ORIGINAL ARTICLE
Doubling of 30-Day Mortality by 90 Days After Esophagectomy A Critical Measure of Outcomes for Quality Improvement Haejin In, MD, MBA, MPH, y Bryan E. Palis, MA,y Ryan P. Merkow, MD, y Mitchell C. Posner, MD, Mark K. Ferguson, MD, David P. Winchester, MD,y and Christopher M. Pezzi, MDyz
Objectives: Our objectives were to (1) compare 30- and 90-day mortality rates after esophagectomy, (2) compare drivers of 30- and 90-day mortality, and (3) examine whether 90-day mortality affects hospital rankings. Background: Operative mortality has traditionally been assessed at 30 days. Ninety-day mortality has been suggested as a more appropriate indicator of quality, particularly after complex cancer surgery. Methods: Esophagectomies for nonmetastatic esophageal cancer patients diagnosed between 2007 and 2011 were identified in the National Cancer Data Base. Mortality rates were examined by patient demographics, tumor characteristics, and hospital procedural volume. Risk-adjusted hierarchical logistic regression models examined hospital performance for mortality. Results: A total of 15,796 esophagectomy patients at 977 hospitals were available for analysis. Ninety-day overall mortality was more than double the 30-day mortality (8.9% vs 4.2%; P < 0.0001). In multivariate analysis, while both 30- and 90-day mortality were associated with patient factors such as age, comorbidity, and hospital volume, only 90-day mortality was influenced by tumor- and management-related variables such as stage, tumor location, and receipt of neoadjuvant therapy. Hospital performance was examined as top 10%, middle 10% to 90%, and lowest 10% as ranked using risk-adjusted odds of mortality. There was moderate correlation between ranking based on 30- and 90-day mortality [weighted k ¼ 0.45 (95% confidence interval, 0.39– 0.52)]. Compared with 30-day mortality rankings, nearly 20% of hospitals changed their ranking category when 90-day mortality rankings were used. Conclusions: Examination of 90-day mortality after esophagectomy reflects cancer patient management decisions and may provide actionable targets for quality improvement. Keywords: 90-day mortality, esophageal cancer, esophagectomy, hospital performance, quality improvement
(Ann Surg 2016;263:286–291)
D
elivery of high-quality cancer care critically relies on the ability to measure and compare quality for continuous improvement. For esophageal cancer, minimizing the morbidity and mortality associated with this complex surgical procedure is essential and a key focus of quality efforts.1 –3 Awareness of surgery-associated
From the Department of Surgery, University of Chicago, Chicago, IL; yDivision of Research and Optimal Patient Care, American College of Surgeons, Chicago, IL; and zDepartment of Surgery, Abington Memorial Hospital, Abington, PA. Disclosure: Dr In’s American College of Surgeons Clinical Scholars in Residence fellowship was supported by the American College of Surgeons and the University of Chicago. The other authors declare no conflicts of interest. Supplemental digital content is available for this article. Direct URL citation appears in the printed text and is provided in the HTML and PDF versions of this article on the journal’s Web site (www.annalsofsurgery.com). Reprints: Haejin In, MD, MBA, MPH, Department of Surgery, University of Chicago, 5841 S Maryland Ave, MC 6040, Chicago, IL 60637. E-mail: [email protected]. Copyright ß 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0003-4932/14/26105-0821 DOI: 10.1097/SLA.0000000000001215
286 | www.annalsofsurgery.com
mortality rates, along with other indices of quality including procedural volume and hospital rankings based on quality metrics, offers an opportunity to develop and implement improvement plans to address these issues. Thirty-day mortality has been the standard generally used for assessing this important outcome and provides a metric for comparing outcomes of research activities and tracking quality improvement efforts. However, with our ever-increasing ability to prolong death through advanced life support systems, there are concerns that comparing the rate of death within 30 days of a procedure may not adequately reflect a patient’s actual outcome.4 In addition, as hospitals are increasingly benchmarked for quality improvement purposes, the choice of metric used for benchmarking is critically important. In this study, the National Cancer Data Base (NCDB) was used to examine 30- and 90-day mortality after esophagectomy for cancer. We hypothesized that 90-day mortality after esophagectomy would be significantly higher than 30-day operative mortality. We also surmised that factors driving mortality would be different between 30- and 90-day mortality and that hospital rankings would differ on the basis of whether 30- or 90-day mortality was used.
METHODS Data Source The NCDB is a joint project of the American College of Surgeons’ Commission on Cancer (CoC) and the American Cancer Society.5 It captures information from approximately 1500 CoCaccredited hospitals and more than 70% of all newly diagnosed malignancies in the United States. The NCDB contains specific details on patient demographics (age, sex, race, payer, zip codelevel median income), tumor characteristics (location, size, grade, stage, histology), treatment course (receipt of chemotherapy and radiation therapy), and vital status (dead or alive) and coded according to Facility Oncology Registry Data Standards (FORDS).6 Mortality is defined as death from any cause between surgery to 30 and 90 days postsurgery regardless of discharge status. Ninety-day mortality includes deaths within 30 days.
Study Population The NCDB was queried to select adult patients (18 years or older) diagnosed with mid-, lower- and gastroesophageal (GE) junction esophageal malignancies between 2007 and 2011, who had undergone an esophagectomy and were administered all or part of their treatment at the reporting facility. Patients from hospitals that were missing follow-up on more than 50% of their patients and patients with unknown vital status (recorded as all-cause alive or dead) were excluded from the analysis. Patient demographics and tumor characteristics were grouped for analytic purposes. Comorbidities were examined using the Charlson/Deyo score7 grouped as 0, 1, 2, and 3 or greater. Patient socioeconomic information was examined as 3 groups by median Annals of Surgery Volume 263, Number 2, February 2016
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 263, Number 2, February 2016
income. Neoadjuvant therapy was defined as patients receiving chemotherapy and/or radiation before surgery. Staging was in accordance with the AJCC (American Joint Committee on Cancer) seventh edition.8 The Collaborative Stage tumor extension was used to forward-convert T and N stage for patients originally staged under the sixth edition guidelines.9 Clinical stage was used for all patients who underwent neoadjuvant therapy. AJCC stage I and II categories were collapsed, as their unadjusted short-term procedure mortality rates were comparable. Tumor location was analyzed as middle thoracic (C15.1, C15.4), GE junction (C15.2, C15.5, C16.0), and esophagus, not otherwise specified (15.9) as classified by International Classification of Disease for Oncology second and third revision codes.10 Average annual procedure volume was derived for each reporting hospital that treated esophageal cancers over the 5-year period of study. Hospital volume was grouped according to volume quartiles for analysis (1–3 cases, 4–9 cases, 10–20 cases, >20 cases).
Statistical Analysis Descriptive statistics were generated for the analytic cohort. Unadjusted 30- and 90-day mortality rates and 95% confidence intervals (CIs) adjusted for clustering at the hospital level were generated by patient, tumor, and hospital characteristics. A x2 test accounting for clustering within hospitals was used to determine statistical significance. A patient-level generalized linear mixed model evaluated predictors of 30- and 90-day mortality adjusted for all examined variables, including patient and tumor characteristics and hospital volume. Model significance for the odds ratios (ORs) were based on the 95% CIs and the P value (P < 0.05). Each hospital was ranked by the adjusted OR of death obtained from the generalized linear mixed model and examined as the top 10% of hospitals (‘‘best performing hospitals’’), the lowest 10% (‘‘worse performing hospitals’’), and middle (‘‘average performing hospitals’’). The agreement between hospital rankings based on 30- and 90-day mortality was assessed by weighted k using Cicchetti-Allison weights. All data analyses were conducted using SAS v9.4 (SAS Institute Inc., Cary, NC).
RESULTS A total of 15,796 nonmetastatic esophageal cancer patients treated with esophagectomy at 977 hospitals were available for analysis. The analytic cohort for examination of 30- and 90-day mortality was 15,443 and 14,802, respectively (Fig. 1).
Middle or Lower esophagus or GE Junction tumors Diagnosed 2007-2011 Limited to esophagectomy cases Adults only (age ≥ 18) Non-duplicated records n=17,281
Patient Characteristics and Unadjusted 30- and 90-Day Overall Mortality Table 1 demonstrates the demographic and clinical characteristics of the cohort and unadjusted 30- and 90-day mortality rates. Patients were generally healthy, white males between the ages of 50 and 79 years with stage I or II adenocarcinoma. There was an even distribution of patients treated between 2007 and 2011. Most had lower esophagus or GE junction cancers. Half of the patients underwent neoadjuvant therapy. Overall 90-day mortality was 8.9% (95% CI, 8.2–9.6) and was more than twice the 30-day mortality rate of 4.2% (95% CI, 3.8–4.7). This greater than doubling of mortality observed for 90-day mortality compared with 30-day mortality was seen in all subcategories (Table 1). There was a progressively higher rate of mortality with increasing age and comorbidities. Medicare patients were observed to have a higher mortality than private insurance patients (30-day mortality: 5.8% vs 3.0%), and mortality was higher for squamous cell carcinoma than for adenocarcinoma (30-day mortality: 5.9% vs 4.0%). This trend was also observed for 90-day mortality. A volume-outcome relationship was observed, and there is at least a doubling in mortality for hospitals with the lowest volumes compared with highest volume hospitals. The 30-day morality rate for hospitals that performed more than 20 procedures was 2.4% (95% CI, 1.6–3.1), whereas that for hospitals that performed 1 to 3 procedures was 6.6% (95% CI, 5.8–7.5). Similarly, 90-day mortality rate was 5.9% (95% CI, 4.5–7.2) compared with 12.9% (95% CI, 11.8–14.1) (Table 1; see Supplemental Digital Content Fig. 2, available at http://links.lww.com/SLA/A768).
Adjusted 30- and 90-Day Overall Mortality After adjusting for all examined variables, the use of either 30- and 90-day mortality demonstrated significance for patient factors including age, comorbidity, insurance status, histologic subtype, and hospital volume (Table 2). Older patients had 2 to 5 times the odds of death than younger patients, and patients with Charlson score 2 and above had a nearly double the odds of death than healthy patients for both 30- and 90-day mortality. Similarly, for both 30- and 90-day mortality, Medicare patients had higher odds of dying than private insurance patients, patients with squamous cell carcinoma had higher odds of dying than patients with adenocarcinoma, and the odds of dying became progressively less in higher volume hospitals. In contrast, only 90-day mortality could demonstrate significance by stage of disease, tumor location, use of neoadjuvant therapy, and median income. Compared with those with stage I/II disease, stage III patients had higher odds of dying 90 days postoperatively (OR 1.26, 95% CI 1.10–1.44). Middle thoracic tumors compared with GE junction cancer patients had higher odds of death (OR ¼ 1.30; 95% CI, 1.04–1.62). Patients who received neoadjuvant therapy were more likely to die than patients who did not (OR ¼ 1.34; 95% CI, 1.18–1.53). Patients living in higher income areas had lower mortality than patients living in the lower income areas (OR ¼ 0.74; 95% CI, 0.63–0.87).
Stage I-III only n=15,805
Comparison of Hospital Rank Using 30- and 90-Day Mortality
Limit to cases from hospitals reporting ≥50% follow up Delete if unknown vital status (dead or alive) 30-day mortality n=15,443 90-day mortality n=14,802
Table 3 shows changes in risk-adjusted hospital rank using 30- and 90- day mortality. Of the 96 hospitals that were ranked as ‘‘best performing hospitals’’ for 30-day mortality, 46 hospitals changed their rank from ‘‘best performing’’ to ‘‘average performing’’ and 1 hospital changed its rank from ‘‘best performing’’ to ‘‘worse performing’’ hospital when 90-day mortality was examined. Of the 96 hospitals that were ranked as ‘‘worst performing hospitals’’ for 30-day mortality, 48 hospitals changed their rank to ‘‘average performing hospitals’’ when 90-day mortality was examined. In
FIGURE 1. Analytic cohort identification. ß
Ninety-day Mortality for Esophagectomy
2015 Wolters Kluwer Health, Inc. All rights reserved.
www.annalsofsurgery.com | 287
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 263, Number 2, February 2016
In et al
TABLE 1. Unadjusted 30- and 90-Day Mortality Rates With 95% CIs, 2007–2011 30-d Mortality
Overall Age, yr 18–49 50–59 60–69 70–79 79 Sex Male Female Race/ethnicity White Black Other Comorbid conditions 0 1 2 3þ Insurance Not insured, Medicaid Private insurancey Medicare Other governmentz Unknown Median income§ Median income 20 cases
n (%)
Mortality Rate [95% CI]
15,443
4.2% [3.8–4.7]
1,411 (9.1) 3,817 (24.7) 5,823 (37.7) 3,599 (23.3) 793 (5.1)
1.8% 2.9% 3.9% 5.7% 10.5%
[1.1–2.5] [2.3–3.5] [3.4–4.5] [4.9–6.6] [8.1–12.9]
90-d Mortality P
n (%)
Mortality Rate [95% CI]
14,802
8.9% [8.2–9.6]
Doubling of 30-Day Mortality by 90 Days After Esophagectomy A Critical Measure of Outcomes for Quality Improvement Haejin In, MD, MBA, MPH, y Bryan E. Palis, MA,y Ryan P. Merkow, MD, y Mitchell C. Posner, MD, Mark K. Ferguson, MD, David P. Winchester, MD,y and Christopher M. Pezzi, MDyz
Objectives: Our objectives were to (1) compare 30- and 90-day mortality rates after esophagectomy, (2) compare drivers of 30- and 90-day mortality, and (3) examine whether 90-day mortality affects hospital rankings. Background: Operative mortality has traditionally been assessed at 30 days. Ninety-day mortality has been suggested as a more appropriate indicator of quality, particularly after complex cancer surgery. Methods: Esophagectomies for nonmetastatic esophageal cancer patients diagnosed between 2007 and 2011 were identified in the National Cancer Data Base. Mortality rates were examined by patient demographics, tumor characteristics, and hospital procedural volume. Risk-adjusted hierarchical logistic regression models examined hospital performance for mortality. Results: A total of 15,796 esophagectomy patients at 977 hospitals were available for analysis. Ninety-day overall mortality was more than double the 30-day mortality (8.9% vs 4.2%; P < 0.0001). In multivariate analysis, while both 30- and 90-day mortality were associated with patient factors such as age, comorbidity, and hospital volume, only 90-day mortality was influenced by tumor- and management-related variables such as stage, tumor location, and receipt of neoadjuvant therapy. Hospital performance was examined as top 10%, middle 10% to 90%, and lowest 10% as ranked using risk-adjusted odds of mortality. There was moderate correlation between ranking based on 30- and 90-day mortality [weighted k ¼ 0.45 (95% confidence interval, 0.39– 0.52)]. Compared with 30-day mortality rankings, nearly 20% of hospitals changed their ranking category when 90-day mortality rankings were used. Conclusions: Examination of 90-day mortality after esophagectomy reflects cancer patient management decisions and may provide actionable targets for quality improvement. Keywords: 90-day mortality, esophageal cancer, esophagectomy, hospital performance, quality improvement
(Ann Surg 2016;263:286–291)
D
elivery of high-quality cancer care critically relies on the ability to measure and compare quality for continuous improvement. For esophageal cancer, minimizing the morbidity and mortality associated with this complex surgical procedure is essential and a key focus of quality efforts.1 –3 Awareness of surgery-associated
From the Department of Surgery, University of Chicago, Chicago, IL; yDivision of Research and Optimal Patient Care, American College of Surgeons, Chicago, IL; and zDepartment of Surgery, Abington Memorial Hospital, Abington, PA. Disclosure: Dr In’s American College of Surgeons Clinical Scholars in Residence fellowship was supported by the American College of Surgeons and the University of Chicago. The other authors declare no conflicts of interest. Supplemental digital content is available for this article. Direct URL citation appears in the printed text and is provided in the HTML and PDF versions of this article on the journal’s Web site (www.annalsofsurgery.com). Reprints: Haejin In, MD, MBA, MPH, Department of Surgery, University of Chicago, 5841 S Maryland Ave, MC 6040, Chicago, IL 60637. E-mail: [email protected]. Copyright ß 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0003-4932/14/26105-0821 DOI: 10.1097/SLA.0000000000001215
286 | www.annalsofsurgery.com
mortality rates, along with other indices of quality including procedural volume and hospital rankings based on quality metrics, offers an opportunity to develop and implement improvement plans to address these issues. Thirty-day mortality has been the standard generally used for assessing this important outcome and provides a metric for comparing outcomes of research activities and tracking quality improvement efforts. However, with our ever-increasing ability to prolong death through advanced life support systems, there are concerns that comparing the rate of death within 30 days of a procedure may not adequately reflect a patient’s actual outcome.4 In addition, as hospitals are increasingly benchmarked for quality improvement purposes, the choice of metric used for benchmarking is critically important. In this study, the National Cancer Data Base (NCDB) was used to examine 30- and 90-day mortality after esophagectomy for cancer. We hypothesized that 90-day mortality after esophagectomy would be significantly higher than 30-day operative mortality. We also surmised that factors driving mortality would be different between 30- and 90-day mortality and that hospital rankings would differ on the basis of whether 30- or 90-day mortality was used.
METHODS Data Source The NCDB is a joint project of the American College of Surgeons’ Commission on Cancer (CoC) and the American Cancer Society.5 It captures information from approximately 1500 CoCaccredited hospitals and more than 70% of all newly diagnosed malignancies in the United States. The NCDB contains specific details on patient demographics (age, sex, race, payer, zip codelevel median income), tumor characteristics (location, size, grade, stage, histology), treatment course (receipt of chemotherapy and radiation therapy), and vital status (dead or alive) and coded according to Facility Oncology Registry Data Standards (FORDS).6 Mortality is defined as death from any cause between surgery to 30 and 90 days postsurgery regardless of discharge status. Ninety-day mortality includes deaths within 30 days.
Study Population The NCDB was queried to select adult patients (18 years or older) diagnosed with mid-, lower- and gastroesophageal (GE) junction esophageal malignancies between 2007 and 2011, who had undergone an esophagectomy and were administered all or part of their treatment at the reporting facility. Patients from hospitals that were missing follow-up on more than 50% of their patients and patients with unknown vital status (recorded as all-cause alive or dead) were excluded from the analysis. Patient demographics and tumor characteristics were grouped for analytic purposes. Comorbidities were examined using the Charlson/Deyo score7 grouped as 0, 1, 2, and 3 or greater. Patient socioeconomic information was examined as 3 groups by median Annals of Surgery Volume 263, Number 2, February 2016
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 263, Number 2, February 2016
income. Neoadjuvant therapy was defined as patients receiving chemotherapy and/or radiation before surgery. Staging was in accordance with the AJCC (American Joint Committee on Cancer) seventh edition.8 The Collaborative Stage tumor extension was used to forward-convert T and N stage for patients originally staged under the sixth edition guidelines.9 Clinical stage was used for all patients who underwent neoadjuvant therapy. AJCC stage I and II categories were collapsed, as their unadjusted short-term procedure mortality rates were comparable. Tumor location was analyzed as middle thoracic (C15.1, C15.4), GE junction (C15.2, C15.5, C16.0), and esophagus, not otherwise specified (15.9) as classified by International Classification of Disease for Oncology second and third revision codes.10 Average annual procedure volume was derived for each reporting hospital that treated esophageal cancers over the 5-year period of study. Hospital volume was grouped according to volume quartiles for analysis (1–3 cases, 4–9 cases, 10–20 cases, >20 cases).
Statistical Analysis Descriptive statistics were generated for the analytic cohort. Unadjusted 30- and 90-day mortality rates and 95% confidence intervals (CIs) adjusted for clustering at the hospital level were generated by patient, tumor, and hospital characteristics. A x2 test accounting for clustering within hospitals was used to determine statistical significance. A patient-level generalized linear mixed model evaluated predictors of 30- and 90-day mortality adjusted for all examined variables, including patient and tumor characteristics and hospital volume. Model significance for the odds ratios (ORs) were based on the 95% CIs and the P value (P < 0.05). Each hospital was ranked by the adjusted OR of death obtained from the generalized linear mixed model and examined as the top 10% of hospitals (‘‘best performing hospitals’’), the lowest 10% (‘‘worse performing hospitals’’), and middle (‘‘average performing hospitals’’). The agreement between hospital rankings based on 30- and 90-day mortality was assessed by weighted k using Cicchetti-Allison weights. All data analyses were conducted using SAS v9.4 (SAS Institute Inc., Cary, NC).
RESULTS A total of 15,796 nonmetastatic esophageal cancer patients treated with esophagectomy at 977 hospitals were available for analysis. The analytic cohort for examination of 30- and 90-day mortality was 15,443 and 14,802, respectively (Fig. 1).
Middle or Lower esophagus or GE Junction tumors Diagnosed 2007-2011 Limited to esophagectomy cases Adults only (age ≥ 18) Non-duplicated records n=17,281
Patient Characteristics and Unadjusted 30- and 90-Day Overall Mortality Table 1 demonstrates the demographic and clinical characteristics of the cohort and unadjusted 30- and 90-day mortality rates. Patients were generally healthy, white males between the ages of 50 and 79 years with stage I or II adenocarcinoma. There was an even distribution of patients treated between 2007 and 2011. Most had lower esophagus or GE junction cancers. Half of the patients underwent neoadjuvant therapy. Overall 90-day mortality was 8.9% (95% CI, 8.2–9.6) and was more than twice the 30-day mortality rate of 4.2% (95% CI, 3.8–4.7). This greater than doubling of mortality observed for 90-day mortality compared with 30-day mortality was seen in all subcategories (Table 1). There was a progressively higher rate of mortality with increasing age and comorbidities. Medicare patients were observed to have a higher mortality than private insurance patients (30-day mortality: 5.8% vs 3.0%), and mortality was higher for squamous cell carcinoma than for adenocarcinoma (30-day mortality: 5.9% vs 4.0%). This trend was also observed for 90-day mortality. A volume-outcome relationship was observed, and there is at least a doubling in mortality for hospitals with the lowest volumes compared with highest volume hospitals. The 30-day morality rate for hospitals that performed more than 20 procedures was 2.4% (95% CI, 1.6–3.1), whereas that for hospitals that performed 1 to 3 procedures was 6.6% (95% CI, 5.8–7.5). Similarly, 90-day mortality rate was 5.9% (95% CI, 4.5–7.2) compared with 12.9% (95% CI, 11.8–14.1) (Table 1; see Supplemental Digital Content Fig. 2, available at http://links.lww.com/SLA/A768).
Adjusted 30- and 90-Day Overall Mortality After adjusting for all examined variables, the use of either 30- and 90-day mortality demonstrated significance for patient factors including age, comorbidity, insurance status, histologic subtype, and hospital volume (Table 2). Older patients had 2 to 5 times the odds of death than younger patients, and patients with Charlson score 2 and above had a nearly double the odds of death than healthy patients for both 30- and 90-day mortality. Similarly, for both 30- and 90-day mortality, Medicare patients had higher odds of dying than private insurance patients, patients with squamous cell carcinoma had higher odds of dying than patients with adenocarcinoma, and the odds of dying became progressively less in higher volume hospitals. In contrast, only 90-day mortality could demonstrate significance by stage of disease, tumor location, use of neoadjuvant therapy, and median income. Compared with those with stage I/II disease, stage III patients had higher odds of dying 90 days postoperatively (OR 1.26, 95% CI 1.10–1.44). Middle thoracic tumors compared with GE junction cancer patients had higher odds of death (OR ¼ 1.30; 95% CI, 1.04–1.62). Patients who received neoadjuvant therapy were more likely to die than patients who did not (OR ¼ 1.34; 95% CI, 1.18–1.53). Patients living in higher income areas had lower mortality than patients living in the lower income areas (OR ¼ 0.74; 95% CI, 0.63–0.87).
Stage I-III only n=15,805
Comparison of Hospital Rank Using 30- and 90-Day Mortality
Limit to cases from hospitals reporting ≥50% follow up Delete if unknown vital status (dead or alive) 30-day mortality n=15,443 90-day mortality n=14,802
Table 3 shows changes in risk-adjusted hospital rank using 30- and 90- day mortality. Of the 96 hospitals that were ranked as ‘‘best performing hospitals’’ for 30-day mortality, 46 hospitals changed their rank from ‘‘best performing’’ to ‘‘average performing’’ and 1 hospital changed its rank from ‘‘best performing’’ to ‘‘worse performing’’ hospital when 90-day mortality was examined. Of the 96 hospitals that were ranked as ‘‘worst performing hospitals’’ for 30-day mortality, 48 hospitals changed their rank to ‘‘average performing hospitals’’ when 90-day mortality was examined. In
FIGURE 1. Analytic cohort identification. ß
Ninety-day Mortality for Esophagectomy
2015 Wolters Kluwer Health, Inc. All rights reserved.
www.annalsofsurgery.com | 287
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery Volume 263, Number 2, February 2016
In et al
TABLE 1. Unadjusted 30- and 90-Day Mortality Rates With 95% CIs, 2007–2011 30-d Mortality
Overall Age, yr 18–49 50–59 60–69 70–79 79 Sex Male Female Race/ethnicity White Black Other Comorbid conditions 0 1 2 3þ Insurance Not insured, Medicaid Private insurancey Medicare Other governmentz Unknown Median income§ Median income 20 cases
n (%)
Mortality Rate [95% CI]
15,443
4.2% [3.8–4.7]
1,411 (9.1) 3,817 (24.7) 5,823 (37.7) 3,599 (23.3) 793 (5.1)
1.8% 2.9% 3.9% 5.7% 10.5%
[1.1–2.5] [2.3–3.5] [3.4–4.5] [4.9–6.6] [8.1–12.9]
90-d Mortality P
n (%)
Mortality Rate [95% CI]
14,802
8.9% [8.2–9.6]
