Abstract
Objective
To identify potential socioeconomic disparities in the procedural choice of patients undergoing surgical aortic valve replacement (SAVR) versus transcatheter aortic valve replacement (TAVR) and in readmission outcomes after SAVR or TAVR.
Methods
The Nationwide Readmissions Database was queried to identify a total of 243,691 patients who underwent isolated SAVR and TAVR between January 2016 and December 2018. Patients were stratified according to a tiered socioeconomic status (SES) metric comprising patient factors including education, literacy, housing, employment, insurance status, and neighborhood median income. Multivariable analyses were used to assess the effect of SES on procedural choice and risk-adjusted readmission outcomes.
Results
SAVR (41.4%; 100,833 of 243,619) was performed less frequently than TAVR (58.6%; 142,786 of 243,619). Lower SES was more frequent among patients undergoing SAVR (20.2% [20,379 of 100,833] vs 19.4% [27,791 of 142,786]; P < .001). Along with such variables as small hospital size, drug abuse, arrhythmia, and obesity, lower SES was independently associated with SAVR relative to TAVR (adjusted odds ratio [aOR], 1.17; 95% confidence interval [CI], 1.11 to 1.24). After SAVR, but not after TAVR, lower SES was independently associated with increased readmission at 30 days (aOR, 1.19; 95% CI, 1.07-1.32), 90 days (aOR, 1.27; 95% CI, 1.15-1.41), and 1 year (adjusted hazard ratio, 1.19; 95% CI, 1.11 to 1.28; P < .05 for all).
Conclusions
Our study findings indicate that socioeconomic disparities exist in the procedural choice for patients undergoing AVR. Patients with lower SES had increased odds of undergoing SAVR, as well as increased odds of readmission after SAVR, but not after TAVR, supporting that health inequities exist in the surgical care of socioeconomically disadvantaged patients.
Key Words: aortic valve replacement, socioeconomic status, socioeconomic determinants, surgical aortic valve replacement, transcatheter aortic replacement
Graphical abstract
Risk-adjusted readmission outcomes after SAVR or TAVR according to socioeconomic status.
Central Message.
Socioeconomic disparities exist in procedural choice and outcomes after surgical or transcatheter aortic valve replacement.
Perspective.
Available reports have indicated discrepancies in transcatheter aortic valve replacement (AVR) use in patients of lower socioeconomic status (SES). We identify and discuss the association between SES and the procedural choice of patients undergoing AVR. Additionally, although SES has been recognized as a risk factor for mortality after surgical AVR, its association with readmission remains underreported, and we explored it in this study.
Socioeconomic status (SES) is being increasingly recognized as a risk factor for adverse outcomes after cardiac surgery.1, 2, 3, 4, 5 Previous studies of single variable surrogates or geographic measures for SES have shown an inverse association between SES and mortality after surgical aortic valve replacement (SAVR)1,3,6; however, this association has not been observed after transcatheter aortic valve replacement (TAVR).1,7 Additionally, available reports have indicated discrepancies in TAVR use in patients with lower SES.8, 9, 10 However, the association between SES and the choice of procedure in patients undergoing aortic valve replacement (AVR) is unknown. Additionally, although mortality outcomes have been reported, the association between SES and readmission after SAVR and TAVR remains underexplored. Our metric builds on previously reported single variable definitions for SES, specifically admission payor and neighborhood income quartile. The use of SES-related International Classification of Diseases, Tenth Revision, Clinical Modification codes (problems related to education/literacy, employment/unemployment, occupational risk factors, physical environment, and housing/economic circumstances) is a novel aspect of our methodology. The Centers for Disease Control and Prevention has recognized the valuable contribution that these additional ICD-10-CM codes can provide to enhance our understanding of social determinants of health. These codes go beyond a simple identification of low-income patients; rather, they identify patients who experience barriers to healthcare. Our metric yields results that support differences in clinical outcomes for low SES patients, which is consistent with other models.
The primary goal of this study was to identify potential socioeconomic disparities in the procedural choice of patients undergoing AVR. We hypothesized that lower SES may be associated with undergoing SAVR instead of TAVR. As a secondary goal, we sought to examine the association between SES and readmission after SAVR or TAVR in a nationally representative cohort.
Methods
Data Collection
In this study, we used data from the Nationwide Readmissions Database (NRD). Annually, approximately 35 million discharges are weighted using a clustered, poststratified design that allows for national estimates of outcomes. As the largest all-payor readmissions database, the NRD links patient index admission and readmissions within a calendar year.11 This study was deemed exempt by the Baylor College of Medicine Institutional Review Board because the data were aggregated and deidentified, as recommended by the Health Insurance Portability and Accountability Act.
Patient Cohort
The NRD was queried to identify and stratify adult patients undergoing isolated SAVR or TAVR between January 1, 2016, and December 31, 2018, using the ICD-10-CM Procedure Coding System (PCS) (Table E1, Figure 1). Patients who underwent concomitant cardiac procedures and those with endocarditis were excluded (Table E1, Figure 1).
Figure 1.
Flowchart diagram showing patient identification, exclusion, and stratification. AVR, Aortic valve replacement; SAVR, surgical aortic valve replacement; TAVR, transcatheter aortic valve replacement; SES, socioeconomic status.
SES
Patients were further stratified according to SES by using a composite, tiered metric composed of individual- and neighborhood-level factors. Individual-level factors included admission payor, as well as education, literacy, housing, employment status, physical environment, and economic circumstances, which were identified using SES-related ICD-10-CM codes (Table E1). The neighborhood-level median household income factor is reported as quartiles by the NRD and is established using the median income of a ZIP code for a calendar year. Median income was used in lieu of mean income to better represent the overall income within the ZIP code. Lower SES was defined as one of the following: (1) having Medicaid or self-pay as the primary payor while being in the lowest- and middle-income neighborhood income quartiles; (2) being a Medicare beneficiary in the lowest neighborhood income quartile; or (3) having an SES-related ICD-10-CM code. Higher SES was defined as living in the highest neighborhood income quartile and having private insurance, self-pay, or Medicare as the primary payor. Patients who did not qualify as lower or higher SES were classified as middle SES.
Patient and Hospital Characteristics
Hospital characteristics examined included teaching status and hospital bed size as defined by the NRD.11 Patient characteristics and comorbidities included age, sex, patient rural/urban classification, urgency of intervention, Elixhauser comorbidity index components, and additional cardiovascular risk factors, such as smoking and hyperlipidemia. The Elixhauser comorbidity index is a weighted composite score that categorizes patient comorbidities on the basis of ICD-10-CM diagnosis codes.12
Outcomes
The primary outcome of this study was procedural choice (SAVR or TAVR). Secondary outcomes included in-hospital mortality, index hospitalization length of stay (LOS), discharge disposition, and readmission at 30 days, 90 days, and at 1 year. Readmissions within a calendar year were determined using Kaplan-Meier analysis as described previously.12
Statistical Analysis
Statistical calculations were performed using R version 4.1.1.13 We adjusted for the complex survey design of the NRD in all statistical calculations using the R package survey.14 Less than 1% of data were missing from any category. Missing data were replaced with the median or mode of the overall cohort for continuous and categorical values, respectively. Continuous data were analyzed using a complex survey-adjusted Kruskal-Wallis rank-sum test and are represented as median (interquartile range). Categorical data were analyzed using a chi-square test with Rao and Scott adjustment and are represented as number (percentage). A survey-adjusted log-rank test was used to determine differences in the rate of readmission between groups within a calendar year. A P value <.05 was considered significant. The independent association of lower SES with 30- or 90-day readmission was assessed with multivariable logistic regression analyses. Independent variables included patient and hospital characteristics that differed significantly among groups on univariate analysis and/or with >5% prevalence, including age, sex, urgent intervention, comorbidities, rural versus urban living, teaching status, and hospital bed size. In all logistic regression models for both TAVR and SAVR, age was included as a covariate with 3 cohorts (age <65 years, 65-79 years, and >80 years). Area under the curve values were used to guide the selection of clinically relevant variables.14 The variable inflation factor was used to assess and remove colinear variables. Schoenfeld residuals were used to assess covariates in the Cox proportional hazard model. Regression results are represented as an adjusted odds ratio (aOR) or an adjusted hazard ratio (aHR) with 95% confidence intervals (CIs) for logistic or Cox proportional hazards, respectively. P values were derived from a survey-adjusted Wald test. Figures were created using Prism 9.5.1 (GraphPad Software) or R version 4.1.1 (R Foundation for Statistical Computing).
Results
Patient Characteristics According to Procedure Type
Over the 3-year study period, a total of 243,619 patients underwent isolated AVR, including 100,833 (41.4%) with SAVR and 142,786 (58.6%) with TAVR (Figure 1). Lower SES was more frequent in patients undergoing SAVR than in patients undergoing TAVR (20.2% vs 19.4%; P < .001) (Table 1). Compared with patients undergoing TAVR, patients undergoing SAVR were younger, less likely to be female and to be treated at large or teaching hospitals, and more likely to reside in rural areas than (P < .05 for all) (Table 1). The comorbidity burden was lower in the patients undergoing SAVR (Elixhauser score P < .001) (Table 1). Although the rates of congestive heart failure, peripheral vascular disease, hypertension, diabetes, chronic obstructive pulmonary disease, renal failure, hyperlipidemia, and pulmonary circulation disorders were higher among the TAVR group, the SAVR group was more likely to have arrhythmias, obesity, illicit drug use, and a history of smoking (P < .001 for all) (Table 1).
Table 1.
Patient and hospital characteristics according to aortic valve replacement procedure type
| Variable | Overall (N = 243,619) | SAVR (N = 100,833; 41.4%) | TAVR (N = 142,786; 58.6%) | P value∗ |
|---|---|---|---|---|
| Age, y, median (IQR) | 76 (67-83) | 68 (60-74) | 81 (75-86) | <.001 |
| Age group, n (%) | <.001 | |||
| <65 y | 46,223 (18.9) | 39,025 (38.7) | 7198 (5.0) | |
| 65-79 y | 103,930 (42.7) | 52,458 (52.0) | 51,472 (36.0) | |
| ≥80 y | 93,466 (38.4) | 9350 (9.3) | 84,116 (58.9) | |
| Female sex, n (%) | 103,893 (42.6) | 38,973 (38.7) | 64,920 (45.5) | <.001 |
| Urgent intervention, n (%) | 44,704 (18.3) | 17,585 (17.4) | 27,119 (19.0) | .02 |
| SES, n (%) | <.001 | |||
| Lower SES | 48,170 (19.8) | 20,379 (20.2) | 27,791 (19.4) | |
| Middle SES | 139,740 (57.3) | 59,134 (58.7) | 80,606 (56.5) | |
| Higher SES | 55,709 (22.9) | 21,320 (21.1) | 34,389 (24.1) | |
| Income quartile, n (%)† | <.001 | |||
| 1 | 50,753 (20.8) | 22,249 (22.1) | 28,504 (20.0) | |
| 2 | 69,877 (28.7) | 29,319 (29.1) | 40,558 (28.4) | |
| 3 | 66,264 (27.2) | 27,218 (27.0) | 39,046 (27.3) | |
| 4 | 56,725 (23.3) | 22,047 (21.9) | 34,678 (24.3) | |
| Primary payor, n (%) | <.001 | |||
| Medicaid | 7190 (3.0) | 5704 (5.7) | 1486 (1.0) | |
| Medicare | 190,311 (78.2) | 60,244 (59.8) | 130,066 (91.2) | |
| Other | 4609 (1.9) | 2137 (2.1) | 2472 (1.7) | |
| Private insurance | 39,500 (16.2) | 31,394 (31.2) | 8106 (5.7) | |
| Self-pay | 1741 (0.7) | 1260 (1.3) | 481 (0.3) | |
| Hospital bed size, n (%) | <.001 | |||
| Large | 177,003 (72.7) | 70,050 (69.5) | 106,954 (74.9) | |
| Medium | 52,073 (21.4) | 22,381 (22.2) | 29,692 (20.8) | |
| Small | 14,543 (6.0) | 8402 (8.3) | 6140 (4.3) | |
| Teaching hospital, n (%) | 210,193 (86.3) | 84,472 (83.8) | 125,722 (88.0) | <.001 |
| Rural patient, n (%) | 20,220 (8.3) | 8830 (8.8) | 11,390 (8.0) | .03 |
| Elixhauser score, median (IQR) | 11 (5-19) | 10 (2-20) | 11 (6-19) | <.001 |
| Congestive heart failure, n (%) | 147,420 (60.5) | 39,910 (39.6) | 107,511 (75.3) | <.001 |
| Arrhythmia, n (%) | 140,140 (57.5) | 58,839 (58.4) | 81,301 (56.9) | <.001 |
| Peripheral vascular disease, n (%) | 51,283 (21.1) | 18,183 (18.0) | 33,100 (23.2) | <.001 |
| Hypertension, n (%) | 207,282 (85.1) | 79,407 (78.8) | 127,874 (89.6) | <.001 |
| Chronic obstructive pulmonary disease, n (%) | 62,023 (25.5) | 21,081 (20.9) | 40,942 (28.7) | <.001 |
| Diabetes mellitus, n (%) | 82,472 (33.9) | 28,347 (28.1) | 54,126 (37.9) | <.001 |
| Renal failure, n (%) | 67,728 (27.8) | 16,086 (16.0) | 51,642 (36.2) | <.001 |
| Obesity, n (%) | 55,177 (22.6) | 27,350 (27.1) | 27,827 (19.5) | <.001 |
| Pulmonary circulation disorder, n (%) | 40,928 (16.8) | 13,726 (13.6) | 27,201 (19.1) | <.001 |
| Hyperlipidemia, n (%) | 165,227 (67.8) | 62,325 (61.8) | 102,902 (72.1) | <.001 |
| Drug abuse, n (%) | 2482 (1.0) | 1856 (1.8) | 626 (0.4) | <.001 |
| Smoking, n (%) | 97,973 (40.2) | 41,724 (41.4) | 56,249 (39.4) | <.001 |
SAVR, Surgical aortic valve replacement; TAVR, transcatheter aortic valve replacement; IQR, interquartile range; SES, socioeconomic status.
Comparison between SES level determined using the Kruskal-Wallis rank-sum test for complex survey samples or the chi-square test with Rao and Scott second-order correction.
Based on patient ZIP code. Values are presented as median (interquartile range) for continuous variables or as number (percentage) for categorical variables. The denominator is indicated at the top of each column unless noted otherwise.
Multivariable Model of Procedural Choice
Based on the result of univariate analysis shown in Table 1, we created a multivariable model with procedural choice as an outcome (TAVR vs SAVR) (Figure 2). The full model is provided in Table E2. Patients with lower SES had a 17% higher risk-adjusted odds of undergoing SAVR relative to TAVR (aOR, 1.17; 95% CI, 1.11-1.24) (Figure 2). Patients age ≥80 years had the greatest odds of receiving TAVR (aOR, 0.03; 95% CI, 0.03-0.04). Additionally, females had risk-adjusted odds in favor of receiving TAVR (aOR, 0.83; 95% CI, 0.8-0.86). Patients requiring urgent intervention (aOR, 1.15; 95% CI, 1.1-1.2), patients with arrhythmia (aOR, 1.86; 95% CI, 1.79-1.93) or obesity (aOR, 1.2; 95% CI, 1.15-1.26), patients with a history of illicit drug use (aOR, 1.94; 95% CI, 1.65-2.28), and patients treated at small hospitals (aOR, 2.19; 95% CI, 2.04-2.36) had odds in favor of receiving SAVR relative to TAVR. Patient rurality (aOR, 0.99; 95% CI, 0.93-1.07) and smoking (aOR, 0.99; 95% CI, 0.95-1.02) were not associated with significant odds of receiving SAVR or TAVR (Table E2).
Figure 2.
Multivariable model of procedural choice (SAVR vs TAVR). Procedural choice as a binary outcome, with TAVR = 0 and SAVR = 1. aIndexed to a higher SES. TAVR, Transcatheter aortic valve replacement; SAVR, surgical aortic valve replacement; SES, socioeconomic status.
Patient Characteristics According to SES in SAVR and TAVR Groups
Patients were further stratified according to SES (Figure 1), and their characteristics were analyzed in the SAVR and TAVR groups. Patients with lower SES undergoing SAVR or TAVR were younger, more likely to be female and to reside in rural areas, and had a higher comorbidity burden compared with patients with a higher SES (P < .001 for all) (Table 2). Patients with lower SES undergoing SAVR or TAVR also were less likely to be treated at a teaching hospital (P < .001) (Table 2). Unique to the SAVR group, patients with lower SES were more likely to require urgent intervention, to have a history of smoking, and to use illicit drugs (P < .001 for all) (Table 2).
Table 2.
Patient and hospital characteristics according to SES
| Variable | SAVR (N = 100,833) |
TAVR (N = 142,786) |
||||||
|---|---|---|---|---|---|---|---|---|
| Higher SES (N = 21,320; 21.1%) | Middle SES (N = 59,134; 58.7%) | Lower SES (N = 20,379; 20.2%) | P value∗ | Higher SES (N = 34,389; 24.1%) | Middle SES (N = 80,606; 56.5%) | Lower SES (N = 27,791; 19.4%) | P value∗ | |
| Age, y, median (IQR) | 69 (61-75) | 67 (60-74) | 68 (58-74) | <.001 | 82 (76-87) | 81 (75-86) | 80 (74-86) | <.001 |
| Age group, n (%) | <.001 | <.001 | ||||||
| <65 y | 7311 (34.3) | 23,591 (39.9) | 8124 (39.9) | 1174 (3.4) | 4041 (5.0) | 1982 (7.1) | ||
| 65-79 y | 11,783 (55.3) | 30,167 (51.0) | 10,508 (51.6) | 11,201 (32.6) | 29,492 (36.6) | 10,779 (38.8) | ||
| ≥80 y | 2227 (10.4) | 5376 (9.1) | 1748 (8.6) | 22,014 (64.0) | 47,073 (58.4) | 15,029 (54.1) | ||
| Female sex, n (%) | 7707 (36.2) | 22,517 (38.1) | 8749 (42.9) | <.001 | 15,161 (44.1) | 36,289 (45.0) | 13,470 (48.5) | <.001 |
| Urgent intervention, n (%) | 3315 (15.6) | 9287 (15.7) | 4982 (24.4) | <.001 | 6557 (19.1) | 15,016 (18.6) | 5546 (20.0) | .26 |
| Income quartile, n (%)† | <.001 | <.001 | ||||||
| 1 | <11 (<1.1)‡ | 6452 (10.9) | 15,796 (77.5) | <11 (<1.1)‡ | 2155 (2.7) | 26,349 (94.8) | ||
| 2 | <11 (<1.1)‡ | 27,033 (45.7) | 2286 (11.2) | <11 (<1.1)‡ | 39,983 (49.6) | 576 (2.1) | ||
| 3 | <11 (<1.1)‡ | 25,648 (43.4) | 1570 (7.7) | <11 (<1.1)‡ | 38,469 (47.7) | 577 (2.1) | ||
| 4 | 21,320 (100.0) | <11 (<1.1)‡ | 727 (3.6) | 34,389 (100.0) | <11 (<1.1)‡ | 288 (1.0) | ||
| Primary payor, n (%) | <.001 | <.001 | ||||||
| Medicaid | <11 (<1.1)‡ | <11 (<1.1)‡ | 5704 (28.0) | <11 (<1.1)‡ | <11 (<1.1)‡ | 1486 (5.3) | ||
| Medicare | 12,940 (60.7) | 33,820 (57.3) | 13,485 (66.2) | 31,749 (92.4) | 72,384 (90.0) | 25,933 (93.3) | ||
| Other | 330 (1.5) | 21 (0.1) | 1786 (3.1) | 324 (0.9) | 2144 (2.6) | <11 (<1.1)‡ | ||
| Private insurance | 7889 (37.0) | 23,443 (39.7) | 61 (0.3) | 2174 (6.3) | 5922 (7.4) | <11 (<1.1)‡ | ||
| Self-pay | 152 (0.7) | <11 (<1.1)‡ | 1108 (5.4) | 125 (0.4) | <11 (<1.1)‡ | 356 (1.3) | ||
| Hospital bed size, n (%) | .002 | .02 | ||||||
| Large | 14,639 (68.7) | 14,541 (71.4) | 40,869 (69.1) | 24,967 (72.6) | 60,703 (75.3) | 21,283 (76.6) | ||
| Medium | 5337 (25.0) | 4310 (21.2) | 12,733 (21.5) | 8394 (24.4) | 15,960 (19.8) | 5338 (19.2) | ||
| Small | 1344 (6.3) | 1528 (7.5) | 5531 (9.4) | 1028 (3.0) | 3943 (4.9) | 1169 (4.2) | ||
| Teaching hospital, n (%) | 19,142 (89.8) | 48,740 (82.4) | 16,590 (81.4) | <.001 | 31,535 (91.7) | 70,242 (87.1) | 23,945 (86.2) | <.001 |
| Rural patient, n (%) | 95 (0.5) | 5523 (9.3) | 3211 (15.8) | <.001 | 117 (0.3) | 6372 (7.9) | 4901 (17.6) | <.001 |
| Elixhauser score, median (IQR) | 11 (2-20) | 10 (2-19) | 11 (3-21) | <.001 | 11 (6-19) | 11 (6-19) | 11 (7-19) | .007 |
| CHF, n (%) | 7622 (35.7) | 22,724 (38.4) | 9564 (46.9) | <.001 | 24,768 (72.0) | 61,020 (75.7) | 21,723 (78.2) | <.001 |
| Arrhythmia, n (%) | 12,899 (60.5) | 34,178 (57.8) | 11,762 (57.7) | <.001 | 20,246 (58.9) | 45,782 (56.8) | 15,274 (55.0) | <.001 |
| PVD, n (%) | 4102 (19.2) | 10,409 (17.6) | 3672 (18.0) | <.001 | 7724 (22.5) | 18,741 (23.2) | 6635 (23.9) | .18 |
| Hypertension, n (%) | 16,431 (77.1) | 46,585 (78.8) | 16,391 (80.4) | <.001 | 30,739 (89.4) | 72,172 (89.5) | 24,963 (89.8) | .65 |
| COPD, n (%) | 3686 (17.3) | 12,034 (20.4) | 5360 (26.3) | <.001 | 8867 (25.8) | 23,118 (28.7) | 8957 (32.2) | <.001 |
| Diabetes mellitus, n (%) | 5225 (24.5) | 16,844 (28.5) | 6278 (30.8) | <.001 | 11,649 (33.9) | 30,999 (38.5) | 11,478 (41.3) | <.001 |
| Renal failure, n (%) | 3159 (14.8) | 9076 (15.3) | 3851 (18.9) | <.001 | 11,653 (33.9) | 29,537 (36.6) | 10,452 (37.6) | <.001 |
| Obesity, n (%) | 5044 (23.7) | 16,786 (28.4) | 5520 (27.1) | <.001 | 5452 (15.9) | 16,401 (20.3) | 5975 (21.5) | <.001 |
| Pulmonary circulation disorder, n (%) | 2587 (12.1) | 7766 (13.1) | 3373 (16.6) | <.001 | 6342 (18.4) | 15,290 (19.0) | 5569 (20.0) | <.04 |
| Hyperlipidemia, n (%) | 11,758 (55.2) | 36,797 (62.2) | 13,859 (68) | <.001 | 25,424 (73.9) | 57,893 (71.8) | 19,585 (70.5) | <.001 |
| Drug abuse, n (%) | 213 (1.0) | 820 (1.4) | 824 (4.0) | <.001 | 123 (0.4) | 325 (0.4) | 178 (0.6) | <.001 |
| Smoking, n (%) | 8159 (38.3) | 24,289 (41.1) | 9276 (45.5) | <.001 | 13,588 (39.52) | 31,612 (39.2) | 11,049 (39.8) | .69 |
| Mechanical valve, n (%) | 2770 (13) | 10,408 (17.6) | 3843 (18.9) | <.001 | NA | NA | NA | NA |
SAVR, Surgical aortic valve replacement; TAVR, transcatheter aortic valve replacement; SES, socioeconomic status; IQR, interquartile range; CHF, congestive heart failure; PVD, peripheral vascular disease; COPD, chronic obstructive pulmonary disease; NA, not applicable.
Comparison between SES level determined using the Kruskal-Wallis rank-sum test for complex survey samples or the χ2 test with Rao and Scott second-order correction.
Based on patient ZIP code.
Observations with cell count <11 reported as <11, according to Healthcare Cost and Utilization Project regulations. Values are presented as median (IQR) for continuous variables or as number (percentage) for categorical variables. The denominator is at the top of the column unless noted otherwise.
Observed Outcomes According to SES After SAVR or TAVR
After SAVR, patients with lower SES had higher rates of in-hospital mortality and longer index hospitalization stays and were less likely to be discharged with home health care compared to patients with higher SES (P < .001 for all) (Table 3). In addition, patients in the SAVR group with lower SES had the highest readmission rates at 30 days and 90 days among all SES levels (Figure 3, Table 3); this persisted at 1 year. Although higher risk-adjusted readmission was observed at 30 days (aOR, 1.19; 95% CI, 1.07-1.32), 90 days (aOR, 1.27; 95% CI, 1.15-1.41), and 1 year (aHR, 1.19; 95% CI, 1.11-1.28) in the SAVR group, the risk of readmission was not independently associated with SES in the TAVR group at 30 days (aOR, 1.01; 95% CI, 0.92-1.09), 90 days (aOR, 0.95; 95% CI, 0.88-1.03), or 1 year (aHR, 1.01; 95% CI, 0.96-1.07) (Figure 4). Full regression models for 30-day, 90-day, and 1-year readmission after SAVR and TAVR are shown in Table E3, Table E4, Table E5, Table E6, Table E7, Table E8.
Table 3.
Observed patient outcomes according to SES
| Outcome | SAVR (N = 100,833) |
TAVR (N = 142,786) |
||||||
|---|---|---|---|---|---|---|---|---|
| Higher SES (N = 19,347) | Middle SES (N = 53,107) | Lower SES (N = 18,263) | P value∗ | Higher SES (N = 30,905) | Middle SES (N = 72,354) | Lower SES (N = 24,960) | P value∗ | |
| In-hospital mortality, n/N (%) | 359/21,320 (1.7) | 1316/59,134 (2.2) | 609/20,379 (3.0) | <.001 | 437/34,389 (1.3) | 1102/80,606 (1.4) | 410/27,791 (1.5) | .32 |
| Index hospitalization LOS, d, median (IQR); mean ± SD | 6 (5-8); 7.9 ± 6.9 | 6 (5-9); 8.1 ± 7.0 | 7 (5-11); 9.8 ± 9.2 | <.001 | 2 (2-5); 4.4 ± 6.0 | 2 (2-4); 4.3 ± 5.7 | 2 (2-5); 4.7 ± 6.4 | .005 |
| Disposition, n (%) | <.001 | .003 | ||||||
| Home-health care | 9748 (50.4) | 23,426 (44.1) | 7709 (42.2) | 8607 (27.9) | 15,807 (21.8) | 5418 (21.7) | ||
| Routine | 6788 (35.1) | 21,339 (40.2) | 6938 (38.0) | 18,616 (60.2) | 47,452 (65.6) | 16,365 (65.6) | ||
| SNF or ICF | 2739 (14.2) | 8126 (15.3) | 3496 (19.1) | 3567 (11.5) | 8855 (12.2) | 3096 (12.4) | ||
| Other | 73 (0.3) | 120 (0.6) | 215 (0.4) | 115 (0.4) | 241 (0.3) | 81 (0.4) | <.001 | |
| 30-d readmission, n (%) | 2161 (11.2) | 5970 (11.2) | 2570 (14.1) | <.001 | 3895 (12.6) | 8939 (12.4) | 3268 (13.1) | .14 |
| 90-d readmission, n/N (%) | 2430/15,931 (15.3) | 7174/43,849 (16.4) | 3045/15,030 (20.3) | <.001 | 5364/24,764 (21.7) | 12,622/57,916 (21.8) | 4446/20,067 (22.2) | .73 |
SAVR, Surgical aortic valve replacement; TAVR, transcatheter aortic valve replacement; SES, socioeconomic status; LOS, length of stay; IQR, interquartile range; SD, standard deviation; SNF, skilled nursing facility; ICF, intermediate care facilities.
Comparison between SES level determined using the Kruskal-Wallis rank-sum test for complex survey samples or the χ2 test with Rao and Scott second-order correction. Values are presented as median (IQR) for continuous variables or as number (percentage) for categorical variables. The denominator is listed at the top of the column unless noted otherwise.
Figure 3.
Observed readmission outcomes after SAVR (left) and TAVR (right) according to SES. SES, Socioeconomic status; SAVR, surgical aortic valve replacement; TAVR, transcatheter aortic valve replacement. ∗P < .05.
Figure 4.
Risk-adjusted readmission outcomes after SAVR and TAVR according to SES. Readmission as a binary outcome with readmitted = 1 and not readmitted = 0. Adjusted odds ratio listed for 30-day readmission and 90-day readmission. The adjusted hazard ratio is listed for 1-year readmission. SES, Socioeconomic status; SAVR, surgical aortic valve replacement; TAVR, transcatheter aortic valve replacement.
Discussion
Both neighborhood- and individual-level factors have been shown to contribute to existent socioeconomic disparities in cardiovascular outcomes.15 Although single-variable surrogates are often reported, they do not capture the multiplexity of a patient's SES.15 In this study, we used a composite tiered SES metric composed of individual- and neighborhood-level factors to reflect a patient's SES level. We showed in a national cohort of patients that lower SES was independently associated with receiving SAVR relative to TAVR. In addition, our data support that lower SES was associated with an increased risk-adjusted odds of readmission after SAVR but not after TAVR (Figure 5). To our knowledge, this study is the first to assess the interaction between SES and procedural choice for patients undergoing AVR. Furthermore, it is the largest, most comprehensive investigation of readmission outcomes after SAVR and TAVR according to patient SES reported to date.
Figure 5.
Socioeconomic disparities in procedural choice and outcomes after aortic valve replacement. Data from the Nationwide Readmissions Database were collected and analyzed for patients who underwent isolated aortic valve replacement during a 3-year study period. Lower SES was independently associated with undergoing SAVR relative to TAVR. Additionally, lower SES was associated with increased risk-adjusted odds of readmission after SAVR, but not after TAVR. SAVR, Surgical aortic valve replacement; SES, socioeconomic status; TAVR, transcatheter aortic valve replacement; AVR, aortic valve replacement. ∗P < .05.
Socioeconomic Disparities in Procedural Choice of Patients Undergoing AVR
We found that patients with lower SES have a 17% increased risk-adjusted odds of undergoing SAVR relative to TAVR compared to patients with higher SES. The adjustment accounted for patient demographic factors, such as age, sex, comorbidities, and rurality, and hospital factors, including teaching status and bed size. One potential explanation for the choice of SAVR over TAVR is that patients with lower SES may present to hospitals without TAVR capabilities. Furthermore, disparities may exist in the TAVR referral process for patients with lower SES. A wide spectrum of social determinants of health may prevent patients who could benefit from TAVR from ultimately undergoing the procedure. These may include inadequate social support and health literacy, inability to afford care, and inability or willingness to utilize public transportation.15 With this in mind, we found that patient rurality was not associated with an increased odds of undergoing SAVR relative to TAVR. This suggests that socioeconomic barriers, but not patient geographical location, contribute to disparities in access to TAVR. In support of this, Nathan and colleagues8 reported that socioeconomic disparities in the use of TAVR persisted despite geographic proximity to a hospital with a TAVR program. Contradictory to this, a report from Damluji and colleagues16 demonstrated a rural-urban divide in the rates of TAVR use; however, these results were not adjusted for SES. Given that our study shows that patients with lower SES are much more likely to reside in rural locations, those latter results may be confounded by the association between lower SES and patient rurality.
Additionally, despite the increasing volume of TAVR procedures and the growing number of sites with TAVR capabilities, we showed that patients treated at smaller hospitals had the highest adjusted odds of undergoing SAVR, and patients treated at teaching hospitals had higher adjusted odds of undergoing TAVR. This is supported by a recent report indicating that new TAVR programs are more likely to be developed at hospitals with a large bed size (aOR, 2.73; 95% CI, 1.53-4.89) and teaching hospitals (aOR, 2.47; 95% CI, 1.93-3.15).9 We also made the universal observation that regardless of AVR type, patients with lower SES were more likely than patients with higher SES to be treated at hospitals of smaller bed size.
Finally, another potential explanation for procedural choice favoring SAVR over TAVR in patients with lower SES is that different SES levels may coincide with different disease pathologies or severities. This is supported by differing patient baseline characteristics among the SES tiers. SAVR may be more appropriate than TAVR for patients with lower SES.
Socioeconomic Disparities in Outcomes After AVR
Concordant with prior reports, we show an inverse relationship between SES and in-hospital mortality after SAVR.1,3,6 Likewise, in agreement with previous studies, no SES-based disparities in mortality were identified among patients undergoing TAVR.1,7 We also found an association between lower SES and longer hospital LOS after SAVR and TAVR, which is supported by previous studies.1,7
In the present study, patients with lower SES were found to have higher rates of readmission after SAVR but not after TAVR at all examined time points. After adjusting for patient demographics, such as age, sex, comorbidities, rurality, valve type (for SAVR), and hospital factors, including teaching status and bed size, this trend was confirmed with greater odds of readmission for patients with lower SES after SAVR but not after TAVR at 30 days, 90 days, and 1 year. These data are strongly suggestive of SAVR-specific socioeconomic disparities. Given the more complex postoperative recovery and follow-up for patients who undergo SAVR versus TAVR, intuitively these disparities are more significant after SAVR. Previous studies have identified lower SES as a strong determinant of poor health literacy.17,18 In turn, reduced health literacy has been associated with reduced medical compliance.17,19 This may partially explain the greater odds of readmission in patients of lower SES after SAVR. Further investment in SES-tailored patient education after SAVR may be warranted to reduce readmissions in this population.20,21 We can only speculate whether TAVR providers identify patients with lower SES through effective social screening and implement more appropriate SES-tailored care plans to prevent readmission. This warrants further analysis to understand the root cause of these differences.
Although we adjusted for valve type in our multivariate logistic regression analysis, the increased use of mechanical valves in patients with lower SES in the SAVR group may contribute to the increased readmission rates in these patients. The greater use of mechanical valves in patients with lower SES in the SAVR group is likely explained by their younger age. However, Dalén and colleagues22 reported that patients with lower SES were more likely to experience bleeding complications after mechanical valve implantation.
Proposed solutions to ameliorate socioeconomic disparities in cardiac surgery outcomes have included the implementation of universal health coverage, although this is not supported by studies performed in Europe.23 We showed that home health care is more likely to be necessary for patients who undergo SAVR compared with patients who undergo TAVR. Furthermore, we observed fewer discharges with home health care in patients with lower SES compared to those with higher SES after SAVR. Home health care has been associated with a reduced readmission rate after cardiac surgery.24 As a result, patients with lower SES might not receive the benefits from discharge with home health after SAVR.
Implementing universal health coverage in turn could increase the rate of discharge with home health care in patients with lower SES, theoretically resulting in decreased readmissions. However, studies from Europe have shown that socioeconomic disparities persist despite study populations that benefit from universal health coverage.22,25 This supports the concept that socioeconomic disparities in outcomes after cardiac surgery are more complex than socioeconomic barriers in health care access. Therefore, it is important that we continue to identify socioeconomic disparities and their causes in cardiac surgery outcomes so that we may develop, disseminate, and implement SES-tailored strategies to close the socioeconomic gap. Minimal progress has been made on addressing these disparities. A variety of interventions are likely necessary to reduce inequities in care. In addition to raising awareness through further study, using a SES metric such as the one proposed in this study may help identify patients of lower SES, allowing for resource allocation and SES-tailored postdischarge support, such as home-health care, cardiac rehab, and SES-tailored discharge protocols. In this context, although patient-level interventions and care customizations may help increase equity in the care of SAVR patients, the multiplexity of cardiovascular health disparities reaches a macro level. Neighborhood socioeconomic characteristics have been associated with an increased risk of cardiovascular disease after adjusting for individual SES characteristics. Outside the scope of this analysis, primary, secondary, and tertiary preventive measures need to be taken at a community level to manage cardiovascular risk factors, and social determinants of health need to be improved within disparate communities.
The results of this study must be interpreted in light of several limitations, including those inherent to a retrospective analysis of a large administrative database. Limitations of using the NRD include dependence on the ICD-10-CM–based derivations of comorbidities. The inclusion of ICD codes in the SES metric and their potential underutilization introduces the possibility of a misclassification error for SES. There is no perfect model for capturing the complexity of SES; ideally, SES would be measured as a continuous variable. However, we believe that the tiered SES metric used in this study, which is a composite of numerous variables, provides a robust method of reporting SES. Another limitation of the NRD is that it lacks granular admission data, such as ejection fraction, disease etiology, cardiac anatomy, lab values, medications, intraoperative details, race/ethnicity, and hospital TAVR/SAVR availability, reducing the ability to capture key clinical and demographic intersectionality data elements. Therefore, as is common with large database studies, we were unable to account for individual hospital practices. Because the NRD does not track patients across calendar years, index admissions in December were excluded from the 30-day readmission analysis, and index admissions from October to December were excluded from the 90-day readmission analysis.12,26 Overall, the NRD generates standardized, robust calculations reflecting national-level in-hospital outcomes and readmissions.
Conclusions
Our study findings suggest that socioeconomic disparities exist in the procedural choice for patients undergoing AVR, with patients of lower SES having increased odds of undergoing SAVR over TAVR. Furthermore, the increased odds of readmission after SAVR, but not after TAVR, in patients with lower SES support the concept that health inequities exist in the surgical care of socioeconomically disadvantaged patients undergoing AVR. Therefore, it is important to investigate the causality of these disparities in future studies to enable the development, dissemination, and implementation of SES-tailored strategies to close the socioeconomic gap in cardiac surgery.
Conflict of Interest Statement
J.S.C. reports participation in clinical studies with and/or consulting for Terumo Aortic, Medtronic, W. L. Gore & Associates, CytoSorbents, Edwards Lifesciences, and Abbott Laboratories and royalties and grant support from Terumo Aortic. M.R.M. serves on an advisory board for Medtronic. S.C. has served on advisory boards for Edwards Lifesciences, La Jolla Pharmaceutical, Eagle Pharmaceuticals, and Baxter Pharmaceuticals. All other authors reported no conflicts of interest.
The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.
Acknowledgments
We thank Nicole Stancel, PhD, ELS(D), of the Department of Scientific Publications at The Texas Heart Institute for her editorial contributions.
Appendix E1
Table E1.
ICD-10-CM/PCS
| ICD-10-CM/PCS codes | Associated procedure/Diagnosis |
|---|---|
| Inclusion codes | |
| ICD-10-PCS 02RF0 | Replacement, aortic valve, open approach |
| ICD-10-PCS 02RF3 | Replacement, aortic valve, percutaneous approach |
| Exclusion codes | |
| ICD-10-CM I20 | Angina pectoris |
| ICD-10-CM I21 | Acute myocardial infarction |
| ICD-10-CM I38 | Endocarditis, valve unspecified |
| ICD-10-CM I33.0 | Acute and subacute infective endocarditis |
| ICD-10-CM I33.9 | Acute and subacute endocarditis, unspecified |
| ICD-10-CM B37.6 | Candidal endocarditis |
| ICD-10-PCS 02QF | Repair, aortic valve |
| ICD-10-PCS 02QG | Repair, mitral valve |
| ICD-10-PCS 02QH | Repair, pulmonary valve |
| ICD-10-PCS 02QJ | Repair, tricuspid valve |
| ICD-10-PCS 02RG | Replacement, mitral valve |
| ICD-10-PCS 02RH | Replacement, pulmonary valve |
| ICD-10-PCS 02RJ | Replacement, tricuspid valve |
| ICD-10-CM I25.4 | Coronary artery aneurysm and dissection |
| ICD-10-PCS 0210 | Bypass, coronary artery, one artery |
| ICD-10-PCS 0211 | Bypass, coronary artery, two arteries |
| ICD-10-PCS 0212 | Bypass, coronary artery, three arteries |
| ICD-10-PCS 0213 | Bypass, coronary artery, four or more arteries |
| ICD-10-PCS 02RX | Replacement of thoracic aorta, ascending/arch |
| ICD-10-PCS 02RW | Replacement of thoracic aorta, descending |
| ICD-10-PCS 02QX | Repair of thoracic aorta, ascending/arch |
| ICD-10-PCS 02QW | Repair of thoracic aorta, descending |
| ICD-10-PCS 04R0 | Replacement of abdominal aorta |
| ICD-10-PCS 04Q0 | Repair of abdominal aorta |
| SES-related codes | |
| ICD-10-CM Z55 | Problems related to education and literacy |
| ICD-10-CM Z56 | Problems related to employment and unemployment |
| ICD-10-CM Z57 | Occupational exposure to risk factors |
| ICD-10-CM Z58 | Problems related to physical environment |
| ICD-10-CM Z59 | Problems related to housing and economic circumstances |
ICD-10-CM/PCS, International Classification of Diseases, Tenth Revision, Clinical Modification/Procedure Coding System; SES, socioeconomic status.
Table E2.
Full regression model for procedure choice
| Risk factor | aOR | 95% CI | P value∗ |
|---|---|---|---|
| Age 65-79 y† | 0.360 | 0.345-0.376 | <.001 |
| Age >80 y† | 0.034 | 0.032-0.036 | <.001 |
| Female sex | 0.829 | 0.800-0.860 | <.001 |
| Urgent intervention | 1.147 | 1.095-1.201 | <.001 |
| Lower SES‡ | 1.174 | 1.112-1.239 | <.001 |
| Middle SES‡ | 1.123 | 1.076-1.172 | <.001 |
| Congestive heart failure | 0.237 | 0.228-0.246 | <.001 |
| Arrhythmia | 1.859 | 1.793-1.927 | <.001 |
| Pulmonary circulation disorder | 0.980 | 0.931-1.026 | .36 |
| Peripheral vascular disease | 0.810 | 0.776-0.846 | <.001 |
| Hypertension | 0.786 | 0.748-0.827 | <.001 |
| Chronic obstructive pulmonary disease | 0.616 | 0.590-0.642 | <.001 |
| Diabetes mellitus | 0.604 | 0.581-0.628 | <.001 |
| Renal failure | 0.517 | 0.495-0.540 | <.001 |
| Obesity | 1.204 | 1.154-1.255 | <.001 |
| Drug abuse | 1.938 | 1.648-2.279 | <.001 |
| Smoking | 0.990 | 0.950-1.023 | .46 |
| Hyperlipidemia | 0.835 | 0.803-0.868 | <.001 |
| Rural patient | 0.990 | 0.927-1.066 | .87 |
| Teaching hospital | 0.641 | 0.611-0.672 | <.001 |
| Hospital bed size | |||
| Medium§ | 1.227 | 1.177-1.279 | <.001 |
| Small§ | 2.190 | 2.035-2.358 | <.001 |
aOR, Adjusted odds ratio; CI, confidence interval; SES, socioeconomic status.
Wald test.
Indexed to <65 years.
Indexed to higher SES.
Indexed to larger bed size. Procedural choice as binary outcome with transthoracic surgical aortic valve replacement (TAVR) = 0 and surgical aortic valve replacement (SAVR) = 1.
Table E3.
Full regression model for 30-day readmission after SAVR
| Risk factor | aOR | 95% CI | P value∗ |
|---|---|---|---|
| Age 65-79 y† | 1.083 | 0.999-1.173 | .052 |
| Age > 80 y† | 1.502 | 1.332-1.694 | <.001 |
| Female sex | 1.120 | 1.043-1.203 | .002 |
| Urgent intervention | 1.063 | 0.970-1.169 | .20 |
| Lower SES‡ | 1.186 | 1.068-1.317 | .001 |
| Middle SES‡ | 0.980 | 0.900-1.070 | .66 |
| Congestive heart failure | 1.191 | 1.106-1.283 | <.001 |
| Arrhythmia | 1.340 | 1.243-1.445 | <.001 |
| Pulmonary circulation disorder | 1.212 | 1.103-1.332 | <.001 |
| Peripheral vascular disease | 1.076 | 0.990-1.174 | .10 |
| Hypertension | 1.120 | 1.021-1.228 | .02 |
| Chronic obstructive pulmonary disease | 1.233 | 1.136-1.339 | <.001 |
| Diabetes mellitus | 1.192 | 1.102-1.288 | <.001 |
| Renal failure | 1.354 | 1.239-1.480 | <.001 |
| Obesity | 1.028 | 0.949-1.114 | .50 |
| Drug abuse | 0.980 | 0.771-1.241 | .86 |
| Smoking | 0.913 | 0.849-0.980 | .01 |
| Hyperlipidemia | 0.901 | 0.835-0.970 | .007 |
| Index hospitalization LOS | 1.022 | 1.016-1.027 | <.001 |
| Rural patient | 0.885 | 0.769-1.019 | .09 |
| Teaching hospital | 0.933 | 0.854-1.020 | .13 |
| Hospital bed size | |||
| Medium§ | 1.012 | 0.932-1.100 | .77 |
| Small§ | 0.921 | 0.802-1.056 | .24 |
| Mechanical valve‖ | 1.226 | 1.119-1.344 | <.001 |
aOR, Adjusted odds ratio; CI, confidence interval; SES, socioeconomic status; LOS, length of stay.
Wald test.
Indexed to <65 years.
Indexed to higher SES.
Indexed to larger bed size.
Indexed to bioprosthetic valve. Readmission as binary outcome with no readmission = 0 and readmission = 1.
Table E4.
Full regression model for 90-day readmission after SAVR
| Risk factor | aOR | 95% CI | P value∗ |
|---|---|---|---|
| Age 65-79y† | 1.116 | 1.034-1.204 | .005 |
| Age> 80y† | 1.561 | 1.390-1.752 | <.001 |
| Female sex | 1.103 | 1.031-1.181 | .005 |
| Urgent intervention | 1.064 | 0.970-1.167 | .18 |
| Lower SES‡ | 1.271 | 1.149-1.406 | <.001 |
| Middle SES‡ | 1.073 | 0.990-1.166 | .10 |
| Congestive heart failure | 1.245 | 1.160-1.336 | <.001 |
| Arrhythmia | 1.275 | 1.187-1.369 | <.001 |
| Pulmonary circulation disorder | 1.279 | 1.169-1.400 | <.001 |
| Peripheral vascular disease | 1.096 | 1.009-1.191 | .03 |
| Hypertension | 1.130 | 1.035-1.235 | .007 |
| Chronic obstructive pulmonary disease | 1.237 | 1.144-1.338 | <.001 |
| Diabetes mellitus | 1.239 | 1.151-1.334 | <.001 |
| Renal failure | 1.317 | 1.209-1.434 | <.001 |
| Obesity | 1.043 | 0.970-1.125 | .28 |
| Drug abuse | 1.077 | 0.859-1.350 | .52 |
| Smoking | 0.960 | 0.899-1.030 | .27 |
| Hyperlipidemia | 0.928 | 0.863-0.997 | .041 |
| Index hospitalization LOS | 1.029 | 1.023-1.035 | <.001 |
| Rural patient | 0.861 | 0.753-0.990 | .03 |
| Teaching hospital | 0.990 | 0.906-1.076 | .77 |
| Hospital bed size | |||
| Medium§ | 1.043 | 0.960-1.128 | .29 |
| Small§ | 0.950 | 0.838-1.082 | .45 |
| Mechanical valve‖ | 1.198 | 1.097-1.307 | <.001 |
aOR, Adjusted odds ratio; CI, confidence interval; SES, socioeconomic status; LOS, length of stay.
Wald test.
Indexed to <65 years.
Indexed to higher SES.
Indexed to larger bed size.
Indexed to bioprosthetic valve. Readmission as binary outcome with no readmission = 0 and readmission = 1.
Table E5.
Full regression model for 1-year readmission after SAVR
| Risk factor | aHR | 95% CI | P value∗ |
|---|---|---|---|
| Age 65-79y† | 1.07 | 1.02-1.12 | .007 |
| Age >80y† | 1.36 | 1.26-1.47 | <.001 |
| Female sex | 1.12 | 1.07-1.16 | <.001 |
| Urgent intervention | 1.13 | 1.06-1.20 | <.001 |
| Lower SES‡ | 1.19 | 1.11-1.28 | <.001 |
| Middle SES‡ | 1.05 | 0.99-1.11 | .13 |
| Congestive heart failure | 1.21 | 1.15-1.27 | <.001 |
| Arrhythmia | 1.25 | 1.19-1.31 | <.001 |
| Pulmonary circulation disorder | 1.19 | 1.12-1.27 | <.001 |
| Peripheral vascular disease | 1.11 | 1.06-1.18 | <.001 |
| Hypertension | 1.11 | 1.04-1.17 | .001 |
| Chronic obstructive pulmonary disease | 1.23 | 1.17-1.30 | <.001 |
| Diabetes mellitus | 1.22 | 1.16-1.29 | <.001 |
| Renal failure | 1.28 | 1.21-1.36 | <.001 |
| Obesity | 1.04 | 0.99-1.09 | .17 |
| Drug abuse | 1.13 | 0.98-1.30 | .10 |
| Smoking | 0.97 | 0.92-1.02 | .18 |
| Hyperlipidemia | 0.93 | 0.88-0.97 | .003 |
| Index hospitalization LOS | 1.02 | 1.01-1.02 | <.001 |
| Rural patient | 0.92 | 0.85-0.99 | .03 |
| Teaching hospital | 0.97 | 0.91-1.04 | .39 |
| Hospital bed size | |||
| Medium§ | 1.03 | 0.97-1.10 | .30 |
| Small§ | 0.98 | 0.89-1.08 | .69 |
| Mechanical valve‖ | 1.11 | 1.04-1.18 | .001 |
aHR, Adjusted hazard ratio; CI, confidence interval; SES, socioeconomic status; LOS, length of stay.
Wald test.
Indexed to <65 years.
Indexed to higher SES.
Indexed to larger bed size.
Indexed to bioprosthetic valve. Readmission as binary outcome with no readmission = 0 and readmission = 1.
Table E6.
Full regression model for 30-day readmission after TAVR
| Risk factor | aOR | 95% CI | P value∗ |
|---|---|---|---|
| Age 65-79 y† | 0.990 | 0.865-1.123 | .83 |
| Age >80 y† | 1.113 | 0.980-1.266 | .10 |
| Female sex | 1.058 | 0.998-1.120 | .06 |
| Urgent intervention | 1.179 | 1.093-1.270 | <.001 |
| Lower SES‡ | 1.005 | 0.923-1.094 | .91 |
| Middle SES‡ | 0.980 | 0.914-1.042 | .46 |
| Congestive heart failure | 1.228 | 1.143-1.320 | <.001 |
| Arrhythmia | 1.278 | 1.205-1.355 | <.001 |
| Pulmonary circulation disorder | 1.055 | 0.990-1.129 | .13 |
| Peripheral vascular disease | 1.098 | 1.030-1.171 | .004 |
| Hypertension | 1.007 | 0.916-1.107 | .88 |
| Chronic obstructive pulmonary disease | 1.164 | 1.094-1.237 | <.001 |
| Diabetes mellitus | 1.164 | 1.098-1.234 | <.001 |
| Renal failure | 1.417 | 1.336-1.502 | <.001 |
| Obesity | 0.980 | 0.909-1.055 | .58 |
| Drug abuse | 1.095 | 0.739-1.622 | .65 |
| Smoking | 0.980 | 0.921-1.037 | .44 |
| Hyperlipidemia | 0.874 | 0.822-0.930 | <.001 |
| Index hospitalization LOS | 1.032 | 1.027-1.037 | <.001 |
| Rural patient | 0.906 | 0.804-1.021 | .11 |
| Teaching hospital | 0.980 | 0.904-1.069 | .69 |
| Hospital bed size | |||
| Medium§ | 0.980 | 0.919-1.056 | .67 |
| Small§ | 1.038 | 0.899-1.199 | .61 |
aOR, Adjusted odds ratio; CI, confidence interval; SES, socioeconomic status; LOS, length of stay.
Wald test.
Indexed to <65 years.
Indexed to higher SES.
Indexed to larger bed size. Readmission as binary outcome with no readmission = 0 and readmission = 1.
Table E7.
Full regression model for 90-day readmission after TAVR
| Risk factor | aOR | 95% CI | P value∗ |
|---|---|---|---|
| Age 65-79 y† | 0.950 | 0.847-1.075 | .44 |
| Age > 80 y† | 1.048 | 0.931-1.179 | .44 |
| Female sex | 1.078 | 1.024-1.135 | .004 |
| Urgent intervention | 1.116 | 1.040-1.198 | .002 |
| Lower SES‡ | 0.950 | 0.881-1.028 | .21 |
| Middle SES‡ | 0.990 | 0.928-1.045 | .62 |
| Congestive heart failure | 1.247 | 1.170-1.328 | <.001 |
| Arrhythmia | 1.272 | 1.207-1.340 | <.001 |
| Pulmonary circulation disorder | 1.090 | 1.024-1.160 | .007 |
| Peripheral vascular disease | 1.126 | 1.062-1.194 | <.001 |
| Hypertension | 1.022 | 0.938-1.113 | .62 |
| Chronic obstructive pulmonary disease | 1.181 | 1.118-1.249 | <.001 |
| Diabetes mellitus | 1.197 | 1.135-1.263 | <.001 |
| Renal failure | 1.399 | 1.327-1.475 | <.001 |
| Obesity | 0.960 | 0.895-1.022 | .19 |
| Drug abuse | 1.430 | 1.031-1.985 | .03 |
| Smoking | 0.980 | 0.931-1.035 | .50 |
| Hyperlipidemia | 0.864 | 0.817-0.914 | <.001 |
| Index hospitalization LOS | 1.034 | 1.028-1.039 | <.001 |
| Rural patient | 0.894 | 0.802-0.997 | .043 |
| Teaching hospital | 0.946 | 0.878-1.019 | .15 |
| Hospital bed size | |||
| Medium§ | 0.997 | 0.936-1.061 | .92 |
| Small§ | 1.029 | 0.907-1.166 | .66 |
aOR, Adjusted odds ratio; CI, confidence interval; SES, socioeconomic status; LOS, length of stay.
Wald test.
Indexed to <65 years.
Indexed to higher SES.
Indexed to larger bed size. Readmission as binary outcome with no readmission = 0 and readmission = 1.
Table E8.
Full regression model for 1-year readmission after TAVR
| Risk factor | aHR | 95% CI | P value∗ |
|---|---|---|---|
| Age 65-79 y† | 0.97 | 0.92-1.03 | .31 |
| Age > 80 y† | 1.03 | 0.97-1.08 | .37 |
| Female sex | 1.06 | 1.02-1.09 | .002 |
| Urgent intervention | 1.15 | 1.09-1.20 | <.001 |
| Lower SES‡ | 1.01 | 0.96-1.07 | .60 |
| Middle SES‡ | 0.98 | 0.94-1.02 | .41 |
| Congestive heart failure | 1.15 | 1.11-1.20 | <.001 |
| Arrhythmia | 1.29 | 1.26-1.33 | <.001 |
| Pulmonary circulation disorder | 1.09 | 1.04-1.14 | <.001 |
| Peripheral vascular disease | 1.11 | 1.07-1.15 | <.001 |
| Hypertension | 1.01 | 0.96-1.07 | .66 |
| Chronic obstructive pulmonary disease | 1.19 | 1.15-1.23 | <.001 |
| Diabetes mellitus | 1.17 | 1.13-1.20 | <.001 |
| Renal failure | 1.32 | 1.27-1.36 | <.001 |
| Obesity | 0.93 | 0.89-0.97 | <.001 |
| Drug abuse | 1.31 | 1.09-1.59 | .005 |
| Smoking | 0.98 | 0.94-1.01 | .19 |
| Hyperlipidemia | 0.90 | 0.87-0.93 | <.001 |
| Index hospitalization LOS | 1.02 | 1.01-1.02 | <.001 |
| Rural patient | 0.90 | 0.84-0.95 | <.001 |
| Teaching hospital | 0.98 | 0.92-1.04 | .52 |
| Hospital bed size | |||
| Medium§ | 1.00 | 0.95-1.06 | .89 |
| Small§ | 1.08 | 0.99-1.19 | .08 |
aHR, Adjusted hazard ratio; CI, confidence interval; SES, socioeconomic status; LOS, length of stay.
Wald test.
Indexed to <65 years.
Indexed to higher SES.
Indexed to larger bed size. Readmission as binary outcome with no readmission = 0 and readmission = 1.
References
- 1.McLeish T., Seadler B.D., Parrado R., Rein L., Joyce D.L. The effect of socioeconomic factors on patient outcomes in cardiac surgery. J Card Surg. 2022;37:5135–5143. doi: 10.1111/jocs.17229. [DOI] [PubMed] [Google Scholar]
- 2.Charles E.J., Mehaffey J.H., Hawkins R.B., Fonner C.E., Yarboro L.T., Quader M.A., et al. Socioeconomic Distressed Communities Index predicts risk-adjusted mortality after cardiac surgery. Ann Thorac Surg. 2019;107:1706–1712. doi: 10.1016/j.athoracsur.2018.12.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Newell P, Asokan S, Zogg C, Prasanna A, Hirji S, Harloff M, et al. Contemporary socioeconomic-based disparities in cardiac surgery: are we closing the disparities gap? J Thorac Cardiovasc Surg. April 22, 2022 [Epub ahead of print]. 10.1016/j.jtcvs.2022.02.061 [DOI] [PubMed]
- 4.Iyengar A., Patrick W.L., Helmers M.R., Kelly J.J., Han J., Williams M.L., et al. Neighborhood socioeconomic status independently predicts outcomes after mitral valve surgery. Ann Thorac Surg. 2023;115:940–947. doi: 10.1016/j.athoracsur.2023.01.002. [DOI] [PubMed] [Google Scholar]
- 5.Strobel R.J., Charles E.J., Mehaffey J.H., Hawkins R.B., Quader M.A., Rich J.B., et al. Effect of socioeconomic distress on risk-adjusted mortality after valve surgery for infective endocarditis. Semin Thorac Cardiovasc Surg. 2022;35:497–507. doi: 10.1053/j.semtcvs.2022.05.007. [DOI] [PubMed] [Google Scholar]
- 6.Bagger J.P., Edwards M.B., Taylor K.M. Influence of socioeconomic status on survival after primary aortic or mitral valve replacement. Heart. 2008;94:182–185. doi: 10.1136/hrt.2007.119172. [DOI] [PubMed] [Google Scholar]
- 7.Mohee K., Protty M.B., Whiffen T., Chase A., Smith D. Impact of social deprivation on outcome following transcatheter aortic valve implantation (TAVI) Open Heart. 2019;6:e001089. doi: 10.1136/openhrt-2019-001089. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Nathan A.S., Yang L., Yang N., Eberly L.A., Khatana S.A.M., Dayoub E.J., et al. Racial, ethnic, and socioeconomic disparities in access to transcatheter aortic valve replacement within major metropolitan areas. JAMA Cardiol. 2022;7:150–157. doi: 10.1001/jamacardio.2021.4641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Nathan A.S., Yang L., Yang N., Khatana S.A.M., Dayoub E.J., Eberly L.A., et al. Socioeconomic and geographic characteristics of hospitals establishing transcatheter aortic valve replacement programs, 2012-2018. Circ Cardiovasc Qual Outcomes. 2021;14:e008260. doi: 10.1161/CIRCOUTCOMES.121.008260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sleder A., Tackett S., Cerasale M., Mittal C., Isseh I., Radjef R., et al. Socioeconomic and racial disparities: a case-control study of patients receiving transcatheter aortic valve replacement for severe aortic stenosis. J Racial Ethn Health Disparities. 2017;4:1189–1194. doi: 10.1007/s40615-016-0325-x. [DOI] [PubMed] [Google Scholar]
- 11.Agency for Healthcare Research and Quality Healthcare cost and utilization project (HCUP). HCUP Nationwide Readmissions Database (NRD). Rockville, MD: Agency for Healthcare Research and Quality, 2016, 2017 and 2018. https://www.ahrq.gov/
- 12.Vera R., Zhang Q., Wall M.J., Jr., Coselli J.S., Rosengart T.K., Chatterjee S., et al. Readmission after bioprosthetic vs mechanical mitral valve replacement in the United States. Ann Thorac Surg. 2022 doi: 10.1016/j.athoracsur.2022.05.064. [Epub ahead of print]. [DOI] [PubMed] [Google Scholar]
- 13.R Core Team. R A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. 2022. https://www.R-project.org/
- 14.Lumley T. Analysis of complex survey samples. https://www.jstatsoft.org/article/view/v009i08
- 15.Schultz W.M., Kelli H.M., Lisko J.C., Varghese T., Shen J., Sandesara P., et al. Socioeconomic status and cardiovascular outcomes: challenges and interventions. Circulation. 2018;137:2166–2178. doi: 10.1161/CIRCULATIONAHA.117.029652. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Damluji A.A., Fabbro M., II, Epstein R.H., Rayer S., Wang Y., Moscucci M., et al. Transcatheter aortic valve replacement in low-population density areas: assessing healthcare access for older adults with severe aortic stenosis. Circ Cardiovasc Qual Outcomes. 2020;13:e006245. doi: 10.1161/CIRCOUTCOMES.119.006245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Maniar H.S., Bell J.M., Moon M.R., Meyers B.F., Marsala J., Lawton J.S., et al. Prospective evaluation of patients readmitted after cardiac surgery: analysis of outcomes and identification of risk factors. J Thorac Cardiovasc Surg. 2014;147:1013–1018. doi: 10.1016/j.jtcvs.2013.10.066. [DOI] [PubMed] [Google Scholar]
- 18.Stormacq C., Van den Broucke S., Wosinski J. Does health literacy mediate the relationship between socioeconomic status and health disparities? Integrative review. Health Promot Int. 2019;34:e1–e17. doi: 10.1093/heapro/day062. [DOI] [PubMed] [Google Scholar]
- 19.Miller T.A. Health literacy and adherence to medical treatment in chronic and acute illness: a meta-analysis. Patient Educ Couns. 2016;99:1079–1086. doi: 10.1016/j.pec.2016.01.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Friedman A.J., Cosby R., Boyko S., Hatton-Bauer J., Turnbull G. Effective teaching strategies and methods of delivery for patient education: a systematic review and practice guideline recommendations. J Cancer Educ. 2011;26:12–21. doi: 10.1007/s13187-010-0183-x. [DOI] [PubMed] [Google Scholar]
- 21.Chudgar N.P., Zhu R., Gray K.D., Chiu R., Carrera A.D., Lang S.J., et al. Implementing a high-value care discharge protocol in patients undergoing CABG reduces readmission. Ann Thorac Surg. 2022;113:1112–1118. doi: 10.1016/j.athoracsur.2021.07.036. [DOI] [PubMed] [Google Scholar]
- 22.Dalén M., Persson M., Glaser N., Sartipy U. Socioeconomic status and risk of bleeding after mechanical aortic valve replacement. J Am Coll Cardiol. 2022;79:2502–2513. doi: 10.1016/j.jacc.2022.04.030. [DOI] [PubMed] [Google Scholar]
- 23.Strobel R.J., Mehaffey J.H., Hawkins R.B., Young A.M., Scott E.J., Quader M., et al. Socioeconomic distress associated with increased use of percutaneous coronary intervention over coronary artery bypass grafting. Ann Thorac Surg. 2023;115:914–921. doi: 10.1016/j.athoracsur.2022.06.040. [DOI] [PubMed] [Google Scholar]
- 24.Hall M.H., Esposito R.A., Pekmezaris R., Lesser M., Moravick D., Jahn L., et al. Cardiac surgery nurse practitioner home visits prevent coronary artery bypass graft readmissions. Ann Thorac Surg. 2014;97:1488–1493. doi: 10.1016/j.athoracsur.2013.12.049. discussion 1493-5. [DOI] [PubMed] [Google Scholar]
- 25.Mackenbach J.P., Stirbu I., Roskam A.J., Schaap M.M., Menvielle G., Leinsalu M., et al. Socioeconomic inequalities in health in 22 European countries. N Engl J Med. 2008;358:2468–2481. doi: 10.1056/NEJMsa0707519. [DOI] [PubMed] [Google Scholar]
- 26.Frankel W.C., Sylvester C.B., Asokan S., Ryan C.T., Zea-Vera R., Zhang Q., et al. Outcomes, cost, and readmission after surgical aortic or mitral valve replacement at safety-net and non-safety-net hospitals. Ann Thorac Surg. 2022;114:703–709. doi: 10.1016/j.athoracsur.2022.01.049. [DOI] [PMC free article] [PubMed] [Google Scholar]