Abstract
Background
Racial and ethnic minority patients undergoing bariatric surgery experience a disproportionate number of unfavorable outcomes. The drivers of these differences have been attributed to differences in comorbidities, socioeconomic factors, access to postoperative care, and cultural considerations. We aimed to explore healthcare disparities among bariatric surgery patients in our institution.
Methods
A retrospective cohort analysis of bariatric surgery patients at a MBSAQIP-accredited bariatric surgery center was conducted over three years. Patient demographics, preoperative comorbidities, and post-operative healthcare utilization and complication rates were stratified according to race. A logistic regression model assessed associations between race and ethnicity and post operative resource utilization, adjusting for demographic, social determinants of health, preoperative risk factors and procedural differences.
Results
A total of 1,408 patients were included, with 60.6% White, 32.8% Black, 0.7% Asian, 5.4% other/unknown. Preoperative risk factors differed slightly across racial groups. Postoperatively, there was minimal difference in the cohorts’ complications but when compared with White patients, Black patients had an increased length of stay (1.55 versus 1.34 days, p<0.001), were more likely to be readmitted (OR=2.64, 95%CI=1.26-5.56, p=0.011), and increased use of ambulatory IV fluid treatment (10% versus 2.7%, p< 0.001).
Conclusions
Despite similar preoperative risk factors and postoperative complications, Black patients experienced a longer length of stay, an increased use of IV fluid treatment, and a higher likelihood of readmission within 30 days after bariatric surgery. Race was associated with both higher length of stay and readmission rate highlighting the need for further investigation into the contributing factors and identification of targets to reduce disparities in bariatric care.
Keywords: Disparity, Race, Ethnicity, Bariatric, Surgical Outcomes
Introduction:
Despite significant advancements in medicine, healthcare disparities continue to persist across all medical fields. In recent years, these disparities have gained increased attention, particularly in the medical and surgical treatment of obesity, a disease that affects 40% of the US population [1]. Racial minority patients are less likely than White patients to undergo weight loss surgery despite the higher rates of obesity in these minority populations [2,3]. Additionally, racial and ethnic minority patients experience a disproportionate number of unfavorable outcomes such as increased rates of readmission, morbidity, unplanned admission, length of stay, and even mortality [4,5]. These differences multifactorial with contributing factors such as socioeconomic factors, access to post-operative care, and cultural considerations, among others.
Research has highlighted the importance of patient education and preoperative counseling in addressing disparities and mitigating unfavorable outcomes for racial and ethnic minority populations, yet much work remains to close the divide [6, 7, 8, 9]. Frequently, these differences are attributed to patient’s medical comorbidities which oftentimes can be a function of socioeconomic status. Additional studies have shown that certain factors such as race itself may be implicated as an independent risk factor for poorer outcomes [4,5,10,11]. There has been limited study data in the bariatric population to explore this issue.
This study aims to elucidate the role of race and ethnicity, as independent risk factors for adverse outcomes 30-days post-surgery, after controlling for preoperative medical comorbidities and other social determinants of health and the potential role of post operative complications in the association between race and ethnicity and outcomes.
Materials and Methods
This retrospective study included adults (≥18 years) who underwent elective bariatric surgery, either sleeve gastrectomy or Roux-en-Y gastric bypass from January 1, 2020, through December 31, 2023. All surgeries were performed by hospital-employed, fellowship-trained bariatric surgeons. Patients who underwent revision or conversion procedures were excluded. The index visit was defined as each patient’s first hospitalization for bariatric surgery within the study time frame. The observation window spanned from January 1, 2019, to March 31, 2024, thereby allowing a one-year lookback period prior to surgery to capture comorbidities, hospitalizations, emergency department (ED) visits, and social determinants of health, as well as a 90-day postoperative follow-up for cases performed in late 2023. The Institutional Review Board approved this protocol, and informed consent was waived due to the retrospective study design and large number of patients.
Data Sources and preparation data were extracted from multiple internal sources, including the electronic health record (EHR), and our local Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program (MBSAQIP) registry. Additional census-based metrics obtained from the American Community Survey, a national survey including data on employment, demographics, and other population-based metrics.
Demographic information included age at the time of surgery, gender, race (White, Black, Asian, or Other), and ethnicity (Hispanic or Non-Hispanic). Clinical variables comprised body mass index (BMI in kg/m2) closest to surgery, as well as comorbidities and risk factors such as diabetes mellitus, hypertension, hyperlipidemia, gastroesophageal reflux disease (GERD), and sleep apnea. Additional patient-level characteristics included insurance type (commercial, Medicare, Medicaid, value-based bundled payment agreements, or self-pay), number of anti-hypertensive medications, immunosuppressive therapy status, and marital status (married/partnership, single/separated, or unknown).
Socioeconomic indicators included the Area Deprivation Index (ADI), a validated NIH funded tool to evaluate resources at a neighborhood level [12], and the Social Vulnerability Index (SVI), a mapping algorithm used to quantify communities experiencing social vulnerability in efforts to decrease health inequities among other social goals [13]. Additional factors included ethnicity, body mass index (BMI) closest to procedure, and insurance type. Preoperative risk status was evaluated using the American Society of Anesthesiologists (ASA) classification (ASA I, ASA II, ASA III, or ASA IV).
The primary outcomes were the occurrence of a 30-day ED visit and 30-day hospital readmission related to bariatric surgery. Secondary and exploratory endpoints included hospital length of stay (LOS), ambulatory intravenous fluid treatment, perioperative complications including superficial or deep incisional surgical site infections, organ/space infections, venous thromboembolism (VTE), blood transfusion, renal insufficiency or failure, cardiac events, unplanned admissions to the ICU, and reoperation and 30-day mortality. Comorbidities were identified from International Classification of Diseases (ICD-10) codes in the EHR.
Descriptive statistics were used to summarize baseline demographics and clinical profiles. Continuous variables, including age at surgery and BMI, were reported as means, standard deviations, medians, interquartile ranges, and full ranges. Categorical variables, such as race and ethnicity, gender, insurance type, and ASA classification, were described using frequencies and percentages. Primary and secondary outcomes were summarized similarly.
Group comparisons of primary outcomes and exploratory outcomes by race were performed using chi-square or Fisher’s exact tests for categorical data, and Kruskal–Wallis tests for continuous data, as appropriate. To assess the association of race with the primary outcomes (30-day ED visits and 30-day readmissions), separate multivariable logistic regression models were fitted, producing odds ratios (OR) with 95% confidence intervals (CI). Adjustments were made for demographic and clinical covariates selected a priori including age, surgery type, insurance type, marital status, ADI, SVI, BMI closest to surgery, diabetes status, GERD, hyperlipidemia, hypertension (including number of anti-hypertension medications), undergoing immune suppressive therapy, and ASA class. For hospital LOS, a log-normal regression model was fitted with the same set of covariates, yielding time ratios (TR) and 95% CIs. Statistical significance was defined as p < 0.05, and all tests were two-sided.
Results
A total of 1,408 patients met the inclusion criteria. Of these, 462 (32.8%) were Black, 861 (61.2%) were White, 10 (0.7%) were Asian, and 75 (5.3%) were classified as “other” race. The overall mean age was 42.49 years (SD 11.04; range, 18–75), with slight variations in mean age among racial subgroups. The majority of patients (83.2%) were female, and 89.1% identified as non-Hispanic. Notably, a large majority of patients classified as “other” identified as Hispanic or Latino (81.3%). Nearly all participants (98.9%) preferred English as their primary language. The mean BMI closest to the date of surgery was 44.4 (SD 6.93), with a median of 43.54 (IQR 39.39–48.33). Commercial insurance was most common (63.9%), followed by value-based bundled payment agreement coverage (27.9%). Fifty-three percent of patients were married or in a partnership. Rates of being single or separated by race were as follows: White, 38.1%, Black, 59.3%, Asian, 50.0%, and “other,” 58.7%. A larger number of patients underwent sleeve gastrectomy (66.3%) which is reflected among all cohorts. Black patients underwent the highest number of sleeve gastrectomy at 76.0% followed by other at 68.9%, White patients at 60.9%, and the lowest rate was seen in Asian patients at 60.0%.
Socioeconomic measures collected included the area deprivation index (ADI) and social vulnerability index (SVI). The ADI is ranked on a scale from 1 to 100, with 100 representing the most disadvantaged neighborhoods. For our cohort, the ADI was as follows: White patients, 42.01; Black patients, 51.76; and “other,” 51.24. The SVI, ranging from 0.0 to 1.0, with 1.0 indicating the most vulnerable populations, showed the following for our cohort: White patients, 0.40; Black patients, 0.52; and “other,” 0.54. The majority of patients (85.1%) were categorized as ASA III. Common comorbidities included sleep apnea (47.4%), hypertension (43.0%), diabetes mellitus (22.4%), GERD (20.2%), and hyperlipidemia (22.4%). The summary statistics for the baseline demographic and clinical characteristics are presented in Table 1.
Table 1.
Baseline demographic and clinical characteristics of the study population
| Characteristic | Total (N=1408) | Black (N=462) | White (N=861) | Asian (N=10) | Other (N=75) |
|---|---|---|---|---|---|
| Age on Date of Surgery | |||||
| Mean (SD) | 42.49 (11.04) | 41.02 (10.74) | 43.59 (11.1) | 40.3 (9.33) | 39.17 (10.98) |
| Median [IQR] | 42 [34, 50] | 40 [33, 49] | 43 [35, 51] | 37.5 [35, 43] | 37 [31.5, 46] |
| Gender, n (%) | |||||
| Female | 1172 (83.2) | 405 (87.7) | 696 (80.8) | 9 (90.0) | 62 (82.7) |
| Male | 236 (16.8) | 57 (12.3) | 165 (19.2) | 1 (10.0) | 13 (17.3) |
| Ethnicity, n (%) | |||||
| Non-Hispanic or Latino | 1255 (89.1) | 446 (96.5) | 788 (91.5) | 9 (90.0) | 12 (16.0) |
| Hispanic or Latino | 145 (10.3) | 14 (3.0) | 70 (8.1) | 0 (0) | 61 (81.3) |
| Refused/Unavailable | 8 (0.6) | 2 (0.4) | 3 (0.3) | 1 (10.0) | 2 (2.7) |
| Height (cm) | |||||
| Mean (SD) | 166.62 (8.78) | 166.2 (8.16) | 167.27 (9.03) | 162.79 (7.27) | 162.18 (8.31) |
| Median [IQR] | 165.1 [160, 172] | 165 [160, 170.2] | 165.1 [160, 172.7] | 161.25 [158.13, 165] | 160 [156.35, 167.5] |
| Highest BMI | |||||
| Mean (SD) | 47.06 (7.4) | 47.59 (7.13) | 46.9 (7.57) | 45.9 (8.32) | 45.72 (6.7) |
| Median [IQR] | 45.73 [41.75, 50.99] | 46.42 [42.63, 51.90] | 45.52 [41.32, 50.73] | 44.33 [41.2, 49.33] | 44 [41.14, 50.10] |
| BMI Closest to Procedure | |||||
| Mean (SD) | 44.4 (6.93) | 45.14 (6.82) | 44.14 (6.98) | 43.23 (8.73) | 43.02 (6.32) |
| Median [IQR] | 43.54 [39.39, 48.33] | 44.17 [40.23, 49.13] | 43.12 [38.99, 47.69] | 39.96 [37.2, 48.08] | 41.94 [39.08, 47.13] |
| Insurance Type, n (%) | |||||
| Bariatric | 393 (27.9) | 85 (18.4) | 286 (33.2) | 4 (40.0) | 18 (24.0) |
| Commercial | 900 (63.9) | 344 (74.5) | 500 (58.1) | 5 (50.0) | 51 (68.0) |
| Medicaid | 56 (4.0) | 15 (3.2) | 38 (4.4) | 1 (10.0) | 2 (2.7) |
| Medicare | 57 (4.0) | 16 (3.5) | 37 (4.3) | 0 (0) | 4 (5.3) |
| Self-pay | 2 (0.1) | 2 (0.4) | 0 (0) | 0 (0) | 0 (0) |
| Marital Status, n (%) | |||||
| Currently Married/Partnership | 755 (53.6) | 187 (40.5) | 532 (61.8) | 5 (50.0) | 31 (41.3) |
| Single/Separated | 651 (46.2) | 274 (59.3) | 328 (38.1) | 5 (50.0) | 44 (58.7) |
| Unknown | 2 (0.1) | 1 (0.2) | 1 (0.1) | 0 (0) | 0 (0) |
| ADI State Rank | |||||
| Mean (SD) | 5.47 (2.81) | 6.32 (2.79) | 4.92 (2.72) | 4.22 (2.39) | 6.66 (2.45) |
| Median [IQR] | 5 [3, 8] | 7 [4, 9] | 5 [2, 7] | 3 [3, 5] | 7 [4.25, 9] |
| Missing, n (%) | 52 (3.7) | 18 (3.9) | 32 (3.7) | 1 (10.0) | 1 (1.3) |
| ADI National Rank | |||||
| Mean (SD) | 45.68 (19.94) | 51.76 (20.84) | 42.01 (18.58) | 37.89 (14.4) | 51.24 (19.77) |
| Median [IQR] | 44 [31, 58] | 50 [35, 66] | 40 [28, 55] | 33 [30, 43] | 50 [37, 62] |
| Missing, n (%) | 52 (3.7) | 18 (3.9) | 32 (3.7) | 1 (10.0) | 1 (1.3) |
| SVI | |||||
| Mean (SD) | 0.45 (0.26) | 0.52 (0.26) | 0.4 (0.25) | 0.33 (0.29) | 0.54 (0.25) |
| Median [IQR] | 0.46 [0.22, 0.65] | 0.55 [0.31, 0.73] | 0.41 [0.17, 0.61] | 0.24 [0.13, 0.31] | 0.57 [0.34, 0.75] |
| Missing, n (%) | 46 (3.3) | 17 (3.7) | 27 (3.1) | 1 (10.0) | 1 (1.3) |
| Albumin (g/dL) | |||||
| Mean (SD) | 4.20 (0.31) | 4.12 (0.33) | 4.24 (0.30) | 4.10 (0.46) | 4.21 (0.19) |
| Median [IQR] | 4.2 [4.0, 4.4] | 4.1 [3.9, 4.3] | 4.2 [4.1, 4.4] | 4.1 [3.7, 4.5] | 4.2 [4.08, 4.4] |
| Missing, n (%) | 874 (62.1) | 290 (62.8) | 524 (60.8) | 6 (60.0) | 54 (73.0) |
| Serum Creatinine (mg/dL) | |||||
| Mean (SD) | 0.85 (0.38) | 0.90 (0.58) | 0.82 (0.23) | 0.69 (0.08) | 0.75 (0.15) |
| Median [IQR] | 0.8 [0.7, 0.91] | 0.8 [0.72, 0.93] | 0.8 [0.7, 0.9] | 0.69 [0.64, 0.69] | 0.71 [0.66, 0.82] |
| Missing, n (%) | 438 (31.1) | 149 (32.3) | 254 (29.5) | 3 (30.0) | 32 (43.2) |
| ASA Classification, n (%) | |||||
| ASA II | 186 (13.2) | 46 (10.0) | 124 (14.4) | 2 (20.0) | 14 (18.9) |
| ASA III | 1,198 (85.1) | 410 (88.7) | 724 (84.0) | 8 (80.0) | 56 (75.7) |
| ASA IV | 24 (1.7) | 6 (1.3) | 14 (1.6) | 0 (0) | 4 (5.4) |
| Surgery type | |||||
| Roux-en-Y Gastric Bypass | 475 (33.7%) | 111 (24.0%) | 337 (39.1%) | 4 (40.0%) | 23 (31.1%) |
| Sleeve Gastrectomy | 933 (66.3%) | 351 (76.0%) | 525 (60.9%) | 6 (60.0%) | 51 (68.9%) |
| Current Smoker, n (%) | 77 (5.5) | 28 (6.1) | 45 (5.2) | 0 (0) | 4 (5.3) |
| Diabetes Mellitus, n (%) | 316 (22.4) | 105 (22.7) | 198 (23.0) | 6 (60.0) | 7 (9.3) |
| Immunosuppressive Therapy, n (%) | 36 (2.6) | 11 (2.4) | 21 (2.4) | 0 (0) | 4 (5.3) |
| History of Severe COPD, n (%) | 4 (0.3) | 1 (0.2) | 2 (0.2) | 0 (0) | 1 (1.3) |
| History of Pulmonary Embolism, n (%) | 25 (1.8) | 10 (2.2) | 14 (1.6) | 0 (0) | 1 (1.3) |
| Sleep Apnea, n (%) | 667 (47.4) | 185 (40.0) | 447 (51.9) | 4 (40.0) | 31 (41.3) |
| GERD, n (%) | 284 (20.2) | 55 (11.9) | 216 (25.1) | 1 (10.0) | 12 (16.0) |
| Previous Foregut Surgery, n (%) | 4 (0.3) | 0 (0) | 4 (0.5) | 0 (0) | 0 (0) |
| History of Myocardial Infarction, n (%) | 7 (0.5) | 3 (0.6) | 4 (0.5) | 0 (0) | 0 (0) |
| Previous PCI/PTCA, n (%) | 11 (0.8) | 7 (1.5) | 4 (0.5) | 0 (0) | 0 (0) |
| Previous Cardiac Surgery, n (%) | 9 (0.6) | 2 (0.4) | 5 (0.6) | 0 (0) | 2 (2.7) |
| Hypertension, n (%) | 605 (43.0) | 205 (44.4) | 376 (43.7) | 3 (30.0) | 21 (28.0) |
| Hyperlipidemia, n (%) | 316 (22.4) | 92 (19.9) | 213 (24.7) | 2 (20.0) | 9 (12.0) |
| Preop Venous Thrombosis Requiring Therapy, n (%) | 38 (2.7) | 15 (3.2) | 23 (2.7) | 0 (0) | 0 (0) |
| Therapeutic Anticoagulation, n (%) | 35 (2.5) | 12 (2.6) | 23 (2.7) | 0 (0) | 0 (0) |
| Venous Stasis, n (%) | 9 (0.6) | 0 (0) | 9 (1.0) | 0 (0) | 0 (0) |
| IVC Filter, n (%) | 2 (0.1) | 0 (0) | 2 (0.2) | 0 (0) | 0 (0) |
| Currently Requiring Dialysis, n (%) | 2 (0.1) | 2 (0.4) | 0 (0) | 0 (0) | 0 (0) |
| Renal Insufficiency, n (%) | 3 (0.2) | 0 (0) | 3 (0.3) | 0 (0) | 0 (0) |
Note: SD, Standard deviation; IQR, Interquartile range; BMI, body mass index; ASA, American Society of Anesthesiologists; ADI, Area Deprivation Index; SVI, Social Vulnerability Index; PCI/PTCA, percutaneous coronary intervention/percutaneous transluminal coronary angioplasty; COPD, chronic obstructive pulmonary disease; GERD, gastroesophageal reflux disease; IVC, inferior vena cava.
Thirty-day intraoperative or postoperative complications stratified by race and ethnicity are presented in Table 2. The percentage of patients experiencing at least one complication ranges from 2.6% (Black) to 10% (Asian). The number of Asian patients is very small (n=10) and the overall difference is not statistically significant (p = 0.485). Urinary tract infections occurred in 0.6% of Black patients, 0.3% of White patients, and 2.7% of “other” patients (p = 0.082). Blood transfusions within 72 hours of incision were significantly more frequent in Asian patients (10%) than in Black (0.6%), White (1%), or “other” (0%) patients (p = 0.016). Venous thromboembolism requiring therapy was seen only in the “other” group (1.3%; p < 0.001). Acute renal failure and pulmonary embolism were infrequent overall. There were no instances of wound disruption, progressive renal insufficiency, or myocardial infarction in the sample. Additionally, less than 0.2% of patients required unplanned admission to the intensive care unit (Table 2).
Table 2.
Comparison of 30-day postoperative occurrence by race.
| Outcome | Black | White | Asian | Other | P-value |
|---|---|---|---|---|---|
| At least one Intraop /Postop Occurrence | 12 (2.6) | 23 (2.7) | 1 (10.0) | 3 (4.0) | 0.485 |
| Superficial Incisional SSI | 0 (0) | 1 (0.1) | 0 (0) | 0 (0) | 0.888 |
| Deep Incisional SSI | 1 (0.2) | 0 (0) | 0 (0) | 0 (0) | 0.562 |
| Organ/Space SSI | 1 (0.2) | 1 (0.1) | 0 (0) | 0 (0) | 0.952 |
| Wound Disruption | 0 (0) | 0 (0) | 0 (0) | 0 (0) | NA |
| Pneumonia | 0 (0) | 1 (0.1) | 0 (0) | 0 (0) | 0.888 |
| Unplanned Intubation | 0 (0) | 1 (0.1) | 0 (0) | 0 (0) | 0.888 |
| Pulmonary Embolism | 1 (0.2) | 1 (0.1) | 0 (0) | 0 (0) | 0.952 |
| On Ventilator ≥48 Hours | 1 (0.2) | 0 (0) | 0 (0) | 0 (0) | 0.562 |
| Progressive Renal Insufficiency | 0 (0) | 0 (0) | 0 (0) | 0 (0) | NA |
| Acute Renal Failure | 2 (0.4) | 0 (0) | 0 (0) | 0 (0) | 0.251 |
| Urinary Tract Infection | 3 (0.6) | 3 (0.3) | 0 (0) | 2 (2.7) | 0.082 |
| Stroke/CVA | 0 (0) | 0 (0) | 0 (0) | 0 (0) | NA |
| Cardiac Arrest Requiring CPR | 0 (0) | 1 (0.1) | 0 (0) | 0 (0) | 0.888 |
| Myocardial Infarction | 0 (0) | 0 (0) | 0 (0) | 0 (0) | NA |
| Blood Transfusion Within 72 Hours from Incision | 3 (0.6) | 9 (1.0) | 1 (10.0) | 0 (0) | 0.016 |
| VTE Requiring Therapy | 0 (0) | 0 (0) | 0 (0) | 1 (1.3) | <0.001 |
| C. diff Colitis | 0 (0) | 0 (0) | 0 (0) | 0 (0) | NA |
| Sepsis | 1 (0.2) | 0 (0) | 0 (0) | 0 (0) | 0.562 |
| Septic Shock | 0 (0) | 0 (0) | 0 (0) | 0 (0) | NA |
| Unplanned Admission to ICU | 1 (0.2) | 3 (0.3) | 0 (0) | 1 (1.3) | 0.509 |
| GI Bleed | 3 (0.6) | 11 (1.3) | 1 (10.0) | 0 (0) | 0.023 |
| Bowel Obstruction | 3 (0.6) | 3 (0.3) | 0 (0) | 0 (0) | 0.794 |
| 30-Day Mortality | 0 (0) | 1 (0.1) | 0 (0) | 0 (0) | 0.888 |
| Anastomotic Staple Line Leak | 0 (0) | 1 (0.1) | 0 (0) | 0 (0) | 0.888 |
Note: SD, Standard deviation; IQR, interquartile range; SSIS, Surgical site infection; CPR, Cardiopulmonary resuscitation; VTE, Venous thromboembolism; ICU, intensive care unit.
The 30-day ED visit and readmission rates by race and ethnicity are presented in Table 3. The 30-day ED visit rates among Black patients were 9.5%, compared to 7.3% among White patients. 30-day mortality was low across all groups (0–0.1%). The multivariable logistic regression model in Table 4 showed that race and ethnicity are not significantly associated with 30-day ED visits. Age and surgery type were independent factors significantly associated with 30-day ED visit status (Table 4). Patients who underwent sleeve gastrectomy had an adjusted odds ratio for 0.38 when compared with patients who underwent gastric bypass (CI 0.24-0.61, p < 0.001). Each additional year of age decreased the likelihood of an ED visit by 3%.
Table 3.
Comparison of outcomes.
| Characteristic | Black (N=462) | White (N=861) | Asian (N=10) | Other (N=75) | P-value |
|---|---|---|---|---|---|
| Hospital Length of Stay (days) | <0.001 | ||||
| Mean (SD) | 1.55 (0.76) | 1.34 (0.71) | 1.6 (1.07) | 1.46 (0.53) | |
| Median [IQR] | 1 [1, 2] | 1 [1, 1.75] | 1 [1, 2] | 1 [1, 2] | |
| IV Treatment, n (%) | 46 (10.0) | 23 (2.7) | 1 (10.0) | 3 (4.0) | <0.001 |
| 30-day ED Visit , n (%) | 44 (9.5) | 63 (7.3) | 0 (0) | 8 (10.8) | 0.313 |
| Number of IV Visits | <0.001 | ||||
| Mean (SD) | 0.05 (0.24) | 0.20 (0.63) | 0.07 (0.25) | - | |
| Median [IQR] | 0 [0, 0] | 0 [0, 0] | 0 [0, 0] | 0 [0, 0] | |
| Intervention, n (%) | 2 (0.4) | 8 (0.9) | 0 (0) | 0 (0) | 0.64 |
| Intervention Type, n (%) | NA | ||||
| EGD | 2 (0.4) | 7 (0.8) | 0 (0) | 0 (0) | |
| Other (Pelvic Mass Biopsy) | 0 (0) | 1 (0.1) | 0 (0) | 0 (0) | |
| Reoperation, n (%) | 3 (0.6) | 5 (0.6) | 0 (0) | 0 (0) | 0.911 |
| 30-day Readmission, n (%) | 21 (4.5) | 18 (2.1) | 0 (0) | 5 (6.8) | 0.021 |
| 30-Day Mortality, n (%) | 0 (0) | 1 (0.1) | 0 (0) | 0 (0) | 0.888 |
| Anastomotic Staple Line Leak, n (%) | 0 (0) | 1 (0.1) | 0 (0) | 0 (0) | 0.888 |
| Blood Transfusion Within 72 Hours from Incision, n (%) | 3 (0.6) | 9 (1.0) | 1 (10.0) | 0 (0) | 0.016 |
| VTE Requiring Therapy, n (%) | 0 (0) | 0 (0) | 0 (0) | 1 (1.3) | <0.001 |
| GI Bleed, n (%) | 3 (0.6) | 11 (1.3) | 1 (10.0) | 0 (0) | 0.023 |
Note: SD, Standard deviation; IQR, Interquartile range; IV, Intravenous; EGD, Esophagogastroduodenoscopy; VTE, Venous thromboembolism; GI, gastrointestinal. The P values hospital length of stays and number of IV visits are from Wilcoxon-rank sum tests, the p-values for categorical outcomes are from Chi-squared/Fisher exact tests.
Table 4.
Results from multivariable logistic regression model of having at least one 30-day Emergency Department visit.
| Variable | OR | 95% CI | p-value |
|---|---|---|---|
| Race | |||
| White (ref) | – | – | – |
| Black | 1.42 | (0.90, 2.23) | 0.133 |
| Asian | 0 | (0.00, ∞) | 0.985 |
| Other | 1.46 | (0.64, 3.31) | 0.364 |
| Age on date of surgery (per 1-year increase) | 0.97 | (0.95, 1.00) | 0.019 |
| Gender | |||
| Female (ref) | – | – | – |
| Male | 0.55 | (0.29, 1.06) | 0.075 |
| BMI closest to procedure (kg/m2, per 1-unit increase) | 0.97 | (0.94, 1.00) | 0.094 |
| Diabetes mellitus | |||
| No (ref) | – | – | – |
| Yes, insulin | 1.28 | (0.49, 3.34) | 0.609 |
| Yes, non-insulin | 1.02 | (0.57, 1.81) | 0.955 |
| Hyperlipidemia | |||
| No (ref) | – | – | – |
| Yes | 0.79 | (0.43, 1.43) | 0.432 |
| Number of anti-hypertensive medications | |||
| 0 (ref) | – | – | – |
| 1 | 1.38 | (0.81, 2.37) | 0.238 |
| 2 | 1.09 | (0.56, 2.11) | 0.805 |
| 3 or more | 1.45 | (0.67, 3.15) | 0.344 |
| GERD | |||
| No (ref) | – | – | – |
| Yes | 1.47 | (0.88, 2.45) | 0.14 |
| Immunosuppressive therapy | |||
| No (ref) | – | – | – |
| Yes | 1.65 | (0.60, 4.54) | 0.336 |
| Insurance type | |||
| Commercial (ref) | – | – | – |
| Medicare | 2.21 | (0.86, 5.66) | 0.098 |
| Medicaid | 0.69 | (0.20, 2.30) | 0.542 |
| Bariatric | 0.91 | (0.57, 1.47) | 0.705 |
| Self-pay | 0 | (0.00, ∞) | 0.993 |
| Marital status | |||
| Currently married/partnership (ref) | – | – | – |
| Single/separated | 1.19 | (0.78, 1.82) | 0.407 |
| Unknown | 0 | (0.00, ∞) | 0.993 |
| ADI national rank (per 1-unit increase) | 1 | (0.99, 1.01) | 0.755 |
| ASA classification | |||
| ASA II – Mild systemic disease (ref) | – | – | – |
| ASA III – Severe systemic disease | 1.92 | (0.94, 3.92) | 0.072 |
| ASA IV – Severe systemic disease threat to life | 2.59 | (0.47, 14.20) | 0.272 |
| Surgery type | |||
| Roux-en-Y Gastric Bypass (ref) | – | – | – |
| Sleeve Gastrectomy | 0.38 | (0.24, 0.61) | <0.001 |
Notes: Reference categories (“ref”) are shown for each factor. OR = odds ratio; CI = confidence interval; ADI = Area Deprivation Index; ASA = American Society of Anesthesiologists.
The thirty-day readmission status varied by race and ethnicity (Table 3), with rates of 4.5% in Black patients, 2.1% in White patients, 0% in Asian patients, and 6.8% in other patients (p = 0.021). The multivariable logistic regression model of (Table 5) showed that race and BMI were significantly associated with a higher 30-day readmission rate. Black patients had more than twice the odds of 30-day readmission (Adjusted OR 2.64, 95% CI 1.25–5.56; p =0.011). “Other” patients, which includes mostly Hispanic individuals, had more than four times higher odds of readmission (Adjusted OR 4.69, 95% CI 1.52–14.53 p =0.007). Surgery type once again played a role in clinical outcomes as patients who underwent sleeve gastrectomy had an odds ratio of 0.4 for readmission when compared to those undergoing gastric bypass (CI 0.19-0.82, p=0.012). Additionally, the increase in BMI closest to the procedure by one kg/m2 was associated with 6% higher odds of 30-day readmission (p=0.011).
Table 5.
Results from multivariable logistic regression model of having at least one 30-day readmission.
| Variable | OR | 95% CI | p-value |
|---|---|---|---|
| Race | |||
| White (ref) | – | – | – |
| Black | 2.64 | (1.25, 5.56) | 0.011 |
| Asian | 0 | (0.00, ∞) | 0.987 |
| Other | 4.69 | (1.52, 14.53) | 0.007 |
| Age on date of surgery (per 1-year increase) | 1.01 | (0.98, 1.05) | 0.457 |
| Gender | |||
| Female (ref) | – | – | – |
| Male | 0.26 | (0.06, 1.12) | 0.071 |
| BMI closest to procedure (kg/m2, per 1-unit increase) | 1.06 | (1.01, 1.11) | 0.011 |
| Diabetes mellitus | |||
| No (ref) | – | – | – |
| Yes, insulin | 0.75 | (0.15, 3.71) | 0.72 |
| Yes, non-insulin | 0.92 | (0.37, 2.30) | 0.859 |
| Hyperlipidemia | |||
| No (ref) | – | – | – |
| Yes | 2.18 | (0.91, 5.23) | 0.08 |
| Number of anti-hypertensive medications | |||
| 0 (ref) | – | – | – |
| 1 | 0.73 | (0.28, 1.89) | 0.517 |
| 2 | 0.8 | (0.29, 2.26) | 0.679 |
| 3 or more | 1.12 | (0.36, 3.49) | 0.839 |
| GERD | |||
| No (ref) | – | – | – |
| Yes | 1.34 | (0.60, 3.00) | 0.468 |
| Immunosuppressive therapy | |||
| No (ref) | – | – | – |
| Yes | 2.86 | (0.76, 10.84) | 0.121 |
| Insurance type | |||
| Commercial (ref) | – | – | – |
| Medicare | 0.67 | (0.14, 3.24) | 0.617 |
| Medicaid | 1.09 | (0.24, 4.91) | 0.91 |
| Bariatric | 0.39 | (0.15, 1.04) | 0.059 |
| Self-pay | 0 | (0.00, ∞) | 0.993 |
| Marital status | |||
| Currently married/partnership (ref) | – | – | – |
| Single/separated | 1.71 | (0.85, 3.44) | 0.133 |
| Unknown | 0 | (0.00, ∞) | 0.994 |
| ADI national rank(per 1-unit increase) | 1 | (0.98, 1.02) | 0.831 |
| ASA classification | |||
| ASA II – Mild systemic disease (ref) | – | – | – |
| ASA III – Severe systemic disease | 0.53 | (0.20, 1.43) | 0.212 |
| ASA IV – Severe systemic disease threat to life | 0.62 | (0.06, 6.84) | 0.698 |
| Surgery type | |||
| Roux-en-Y Gastric Bypass (ref) | – | – | – |
| Sleeve Gastrectomy | 0.4 | (0.19, 0.82) | 0.012 |
Notes: Reference categories (“ref”) are shown for each factor. OR = odds ratio; CI = confidence interval; ADI = Area Deprivation Index; ASA = American Society of Anesthesiologists.
Hospital Length of Stay (LOS) differed significantly by race and ethnicity (p < 0.001, Table 3). White patients had the shortest mean LOS (1.34 days, SD 0.71), followed by the “other” (1.46 days, SD 0.53), Black (1.55 days, SD 0.76), and Asian (1.60 days, SD 1.07) patients. Log-normal regression (Table 6) indicated that Black patients had a LOS 13% longer than the LOS of White patients (time ratio 1.13, 95% CI 1.08–1.1; p < 0.001). In addition to Black patients, and “other” patients also had longer stays (time ratio 1.1, 95% CI 1.01–1.21; p = 0.036). Other factors influencing length of stay include Commercial insurance (time ratio 1.08, 95% CI 1.03-1.13, p<0.002). Patient factors predictive of shorter stays included sleeve gastrectomy (time ratio 0.95, 95% CI 0.90-1.00, p=0.033) and ASA III, indicative of severe systemic disease (time ratio 0.93, 95% CI 0.87-0.99, p=0.028)
Table 6.
Results from multivariable log normal regression model of hospital length of stay.
| Variable | Time ratio (e^β) | 95% CI | p-value |
|---|---|---|---|
| Race | |||
| White (ref) | – | – | – |
| Black | 1.13 | (1.08, 1.18) | <0.001 |
| Asian | 1.15 | (0.89, 1.48) | 0.274 |
| Other | 1.1 | (1.01, 1.21) | 0.036 |
| Age on date of surgery (per 1-year increase) | 1 | (1.00, 1.00) | 0.608 |
| Gender | |||
| Female (ref) | – | – | – |
| Male | 0.97 | (0.91, 1.02) | 0.225 |
| BMI closest to procedure (kg/m2, per 1-unit increase) | 1 | (1.00, 1.01) | 0.136 |
| Diabetes mellitus | |||
| No (ref) | – | – | – |
| Yes, insulin | 1.05 | (0.95, 1.17) | 0.355 |
| Yes, non-insulin | 1.02 | (0.96, 1.08) | 0.612 |
| Hyperlipidemia | |||
| No (ref) | – | – | – |
| Yes | 0.98 | (0.92, 1.04) | 0.463 |
| Hypertension (yes) | 1.03 | (0.98, 1.08) | 0.298 |
| GERD (yes) | 1.03 | (0.97, 1.10) | 0.305 |
| Immunosuppressive therapy (yes) | 1.1 | (0.97, 1.25) | 0.151 |
| Insurance type | |||
| Commercial | 1.08 | (1.03, 1.13) | 0.002 |
| Medicaid | 1.02 | (0.91, 1.13) | 0.787 |
| Medicare | 1.11 | (0.99, 1.24) | 0.079 |
| Self-pay | 1.1 | (0.64, 1.87) | 0.736 |
| Marital status | |||
| Currently married/partnership (ref) | – | – | – |
| Single/separated | 1.01 | (0.97, 1.06) | 0.579 |
| Unknown | 1.33 | (0.78, 2.28) | 0.29 |
| ADI national rank (per 1-unit increase) | 1 | (1.00, 1.00) | 0.889 |
| ASA classification | |||
| ASA II – Mild systemic disease (ref) | – | – | – |
| ASA III – Severe systemic disease | 0.93 | (0.87, 0.99) | 0.028 |
| ASA IV – Severe systemic disease threat to life | 0.87 | (0.72, 1.03) | 0.112 |
| Surgery type | |||
| Roux-en-Y Gastric Bypass (ref) | – | – | – |
| Sleeve Gastrectomy | 0.95 | (0.90, 1.00) | 0.033 |
Note: Notes: Reference categories (“ref”) are shown for each factor. For insurance type, all other types are used as reference, for example, for Commercial all non-commercial are reference. Time ratios (e^β) >1 denote a longer length of stay; <1 denote a shorter stay. ADI = Area Deprivation Index; ASA = American Society of Anesthesiologists.
Mortality within 30 days of surgery was very low, with only one White patient (0.1%) experiencing a 30-day postoperative death (p = 0.888) (Table 3). Reoperation rates were also low across all racial groups (0–0.6%, p = 0.911). Blood transfusions (beyond the 72-hour window) were infrequent. Ambulatory intravenous fluid (IV) treatment was used more often by Black patients compared to White patients. Forty-six Black patients (10%) utilized ambulatory IV treatment, compared to 23 White patients (2.7%) (p < 0.001).
Discussion:
Our institution recognized potential disparities in patient outcomes following bariatric surgery and sought to identify risk factors. We initially identified higher rates of 30-day readmissions, and longer LOS in racial and ethnic minority patients hypothesizing that underlying medical comorbidities or higher complication rates would lead to the increased utilization of postoperative resources.
Our results revealed that preoperative risk factors were largely similar across all racial and ethnic minority groups. Regarding postoperative complications, the groups showed minimal differences. A notable difference in our patients is a lower rate of gastric bypass in Black patients. Previous studies have demonstrated gastric bypass having a lower risk of readmission compared with that of sleeve gastrectomy [14]. Given our Black cohort has the lowest rate of gastric bypass this would theorize a lower rate of readmission. Despite these characteristics of the groups, our study found that Black patients exhibited a longer LOS, increased use of IV fluid treatment, and increased likelihood of readmission within 30 days. After conducting rigorous analysis, we identified that race and ethnicity truly played a crucial association with these differences. Interestingly, we identified that age acted as a protective factor for ED visits, and that gastric bypass was a factor predicting increased ED visitations, consistent with previous literature [15]. After analyzing our cohorts considering these preoperative risk factors, we found that Black patients used more postoperative resources, including longer hospital stays, more frequent 30-day rehospitalizations, and higher use of outpatient IV infusion centers. Given these findings, we hypothesized that social determinants of health and other key components related to patient navigation of the healthcare system may be influencing outcomes. Our analysis of patient social determinants of health was multifaceted and included their partner status, and SVI and ADI rankings. ADI and SVI were found to be higher for Black patients. Additionally, Black patients were more likely to be single. Based on these data points, our Black cohort of patients were more likely predisposed to unfavorable social determinants of health. Despite these findings, our regression models did not show that these SDOH independently influenced ED visits, LOS, or readmission rates. However, including partner status, ADI, and SVI in the model did not explain the strong association between race and outcomes, as evidenced by the significant adjusted odds ratio for 30-day readmission and time-ratio for LOS.
Race and social determinants of health in bariatrics remain an understudied field that is drawing more attention in recent years. In 2021, Zhao et al published a systematic review of 52 studies which ultimately found less weight loss in Black patients when compared with non-Hispanic White patients but no association between race and the resolution of comorbidities, an important outcome in bariatric patients [11]. In 2019, Wood et al published in Jama Surgery a cohort of over 7000 patients where Black patients were found to have a higher rate of any complications, ED visit rate, readmission rate, and a longer LOS [10].
To further elucidate the differences in race, we completed an analysis adjusting for SVI, ADI, and marital status. These well-established metrics work as proxies to attempt to adjust for social determinants of health. Despite these adjustments, Black patients continued to demonstrate higher resource utilization. This raises an important consideration that once accounting for indicators of the patient’s likelihood to succeed in surgical treatment of obesity, one key component that determines the patient utilization of resources is race. Further work is needed to further parse out other aspects of patient postoperative care such as their understanding of their own success in patient reported outcomes measures as well as more qualitative in-depth interviews or focus group analyses. However, from this level, the data shows that in our cohort patients had higher resource utilization despite all adjustments, near similar medical comorbidities and post-operative complications.
Black patients were using ambulatory IV fluid treatment at a nearly fourfold higher rate than White patients. Despite higher use of this care tool, Black patients continued to have significantly higher resource utilization as a whole which warrants further investigation into the outcomes of the specific patients who receive outpatient IV fluid treatment as well as the ability of those who visit the ED or get readmitted to navigate the healthcare system and receive this therapy. Challenges in healthcare navigation are systemic, especially in racial minority patients. While the proxies of ADI and SVR give insight to the patient resources of different zip codes, there remains significant work to be done to improve access to care. Many of the patients included in this study are local to the healthcare system’s catchment area in Northern Delaware, near the city of Wilmington where 23% of the population falls below the poverty line, multiple large areas meet criteria for low supermarket access, and nearly half of the population lack effective transportation support [16]. Within this region alone, there is significant work needed on a system level to improve access to care for patients, at a provider level to continue advocacy work, and to ensure that the community and its leaders have adequate purchase and voice.
Improved communication at the patient level may enhance understanding of postoperative care expectations, leading to a smoother recovery and ultimately supporting more equitable and efficient resource utilization. Numerous studies show the lack of understanding that patients have both prior to and after having a procedure performed [17]. Continuing to improve cultural competency among healthcare providers and strengthening partnerships with community leaders could empower patients and increase satisfaction with care. Working with both patient reported outcome measures and patient reported experience measures may further help to elucidate these differences. These measures may show even further insight into the perception of care; an ever more important aspect in understanding patients’ navigation of the healthcare system, particularly in underserved populations.
There are several limitations to this study, including its single-institution, retrospective design. While this study does highlight different racial groups’ utilization of healthcare resources, further studies on larger databases or multi-center studies may continue to elucidate this relationship. Within our analysis we used ADI/SVI as social determinants of health, and these variables are measured at the census tract level and not at the individual level. More granular measures of social determinants of health such as individual surveys asking about food insecurity, ease of transportation, and access to medications are warranted. Future studies should be carried out to more fully understand a patient’s social determinants of health.
Ongoing research is crucial to the advancement of bariatric and obesity medicine as well as more generally in healthcare to better support our patients and ensure that they receive the care that they need in an effective and equitable way. Our study showcases the impact of disparities on healthcare. To better understand and address these problems we must continually improve the health care for our patients and continue to fight disparities and achieve equity in healthcare.
Acknowledgements
Christiana Care Surgery Strategic Aspiration Workgroup: Brittany Anderson, Dan Grawl, Patty Blair, Read Siry, Drew Brady, Eric Johnson, Madalene Zale, William Bryant, Theresa Mead, Stephanie De Mel, Erin Booker, Jacqueline Ortiz, Mike Clagett, Nick Patel, Jennifer Czerwinski, Matthew Powell, Sherri Ferry, Claudia Reyes-Hull.
The authors would like to thank James T. Laughery for his contributions to data collection.
Work supported in part by an Institutional Development Award (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health under grant number U54-GM104941 (PI: Hicks) and the State of Delaware. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
Disclosures
Dr. Halbert has received honorarium for consulting with Medtronic.
Connor Magura, Keshab Subedi, Claudine Jurkovitz, Elizabeth McCarthy, Arielle Brackett, Robin Ellis, John Getchell, Matthew Rubino have no disclosures.
References
- 1-.Li M, Gong W, Wang S, Li Z (2022) Trends in body mass index, overweight and obesity among adults in the USA, the NHANES from 2003 to 2018: a repeat cross-sectional survey. BMJ Open, 12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2-.Byrd AS, Toth AT, Stanford FC (2018) Racial disparities in obesity treatment. Curr Obes Rep 7:130–138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3-.Johnson VR, Acholonu NO, Dolan AC, Krishnan A, Wang EH, Stanford FC (2021) Racial disparities in obesity treatment among children and adolescents. Curr Obes Rep, 10:342–350. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4-.Nafiu OO, Mpody C, Michalsky MP, Tobias JD (2021) Unequal rates of postoperative complications in relatively healthy bariatric surgical patients of White and black race. Surg Obes Relat Dis, 17:1249–1255 [DOI] [PubMed] [Google Scholar]
- 5-.Welsh LK, Luhrs AR, Davalos G, Diaz R, Narvaez A, Perez JE, Lerebours R, Kuchibhatla M, Portenier DD, Guerron AD (2020) Racial disparities in bariatric surgery complications and mortality using the MBSAQIP data registry. Obes Surg, 30:3099–3110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6-.Reducing disparities in health care | Causes of health disparity | AMA; (n.d.) Available at: https://www.ama-assn.org/delivering-care/health-equity/reducing-disparities-health-care. Oct 23, 2023; Accessed 12 January 2025 [Google Scholar]
- 7-.Smedley BD, Stith AY, Nelson AR (2003) Unequal treatment: confronting racial and ethnic disparities in health care. National Academies Press, Washington, DC. [PubMed] [Google Scholar]
- 8-.Turcotte JJ, Brennan JC, Holbert SE, Dolle SS, King PJ (2022) Enhanced preoperative education pathways: a step toward reducing disparities in total joint arthroplasty outcomes. J Arthroplasty, 37:1233–1240 [DOI] [PubMed] [Google Scholar]
- 9 -.American Medical Association (n.d.) Reducing health care disparities: understanding and addressing inequities in our health care system. Available at: https://www.ama-assn.org/sites/ama-assn.org/files/corp/media-browser/public/public-health/reducing-health-care-disparities-report_1.pdf. March 2014; Accessed 12 March 2025.
- 10 -.Wood MH, Carlin AM, Ghaferi AA, Varban OA, Hawasli A, Bonham AJ, Birkmeyer NJ, Finks JF (2019) Association of race with bariatric surgery outcomes. JAMA Surg, 154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11-.Zhao J, Samaan JS, Abboud Y, Samakar K (2021) Racial disparities in bariatric surgery postoperative weight loss and co-morbidity resolution: a systematic review. Surg Obes Relat Dis, 17:1799–1823. [DOI] [PubMed] [Google Scholar]
- 12-.Forefront Group (2023). The Area Deprivation Index is the most scientifically validated social exposome tool available for policies advancing health equity. DOI: 10.1377/FOREFRONT.20230714.676093, July 20, 2023. [DOI] [Google Scholar]
- 13-.Social Vulnerability Index | Place and Health - Geospatial Research, Analysis, and Services Program (GRASP) | ATSDR; (n.d.) Available at: https://www.atsdr.cdc.gov/place-health/php/svi/index.html. July 22, 2024; Accessed 12 January 2025. [Google Scholar]
- 14-.Ahmed MA, Zhang Z, Ramamoorthy V, Saxena A, Rubens M, Appunni S, McGranaghan P, Jalal AH, Veledar E. (2023) Difference in 30-Day Readmission Rates After Laparoscopic Sleeve Gastrectomy Versus Laparoscopic Roux-En-Y Gastric Bypass: a Propensity Score Matched Study Using ACS NSQIP Data (2015–2019). Obes Surg. 33:1040–1048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15-.Obeid T, Krishnan A, Abdalla G, Schweitzer M, Magnuson T, Steele KE (2016). GERD Is Associated with Higher Long-Term Reoperation Rates After Bariatric Surgery. J. Gastrointest. Surg, 20: 119–124. [DOI] [PubMed] [Google Scholar]
- 16-.Moselle A (2024) Racial inequities found in Delaware’s regional transportation network. WHYY. [Google Scholar]
- 17-.Schenker Y, Fernandez A, Sudore R, Schillinger D (2011). Interventions to improve patient comprehension in informed consent for medical and surgical procedures: a systematic review. Med Decis Making, 31:151–173. [DOI] [PMC free article] [PubMed] [Google Scholar]