Lung resection outcomes and costs in Washington State: a case for regional quality improvement.

A regional quality improvement effort does not exist for thoracic surgery in the United States. To initiate the development of one, we sought to describe temporal trends and hospital-level variability in associated outcomes and costs of pulmonary resection in Washington (WA) State. A cohort study (2000-2011) was conducted of operated-on lung cancer patients. The WA State discharge database was used to describe outcomes and costs for operations performed at all nonfederal hospitals within the state. Over 12 years, 8,457 lung cancer patients underwent pulmonary resection across 49 hospitals. Inpatient deaths decreased over time (adjusted p-trend=0.023) but prolonged length of stay did not (adjusted p-trend=0.880). Inflation-adjusted hospital costs increased over time (adjusted p-trend<0.001). Among 24 hospitals performing at least 1 resection per year, 5 hospitals were statistical outliers in rates of death (4 lower and 1 higher than the state average), and 13 were outliers with respect to prolonged length of stay (7 higher and 6 lower than the state average) and costs (5 higher and 8 lower than the state average). When evaluated for rates of death and costs, there were hospitals with fewer deaths/lower costs, fewer deaths/higher costs, more deaths/lower costs, and more deaths/higher costs. Variability in outcomes and costs over time and across hospitals suggest opportunities to improve the quality and value of thoracic surgery in WA State. Examples from cardiac surgery suggest that a regional quality improvement collaborative is an effective way to meaningfully and rapidly act upon these opportunities.

Farjah F ，Varghese TK ，Costas K ，Krishnadasan B ，Farivar AS ，Hubka M ，Louie BE ，Backhus LM ，Chong N ，Gorden J ，Cheng AM ，He H ，Flum DR ，Low D ，Aye R ，Vallières E ，Mulligan MS ，Wood DE ... -  《-》

Inefficiency as the major driver of excess costs in lung resection.

Risk-adjusted outcomes of surgical care are important for quality and cost assessments. Although cardiac surgery is commonly studied, risk-adjusted analysis of excess costs of lung resection has not been pursued. We used 2002 to 2005 National Inpatient Sample of the Healthcare Cost and Utilization Project data to evaluate adverse outcomes and costs in elective lung resections in hospitals with more than 20 cases during that period. Adverse outcomes were inpatient death or excessive risk-adjusted postoperative stay. Logistic models were defined to predict adverse outcomes. Linear models were designed to predict costs. Hospital-specific adverse outcome rates and costs were measured to define performance outliers. Cost-effective reference hospitals were used to define total excess costs. Among 12,182 patients at 215 hospitals undergoing lung resection, there were 336 inpatient deaths (2.8%) and 880 live discharges with prolonged risk-adjusted postoperative stay (7.2%). Predictive models for mortality and risk-adjusted postoperative stay had C statistics of 0.773 and 0.643, respectively. There were 11 ineffective hospitals (5.1%) with excessive adverse outcomes (P < .005) and 34 inefficient hospitals (15.8%) meeting quality measures but with higher than predicted costs (P < .0005). Ineffective hospitals had costs $1020 per case lower than predicted. Inefficient hospitals had costs $9978 higher than predicted. Inefficiency is the major factor in excess inpatient costs associated with lung resection in this model. Although refinements in databases, including total physician costs and postdischarge adverse event costs, will alter models, excess costs of lung resection appear to be driven by inefficiency, not adverse outcomes.

Fry DE ，Pine M ，Jones BL ，Meimban RJ ... -  《-》

Ninety-day costs of video-assisted thoracic surgery versus open lobectomy for lung cancer.

Complications after pulmonary resection lead to higher costs of care. Video-assisted thoracoscopic surgery (VATS) for lobectomy is associated with fewer complications, but lower inpatient costs for VATS have not been uniformly demonstrated. Because some complications occur after discharge, we compared 90-day costs of VATS lobectomy versus open lobectomy and explored whether differential health care use after discharge might account for any observed differences in costs. A cohort study (2007-2011) of patients with lung cancer who had undergone resection was conducted using MarketScan-a nationally representative sample of persons with employer-provided health insurance. Total costs reflect payments made for inpatient, outpatient, and pharmacy claims up to 90 days after discharge. Among 9,962 patients, 31% underwent VATS lobectomy. Compared with thoracotomy, VATS was associated with lower rates of prolonged length of stay (PLOS) (3.0% versus 7.2%; p<0.001), 90-day emergency department (ED) use (22% versus 24%; p=0.005), and 90-day readmission (10% versus 12%; p=0.026). Risk-adjusted 90-day costs were $3,476 lower for VATS lobectomy (p=0.001). Differential rates of PLOS appeared to explain this cost difference. After adjustment for PLOS, costs were $1,276 lower for VATS, but this difference was not significant (p=0.125). In the fully adjusted model, PLOS was associated with the highest cost differential (+$50,820; p<0.001). VATS lobectomy is associated with lower 90-day costs--a relationship that appears to be mediated by lower rates of PLOS. Although VATS may lead to lower rates of PLOS among patients undergoing lobectomy, observational studies cannot verify this assertion. Strategies that reduce PLOS will likely result in cost-savings that can increase the value of thoracic surgical care.

Farjah F ，Backhus LM ，Varghese TK ，Mulligan MS ，Cheng AM ，Alfonso-Cristancho R ，Flum DR ，Wood DE ... -  《-》

Video-assisted thoracic surgery lobectomy cost variability: implications for a bundled payment era.

In 2013, the Centers for Medicare and Medicaid Services began its Bundled Payments for Care Improvement Initiative. If payments are to be bundled, surgeons must be able to predict which patients are at risk for more costly care. We aim to identify factors driving variability in hospital costs after video-assisted thoracic surgery (VATS) lobectomy for lung cancer. Our institutional Society of Thoracic Surgeons data were queried for patients undergoing VATS lobectomy for lung cancer during fiscal years 2010 to 2011. Clinical outcomes data were linked with hospital financial data to determine operative and postoperative costs. Linear regression models were created to identify the impact of preoperative risk factors and perioperative outcomes on cost. One hundred forty-nine VATS lobectomies for lung cancer were reviewed. The majority of patients had clinical stage IA lung cancer (67.8%). Median length of stay was 4 days, with 30-day mortality and morbidity rates of 0.7% and 37.6%, respectively. Mean operative and postoperative costs per case were $8,492.31 (±$2,238.76) and $10,145.50 (±$7,004.71), respectively, resulting in an average overall hospital cost of $18,637.81 (±$8,244.12) per patient. Patients with chronic obstructive pulmonary disease and coronary artery disease, as well as postoperative urinary tract infections and blood transfusions, were associated with statistically significant variability in cost. Variability in cost associated with VATS lobectomy is driven by assorted patient and clinical variables. Awareness of such factors can help surgeons implement quality improvement initiatives and focus resource utilization. Understanding risk-adjusted clinical-financial data is critical to designing payment arrangements that include financial and performance accountability, and thus ultimately increasing the value of health care.

Medbery RL ，Perez SD ，Force SD ，Gillespie TW ，Pickens A ，Miller DL ，Fernandez FG ... -  《-》

Failure to rescue and pulmonary resection for lung cancer.

Failure to rescue is defined as death after an acute inpatient event and has been observed among hospitals that perform general, vascular, and cardiac surgery. This study aims to evaluate variation in complication and failure to rescue rates among hospitals that perform pulmonary resection for lung cancer. By using the Society of Thoracic Surgeons General Thoracic Surgery Database, a retrospective, multicenter cohort study was performed of adult patients with lung cancer who underwent pulmonary resection. Hospitals participating in the Society of Thoracic Surgeons General Thoracic Surgery Database were ranked by their risk-adjusted, standardized mortality ratio (using random effects logistic regression) and grouped into quintiles. Complication and failure to rescue rates were evaluated across 5 groups (very low, low, medium, high, and very high mortality hospitals). Between 2009 and 2012, there were 30,000 patients cared for at 208 institutions participating in the Society of Thoracic Surgeons General Thoracic Surgery Database (median age, 68 years; 53% were women, 87% were white, 71% underwent lobectomy, 65% had stage I). Mortality rates varied over 4-fold across hospitals (3.2% vs 0.7%). Complication rates occurred more frequently at hospitals with higher mortality (42% vs 34%, P < .001). However, the magnitude of variation (22%) in complication rates dwarfed the 4-fold magnitude of variation in failure to rescue rates (6.8% vs 1.7%, P < .001) across hospitals. Variation in hospital mortality seems to be more strongly related to rescuing patients from complications than to the occurrence of complications. This observation is significant because it redirects quality improvement and health policy initiatives to more closely examine and support system-level changes in care delivery that facilitate early detection and treatment of complications.

Farjah F ，Backhus L ，Cheng A ，Englum B ，Kim S ，Saha-Chaudhuri P ，Wood DE ，Mulligan MS ，Varghese TK ... -  《-》