Patients were classified as those who underwent initial evaluation before (November 23, 2005, to October 31, 2008) or after (November 1, 2008, to December 31, 2013) implementation of the HPB-SP. A, Clinical and surgical case volume (minimum and maximum in each category) before and after implementation of the HPB-SP. B, Surgical volume (minimum and maximum in each category) for liver and pancreatic/periampullary tumors before and after implementation of the HPB-SP. Whiskers indicate 95% confidence intervals.
All models evaluated the association between implementation of the hepatopancreaticobiliary surgical program (HPB-SP) and postoperative adverse events (n = 240). Model 1 depicts univariate analysis. Model 2 depicts multivariate analysis adjusted for patient and tumor characteristics (including age, body mass index, comorbidity, cirrhosis, and tumor site). Model 3 depicts multivariate analysis adjusted for the characteristics in model 2 and intraoperative variables (including estimated blood loss, operating time, and transfusion).
Customize your JAMA Network experience by selecting one or more topics from the list below.
Lau K, Salami A, Barden G, et al. The Effect of a Regional Hepatopancreaticobiliary Surgical Program on Clinical Volume, Quality of Cancer Care, and Outcomes in the Veterans Affairs System. JAMA Surg. 2014;149(11):1153–1161. doi:10.1001/jamasurg.2014.1711
Malignant neoplasms of the hepatopancreaticobiliary (HPB) system constitute a significant public health problem worldwide. Treatment coordination for these tumors is challenging and can result in substandard care. Referral centers for HPB disease have been used as a strategy to improve postoperative outcomes, but their effect on accomplishing regionalization of care and improving quality of cancer care is not well known.
To evaluate the effect of implementing a multidisciplinary HPB surgical program (HPB-SP) on regionalization of care, the quality of cancer care, and surgical outcomes within an integrated health care system.
Design, Setting, and Participants
We designed a retrospective cohort study in a tertiary referral Veterans Affairs (VA) medical center within an 8-state designated VA health care region from November 23, 2005, through December 31, 2013. We compared patients with HPB tumors undergoing evaluation by the surgical oncology service before and after implementation of the HPB-SP on November 1, 2008.
Implementation of the HPB-SP to improve access to specialized, multidisciplinary cancer care for veterans across the region.
Main Outcomes and Measures
Clinical and surgical volume, proportion of patients undergoing a comprehensive multidisciplinary evaluation, and postoperative adverse events included as a composite outcome defined by occurrence of postoperative mortality, severe complications, and/or reoperation.
We identified 516 patients referred to the surgical oncology service. Establishment of the HPB-SP resulted in significant increases in regional referrals (17.3% vs 44.4%; P < .001), median monthly clinic visits (5 vs 20; P < .001), and median number of HPB surgical procedures (3 vs 9; P = .003) per quarter. Multidisciplinary assessment increased from 52.6% to 70.0% (P < .001). When we compared patients with hepatocellular carcinoma before (n = 55) and after (n = 131) implementation, more patients received any treatment (35 [63.6%] vs 109 [83.2%]; P = .004) with increased use of liver resection (0 vs 20 [15.3%]; P = .002), percutaneous ablation (0 vs 15 [11.5%]; P = .009), and oncosurgical strategies (0 vs 16 [12.2%]; P = .007) after implementation. Among patients with colorectal liver metastases (29 before vs 76 after implementation), a significant shift occurred from use of ablations (5 [17.2%] vs 3 [3.9]%; P = .02) to resections (6 [20.7%] vs 40 [52.6%]; P = .003), and use of perioperative chemotherapy increased (5 of 11 [45.5%] vs 33 of 43 [76.7%]; P = .01). The HPB-SP was associated with lower odds of postoperative adverse events, even after adjusting for important covariates (odds ratio, 0.29 [95% CI, 0.12-0.68]; P = .005), and a high rate of margin-negative liver (94.6%) and pancreatic (90.0%) resections.
Conclusions and Relevance
The development of an HPB-SP led to regionalization of care and improved quality of cancer care and surgical outcomes. Establishment of regional programs within the VA system can help improve the quality of care for patients presenting with complex cancers requiring subspecialized care.
Malignant tumors of the liver, pancreas, and biliary system constitute a significant public health problem worldwide. The American Cancer Society estimates 90 260 new diagnoses and 66 220 deaths attributable to cancers of the hepatopancreaticobiliary (HPB) system will occur during 2014.1 Incidence of HPB tumors and related mortality have increased over time, largely owing to a lack of effective screening modalities and limited curative treatments available for selected HPB tumors. Further, owing to the need for a multidisciplinary treatment approach, treatment coordination is challenging and can result in substandard delivery of care.2 Similarly, surgical treatment represents the only curative option for these patients; such surgical procedures are complex and associated with a high risk for perioperative complications and mortality when performed outside well-delineated standards.3 Based on these issues, surgeons and cancer care providers alike have focused on establishing disease-based referral centers for HPB tumors with organized multidisciplinary care as a means to improve the overall outcomes after cancer treatment.
Previous studies have demonstrated the advantages of multidisciplinary disease-based care in improving surgical and oncologic outcomes for patients with HPB and other tumors.4-7 However, limited data are available concerning the effect of HPB centers on quality of cancer care and regionalization of care. Specifically, the effect of developing an HPB center on regionalization of care across an integrated health care network, such as the Department of Veterans Affairs (VA) system, is unknown, and the effect of establishing an HPB surgical program (HPB-SP) on the quality of the process of cancer care and postoperative outcomes in such a system has not been studied. This study aims to use direct clinic- and patient-level data to evaluate the effect of implementing a multidisciplinary HPB-SP on regionalization of care, the quality of the cancer care process, and surgical outcomes within the VA, the largest integrated health care system in the United States. We hypothesized that implementation of a regional HPB-SP within the VA system would result in increased regional referrals and overall clinical and surgical volume and improved quality of care and surgical outcomes.
We designed a retrospective cohort study to evaluate the impact of implementing an HPB-SP for patients with HPB tumors. This study was approved by the institutional review board of the Baylor College of Medicine and the Research and Development Committee of the Michael E. DeBakey Veterans Affairs Medical Center (MEDVAMC). Written informed consent was obtained from eligible patients prior to enrollment.
The study was conducted at the MEDVAMC in Houston, Texas, which is 1 of 10 VA medical centers within the South Central VA Veterans Integrated Service Network (VISN16). The VISN16 is one of the largest designated regions within the VA health care system, providing care across 8 different states and serving more than 1.8 million veterans. The MEDVAMC is one of the largest, high-complexity VA medical centers in the United States; cancer care is provided through an American College of Surgeons Commission on Cancer–accredited cancer program and is the only VA hospital with the necessary infrastructure and expertise to provide multidisciplinary cancer care for malignant HPB disease in the VISN16 region.
A multidisciplinary HPB-SP was established at the MEDVAMC on November 1, 2008. Before implementation, patients with HPB tumors referred for surgical evaluation were seen by the MEDVAMC surgical oncology service during a weekly general surgical oncology–designated clinic. Preoperative multidisciplinary assessment relied on subspecialty referrals within the MEDVAMC or case presentation during a weekly general tumor board conference, as per physician discretion. Perioperative care, including the preoperative workup, and intraoperative management and postoperative care were led by the operating surgeon on a case-by-case basis.
The primary goal of the HPB-SP was to improve access to high-quality care for veterans with HPB tumors across the region. As such, the establishment of this program emerged with an emphasis to provide specialized, multidisciplinary cancer care and to expand beyond the traditional local reach throughout the whole VISN16 region. The program focused on patients with primary and metastatic HPB tumors, and tumors were managed generally in accordance with National Comprehensive Cancer Network guidelines8 for each corresponding cancer type. The program was implemented in a stepwise fashion to facilitate development of the critical components of structure and process of care over time. The structure of the program included the physical and human resources of the hospital to support perioperative and comprehensive cancer care for patients with HPB tumors, including standardized weekly tumor board conferences. The process of care for patients with HPB tumors was redesigned. Quiz Ref IDAs a first step, the HPB-SP assembled a multidisciplinary team at MEDVAMC, representing specialties from different fields involved in the care of these HPB tumors who agreed to a liberal policy on presenting local HPB cases during the weekly MEDVAMC general tumor board meeting, when the team would convene. To further support patient flow to and from MEDVAMC tumor boards, standardized patient flow processes were developed and centered around the HPB-SP. Structure and process of care evolved over time as the HPB-SP grew, while maintaining the same principles of multidisciplinary care and regional reach; as the volume of cases discussed during tumor boards increased (primarily owing to hepatocellular carcinoma [HCC] volume), an HCC-specific weekly tumor board was developed. Concurrently with the initial local approach used to ensure multidisciplinary evaluation, the HPB-SP led an outreach effort with visits to give HPB-related educational talks and to nurture network building with other VA hospitals in the region. Subsequently, a regional virtual tumor board program was developed to facilitate multidisciplinary discussion of HPB patients referred from these distant VA hospitals during the MEDVAMC tumor boards.9,10 Before HPB-SP implementation, patients referred for surgical evaluation were seen during the surgical oncology clinic. However, shortly after HPB-SP implementation, an HPB-dedicated surgical clinic was created at MEDVAMC to facilitate focused specialized care. Standardized perioperative and surveillance protocols were developed to guide care through the clinic visits.
All patients with HPB tumors diagnosed by imaging and/or biopsy results who underwent evaluation by the MEDVAMC surgical oncology service from November 23, 2005, through December 31, 2013, were eligible for study inclusion. Patients were identified and information was retrieved from an institutional HPB database by trained personnel (K.L., A.S., G.B., and S.K.) using standardized predefined algorithms, with discrepancies resolved by an experienced HPB surgeon (D.A.A.).
Patients were classified in 2 groups and compared by whether they underwent initial evaluation before or after HPB-SP implementation (November 1, 2008). Clinic- and patient-level data were collected during the study period. Clinic-level data included total, new, and follow-up HPB clinic visits. Patient-level data included baseline demographic and clinical characteristics. Tumor type was retrieved and classified as hepatobiliary or pancreatic/periampullary. Information regarding the process of care was recorded, including referring service, referring institution (MEDVAMC [local] vs VISN16 [regional]), and whether multidisciplinary assessment was performed. Treatment data included the type, timing, and sequence of the treatments provided. Postoperative outcomes data included the occurrence of postoperative complications (identified and recorded using the VA Surgical Quality Improvement Program definitions11), severity of complications (Clavien-Dindo classification system),12 reoperation within 60 days, unplanned readmission within 90 days, and death (within 90 days for liver procedures and 30 days for pancreatic procedures). The surgical margins were classified as R0 (negative), R1 (positive/microscopic), and R2 (positive/macroscopic) for procedures with a surgical specimen using information from the operative notes and pathology reports.
The primary outcomes of interest were clinical- and surgical-case volume, the proportion of patients undergoing evaluation in a comprehensive multidisciplinary setting (according to National Comprehensive Cancer Network guidelines8), and postoperative adverse events (POP-AEs); POP-AE was a composite outcome defined by the presence of postoperative mortality, severe postoperative complications (Clavien-Dindo grades III-V), and/or reoperation within 60 days. Outcomes were compared by study period based on the HPB-SP implementation date. Univariate and multivariate analyses were performed to examine the effect of HPB-SP on POP-AEs. Secondary outcomes included the effect of HPB-SP on referral patterns, quality of cancer care (treatment type and use of oncosurgical strategies), and other postoperative surgical (eg, length of stay, readmission) and cancer-related (eg, margin-negative resections) outcomes.
We used descriptive statistics to examine the cohort; proportions were used to describe categorical variables, and means (standard deviation) or medians (range) were used to describe continuous data. Differences in binary and categorical variables between the 2 groups were assessed using the 2-sided χ2 test. Differences in continuous variables were assessed using the 2-sided t test for normally distributed variables and the Mann-Whitney test for nonparametric variables. Univariate (model 1) and stepwise multivariate logistic regression models were used to examine the association between HPB-SP implementation and POP-AEs while adjusting for important baseline patient and tumor characteristics (model 2) and intraoperative variables (model 3). Criteria for inclusion in the multivariable models were clinically and statistically driven (P < .20 in univariate models for the outcome). Significance for all analysis was defined as P < .05. All statistical analysis was performed using commercially available software (STATA, version 12; StataCorp).
A total of 1381 clinic visits took place during the study period; 1181 (85.5%) occurred after the HPB-SP implementation. Quiz Ref IDThe establishment of the HPB-SP was associated with an overall significant increase in clinic volume with respect to the median numbers of monthly clinic visits (5.0 vs 20.0; P < .001) and new (4.0 vs 6.0; P < .001) and follow-up (1.5 vs 13.0; P < .001) visits (Figure 1A). Implementation of the HPB-SP also resulted in a significant increase in referrals from regional VA facilities (17.3% vs 44.4%; P < .001).
In all, 240 HPB operations were performed during the study period, including 180 (75.0%) after the HPB-SP implementation. The establishment of the HPB-SP resulted in an overall increase in the median numbers of HPB operations per quarter (3.0 vs 9.0; P = .003) (Figure 1A), liver operations per quarter (2.5 vs 7.0; P < .001), and liver resections overall (0.5 vs 4.0; P < .001) (Figure 1B). We found a nonsignificant trend toward an increased median number of pancreatic operations per quarter (1.5 vs 3.0; P = .09) after HPB-SP implementation (Figure 1B).
A total of 516 HPB patients referred to the MEDVAMC surgical oncology service were identified and included in the study; 383 (74.2%) underwent evaluation after HPB-SP implementation. Differences in baseline characteristics between patients undergoing evaluation before and after HPB-SP implementation are described in Table 1. Quiz Ref IDHepatocellular carcinoma constituted the highest proportion of tumor diagnoses seen (n = 186 [36.0%]), followed by pancreatic/periampullary tumors (n = 162 [31.4%]) and colorectal liver metastases (n = 105 [20.3%]).
The establishment of an HPB-SP resulted in a significant increase in comprehensive multidisciplinary assessment for each corresponding tumor type (52.6% vs 70.0% overall; P < .001) (Table 2). Patients who underwent evaluation after HPB-SP implementation had a greater likelihood of completing a comprehensive multidisciplinary assessment (odds ratio, 2.10 [95% CI, 1.40-3.14]; P < .001). This finding also held for patients with hepatobiliary tumors specifically (54.2% vs 72.1%; P = .001), and we observed a similar trend (nonsignificant) for those with pancreatic/periampullary tumors (48.7% vs 65.6%; P = .06) (data not shown).
We observed a significant improvement in the treatment provided to patients after the HPB-SP implementation (Table 2). Specifically, for patients with HCC, delivery of any treatment increased from 63.6% to 83.2% (P = .004), with increased use of percutaneous ablation (0 vs 11.5%; P = .009), liver resection (0 vs 15.3%; P = .002), and preoperative portal vein embolization to allow major liver resection (0 vs 12.2%; P = .007). Quiz Ref IDLikewise, among patients with colorectal liver metastases, a significant shift occurred from the use of ablation to the use of resection; the proportion of patients undergoing liver resections increased from 20.7% to 52.6% (P = .003), whereas those undergoing ablations decreased from 17.2% to 3.9% (P = .02). Similarly, the use of perioperative chemotherapy increased significantly after implementation of the HPB-SP (5 of 11 [45.5%] vs 33 of 43 [76.7%]; P = .01).
Postoperative adverse events occurred in 42 patients (17.5%) and were significantly more common before establishment of the HPB-SP (28.3% vs 13.9%; P = .01) (Table 3). Quiz Ref IDLogistic regression analyses showed that implementation of the HPB-SP was associated with significantly lower odds of POP-AEs even after adjusting for important covariates (model 3 odds ratio, 0.29 [95% CI, 0.12-0.68]; P = .005) (Figure 2).
Similarly, 90-day postoperative mortality after hepatobiliary operations dropped from 5.3% to 2.5% (P = .39) after HPB-SP implementation. Likewise, pancreatic operations had a trend toward decreased 30-day postoperative mortality from 4.5% to 0 (P = .10). Significant decreases were observed in the median length of stay (15 vs 10 days; P = .01) and the occurrence of severe postoperative complications (Clavien-Dindo grades III-V) for pancreatic operations (45.5% vs 15.3%; P = .004) after HPB-SP implementation. In addition, HPB-SP implementation was associated with a high proportion of margin-negative (R0) resections after liver (94.6%) and pancreatic (90.0%) resections (Table 4).
This study highlights the effect of implementing an HPB-SP on regionalization, the quality of the cancer care process, and surgical outcomes after liver and pancreatic operations within an integrated health care system in the United States. We demonstrated a 2.5-fold increase in referrals from regional VA facilities, a 4-fold increase in monthly clinic visits, and a 3-fold increase in quarterly surgical procedures after HPB-SP implementation. Further, the proportion of patients receiving a comprehensive multidisciplinary assessment increased significantly, with patients undergoing evaluation by the HPB-SP having 2.1 higher odds of accomplishing such a standard. We also observed an increase in the use of curative procedures and adherence to evidence-based treatment recommendations. Last, HPB-SP implementation was associated with significantly lower odds of POP-AEs for the whole cohort and high rates of margin-negative resection.
With the development of an organized HPB-SP, we found a dramatic increase in the clinical case volume for HPB cases in regard to clinic visits and surgical volume overall. Chang et al6 previously reported a similar effect of establishing an HCC disease-based multidisciplinary team on clinical volume and found a 2-fold increase in the overall volume during a 3-year period in another major VA institution. Such change was presumed to be related to increased referrals, although the referral patterns were never examined. The 4-fold increase in clinical volume observed in our study is a result of organized efforts to maximize referrals locally and at a regional level. Our organized efforts resulted in a broader impact and improved access to HPB care at a larger regional level. Regionalization of HPB care to high-volume centers is a well-established strategy to minimize postoperative mortality and improve overall patient outcomes.13-15 However, these efforts are limited by increased travel-related costs to the health care system, and they constitute a significant burden to patients.16 Novel strategies to support regionalization of care, such as the regional virtual tumor board program implemented with our HPB-SP, contribute by establishing efficient referral mechanisms while minimizing the burden to patients and the health care system9,10; the virtual tumor board program was critical in accomplishing success in the regionalization of HPB care for our region. As the largest integrated health care system in the United States, the VA is uniquely positioned to implement regionalization of care for HPB patients by establishing such regional-based support programs (eg, virtual tumor board program) and avoiding challenges derived from health care coverage by different payer systems.
One of the most remarkable findings from this study was the effect of the HPB-SP on the quality of cancer care, a benefit beyond the traditional volume-outcome association described for high-risk surgical procedures that has driven regionalization initiatives. Multidisciplinary assessment and treatment of primary and secondary malignant HPB neoplasms is essential, the current standard of care, and associated with improved care and overall outcomes.6,7,17-19 Few studies have examined the specific effect of multidisciplinary teams on the quality of cancer care. Pawlik et al17 described a multidisciplinary panel that convened on a single day to review the medical records of patients treated at an outside institution. They found this multidisciplinary assessment model resulted in changes in treatment recommendations in 25% of patients. Our study focused on examining the ability of the HPB-SP to accomplish such multidisciplinary assessment. Our emphasis on redesigning the process of care significantly affected the ability to receive appropriate multidisciplinary evaluation, which directly translated into a higher proportion of evidence-based treatment, in particular for patients with HCC and colorectal liver metastases, for whom liver resection became the primary treatment strategy in well-selected patients20,21 and in whom the use of oncosurgical strategies associated with improved outcomes significantly increased.22-25
Last, the establishment of an HPB-SP was associated with significantly lower odds of POP-AEs, including mortality, severe complications, and reoperations after liver and pancreatic procedures, and with a high rate of margin-negative resections. These findings further support previous reports emphasizing the association of volume with outcomes (mortality) for high-risk operations,3,26 including HPB procedures.13,27,28 However, beyond volume, standardization of perioperative care within the HPB-SP framework likely contributed to improved surgical outcomes, although future studies are needed to document such an association.29
This study has several limitations. The retrospective nature of the study is subject to selection bias, although we observed no significant differences between the 2 patient groups. Also, the changes observed over time might have been owing to unmeasured confounders; however, this change is unlikely given that relevant changes were related to those derived from the HPB-SP implementation.
We found the establishment of an HPB-SP to be associated with regionalization of care, improved multidisciplinary assessment and evidence-based treatment, improved surgical outcomes, and excellent cancer-related surgical care. This study provides evidence regarding the benefit of regionalization of care for high-complexity procedures within the VA system. The VA structure, as an integrated health care system with widespread coverage organized through established regional networks, represents the ideal setting to implement regionalization efforts with HPB-SP across the United States. Establishment of HPB surgical programs in accordance to the blueprint we have outlined can help to improve the process of regionalization while improving the quality of care and outcomes for patients with HPB tumors in the VA system. Future efforts should focus on examining differences in HPB care beyond those traditionally described in the volume-outcome literature, with a special focus on the effect on quality of cancer care. Further, differences in HPB care in the VA system outside these standards need to be explored to help support policy changes at a national level.
Accepted for Publication: June 5, 2014.
Corresponding Author: Daniel A. Anaya, MD, Houston Veterans Affairs Center for Innovations in Quality, Effectiveness, and Safety, Michael E. DeBakey Veterans Affairs Medical Center, 2002 Holcombe Blvd, Ste 152, Houston, TX 77030 (firstname.lastname@example.org).
Published Online: September 10, 2014. doi:10.1001/jamasurg.2014.1711.
Author Contributions: Dr Anaya had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Anaya.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Lau, Salami, Castillo, Anaya.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Salami, Anaya.
Obtained funding: Anaya.
Administrative, technical, and material support: Castillo, Poppelaars, Anaya.
Study supervision: Anaya.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported in part by grant P00494-S1 from the Office of Rural Health–South Central VA Veterans Integrated Service Network Clinical Systems Program Office, Telehealth and Rural Access Program. This material is also based on work supported in part by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, and grant HFP90-020 from the Houston VA Health Services Research and Development Center of Excellence.
Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The views expressed in this article are those of the authors and do not necessarily reflect the position or the policy of the Department of Veterans Affairs of the US Government or of Baylor College of Medicine.
Previous Presentation: This paper was presented at the 38th Annual Surgical Symposium of the Association of VA Surgeons; April 5, 2014; New Haven, Connecticut.
Create a personal account or sign in to: