If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Cardiovascular Outcomes Research Laboratories, David Geffen School of Medicine at UCLA, Los Angeles, CalifDivision of Cardiac Surgery, Department of Surgery, David Geffen School of Medicine at UCLA, Los Angeles, Calif
Safety-net hospitals (SNHs) have previously been associated with inferior outcomes and greater resource use. However, this relationship has not been explored in the contemporary setting of pulmonary lobectomy. In the present national study we characterized the association between SNHs and mortality, complications, and resource use.
Methods
All adults (18 years of age or older) who underwent elective lobectomy for lung cancer were identified within the 2010 to 2019 Nationwide Readmissions Database. Hospitals in the highest quartile of safety-net burden were categorized as SNHs, and others non-SNHs. Multivariable regressions were developed to assess the independent association between safety-net status and outcomes of interest.
Results
Of an estimated 282,011 patients who met inclusion criteria, 41,015 (14.5%) were treated at SNHs. Patients at SNHs were younger but as commonly female, compared with non-SNHs. After multivariable adjustment, there was no association between SNHs and mortality. However, treatment at SNHs was linked to higher odds of pneumonia (adjusted odds ratio [AOR], 1.11; 95% CI, 1.02-1.21) and prolonged ventilation (AOR, 1.36; 95% CI, 1.11-1.66), as well as infectious (AOR, 1.24; 95% CI, 1.08-1.43), intraoperative (AOR, 1.22; 95% CI, 1.07-1.39), and overall complications (AOR, 1.07; 95% CI, 1.01-1.14). Patients at SNHs also showed a greater need for a blood transfusion (AOR, 1.37; 95% CI, 1.23-1.53). In addition, SNHs were associated with increased length of stay (+0.33 days; 95% CI, 0.17-0.48) and greater costs (+$4130; 95% CI, 3.34-4.92), relative to non-SNHs.
Conclusions
Hospital safety-net status was associated with greater odds of perioperative complications and greater health care expenditure. Further investigation is necessary uncover the mechanisms contributing to these complications and eradicate persistent disparities in lobectomy.
Hospital safety-net status is associated with greater rates of complications and costs after lobectomy for lung cancer. Further study is needed to uncover contributing factors and persistent disparities.
Safety-net hospitals treat some of society's most vulnerable populations but have previously been associated with greater mortality and morbidity across a diverse set of operations. Using a nationally representative cohort of pulmonary lobectomy for lung cancer, the present study showed that safety-net hospital status was linked with increased complication rates and resource use.
See Commentary on page 1585.
Safety-net hospitals (SNHs) provide critically needed care for the underinsured and uninsured, whether by legal mandate or an adopted hospital mission.
Seeking to identify systemic and hospital characteristics that might perpetuate socioeconomic health disparities, a large body of literature has been published on clinical and financial benchmarking metrics at SNHs.
reported that patients at SNHs faced greater rates of mortality and complications. Upon further examination, this study posited advanced disease presentation, reduced hospital efficiency, and lack of resources as putative factors for this observation. Moreover, SNHs might be subjected to disproportionate financial penalties because of relatively worse outcomes, in particular for complex procedures.
Indeed, several groups have noted that SNHs are trapped in a vicious cycle of inferior outcomes, reduced reimbursement rates, and subsequent reductions in quality of care because of this lack of resources.
Considered the standard of care for localized lung cancer, pulmonary lobectomy is a complex operation, requiring attention to proper preoperative, intraoperative, and postoperative management. To achieve optimal outcomes, multidisciplinary care is warranted, often involving expertise of thoracic surgeons, medical and radiation oncologists, critical care hospitalists, and respiratory therapists, among others.
investigated outcomes at SNHs, and reported greater complication rates and health care expenditures. However, their study was limited in scope and used decade-old data, limiting its generalizability to the contemporary landscape of lung cancer screening and early-stage tumor detection.
With policy and practice implications, in the present, decade-long national study we characterized short-term clinical and financial outcomes among patients who underwent pulmonary lobectomy at SNHs. We hypothesized SNHs to be associated with increased mortality, complications, and resource use.
Methods
All elective adult (aged 18 years or older) hospitalizations for pulmonary lobectomy for lung cancer were identified in the 2010 to 2019 Nationwide Readmissions Database (NRD) using International Classification of Diseases, Ninth/Tenth Revision diagnosis and procedure codes.
Maintained by the Healthcare Cost and Utilization Project, the NRD is the largest, all-payer readmissions database and provides accurate estimates for nearly 60% of all hospitalizations in the United States. Each record in the NRD is given a unique identifier, allowing subsequent readmissions to be linked within a single calendar year.
Patients who underwent nonelective lobectomy or lobectomy for indications other than lung cancer were excluded from further analysis. Records with missing key data including age, sex, mortality, and insurance information (3.5%) were similarly not considered (Figure 1).
Figure 1Study flow chart of survey-weighted estimates. Of 372,175 lobectomies identified in the 2010 to 2019 Nationwide Readmissions Database, 282,011 patients underwent resection for lung cancer. Of these, 41,015 (14.5%) received treatment at safety-net hospitals (SNH). All estimates represent survey-weighted methodology.
Patient and hospital characteristics were ascertained in accordance with the Healthcare Cost and Utilization Project data dictionary and included age, sex, income quartile, primary payer, and hospital teaching status.
International Classification of Diseases, Ninth/Tenth Revision diagnosis codes were used to identify specific comorbidities. Operative approach was similarly defined as open or minimally invasive, which included video-assisted and robotic-assisted thoracoscopic surgery. Perioperative complications, such as acute kidney injury and blood transfusion were defined using International Classification of Diseases, Ninth/Tenth Revision codes. Furthermore, cardiac (cardiac arrest, ventricular fibrillation, ventricular tachycardia, tamponade, myocardial infarction), cerebrovascular (intracranial hemorrhage, acute ischemic complications, stroke), infectious (sepsis, surgical site infection), intraoperative (accidental puncture, hemorrhage, nerve injury), and respiratory (pneumonia, acute respiratory distress syndrome, respiratory failure, pneumothorax, prolonged ventilation) complications were grouped.
Hospitalization costs were calculated through application of hospital-specific cost-to-charge ratios to overall charges and inflation adjustment to the 2019 Personal Healthcare Price Index.
Hospitals were categorized into low-, medium-, or high-volume tertiles on the basis of their annual pulmonary lobectomy caseload. Following the definitions set forth by the Agency for Healthcare Research and Quality and as previously reported,
the proportion of Medicaid or self-pay (uninsured) admissions was calculated for each hospital, with the institutions in the top quartile considered SNHs.
The primary outcome of this study was in-hospital mortality. Secondary end points included perioperative complications, hospitalization costs, length of stay (LOS), and unplanned readmissions within 30 days of index discharge.
Continuous variables are reported as median with interquartile range and categorical variables are presented as percentages (%). Medians and proportions were compared across groups using the Mann–Whitney U and χ2 tests, respectively. Multivariable logistic and linear regression models were fit to evaluate independent associations between SNHs and study end points. Covariates were selected using elastic net regularization, which minimizes collinearity through a penalized least squares methodology.
The receiver operating characteristic was used to evaluate models. Cuzick's nonparametric test was applied to determine significance of temporal trends (nptrend).
Adjusted outcomes are reported as adjusted odds ratio (AOR) or β-coefficient (β) with 95% CI for binary and continuous end points, respectively. All statistical analyses were performed using Stata version 16.1 (StataCorp). This study was deemed exempt from full review by the institutional review board at the University of California, Los Angeles.
Results
Of an estimated 282,011 elective hospitalizations entailing lobectomy for lung cancer, 14.5% were treated at SNHs. Over the study period, the proportion of patients treated at SNHs increased from 10.4% in 2010 to 15.3% in 2019 (nptrend < .001; Figure 2). On average, patients at SNHs were younger (68 vs 69 years; P < .001) and had a lower Elixhauser Comorbidity Index (3.55 vs 3.61; P = .007), but were similar in sex distribution, compared with non-SNHs (Table 1). Specifically, SNH patients more frequently suffered from liver disease, hypertension, and diabetes. Rates of minimally invasive surgery utilization were equal between SNHs and non-SNHs (52.5 vs 51.5%; P = .062). Those treated at SNHs were more frequently classified in the lowest income quartile (31.4% vs 23.0%; P < .001). Patients in the SNH cohort were less commonly treated at high-volume hospitals (72.4% vs 84.3%; P < .001), but more often received care at metropolitan teaching institutions (85.5% vs 75.2%; P < .001), compared with the non-SNH cohort.
Figure 2Annual trends in lobectomy caseload at safety-net hospitals (SNH) and non-SNH. The proportion of patients who underwent lobectomy for lung cancer at SNH increased across the study time period, from 10.4% (n = 2818) operations in 2010 to 15.3% (n = 4690) in 2019; nptrend < .001. Mortality rates declined at SNH (4.26% to 4.20%; nptrend < .001) and non-SNH (4.33% to 3.35%; nptrend < .001).
Bivariate comparison of outcomes is shown in Table 2. Patients at SNHs and non-SNHs faced similar rates of in-hospital mortality (1.4% vs 1.4%; P = .907). Further, mortality rates declined over the study period at SNHs and non-SNHs (Figure 2). However, the SNH cohort faced greater rates of infectious complications (1.9% vs 1.6%; P = .012), pneumonia (6.4% vs 5.8%; P = .011), and blood transfusions (6.6% vs 5.8%; P = .026). The adjusted risk of infectious complications decreased from 2010 to 2019 at SNHs (2.49% to 1.22%; nptrend < .001) and non-SNHs (1.84% to 1.14%; P < .001; Figure 3). In addition, the SNH cohort experienced longer LOS and higher hospitalization costs, compared with the non-SNH cohort. Patients at SNHs and non-SNHs experienced similar rates of nonhome discharge (7.4% vs 7.8%; P = .235) and 30-day, nonelective readmission (7.8% vs 8.0%; P = .390).
Table 2Clinical outcomes at SNH compared with non-SNH
Unadjusted
P value
Adjusted multivariable regression
Non-SNH (%)
SNH (%)
SNH (AOR)
95% CI
Clinical outcomes
In-hospital mortality
1.4
1.4
.91
1.07
0.92-1.25
Infectious complications
1.6
1.9
.01
1.24
1.08-1.43
Intraoperative complications
2.3
2.5
.17
1.22
1.07-1.39
Respiratory complications
19.6
20.1
.46
1.07
1.00-1.15
Blood transfusion
5.8
6.6
.03
1.37
1.23-1.53
Cerebrovascular complications
0.5
.05
.31
1.32
1.02-1.71
Any complication
25.4
25.7
.59
1.07
1.01-1.14
Non-home discharge
7.8
7.4
.24
1.04
0.95-1.13
Nonelective 30-d readmission
8.0
7.8
.39
1.01
0.94-1.09
Resource utilization
Length of stay ± SD, d
6.68 ± 6.08
6.98 ± 7.29
.01
0.33
0.17-0.48
Cost (IQR), per USD $1000
22.0 (16.7-30.1)
25.1 (18.9-33.9)
.01
4.13
3.34-4.92
Outcomes reported as proportions (%) or as adjusted odds ratio with 95% CI. SNH, Safety-net hospitals; AOR, adjusted odds ratio; CI, confidence interval; SD, standard deviation; IQR, interquartile range.
Figure 3Temporal trends in adjusted risk of infectious complications at safety-net hospitals (SNH) and non-SNH. Adjusted risk of infectious complications decreased from 2010 to 2019 at SNH (2.49% to 1.22%; nptrend = .001) and non-SNH (1.84% to 1.14%; nptrend < .001). Notably, the difference in adjusted risk between SNH and non-SNH decreased across the study time course.
After risk adjustment, safety-net status remained associated with several key study outcomes. Despite no association with mortality, treatment at SNHs was associated with greater odds of pneumonia (AOR, 1.11; 95% CI, 1.02-1.21) and prolonged mechanical ventilation (AOR, 1.36; 95% CI, 1.11-1.66). SNH status was further linked with a 24% increase in relative odds of infectious complications (95% CI, 1.08-1.43) and a 32% increase in cerebrovascular complications (AOR, 1.32; 95% CI, 1.02-1.71; Figure 4). In addition, treatment at SNHs was linked to a 0.3-day increment in LOS (95% CI, 0.2-0.5) and $4130 in attributable hospitalization costs (95% CI, 3300-4900). Odds of 30-day nonelective readmission were unaltered on the basis of SNH status (Tables 2 and E1).
Figure 4Effect of hospital safety-net status on risk-adjusted outcomes. After risk adjustment, patients who underwent lobectomy at a safety-net hospital (SNH) faced greater odds of in-hospital mortality, complications, and non-home discharge. Reference: non–safety-net hospitals. Error bars represent 95% CI.
In the present study, we used a nationally representative cohort of pulmonary lobectomy to characterize the effect of SNHs on clinical and financial end points. First, we observed that an increasing proportion of lobectomies are being performed at SNHs each year. Moreover, in congruence with previous work, SNHs were generally large urban teaching hospitals that serve a predominantly lower income, younger population.
Controlling for these important differences, as well as hospital volume, we identified that safety-net status was independently associated with greater odds of pneumonia and prolonged ventilation, among other complications. Importantly, SNH status was also linked to longer lengths of stay and greater hospital expenditures (Figure 5). Several of these findings warrant further discussion.
Figure 5Lobectomy at safety-net hospitals (SNH) is associated with inferior outcomes, relative to non-SNH. Patients who underwent lobectomy for lung cancer at SNH incur greater complications and resource use, compared with non-SNH. Because the proportion of patients treated at SNH increased over the study time course, more work must be done to eradicate continued disparities in outcomes.
A large body of literature has been published on the effect of SNH status on a myriad of postoperative outcomes ranging from major morbidity to nonelective readmission.
Care of patients undergoing vascular surgery at safety net public hospitals is associated with higher cost but similar mortality to nonsafety net hospitals.
In this work, we found that SNH status was not associated with greater odds of in-hospital mortality. This observation is in accordance with previous investigations in the contexts of esophagectomy, surgical aortic valve replacement, and vascular surgery procedures.
Care of patients undergoing vascular surgery at safety net public hospitals is associated with higher cost but similar mortality to nonsafety net hospitals.
In-hospital outcomes of ST-segment elevation myocardial infarction complicated with cardiogenic shock at safety-net hospitals in the United States (from the Nationwide Inpatient Sample).
The homogeneity of outcomes across safety net burden might be attributable to improvements in surgical technique and care, as well as the adoption of standardized protocols to prevent failure to rescue and the more widespread use of minimally invasive approaches that reduce chest wall trauma and postoperative pain.
As such, it is unsurprising that safety-net status would not appreciably affect mortality rates.
In this investigation, we identified an increase in hospitalization costs at safety-net centers. Interestingly, although the SNH cohort experienced longer LOS, the 0.33 day increment observed after multivariable adjustment is not clinically relevant and would not account for >$4000 increment in costs. Therefore, our findings allude to the hypothesis that expenditure might stem from inefficient care, whether due to understaffed, under-resourced teams, or limited access to posthospital care and rehabilitation facilities.
cite several other potential contributors, including greater incidence of surgical complications and poor performance on Surgical Care Improvement Project Measures. Further, patients treated at SNHs often have greater social needs and present with more advanced disease and undiagnosed comorbidities. As such, SNHs might face greater expenditure requirements at baseline that limit available resources for care-improvement and cost-containment efforts.
Consistent with previous work, we identified SNH status to be linked with greater odds of intraoperative and postoperative complications, most notably a 24% greater relative risk of infectious complications. Because our multivariable models were adjusted for clinically relevant patient, hospital, and operative factors, these adverse outcomes appear to be related to structural and systemic factors inherent to SNHs. Although the exact causes of these infection rate disparities remain elusive, some have proposed they stem from inferior surgical quality due to delays in diagnosis or treatment, reduced adherence to quality metrics, and inefficient care pathways at SNHs.
Our temporal trend analysis reveals that infection rates have decreased at SNHs and non-SNHs over the study time course, with a diminishing gap between SNHs and non-SNHs. We posit that broader policy change in infection control or response might be implicated, considering that gradual reductions in infection rates were observed over the course of the decade at SNHs and non-SNHs. Indeed, these findings might, in part, be attributable to increased adoption of multidisciplinary bundles aimed at reducing postoperative infections such as those recommended by the Surgical Care Improvement Project.
Improving surgical site infections: using National Surgical Quality Improvement Program data to institute Surgical Care Improvement Project protocols in improving surgical outcomes.
The use of report cards and outcome measurements to improve the safety of surgical care: the American College of Surgeons National Surgical Quality Improvement Program.
Nonetheless, further investigation is necessary to identify the factors contributing to these improvements in infection rates and mitigate the increased clinical burden faced by patients at SNHs.
Patient-level consequences aside, however, disparities in rates of postoperative infection might be financially deleterious for SNHs. Linking hospital performance to national infection rate benchmarks, the Affordable Care Act penalizes low-performing centers—which are more often SNHs than non-SNHs—by docking payments.
Because SNHs already rely on low operating margins, additional financial losses might unintentionally jeopardize the safety net and reduce the quality of patient care. Because SNHs play an essential role in the care of vulnerable, under-resourced individuals who face pervasive structural and systemic factors that relentlessly erode at their health, some have proposed that the Centers for Medicare and Medicaid Services redesign these programs to include social risk factor adjustments and reduce inequity in penalty disbursement.
Although we observed that infection rates at SNHs and non-SNHs appear to be declining, continued disparities in outcomes and expenditures strongly underscore the need for additional policies and programs to uncover and eliminate ongoing care inequities at SNHs. Ultimately, systemic reforms are needed to end the vicious cycle of inferior outcomes, reimbursement penalties, and quality reductions that entrap SNHs. By shedding light on these persistent disparities, we hope to spur conversation, policy change, and improvements in care (Video 1).
This study has several important limitations. Because of the retrospective design of the present study causative conclusions cannot be made. Moreover, the NRD is an administrative database that lacks granular physiologic, laboratory, and radiographic information. Lung cancer staging data were unavailable, as was information about pulmonary function, extent of smoking history, tumor size, or metastasis. International Classification of Diseases coding of diagnosis and procedures can vary across providers, hospitals, and regions. Granular hospital-level data, such as nurse to patient ratios, standardized care pathways, the availability of specific general thoracic surgery units, or specialized unit staffing, are also not available in the NRD. Additionally, center- and surgeon-level specialization is not reported, but would be interesting to consider for future study. Despite these restrictions, the NRD is the largest all-payer database that allows for nationally representative findings.
Conclusions
Treatment at a SNH was associated with greater complications and resource use after lobectomy for lung cancer. Yet, SNH status had no effect on in-hospital mortality or readmissions. Because of the important role SNHs play in treating some of society's most vulnerable individuals, our findings prompt further studies on disparities in patient outcomes to improve the value, quality, and equity of care at these institutions.
Conflict of Interest Statement
The 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.
Relevance and implications for safety-net centers. Systemic reforms are needed to help safety-net centers escape the vicious cycle of financial penalties for inferior outcomes that lead to further reductions in quality of care. Video available at: https://www.jtcvs.org/article/S0022-5223(22)01026-1/fulltext.
Relevance and implications for safety-net centers. Systemic reforms are needed to help safety-net centers escape the vicious cycle of financial penalties for inferior outcomes that lead to further reductions in quality of care. Video available at: https://www.jtcvs.org/article/S0022-5223(22)01026-1/fulltext.
Appendix E1
Table E1Adjusted clinical outcomes at SNH compared with non-SNH for privately insured patients
Adjusted multivariable regression
AOR (95% CI)
Clinical outcomes
In-hospital mortality
1.37 (0.94-2.00)
Infectious complications
1.47 (1.08-2.00)
Intraoperative complications
1.18 (0.91-1.54)
Respiratory complications
1.06 (0.95-1.18)
Blood transfusion
1.29 (1.06-1.57)
Cerebrovascular complications
2.11 (1.18-3.79)
Any complication
1.05 (0.95-1.16)
Non-home discharge
1.05 (0.82-1.34)
Nonelective 30-d readmission
1.03 (0.89-1.20)
Resource utilization
Length of stay, d
0.19 (−0.05 to 0.42)
Cost (USD per $1000)
3.14 (2.08-4.20)
When evaluating patients who are privately insured, treatment at SNH remains associated with greater risk-adjusted odds of postoperative complications and resource use. AOR, Adjusted odds ratio ; IQR, interquartile range.
Care of patients undergoing vascular surgery at safety net public hospitals is associated with higher cost but similar mortality to nonsafety net hospitals.
In-hospital outcomes of ST-segment elevation myocardial infarction complicated with cardiogenic shock at safety-net hospitals in the United States (from the Nationwide Inpatient Sample).
Improving surgical site infections: using National Surgical Quality Improvement Program data to institute Surgical Care Improvement Project protocols in improving surgical outcomes.
The use of report cards and outcome measurements to improve the safety of surgical care: the American College of Surgeons National Surgical Quality Improvement Program.
Safety-net hospitals (SNHs) are defined as institutions that “provide care to a substantial share of vulnerable patients regardless of their ability to pay.”1 Many consider such hospitals as providers of last resort and assume these institutions have worse outcomes and increased overall costs. In our current era of thoracic surgery, elective resection for lung cancer is associated with low morbidity and mortality,2,3 and such outcomes should be available for all patients, regardless of financial or social status.
01026-1&id=fx1.jpg){kind=link}
01026-1&id=fx2.jpg){kind=link}
01026-1&id=gr1.jpg){kind=link}
01026-1&id=gr2.jpg){kind=link}
01026-1&id=gr3.jpg){kind=link}
01026-1&id=gr4.jpg){kind=link}
01026-1&id=gr5.jpg){kind=link}