predicted the primary outcome (30-day readmission) and resulted in OR=1.25 (95% CI=1.14, 1.37, p<0.001). Also, black men were as likely as white men to be admitted to decentralized/independent hospitals, OR=0.88 (95% CI=0.25, 3.07, p=0.84). When controlling only for system type, black men were more likely to be readmitted after AMI than white men, OR=1.25 (95% CI=1.14, 1.38, p<0.001). In the adjusted models that controlled for patient-level characteristics (age, state, insurance, Elixhauser Index, and length of stay) and hospital-level characteristics (AMI annual volume, bed size, and year of data), there was a positive relationship between race and 30-day readmission OR=1.13 (95% CI=1.03, 1.25, p=0.046). Similar to the unadjusted model, black men were as likely to be admitted to independent or decentralized hospitals as white men OR=1.02 (95% CI=0.21, 5.10, p=0.98). When controlling for system characteristics in the adjusted model, black men appeared again more likely than white men to experience readmission, OR=1.14 (95% CI=1.03, 1.25).
Table 3.
Hierarchical Logistic Regression Models Predicting 30-day Readmissions After AMI, Unadjusted and Adjusteda
| Logistic regression models | 30 day readmission after AMI
|
|
|---|---|---|
| OR (95% CI) | p-value | |
| Unadjusted | ||
| Race black vs white predicting readmission | 1.25 (1.14, 1.37) | <0.001 |
| Race black vs white predicting health system-independent | 0.88 (0.25, 3.07) | 0.84 |
| Race black vs white predicting readmission, controlling for health system-independent | 1.25 (1.14, 1.38) | <0.001 |
| Adjusted | ||
| Race black vs white predicting readmission | 1.13 (1.03, 1.25) | 0.046 |
| Race black vs white predicting health system-independent | 1.02 (0.21, 5.10) | 0.98 |
| Race black vs white predicting readmission, controlling for health system-independent | 1.14 (1.03, 1.25) | 0.037 |
Note: Boldface indicates statistical significance (p<0.05).
Controlling for patient-age, state, insurance, Elixhauser Index, length of stay and hospital-AMI annual volume, bed size, and year of data.
AMI, acute myocardial infarction
DISCUSSION
The objective of this study was to explore the role of hospital system affiliation and its impact on 30-day readmissions after AMI in black men. In this study, black men had significantly higher readmission than men of other races in decentralized/independent hospitals; there was no significant difference by race in centralized hospitals. In unadjusted hierarchic logistic regression models, black race was associated with significantly higher odds of 30-day readmission after AMI than whites; these higher odds remained after controlling for patient and hospital-level characteristics. In the unadjusted mediation analysis, black race significantly predicted 30-day readmission after AMI, black men are as likely as white men to be admitted to decentralized/independent hospitals, and black men are more likely than white men to be readmitted when controlling for system type. After adjusting for patient and hospital-level characteristics, black race significantly predicted 30-day readmission after AMI, black men are as likely as white men to be admitted to decentralized/independent hospitals, and black men are more likely to be readmitted than white men when controlling for system type.
Previous research supports the findings that there are higher readmissions for blacks. For instance, a study of 30-day readmission for Medicare beneficiaries found black patients had higher readmission than white patients after hospitalization for AMI.24 Similarly, a study of Medicare and Medicare Advantage patients in New York showed that black patients had a higher risk for readmission post-surgery discharge for coronary artery bypass graft than white patients.32
Controlling for patient characteristics in this study slightly lowered the strength of the association between race and readmission, and the association remained statistically significant. Prior studies have also examined the role of patient characteristics in reducing readmissions. A 2013 study found income inequality was associated with increased risk of readmission.33 Tobacco use has also been shown to increase the risk of readmission after AMI.34
This is the first study to compare associations and the potential mediating relationship of hospital system affiliation with 30-day readmissions after AMI by race. However, results do not support the assertion that the type of system to which a hospital belongs (centralized versus decentralized/independent) mediates the relationship between race and readmissions for AMI; controlling for hospital centralization does not reduce the odds of readmission as hypothesized. This was surprising given the extensive literature suggesting that the extent to which health system affiliation affects patient outcomes is highly dependent on how closely hospital operations are tied to the health system.35 Theoretically, hospitals that operate within health systems should experience improved coordination and information transfer between hospitals in multihospital systems35 that would result in better patient outcomes. Hospitals affiliated with health systems have better quality of care than independent hospitals.36
Several prior studies have examined the relationship between hospital characteristics, health system affiliation, and patient outcomes. A study by Joynt et al.37 found higher 30-day readmission for heart failure among patients who were discharged from hospitals that were resource poor, both financially and clinically; publicly owned; located in counties with low median income; hospitals with no cardiac capabilities and low nursing staff; and smaller hospitals. Hospitals in centralized health systems or that become more centralized may experience larger reductions in mortality over time than hospitals that remain freestanding.38
Limitations
A strength is that this paper combines observations from two high-quality, comprehensive databases that allowed an analyst not only to control for individual-level characteristics but also health system and community-level factors. However, there are several limitations worth noting. Although the HCUP SID data offer comprehensive portraits of hospital discharges within each state, this analysis was limited to those states reporting both race and readmission data. In addition, the data do not contain information on post-discharge mortality. Within the decentralized/independent and centralized groups, there are likely many different organizational strategies and tools to reduce readmissions. The analysis combined the most centralized categories of hospitals into one group and the least centralized hospitals together with independent hospitals in a second group. There are other categorizations possible, though the one employed is conservative in the sense that it used hospitals in the full range of system types rather than those at the extreme ends of the centralization spectrum. The analysis only included data from four states (Arkansas, California, North Carolina, and Wisconsin). Although these include different geographic regions, provider markets and patient populations may not be representative of the entire U.S. Because disparities differ greatly across regions and for different procedures,39 results of this study cannot be generalized to the U.S. as a whole.
CONCLUSIONS
These results reaffirm a troubling finding, that black men are more likely than white men to experience readmission after AMI. This suggests that effective interventions be developed and targeted to reduce readmissions among this vulnerable group of patients. Future evaluation of quality and readmissions among black men should consider the centralization status of the hospitals in which these patients are treated in order to test for correlations.
Acknowledgments
Research reported in this publication was supported by the Center for Healthy African American Men through Partnerships, funded by the National Institute of Minority Health and Heath Disparities through a grant from the NIH under award number U54MD008620. The content is solely the responsibility of the authors and does not necessarily represent the official views of NIH.
JW and AH developed the original investigation concept; JW, SJ, YF, and AH guided or executed the data analysis; JW created all tables and figures; manuscript preparation was led by JW and AH; and all authors contributed to interpretation of data and writing the manuscript. Only SJ and YF had full access to all study data as these were obtained as part of award funding; the corresponding author assumes final responsibility for the decision to submit for publication. The funder had no role in the study design, data collection, data analysis, data interpretation, or writing of the report.
This article is part of a supplement entitled African American Men’s Health: Research, Practice, and Policy, which is sponsored by the National Institutes of Health.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
No financial disclosures were reported by the authors of this paper.
References
- 1.Carroll DL, Rankin SH, Cooper BA. The effects of a collaborative peer advisor/advanced practice nurse intervention: cardiac rehabilitation participation and rehospitalization in older adults after a cardiac event. J Cardiovasc Nurs. 2007;22(4):313–319. doi: 10.1097/01.JCN.0000278955.44759.73. [DOI] [PubMed] [Google Scholar]
- 2.Young W, Rewa G, Goodman SG, et al. Evaluation of a community-based inner-city disease management program for postmyocardial infarction patients: a randomized controlled trial. CMAJ. 2003;169(9):905–910. [PMC free article] [PubMed] [Google Scholar]
- 3.Girotti ME, Shih T, Revels S, Dimick JB. Racial disparities in readmissions and site of care for major surgery. J Am Coll Surg. 2014;218(3):423–430. doi: 10.1016/j.jamcollsurg.2013.12.004. [DOI] [PubMed] [Google Scholar]
- 4.Singh JA, Lu X, Ibrahim S, Cram P. Trends in and disparities for acute myocardial infarction: an analysis of Medicare claims data from 1992 to 2010. BMC Med. 2014;12:190. doi: 10.1186/s12916-014-0190-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Jha AK, Orav EJ, Li Z, Epstein AM. Concentration and quality of hospitals that care for elderly black patients. Arch Intern Med. 2007;167(11):1177–1182. doi: 10.1001/archinte.167.11.1177. [DOI] [PubMed] [Google Scholar]
- 6.Kahn KL, Pearson ML, Harrison ER, et al. Health care for black and poor hospitalized Medicare patients. JAMA. 1994;271(15):1169–1174. doi: 10.1001/jama.1994.03510390039027. [DOI] [PubMed] [Google Scholar]
- 7.Skinner J, Chandra A, Staiger D, Lee J, McClellan M. Mortality after acute myocardial infarction in hospitals that disproportionately treat black patients. Circulation. 2005;112(17):2634–2641. doi: 10.1161/CIRCULATIONAHA.105.543231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.American Hospital Association. Fast Facts on US Hospitals. 2017 Published 2017. [Google Scholar]
- 9.Bazzoli GJ, Shortell SM, Dubbs N, Chan C, Kralovec P. A taxonomy of health networks and systems: bringing order out of chaos. Health Serv Res. 1999;33(6):1683–1717. [PMC free article] [PubMed] [Google Scholar]
- 10.Carey K. Hospital cost efficiency and system membership. Inquiry. 2003;40(1):25–38. doi: 10.5034/inquiryjrnl_40.1.25. [DOI] [PubMed] [Google Scholar]
- 11.Burns LR, McCullough JS, Wholey DR, Kruse G, Kralovec P, Muller R. Is the system really the solution? Operating costs in hospital systems. Med Care Res Rev. 2015;72(3):247–272. doi: 10.1177/1077558715583789. [DOI] [PubMed] [Google Scholar]
- 12.Ford EW, Short JC. The impact of health system membership on patient safety initiatives. Health Care Manage Rev. 2008;33(1):13–20. doi: 10.1097/01.HMR.0000304496.89684.7f. [DOI] [PubMed] [Google Scholar]
- 13.Madison K. Multihospital system membership and patient treatments, expenditures, and outcomes. Health Serv Res. 2004;39(4 pt 1):749–770. doi: 10.1111/j.1475-6773.2004.00256.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bazzoli GJ, Shortell SM, Ciliberto F, Kralovec PD, Dubbs NL. Tracking the changing provider landscape: implications for health policy and practice. Health Aff (Millwood) 2001;20(6):188–196. doi: 10.1377/hlthaff.20.6.188. [DOI] [PubMed] [Google Scholar]
- 15.Shortell SM, Bazzoli GJ, Dubbs NL, Kralovec P. Classifying health networks and systems: managerial and policy implications. Health Care Manage Rev. 2000;25(4):9–17. doi: 10.1097/00004010-200010000-00002. [DOI] [PubMed] [Google Scholar]
- 16.Coleman EA, Parry C, Chalmers S, Min S. The care transitions intervention: Results of a randomized controlled trial. Arch Intern Med. 2006;166(17):1822–1828. doi: 10.1001/archinte.166.17.1822. [DOI] [PubMed] [Google Scholar]
- 17.Voss R, Gardner R, Baier R, Butterfield K, Lehrman S, Gravenstein S. The care transitions intervention: translating from efficacy to effectiveness. Arch Intern Med. 2011;171(14):1232–1237. doi: 10.1001/archinternmed.2011.278. [DOI] [PubMed] [Google Scholar]
- 18.Robinson JC, Dratler S. Corporate structure and capital strategy at Catholic Healthcare West. Health Aff (Millwood) 2006;25(1):134–147. doi: 10.1377/hlthaff.25.1.134. [DOI] [Google Scholar]
- 19.Blair IV, Havranek EP, Price DW, et al. Assessment of biases against Latinos and African Americans among primary care providers and community members. Am J Public Health. 2013;103(1):92–98. doi: 10.2105/AJPH.2012.300812. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Blair IV, Steiner JF, Hanratty R, et al. An investigation of associations between clinicians’ ethnic or racial bias and hypertension treatment, medication adherence and blood pressure control. J Gen Intern Med. 2014;29(7):987–995. doi: 10.1007/s11606-014-2795-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Chae DH, Nuru-Jeter AM, Adler NE. Implicit racial bias as a moderator of the association between racial discrimination and hypertension: a study of midlife African American men. Psychosom Med. 2012;74(9):961–964. doi: 10.1097/PSY.0b013e3182733665. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Chapman EN, Kaatz A, Carnes M. Physicians and implicit bias: how doctors may unwittingly perpetuate health care disparities. J Gen Intern Med. 2013;28(11):1504–1510. doi: 10.1007/s11606-013-2441-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Green AR, Carney DR, Pallin DJ, et al. Implicit bias among physicians and its prediction of thrombolysis decisions for black and white patients. J Gen Intern Med. 2007;22(9):1231–1238. doi: 10.1007/s11606-007-0258-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Joynt KE, Orav EJ, Jha AK. Thirty-day readmission rates for Medicare beneficiaries by race and site of care. JAMA. 2011;305(7):675–681. doi: 10.1001/jama.2011.123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.HCUP Databases. Healthcare Cost and Utilization Project (HCUP) www.hcup-us.ahrq.gov/sidoverview.jsp. Published 2017. Accessed May 25, 2018.
- 26.American Hospital Association. AHA Data. www.aha.org/research/rc/stat-studies/data-and-directories.shtml. Published 2017. Accessed November 3, 2017.
- 27.Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8–27. doi: 10.1097/00005650-199801000-00004. [DOI] [PubMed] [Google Scholar]
- 28.Moore BJ, White S, Washington R, Coenen N, Elixhauser A. Identifying increased risk of readmission and in-hospital mortality using hospital administrative data: the AHRQ Elixhauser Comorbidity Index. Med Care. 2017;55(7):698–705. doi: 10.1097/MLR.0000000000000735. [DOI] [PubMed] [Google Scholar]
- 29.Moore B, Karaca Z, Henke RM, Wong H. Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Rockville, MD: 2015. Characteristics and quality of inpatient stays at hospitals affiliated with health systems, 2009-2012 Statistical Brief #197. [Google Scholar]
- 30.Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51(6):1173–1182. doi: 10.1037/0022-3514.51.6.1173. [DOI] [PubMed] [Google Scholar]
- 31.Kenny DA. Mediation. http://davidakenny.net/cm/mediate.htm#BK. Published 2016. Accessed May 25, 2018.
- 32.Li Y, Cen X, Cai X, Thirukumaran CP, Zhou J, Glance LG. Medicare advantage associated with more racial disparity than traditional Medicare for hospital readmissions. Health Aff (Millwood) 2017;36(7):1328–1335. doi: 10.1377/hlthaff.2016.1344. [DOI] [PubMed] [Google Scholar]
- 33.Lindenauer PK, Lagu T, Rothberg MB, et al. Income inequality and 30 day outcomes after acute myocardial infarction, heart failure, and pneumonia: retrospective cohort study. BMJ. 2013;346:f521. doi: 10.1136/bmj.f521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Oberg EB, Fitzpatrick AL, Lafferty WE, LoGerfo JP. Secondary prevention of myocardial infarction with nonpharmacologic strategies in a Medicaid cohort. Prev Chronic Dis. 2009;6(2):A52. [PMC free article] [PubMed] [Google Scholar]
- 35.Madison K. Multihospital system membership and patient treatments, expenditures, and outcomes. Health Serv Res. 2004;39(4 Pt 1):749–769. doi: 10.1111/j.1475-6773.2004.00256.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Henke RM, Karaca Z, Moore B, et al. Impact of health system affiliation on hospital resource use intensity and quality of care. Health Serv Res. 2018;53(1):63–86. doi: 10.1111/1475-6773.12631. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Joynt KE, Jha AK. Who has higher readmission rates for heart failure, and why? Implications for efforts to improve care using financial incentives. Circ Cardiovasc Qual Outcomes. 2011;4(1):53–59. doi: 10.1161/CIRCOUTCOMES.110.950964. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Chukmaitov A, Harless DW, Bazzoli GJ, Carretta HJ, Siangphoe U. Delivery system characteristics and their association with quality and costs of care: implications for accountable care organizations. Health Care Manage Rev. 2015;40(2):92–103. doi: 10.1097/HMR.0000000000000014. [DOI] [PubMed] [Google Scholar]
- 39.Baicker K, Chandra A, Skinner JS, Wennberg JE. Who you are and where you live: how race and geography affect the treatment of medicare beneficiaries. Health Aff (Millwood) 2004 doi: 10.1377/hlthaff.var.33. Suppl Variation: VAR33-44. [DOI] [PubMed]