Abstract
Background
Community-level sociodemographic factors and hospital quality are associated with access to cancer care and outcomes.
Methods
Using Medicare data (2016–2018), we evaluated the association between social determinants of health (SDoH) and hospital quality with 30-day mortality and readmission after selected elective cancer surgery. Separate multivariable logistic regression models sequentially adjusted for comorbidities and SDoH factors (Social Vulnerability Index (SVI) and Distressed Communities Index (DCI)) and then Hospital Star Rating.
Results
Among 16,869 patients, the “At Risk” DCI group and higher SVI were associated with higher 30-day mortality, and higher SVI was associated with higher odds of 30-day readmission and mortality. Subsequent models demonstrated that higher Hospital Star Rating was associated with lower odds of 30-day mortality and readmission and SDoH factors lost significance after adjusting for Hospital Star Rating.
Conclusions
Hospital quality may have a greater impact on short-term outcomes than SDoH factors.
Keywords: Social Determinants of Health, Hospital Star Rating, Cancer surgery, short-term outcome
Graphical Abstract
INTRODUCTION
Hospital quality has been found to be associated with surgical outcomes. (1) There is emerging data that county-level social vulnerability and zip-code level community distress are also associated with surgical outcomes. (2–5) It is unclear if the association of these factors with short-term outcomes following cancer operations is additive. Medicare beneficiaries, due to their older age and higher prevalence of comorbidities, might be more vulnerable to the potential negative effects from disparities in access to cancer care. Elective cancer surgical procedures might serve as a critical moment in patients’ lives for studying how hospital quality and community-level social determinants of health (SDoH) are associated with short-term surgical outcomes. SDoH indices were developed to assess how broader environmental factors influence a population’s ability to respond to disasters or serious health conditions. SDoH factors include socioeconomic status (SES), race/ethnicity, education, employment, and access to transportation, which can impact the health of a population living a specific community/geographic unit. (6–10)
Two key geographic measures of SDoH that have been previously reported in the literature are the county-level Social Vulnerability Index (SVI) and the zip code-level Distressed Communities Index (DCI). (11,12) The SVI is calculated based on 16 variables across 4 subgroups: SES, household characteristics, racial and ethnic minority status, and household type & transportation. (11) On the other hand, the DCI incorporates seven main factors—education level, housing vacancy rates, unemployment rates, poverty rates, income, changes in employment, and changes in establishment—to assess disparities at the zip code level. (12) Recent studies have shown that individuals living in more vulnerable areas face a higher risk of developing chronic diseases and that they have worse outcomes after complex surgical procedures, including cancer surgical procedures. (4,5,13–17) In addition, the hospital quality where patients receive care plays a significant role in patient outcomes. (18,19) The Overall Hospital Star Rating, calculated by the Centers for Medicare and Medicaid Services (CMS), assesses hospital quality and has been found to be useful in discriminating between hospitals regarding the quality of care that patients receive (20). This rating is based on five components: mortality, safety, readmission, patient experience, and timely & effective care. Only 10.1% of hospitals are categorized as 5-star. Although the CMS star rating is not disease- or population-specific, it has been shown to correlate with safety across various conditions and care settings.(18) This information could be useful for patients and referring physicians when selecting where patients receive their care. Additionally, discriminating hospitals according to quality of care they provide could also inform strategies and/or interventions to improve the quality of lower performing hospitals.
The aim of this study was to compare short-term postoperative outcomes (30-day mortality and 30-day readmission) among Medicare beneficiaries who underwent elective cancer surgery based on the CMS overall Hospital Star Rating of the hospital where they received care and the county-level SVI and zip-code level DCI of where they lived. Additionally, this study aimed to evaluate whether hospital quality can mitigate the adverse effects of social vulnerability on short-term postoperative outcomes among patients from disadvantaged communities.
METHODS
The Medicare data was used for this retrospective analysis to identify patients who received care between January 1, 2016, and December 31, 2018. This study was approved by the Mass General Brigham Institutional Review Board. This study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
The Inpatient file was used to identify patients who were greater than or equal to 65 years of age at the time of an admission for an elective cancer surgical procedure that included liver resection, pancreatectomy, esophagectomy, and gastrectomy. Patients were identified using ICD10-CM codes to identify a cancer diagnosis and then an ICD10-PCS code (supplementary 1) to identify corresponding surgical procedure for the identified cancer diagnosis. For patients who underwent more than one surgery during the study period (2016–2018), we only kept the first operation that was listed. Same-day surgical procedures, in which the patient was discharged on the same day as the date of surgery, were excluded. We excluded patients who received surgery after November 30, 2018, so we could account for our 30-day outcomes of interest. We also excluded patients that left the hospital against medical advice, were still admitted at the hospital as of December 31, 2018, or who were reported as having discontinued their care.
The Medicare data was merged with the Hospital Rating Star Data (Version 2021 Apr) based on provider ID. The data were then linked to county level SVI data (year 2018) using FIPS code, with a crosswalk used to map social security administration (SSA) code from Medicare to FIPS code. (21) The data was then merged with DCI data by zip code.
Outcomes of Interest
The 30-day readmission was defined as any unplanned inpatient readmission within 30 days from the discharge date of the index operation. The 30-day postoperative mortality was defined as death that occurred during the index admission (i.e., in-hospital mortality) and following surgery or within 30 days from the date of surgery (irrespective of location of death).
Exposures of Interest
We used two measures of SDoH: 1) county-level Social Vulnerability Index (SVI), measured on a scale from 0 (least vulnerable) to 1 (most vulnerable); and 2) zip-code Distressed Communities Index (DCI), which categorized patients into five groups: prosperous, comfortable, mid-tier, at risk, and distressed. The CMS Overall Hospital Star Rating ranged from 1 (worst) to 5 (best).
Covariates
Covariates included age, sex, race, operation type, dementia (22) (see Supplementary 2 for list of ICD-10-CM codes for dementia), and Charlson Comorbidity Index (CCI). CCI was calculated using the comorbidity package from R and was based on the diagnosis code from index surgery admission. (23) The cancer diagnoses were excluded from the CCI score because cancer was the index condition for our study cohort. The CCI score was categorized as 0, 1, and ≥2.
Statistical Analysis
We performed descriptive statistics using the median (IQR) for continuous variables and counts and percentages for categorical variables. For comparing continuous variables, the Mann-Whitney test or Kruskal-Wallis test was employed as appropriate, and the Chi-square test was used for categorical variables. Short-term outcomes were evaluated using separate multivariable logistic regression models that sequentially adjusted for comorbidities and each SDoH (Models 1 and 3) and then the Hospital Star Rating (Models 2 and 4). Model 5 was adjusted for Hospital Star Rating and other covariates without adjusting for SVI and DCI. An ordinal logistic regression model was used to evaluate the association between each SDoH factor and Hospital Star Rating. A two-sided P-value of <0.05 was selected as signifying statistical significance.
The SAS 9.4 statistical program was used for data cleaning and Stata 18 was used for analysis. CCI scores were calculated using R version 4.4.1.
RESULTS
Among 16,869 patients who met the inclusion criteria (Figure 1), 12.5%, 14.8%, 24.1%, and 48.6% underwent esophagectomy, gastrectomy, liver resection, or pancreatectomy, respectively. The 30-day mortality and readmission rates for the whole cohort were 4.1% and 21.9%, respectively. Older age, male sex, open surgery, and higher CCI score were associated with higher incidence of 30-day mortality (Table 1). Factors associated with higher incidence of 30-day readmission included older age, male sex, race, open surgery, and higher CCI score. Although patients who received care at 5-star hospitals experienced lower rates of 30-day mortality compared to 1-star hospital (3.03% vs 6.62%, p <0.001), similar findings were observed for 30-day readmission rates. Patients who underwent surgery at 5-star hospitals also had lower 30-day readmission rates (19.97% vs 24.46%, p < 0.001). There were no significant differences in 30-day mortality or readmission across DCI groups. Similarly, SVI did not differ between patients who experienced 30-day mortality or readmission and those who did not.
Figure 1.
Consort diagram for patients who met the inclusion criteria
Table 1.
Comparison of baseline characteristics between groups stratified by outcomes.
| 30-day mortality | 30-day readmission | |||||
|---|---|---|---|---|---|---|
| Survived 16,171 (95.86%) | Deceased 698 (4.14%) | P | Non readmitted 13,171 (78.08%) | Readmitted 3,698 (21.92%) | P | |
| Age* Median (IQR) | 72 (68–77) | 75 (70–79) | <0.001 | 72 (68–77) | 73 (69–78) | <0.001 |
| Gender ** | <0.001 | 0.001 | ||||
| Male | 9,325 (95.35%) | 455 (4.65%) | 7,551 (77.21%) | 2,229 (22.79%) | ||
| Female | 5,846 (96.57%) | 243 (3.43%) | 5,620 (79.28% | 1,469 (20.72%) | ||
| Race ** | 0.088 | 0.001 | ||||
| White | 13,465 (95.74%) | 599 (4.26%) | 10,910 (77.57%) | 3,154 (22.43%) | ||
| Black | 1,255 (95.73%%) | 56 (4.27%) | 1,034 (78.87%) | 277 (21.13%) | ||
| Asian | 451 (97.62%) | 11 (2.38%) | 390 (84.42%) | 72 (15.58%) | ||
| Hispanic | >224 (>95.32%) | <11 (<4.68%) | 194 (82.55%) | 41 (17.45%) | ||
| North American Native | >53 (>82.81%) | <11 (<17.19%) | 49 (76.56%) | 15 (23.44%) | ||
| Other | 365 (96.05%) | 15 (3.95%) | 299 (78.68% | 81 (21.32%) | ||
| Unknown | >342 (>96.88) | <11 (<3.12%) | 295 (83.57%) | 58 (16.43%) | ||
| Operation type ** | 0.001 | <0.001 | ||||
| Open | 13,053 (95.62%) | 598 (4.38%) | 10,531 (77.14%) | 3,120 (22.86%) | ||
| Minimally invasive | 3,118 (96.89%) | 100 (3.11%) | 2,640 (82.04%) | 578 (17.96%) | ||
| Type of surgery ** | <0.001 | <0.001 | ||||
| Esophagectomy | 1,962 (93.38%) | 139 (6.62%) | 1,560 (74.25%) | 541 (25.75%) | ||
| Gastrectomy | 2,382 (95.20%) | 120 (4.80%) | 1,996 (79.78%) | 506 (20.22%) | ||
| Liver resection | 3,934 (96.85%) | 128 (3.15%) | 3,419 (84.17%) | 643 (15.83%) | ||
| Pancreatectomy | 7,893 (96.21%) | 311 (3.79%) | 6,196 (75.52%) | 2,008 (24.48%) | ||
| CCI ** | <0.001 | <0.001 | ||||
| 0 | 7,038 (97.09%) | 211 (2.91%) | 5,822 (80.31%) | 1,427 (19.69%) | ||
| 1 | 4,923 (96.23%) | 193 (3.77%) | 3,988 (77.95%) | 1,128 (22.05%) | ||
| ≥2 | 4,210 (93.47%) | 294 (6.53%) | 3,361 (74.62%) | 1,143 (25.38%) | ||
| Dementia ** | 0.098 | 0.052 | ||||
| No | 15,946 (95.89%) | 683 (4.11%) | 12,996 (78.15%) | 3,633 (21.85%) | ||
| Yes | 225 (93.75%) | 15 (6.25%) | 175 (72.92%) | 65 (27.08%) | ||
| Hospital Star Rating ** | <0.001 | <0.001 | ||||
| Star 1 | 607 (93.38%) | 43 (6.62%) | 491 (75.54%) | 159 (24.46%) | ||
| Star 2 | 2,020 (95.06%) | 105 (4.94%) | 1,660 (78.12%) | 465 (21.88%) | ||
| Star 3 | 4,107 (95.20%) | 207 (4.80%) | 3,282 (76.08%) | 1,032 (23.92%) | ||
| Star 4 | 6,042 (96.23%) | 237 (3.77%) | 4,936 (78.61%) | 1,343 (21.39%) | ||
| Star 5 | 3,395 (96.97%) | 106 (3.03%) | 2,802 (80.03%) | 699 (19.97%) | ||
| DCI ** | 0.21 | 0.43 | ||||
| Prosperous | 4,625 (96.23%) | 181 (3.77%) | 3,773 (78.51%) | 1,033 (21.49%) | ||
| Comfortable | 3,795 (96.05%) | 156 (3.95%) | 3,114 (78.82%) | 837 (21.18%) | ||
| Mid-tier | 3,096 (95.91%) | 132 (4.09%) | 2,510 (77.76%) | 718 (22.24%) | ||
| At risk | 2,657 (95.30%) | 131 (4.70%) | 2,157 (77.37%) | 631 (22.63%) | ||
| Distressed | 1,998 (95.32%) | 98 (4.68%) | 1,617 (77.15%) | 479 (22.85%) | ||
| SVI* Median (IQR) | 0.50 (0.28–0.69) | 0.52 (0.31–0.70) | 0.09 | 0.50 (0.28–0.68) | 0.51 (0.29–0.69) | 0.10 |
CCI: Charlson Comorbidity Index; IQR: Inter Quantile Range
Mann-Whitney test
Chi-test
The unadjusted model demonstrated that patients from “At Risk” communities had higher odds of 30-day mortality (OR 1.26, 95% CI 1.00–1.58, p=0.049, compared to the “Prosperous” group, which was designated as the reference; Table 2). However, there was no statistically significant difference between DCI groups and 30-day readmission. The SVI was neither associated with 30-day mortality nor with 30-day readmission. However, patients who underwent surgery at hospitals rated as 5 stars had lower odds of 30-day mortality (OR 0.44, 95% CI 0.31–0.63, p<0.001) and readmission (OR 0.77, 95% CI 0.63–0.94, p=0.01).
Table 2.
Unadjusted odds ratio for the association between SVI, DCI, and Hospital Star Rating with 30-day mortality and 30-day readmission
| 30-day mortality | 30-day readmission | ||||||
|---|---|---|---|---|---|---|---|
| Unadjusted OR | 95% CI | P | Unadjusted OR | 95% CI | P | ||
| DCI | Prosperous | Reference | Reference | ||||
| Comfortable | 1.05 | 0.84–1.31 | 0.66 | 0.98 | 0.88–1.09 | 0.72 | |
| Mid-tier | 1.09 | 0.87–1.37 | 0.46 | 1.04 | 0.94–1.16 | 0.42 | |
| At risk | 1.26 | 1.00–1.58 | 0.049 | 1.07 | 0.95–1.19 | 0.24 | |
| Distressed | 1.25 | 0.97–1.61 | 0.08 | 1.08 | 0.96–1.22 | 0.21 | |
| SVI | 1.33 | 0.98–1.79 | 0.07 | 1.13 | 0.98–1.31 | 0.09 | |
| Hospital Star Rating | Star 1 | Reference | Reference | ||||
| Star 2 | 0.73 | 0.51–1.06 | 0.098 | 0.86 | 0.70–1.06 | 0.17 | |
| Star 3 | 0.71 | 0.51–0.99 | 0.049 | 0.97 | 0.80–1.18 | 0.76 | |
| Star 4 | 0.55 | 0.39–0.77 | 0.001 | 0.84 | 0.69–1.01 | 0.07 | |
| Star 5 | 0.44 | 0.31–0.63 | <0.001 | 0.77 | 0.63–0.94 | 0.01 | |
Distribution of DCI/SVI between Hospital Star Rating groups were showed in supplementary 3 and 4. Patients receiving their care at the highest rated hospitals (Hospital Star Rating of 5), had the lowest median SVI score (i.e., patients living in less socially vulnerable communities). With regards to DCI, patients receiving their care at the highest rated hospitals had the highest proportion of patients residing in prosperous communities.
Five multivariable models adjusted for all variables are presented in Table 3. In model 1 (without Hospital Star Rating adjustment), the “At Risk” DCI group was associated with higher 30-day mortality (OR 1.28, 95% CI 1.01–1.60, p=0.04). This model showed no significant difference in 30-day readmission among DCI groups. After adjusting for Hospital Star Rating in model 2, there were no significant differences observed between DCI groups and 30-day mortality or 30-day readmission. This model demonstrated that patients who received care at hospitals rated as 3, 4, or 5 stars had lower odds of experiencing 30-day mortality compared to patients who received care at hospitals rated as 1 star. Moreover, compared to patients treated at hospitals with a 1-star rating, patients treated at hospitals rated as a 5-star had lower odds of 30-readmission (OR 0.77, 95% CI 0.63–0.95, P=0.01).
Table 3.
Adjusted odds ratio for the association between DCI and SVI with 30-day mortality and readmission
| 30-day mortality | 30-day readmission | |||
|---|---|---|---|---|
| OR (95% CI) * | p-value | OR (95% CI) * | p-value | |
| DCI Model 1 (Multivariable model without adjusting for Hospital Star Rating) | ||||
| DCI | ||||
| Prosperous | 1 (Ref.) | 1 (Ref.) | ||
| Comfortable | 1.04 (0.83–1.29) | 0.73 | 0.98 (0.88–1.08) | 0.67 |
| Mid-tier | 1.08 (0.85–1.36) | 0.51 | 1.04 (0.94–1.16) | 0.43 |
| At risk | 1.28 (1.01–1.60) | 0.04 | 1.07 (0.96–1.20) | 0.20 |
| Distressed | 1.28 (0.99–1.65) | 0.06 | 1.10 (0.97–1.25) | 0.12 |
| DCI Model 2 (Multivariable model that adjusts for Hospital Star Rating) | ||||
| DCI | ||||
| Prosperous | 1 (Ref.) | 1 (Ref.) | ||
| Comfortable | 1.03 (0.82–1.28) | 0.80 | 0.97 (0.88–1.08) | 0.59 |
| Mid-tier | 1.05 (0.83–1.32) | 0.69 | 1.03 (0.92–1.15) | 0.59 |
| At risk | 1.21 (0.96–1.52) | 0.11 | 1.05 (0.94–1.19) | 0.35 |
| Distressed | 1.15 (0.89–1.49) | 0.29 | 1.07 (0.94–1.21) | 0.32 |
| Hospital Star Rating | ||||
| Star 1 | 1 (Ref.) | 1 (Ref.) | ||
| Star 2 | 0.70 (0.48–1.01) | 0.06 | 0.85 (0.69–1.04) | 0.12 |
| Star 3 | 0.70 (0.50–0.99) | 0.04 | 0.98 (0.80–1.18) | 0.81 |
| Star 4 | 0.55 (0.39–0.77) | 0.001 | 0.84 (0.69–1.02) | 0.08 |
| Star 5 | 0.44 (0.30–0.63) | <0.001 | 0.77 (0.63–0.95) | 0.01 |
| SVI Model 3 (Multivariable model without adjusting for Hospital Star Rating) | ||||
| SVI | 1.38 (1.01–1.88) | 0.04 | 1.17 (1.01–1.36) | 0.03 |
| SVI Model 4 (Multivariable model that adjusts for Hospital Star Rating) | ||||
| SVI | 1.19 (0.87–1.63) | 0.26 | 1.13 (0.97–1.31) | 0.11 |
| Hospital Star Rating | ||||
| Star 1 | 1 (Ref.) | 1 (Ref.) | ||
| Star 2 | 0.71 (0.49–1.04) | 0.08 | 0.86 (0.70–1.06) | 0.15 |
| Star 3 | 0.71 (0.50–1.01) | 0.06 | 0.99 (0.81–1.20) | 0.91 |
| Star 4 | 0.55 (0.39–0.78) | 0.001 | 0.85 (0.70–1.03) | 0.10 |
| Star 5 | 0.44 (0.30–0.64) | <0.001 | 0.78 (0.64–0.96) | 0.02 |
| Model 5 (Multivariable model that adjusts for Hospital Star Rating without adjusting for SVI/DCI) | ||||
| Hospital Star Rating | ||||
| Star 1 | 1 (Ref.) | 1 (Ref.) | ||
| Star 2 | 0.69 (0.48–1.01) | 0.054 | 0.84 (0.69–1.04) | 0.11 |
| Star 3 | 0.69 (0.49–0.98) | 0.04 | 0.97 (0.80–1.17) | 0.75 |
| Star 4 | 0.53 (0.38–0.75) | <0.001 | 0.83 (0.69–1.00) | 0.054 |
| Star 5 | 0.42 (0.29–0.61) | <0.001 | 0.76 (0.62–0.93) | 0.01 |
All models were adjusted with age, gender, race, operation type, CCI, and dementia.
Model 3 demonstrated that higher SVI was associated with higher odds of 30-day mortality and 30-day readmission. However, SVI was no longer significant after adjusting for Hospital Star Rating in model 4. Similar to model 2, patients who underwent surgery at higher-rated hospitals had lower odds of 30-day mortality (Star 5 vs. Star 1: OR 0.44, 95% CI 0.30–0.64, P<0.001) and 30-day readmission (Star 5 vs. Star 1: OR 0.78, 95% CI 0.64–0.96, P=0.02). In model 5 (without adjusting for SVI and DCI), receiving care at higher-rated hospital was associated with lower rates of 30-readamission and mortality.
Patients from distressed communities (p < 0.05 for all DCI groups) or from more socially vulnerable counties (p <0.001) had lower odds of receiving surgery at higher quality rated hospitals (Table 4).
Table 4.
Ordinal Logistic Regression Analysis of Hospital Star Rating and each SDoH variable (SVI and DCI)
| OR | 95% CI | P | ||
|---|---|---|---|---|
| DCI | Prosperous | Reference | ||
| Comfortable | 0.91 | 0.84–0.98 | 0.01 | |
| Mid-tier | 0.75 | 0.69–0.81 | <0.001 | |
| At risk | 0.58 | 0.53–0.63 | <0.001 | |
| Distressed | 0.35 | 0.32–.039 | <0.001 | |
| SVI | 0.32 | 0.29–0.36 | <0.001 |
DISCUSSION
In this large, national retrospective study of Medicare beneficiaries undergoing major elective cancer operations, we found that community-level SDoH factors (e.g., DCI and SVI) and hospital quality were associated with short-term outcomes after select elective cancer operations. Patients from more vulnerable counties and more distressed communities were associated with worse 30-day mortality and 30-day readmission. However, after adjusting for hospital quality, as reported by the CMS Hospital Star Rating system, both the DCI and SVI were not significantly associated with 30-day mortality or readmission and hospital rating remained significantly associated with short-term outcomes. This result suggests that the quality of the hospital where the surgery was performed could play a more important role in determining short-term outcomes following surgery than the community-level SDoH factors based on where patients live. Patients treated at hospitals with a 5-star rating had lower odds of experiencing 30-day mortality and readmission compared to those who underwent an operation at 1-star rated hospitals. This result validates CMS Hospital Star Rating scoring system. Additionally, we found that patients from more vulnerable or distressed communities were less likely to receive care at high-quality hospitals. This disparity raises important concerns about barriers to receiving high-quality surgical care, especially among vulnerable cancer patients. Additionally, hospital quality, as measured by the CMS Hospital Star Rating system, could be a mediating factor between community disadvantage and postoperative outcomes.
The disparities initially represented by SDoH reflected differences at the community level, or “where you live”. However, SDoH alone cannot fully predict patient outcomes. The hospital star rating- representing “where you go”- has been consistently associated with patient outcomes. Nonetheless, this does not imply that community-level disparities are unimportant. In fact, patients from communities with greater disparities are less likely to receive high-quality care. (24–26)
Recent studies have highlighted the association between county-level and zip-code level SDoH factors and outcomes after cancer surgery. Patients from disadvantaged communities face adverse outcomes. (17,27–30) A study by Azap et al. found that Medicare beneficiaries who underwent elective hepatopancreatic surgery and were from low SVI communities (i.e., less socially vulnerable) were significantly more likely to achieve textbook outcomes compared to those from average or high-SVI communities. (17) In addition, hospital quality metrics play a critical role in patient outcomes following surgery. Hospitals with better infrastructure, highly trained surgical teams, and high-quality postoperative care are associated with improved patient outcomes. (31,32) However, our knowledge about the interaction between SDoH and hospital quality and their influence on short-term outcomes after cancer surgery is limited. A study conducted on patients who underwent pancreatectomy or hepatectomy showed that the risk of postoperative complications among patients from vulnerable counties was better mitigated by high-quality hospitals. (33) Another study that included 26,937 patients who underwent cancer surgery showed that patients living in communities with higher SVI had lower odds of undergoing surgery at high-volume hospitals. (34)
Community-level vulnerability can impact patient outcomes, yet these factors are beyond the control and scope of what surgeons and healthcare providers can directly manage. (35) Although increasing access to higher-rated hospitals for patients from vulnerable segments of society could have a significant impact on outcomes, the capacity to treat more patients is limited at top-rated healthcare systems that are currently operating at or near full capacity. Furthermore, patients might face barriers to seeking care at highly rated hospitals, including long traveling distance, lack of reliable or affordable transportation, or insurance restrictions. (36,37) Despite the challenges in directly reducing community-level vulnerabilities, we need to focus on innovative strategies to improve cancer care at lower-quality-rated hospitals. This should include cross-collaboration among institutions of varying quality ratings that ensures that more patients initially seen at lower rated hospitals can have access to second opinions at higher-rated hospitals that are geographically accessible to these patients. (38) Additionally, virtual interdisciplinary rounds between physicians from varying quality rated hospitals could help improve decision-making in complex situations that might require expertise and care that is not available at their local institutions. Ultimately, the goal should be to improve cancer care, especially at currently low rated hospitals.
A study focusing on cross-collaboration between an academic medical center and four local hospitals in low-resource areas, centered on congenital heart disease, demonstrated that this collaboration reduced the rate of delayed surgeries in local hospitals compared to before its implementation. Postoperative outcomes also improved in local hospitals following the collaboration. Overall, the study suggests that cross-collaboration between local hospitals and academic centers can improve patient outcomes, help reduce socioeconomic disparities, and mitigate barriers to accessing higher-quality care (39).
In addition, to improve equitable access to higher-rated hospitals, these institutions should adopt multi-level strategies to better serve patients from vulnerable communities. Practical approaches include providing transportation and lodging assistance to reduce logistical barriers to care. In addition, formal collaborations between high-performing and safety-net hospitals are essential. Policies implemented at the hospital level can have a lasting impact, particularly for patients from more vulnerable populations. Overall, a multi-level approach that integrates supportive services, cross-institutional collaboration, and thoughtful policy can strengthen the role of hospitals in advancing health equity (40).
Limitations
This study has certain limitations. First, this study only included patients who were older than 65 years and who had Medicare insurance. This limits the generalizability of the finding to patients that are younger, have other types of insurance, or who are uninsured. Second, the SDoH factors that we assessed, DCI and SVI, were initially developed to evaluate the impact of natural disasters or adverse health outcomes in vulnerable communities. Although these SDoH factors have not been validated to be used in studying their association with short-term outcomes following cancer surgical procedures, many previous studies have used these in their analysis. Third, there are potential confounders like frailty that we did not have access to and therefore could not adjust for. Fourth, as cancer staging data were not available in this dataset, we focused on short-term outcomes, which are less likely to be impacted by cancer stage, to mitigate this limitation. Fifth, CMS Hospital Star Ratings are updated quarterly; however, the April 2021 release was applied uniformly across all patient encounters. While this cross-sectional approach may not fully capture temporal changes in hospital quality over the 2016–2018 study period, applying different rating versions to individual encounters was not feasible given the retrospective nature and large scale of this study. Sixth, while CMS Hospital Star Ratings provide a broad measure of hospital quality across a wide range of institutions, they do not capture cancer-specific quality metrics. Future studies incorporating American College of Surgeons Commission on Cancer (CoC) accreditation status alongside CMS Hospital Star Ratings may offer a comprehensive assessment of hospital quality.
Conclusions
The lack of significant associations between DCI/SVI and 30-day mortality and 30-day readmission following elective cancer surgery after adjusting for Hospital Star Rating suggests that the quality of the hospital where patients receive surgery might have a greater impact on short-term outcomes than community-level SDoH factors. Additionally, patients from more distressed communities and socially vulnerable counties were less likely to receive surgery at higher rated hospitals. These findings underscore the importance of expanding access to higher rated hospitals for all patients. This might be achievable by promoting and supporting cross-collaboration between hospitals of varying hospital quality rating so that patients seen at lower rated hospitals, potentially due to proximity or other factors, also have access to expertise and resources at higher rated hospitals.
Supplementary Material
Highlights.
DCI and SVI were not associated with 30-day mortality or readmission.
Hospital Star Rating had a greater impact on short-term outcomes than SDoH.
Patients from vulnerable communities were less likely to access high-quality hospitals.
Funding:
Dr. Joel S. Weissman was supported by National Institutes of Health’s National Institutes on Aging, grant R01AG067507.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflicts of Interest:
The authors declare that they have no conflicts of interest.
Declaration of interests
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:
Joel S. Weissman reports financial support was provided by National Institutes of Health. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References
- 1.Mehta R, Paredes AZ, Tsilimigras DI, Farooq A, Sahara K, Merath K, et al. CMS hospital compare system of star ratings and surgical outcomes among patients undergoing surgery for cancer: do the ratings matter? Ann Surg Oncol. 2020. Sep;27(9):3138–46. [DOI] [PubMed] [Google Scholar]
- 2.Diaz A, Hyer JM, Barmash E, Azap R, Paredes AZ, Pawlik TM. County-level Social Vulnerability is Associated With Worse Surgical Outcomes Especially Among Minority Patients. Ann Surg. 2021. Dec 1;274(6):881–91. [DOI] [PubMed] [Google Scholar]
- 3.Councell KA, Polcari AM, Nordgren R, Skolarus TA, Benjamin AJ, Shubeck SP. Social vulnerability is associated with advanced breast cancer presentation and all-cause mortality: a retrospective cohort study. Breast Cancer Res. 2024. Dec 3;26(1):176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Masoud SJ, Seo JE, Singh E, Woody RL, Muhammed M, Webster W, et al. Social Vulnerability Index and Survivorship after Colorectal Cancer Resection. J Am Coll Surg. 2024. Apr 1;238(4):693–706. [DOI] [PubMed] [Google Scholar]
- 5.Carmichael H, Dyas AR, Bronsert MR, Stearns D, Birnbaum EH, McIntyre RC, et al. Social vulnerability is associated with increased morbidity following colorectal surgery. Am J Surg. 2022. Jul;224(1 Pt A):100–5. [DOI] [PubMed] [Google Scholar]
- 6.Braveman P, Egerter S, Williams DR. The social determinants of health: coming of age. Annu Rev Public Health. 2011;32:381–98. [DOI] [PubMed] [Google Scholar]
- 7.Hughes Halbert C Social determinants of health and cancer care: where do we go from here? J Natl Cancer Inst. 2022. Dec 8;114(12):1564–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lara-Morales A, Soto-Ruiz N, Agudelo-Suárez AA, García-Vivar C. Social determinants of health in post-treatment cancer survivors: Scoping review. Eur J Oncol Nurs. 2024. Jun;70:102614. [DOI] [PubMed] [Google Scholar]
- 9.Tran T, Rousseau MA, Farris DP, Bauer C, Nelson KC, Doan HQ. The social vulnerability index as a risk stratification tool for health disparity research in cancer patients: a scoping review. Cancer Causes Control. 2023. May;34(5):407–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Diaz A, Pawlik TM. Social determinants of health and surgical care. Curr Probl Surg. 2023. Apr;60(4):101300. [DOI] [PubMed] [Google Scholar]
- 11.Social Vulnerability Index | Place and Health - Geospatial Research, Analysis, and Services Program (GRASP) | ATSDR; [Internet]. [cited 2024 Dec 22]. Available from: https://www.atsdr.cdc.gov/place-health/php/svi/index.html [Google Scholar]
- 12.2024 DCI Interactive Map - Economic Innovation Group [Internet]. [cited 2024 Dec 22]. Available from: https://eig.org/distressed-communities/?geo=states&lat=38.01&lon=-96.42&z=4
- 13.Shah SK, Khan AA, Basu S, Alex GC, Geissen NM, Liptay MJ, et al. Social vulnerability index is associated with major morbidity after lung resection. Annals of Thoracic Surgery Short Reports. 2025. Mar;3(1):144–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Stuart CM, Dyas AR, Byers S, Velopulos C, Randhawa S, David EA, et al. Social vulnerability is associated with increased postoperative morbidity following esophagectomy. J Thorac Cardiovasc Surg. 2023. Oct;166(4):1254–61. [DOI] [PubMed] [Google Scholar]
- 15.Jain V, Al Rifai M, Khan SU, Kalra A, Rodriguez F, Samad Z, et al. Association between social vulnerability index and cardiovascular disease: a behavioral risk factor surveillance system study. J Am Heart Assoc. 2022. Aug 2;11(15):e024414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Nguyen TN, Ngangue P, Bouhali T, Ryan BL, Stewart M, Fortin M. Social Vulnerability in Patients with Multimorbidity: A Cross-Sectional Analysis. Int J Environ Res Public Health. 2019. Apr 8;16(7). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Azap RA, Paredes AZ, Diaz A, Hyer JM, Pawlik TM. The association of neighborhood social vulnerability with surgical textbook outcomes among patients undergoing hepatopancreatic surgery. Surgery. 2020. Nov;168(5):868–75. [DOI] [PubMed] [Google Scholar]
- 18.Papageorge MV, Resio BJ, Monsalve AF, Canavan M, Pathak R, Mase VJ, et al. Navigating by stars: using CMS star ratings to choose hospitals for complex cancer surgery. JNCI Cancer Spectr. 2020. Oct;4(5):pkaa059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kaye DR, Norton EC, Ellimoottil C, Ye Z, Dupree JM, Herrel LA, et al. Understanding the relationship between the Centers for Medicare and Medicaid Services’ Hospital Compare star rating, surgical case volume, and short-term outcomes after major cancer surgery. Cancer. 2017. Nov 1;123(21):4259–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Hospitals - Overall hospital quality star rating | Provider Data Catalog; [Internet]. [cited 2024 Dec 22]. Available from: https://data.cms.gov/provider-data/topics/hospitals/overall-hospital-quality-star-rating [Google Scholar]
- 21.SSA to Federal Information Processing Series (FIPS) State and County Crosswalk | NBER; [Internet]. [cited 2025 Jun 9]. Available from: https://www.nber.org/research/data/ssa-federal-information-processing-series-fips-state-and-county-crosswalk [Google Scholar]
- 22.Moura LMVR, Festa N, Price M, Volya M, Benson NM, Zafar S, et al. Identifying Medicare beneficiaries with dementia. J Am Geriatr Soc. 2021. Aug;69(8):2240–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Gasparini A comorbidity: An R package for computing comorbidity scores. JOSS. 2018. Mar 29;3(23):648. [Google Scholar]
- 24. .;. NANAABDC NAHealth affairsK Baicker, Chandra A, Skinner JS, Wennberg JE. Who You Are And Where You Live: How Race And Geography Affect The Treatment Of Medicare Beneficiaries: There is no simple story that explains the regional …. Health affairs. 2004; [DOI] [PubMed] [Google Scholar]
- 25.Dimick J, Ruhter J, Sarrazin MV, Birkmeyer JD. Black patients more likely than whites to undergo surgery at low-quality hospitals in segregated regions. Health Aff (Millwood). 2013. Jun;32(6):1046–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Barnato AE, Lucas FL, Staiger D, Wennberg DE, Chandra A. Hospital-level racial disparities in acute myocardial infarction treatment and outcomes. Med Care. 2005. Apr;43(4):308–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Valencia CI, Wightman P, Morrill KE, Hsu C-H, Arif-Tiwari H, Kauffman E, et al. Neighborhood social vulnerability and disparities in time to kidney cancer surgical treatment and survival in Arizona. Cancer Med. 2024. Feb;13(3):e7007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Dalmacy DM, Tsilimigras DI, Hyer JM, Paro A, Diaz A, Pawlik TM. Social vulnerability and fragmentation of postoperative surgical care among patients undergoing hepatopancreatic surgery. Surgery. 2022. Apr;171(4):1043–50. [DOI] [PubMed] [Google Scholar]
- 29.Hyer JM, Tsilimigras DI, Diaz A, Mirdad RS, Azap RA, Cloyd J, et al. High Social Vulnerability and “Textbook Outcomes” after Cancer Operation. J Am Coll Surg. 2021. Apr;232(4):351–9. [DOI] [PubMed] [Google Scholar]
- 30.Stuart CM, Dyas AR, Byers S, Velopulos CG, Randhawa SK, David EA, et al. Social vulnerability is associated with post-operative morbidity following robotic-assisted lung resection. J Thorac Dis. 2023. Nov 30;15(11):5931–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Begg CB, Cramer LD, Hoskins WJ, Brennan MF. Impact of hospital volume on operative mortality for major cancer surgery. JAMA. 1998. Nov 25;280(20):1747–51. [DOI] [PubMed] [Google Scholar]
- 32.GlobalSurg Collaborative and NIHR Global Health Research Unit on Global Surgery. Effects of hospital facilities on patient outcomes after cancer surgery: an international, prospective, observational study. Lancet Glob Health. 2022. Jul;10(7):e1003–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Diaz A, Beane JD, Hyer JM, Tsilimigras D, Pawlik TM. Impact of hospital quality on surgical outcomes in patients with high social vulnerability: Association of textbook outcomes and social vulnerability by hospital quality. Surgery. 2022. Jun;171(6):1612–8. [DOI] [PubMed] [Google Scholar]
- 34.Diaz A, Chavarin D, Paredes AZ, Tsilimigras DI, Pawlik TM. Association of Neighborhood Characteristics with Utilization of High-Volume Hospitals Among Patients Undergoing High-Risk Cancer Surgery. Ann Surg Oncol. 2021. Feb;28(2):617–31. [DOI] [PubMed] [Google Scholar]
- 35.Chakravarti S, Kuo CC, Oak A, Ranganathan S, Jimenez AE, Kazemi F, et al. The Socioeconomic Distressed Communities Index Predicts 90-Day Mortality Among Intracranial Tumor Patients. World Neurosurg. 2024. Jun;186:e552–65. [DOI] [PubMed] [Google Scholar]
- 36.Endo Y, Woldesenbet S, Kawashima J, Tsilimigras DI, Rashid Z, Catalano G, et al. Impact of hospital volume and facility characteristics on postoperative outcomes after hepatectomy: A mediation analysis. Surgery. 2024. Dec;176(6):1653–60. [DOI] [PubMed] [Google Scholar]
- 37.Resio BJ, Chiu AS, Hoag JR, Brown LB, White M, Omar A, et al. Motivators, barriers, and facilitators to traveling to the safest hospitals in the united states for complex cancer surgery. JAMA Netw Open. 2018. Nov 2;1(7):e184595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Molina G, Ferrone CR, Qadan M. Failure to Refer Patients with Colorectal Liver Metastases to a Multidisciplinary Oncology Team Should be a “Never-Event”. J Natl Med Assoc. 2020. Oct;112(5):553–5. [DOI] [PubMed] [Google Scholar]
- 39. .Shi G, Chen S, Shentu J, Liu N, Luo H, Li C, et al. Partnership model of regionalized care for congenital heart disease in resource-limited settings: results from the ASSIST Project. Circulation. 2025; [DOI] [PubMed] [Google Scholar]
- 40.Rangachari P, Thapa A. Impact of hospital and health system initiatives to address Social Determinants of Health (SDOH) in the United States: a scoping review of the peer-reviewed literature. BMC Health Serv Res. 2025. Mar 5;25(1):342. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.