Surgical readmissions: results of integrating pre-, peri- and postsurgical care

AIMS

To explore the feasibility of recruiting surgical oncology patients and implementing a surgical integrated discharge (SID) programme led by advanced practice providers (APP).

BACKGROUND

Burden of illness and complexity of treatment regimen makes it challenging for surgical oncology patients to participate in research. Surgical oncology nurses may have the necessary expertise to overcome this problem.

DESIGN

Controlled longitudinal prospective observational study.

METHODS

The SID programme included multidisciplinary care coordination, regular communication among APPs and proactive postdischarge follow-up. Administrative databases were used to identify matching historical controls (n = 113) and evaluate programme outcomes.

RESULTS

Patient enrolment was 84%. The main challenges for the programme implementation included incompatible health information systems among care settings, variation in care processes among hospital units and need for provider behaviour change.

CONCLUSIONS

Most surgical oncology patients are willing to participate in outcomes programmes when contacted by familiar clinical personnel but programme implementation requires leadership support, communication among care teams and training and infrastructure.

High Variability in Nosocomial Clostridium difficile Infection Rates Across Hospitals After Colorectal Resection

BACKGROUND

Hospital-acquired Clostridium difficile infection is associated with adverse patient outcomes and high medical costs. The incidence and severity of C. difficile has been rising in both medical and surgical patients.

OBJECTIVE

Our aim was to assess risk factors and variation associated with the development of nosocomial C. difficile colitis among patients undergoing colorectal resection.

DESIGN

This was a retrospective cohort study.

SETTINGS

The study included segmental colectomy and proctectomy cases in New York State from 2005 to 2013.

PATIENTS

The study cohort included 150,878 colorectal resections. Patients with a documented previous history of C. difficile infection or residence outside of New York State were excluded.

MAIN OUTCOME MEASURES

A diagnosis of C. difficile colitis either during the index hospital stay or on readmission within 30 days was the main measure.

RESULTS

C. difficile colitis occurred in 3323 patients (2.2%). Unadjusted C. difficile colitis rates ranged from 0% to 11.3% among surgeons and 0% to 6.8% among hospitals. After controlling for patient, surgeon, and hospital characteristics using mixed-effects multivariable analysis, significant unexplained variation in C. difficile rates remained present across hospitals but not surgeons. Patient factors explained only 24% of the total hospital-level variation, and known surgeon and hospital-level characteristics explained an additional 8% of the total hospital-level variation. Therefore, ≈70% of the hospital variation in C. difficile infection rates remained unexplained by captured patient, surgeon, and hospital factors. Furthermore, there was an ≈5-fold difference in adjusted C. difficile rates across hospitals.

LIMITATIONS

A limited set of hospital and surgeon characteristics was available.

CONCLUSIONS

Colorectal surgery patients appear to be at high risk for C. difficile infection, and alarming variation in nosocomial C. difficile infection rates currently exists among hospitals after colorectal resection. Given the high morbidity and cost associated with C. difficile colitis, adopting institutional quality improvement programs and maintaining strict prevention strategies are of the utmost importance.

Evaluating the Prognostic Role of Elevated Preoperative Carcinoembryonic Antigen Levels in Colon Cancer Patients: Results from the National Cancer Database

OBJECTIVES

Carcinoembryonic antigen (CEA) is a reliable tumor marker for the management and surveillance of colon cancer. However, limitations in previous studies have made it difficult to elucidate whether CEA should be established as a prognostic indicator. This study examines the association between elevated preoperative CEA levels and overall survival in colon cancer patients using a national population-based registry.

METHODS

Stage I-III colon cancer patients were identified from the 2004-2006 National Cancer Database. A multivariable Cox proportional hazards model was used to estimate the association between elevated CEA levels and overall survival after controlling for important patient, hospital, and tumor characteristics. A Monte Carlo Markov Chain was used to impute the large degree of missing CEA data. All models controlled for the propensity score in order to account for selection bias.

RESULTS

A total of 137,381 patients met the inclusion criteria. Overall, 34 % of patients had an elevated CEA level and 66 % had a normal CEA level, with a median survival of 70 and 100 months, respectively. Patients with an elevated CEA level had a 62 % increase in the hazard of death (HR 1.62, 95 % CI 1.53-1.74) compared with patients with a normal CEA level.

CONCLUSIONS

Preoperative CEA was an independent predictor of overall survival across all stages. The results support recommendations to include CEA levels as another high-risk feature that clinicians can use to counsel patients on adjuvant chemotherapy, especially for stage II patients.

Watch and Wait?--Elevated Pretreatment CEA Is Associated with Decreased Pathological Complete Response in Rectal Cancer

INTRODUCTION

Between 10 and 30% of rectal cancer patients experience pathological complete response after neoadjuvant treatment. However, physiological factors predicting which patients will experience tumor response are largely unknown. Previous single-institution studies have suggested an association between elevated pretreatment carcinoembryonic antigen and decreased pathological complete response.

METHODS

Clinical stage II-III rectal cancer patients undergoing neoadjuvant chemoradiotherapy and surgical resection were selected from the 2006-2011 National Cancer Data Base. Multivariable analysis was used to examine the association between elevated pretreatment carcinoembryonic antigen and pathological complete response, pathological tumor regression, tumor downstaging, and overall survival.

RESULTS

Of the 18,113 patients meeting the inclusion criteria, 47% had elevated pretreatment carcinoembryonic antigen and 13% experienced pathological compete response. Elevated pretreatment carcinoembryonic antigen was independently associated with decreased pathological complete response (OR = 0.65, 95% CI = 0.52-0.77, p < 0.001), pathological tumor regression (OR = 0.74, 95% CI = 0.67-0.70, p < 0.001), tumor downstaging (OR = 0.77, 95% CI = 0.63-0.92, p < 0.001), and overall survival (HR = 1.45, 95% CI = 1.34-1.58, p < 0.001).

CONCLUSION

Rectal cancer patients with elevated pretreatment carcinoembryonic antigen are less likely to experience pathological complete response, pathological tumor regression, and tumor downstaging after neoadjuvant treatment and experience decreased survival. These patients may not be suitable candidates for an observational "watch-and-wait" strategy. Future prospective studies should investigate the relationships between CEA levels, neoadjuvant treatment response, recurrence, and survival.

Significant Variation in Blood Transfusion Practice Persists following Upper GI Cancer Resection

PURPOSE

Perioperative blood transfusions are costly and linked to adverse clinical outcomes. We investigated the factors associated with variation in blood transfusion utilization following upper gastrointestinal cancer resection and its association with infectious complications.

METHODS

The Statewide Planning and Research Cooperative System was queried for elective esophagectomy, gastrectomy, and pancreatectomy for malignancy in NY State from 2001 to 2013. Bivariate and hierarchical logistic regression analyses were performed to assess the factors associated with receiving a perioperative allogeneic red blood cell transfusion. Additional multivariable analysis examined the relationship between transfusion and infectious complications.

RESULTS

Among 14,875 patients who underwent upper GI cancer resection, 32 % of patients received a perioperative blood transfusion. After controlling for patient, surgeon, and hospital-level factors, significant variation in transfusion rates was present across both surgeons (p < 0.0001) and hospitals (p < 0.0001). Receipt of a blood transfusion was also independently associated with wound infection (OR = 1.68, 95% CI = 1.47 and 1.91), pneumonia (OR = 1.98, 95% CI = 1.74 and 2.26), and sepsis (OR = 2.49, 95% CI = 2.11 and 2.94).

CONCLUSION

Significant variation in perioperative blood transfusion utilization is present at both the surgeon and hospital level. These findings are unexplained by patient-level factors and other known hospital characteristics, suggesting that variation is due to provider preferences and/or lack of standardized transfusion protocols. Implementing institutional transfusion guidelines is necessary to limit unwarranted variation and reduce infectious complication rates.

The pitfalls of inguinal herniorrhaphy: Surgeon volume matters

BACKGROUND

There is currently little information regarding the impact of procedure volume on outcomes after open inguinal hernia repair in the United States. Our hypothesis was that increasing procedure volume is associated with lesser rates of reoperation and resource use.

METHODS

The database of the Statewide Planning and Research Cooperative System was queried for elective open initial inguinal hernia repairs performed in New York State from 2001 to 2008 via the use of International Classification of Diseases, 9th Revision and Current Procedural Terminology codes. Surgeon and hospital procedure volumes were grouped into tertiles based on the number of open inguinal hernia repairs performed per year. Bivariate, hierarchical mixed effects Cox proportional-hazards, and negative binomial regression analyses were performed assessing for factors associated with reoperation for recurrence, procedure time, and downstream total charges.

RESULTS

Among 151,322 patients who underwent open inguinal hernia repair, the overall rate of reoperation for recurrence within 5 years was 1.7% with a median time to reoperation of 1.9 years. An inverse relationship was seen between surgeon volume and reoperation rate, procedure time, and health care costs (P < .001). After we controlled for surgeon, facility, operative and patient characteristics, low-volume surgeons (<25 repairs/year) had greater rates of reoperation (hazard ratio 1.23,95% confidence interval [95% CI] 1.11-1.36), longer procedure times (incidence rate ratio 1.22, 95% CI 1.21-1.24), and greater downstream costs (incidence rate ratio 1.13,95% CI 1.10-1.17) than high-volume surgeons (≥25 repairs/year).

CONCLUSION

Surgeon volume <25 cases per year for open inguinal hernia repair was independently associated with greater rates of reoperation for recurrence, worse operative efficiency, and greater health care costs. Referral to surgeons who perform ≥25 inguinal hernia repairs per year should be considered to decrease reoperation rates and resource use.

Extended Intervals after Neoadjuvant Therapy in Locally Advanced Rectal Cancer: The Key to Improved Tumor Response and Potential Organ Preservation

BACKGROUND

Many rectal cancer patients experience tumor downstaging and some are found to achieve a pathologic complete response (pCR) after neoadjuvant chemoradiotherapy (nCRT). Previous data suggest that there is an association between the time interval from nCRT completion to surgery and tumor response rates, including pCR. However, these studies have been primarily from single institutions with small sample sizes. The aim of this study was to examine the relationship between a longer interval after nCRT and pCR in a nationally representative cohort of rectal cancer patients.

STUDY DESIGN

Clinical stage II to III rectal cancer patients undergoing nCRT with a documented surgical resection were selected from the 2006 to 2011 National Cancer Data Base. Multivariable logistic regression analysis was used to assess the association between the nCRT-surgery interval time (<6 weeks, 6 to 8 weeks, >8 weeks) and the odds of pCR. The relationship between nCRT-surgery interval, surgical morbidity, and tumor downstaging was also examined.

RESULTS

Overall, 17,255 patients met the inclusion criteria. An nCRT-surgery interval time >8 weeks was associated with higher odds of pCR (odds ratio [OR] 1.12, 95% CI 1.01 to 1.25) and tumor downstaging (OR 1.11, 95% CI 1.02 to 1.25). The longer time delay was also associated with lower odds of 30-day readmission (OR 0.82, 95% CI 0.70 to 0.92).

CONCLUSIONS

An nCRT-surgery interval time >8 weeks results in increased odds of pCR, with no evidence of associated increased surgical complications compared with an interval of 6 to 8 weeks. These data support implementation of a lengthened interval after nCRT to optimize the chances of pCR and perhaps add to the possibility of ultimate organ preservation (nonoperative management).

Timing of discharge: a key to understanding the reason for readmission after colorectal surgery

PURPOSE

There is a growing interest in surgery regarding the balance between appropriate hospital length of stay (LOS) and prevention of unnecessary readmissions. This study examines the relationship between postoperative LOS and unplanned readmission after colorectal resection, exploring whether patients discharged earlier have different readmission risk profiles.

METHODS

Patients undergoing colorectal resection were selected by Common Procedural Terminology (CPT) code from the 2012 ACS National Surgical Quality Improvement Program (NSQIP) database. Patients were stratified by LOS quartile. Kaplan-Meier analysis was used to examine characteristics associated with 30-day unplanned readmission. Factors with a p < 0.1 were included in the Cox proportional hazards model. Subsequently, chi-square analysis compared LOS, patient, and perioperative factors with the primary reason for readmission. Factors with a p < 0.2 were included in a multivariable logistic regression for each readmission reason.

RESULTS

For 33,033 patients undergoing colorectal resection, the overall 30-day unplanned readmission rate was 11 %. After adjusting for patient and perioperative factors, a postoperative LOS ≥8 days was associated with a 55 % increase in the relative hazard of readmission. Patients with a ≤3-day LOS were more likely to be readmitted with ileus/obstruction (odds ratio (OR): 1.8, p = 0.001) and pain (OR: 2.2, p = 0.007). LOS was not significantly associated with readmission for intraabdominal infection or medical complications.

CONCLUSIONS

Patients with longer LOS and complicated hospital courses continue to be high risk post-discharge, while straightforward early discharges have a different readmission risk profile. More targeted readmission prevention strategies are critical to focusing resource utilization for colorectal surgery patients.

Parastomal hernia: a growing problem with new solutions

Parastomal hernia is one of the most common complications following stoma creation and its prevalence is only expected to increase. It often leads to a decrease in the quality of life for patients due to discomfort, pain, frequent ostomy appliance leakage, or peristomal skin irritation and can result in significantly increased healthcare costs. Surgical technique for parastomal hernia repair has evolved significantly over the past two decades with the introduction of new types of mesh and laparoscopic procedures. The use of prophylactic mesh in high-risk patients at the time of stoma creation has gained attention in lieu of several promising studies that have emerged in the recent days. This review will attempt to demonstrate the burden that parastomal hernias present to patients, surgeons, and the healthcare system and also provide an overview of the current management and surgical techniques at both preventing and treating parastomal hernias.

Defining high risk: cost-effectiveness of extended-duration thromboprophylaxis following major oncologic abdominal surgery

PURPOSE

Extended-duration thromboprophylaxis (EDTPPX) is the practice of prescribing antithrombotic therapy for 21 days after discharge, commonly used in surgical patients who are at high risk for venothromboembolism (VTE). While guidelines recommend EDTPPX, criteria are vague due to a paucity of data. The criteria can be further informed by cost-effectiveness thresholds. This study sought to determine the VTE incidence threshold for the cost-effectiveness of EDTPPX compared to inpatient prophylaxis.

METHODS

A decision tree was used to compare EDTPPX for 21 days after discharge to 7 days of inpatient prophylaxis with base case assumptions based on an abdominal oncologic resection without complications in an otherwise healthy individual. Willingness to pay was set at $50,000/quality-adjusted life year (QALY). Sensitivity analyses were performed to assess uncertainty within the model, with particular interest in the threshold for cost-effectiveness based on VTE incidence.

RESULTS

EDTPPX was the dominant strategy when VTE probability exceeds 2.39 %. Given a willingness to pay threshold of $50,000/QALY, EDTPPX was the preferred strategy when VTE incidence exceeded 1.22 and 0.88 % when using brand name or generic medication costs, respectively.

CONCLUSIONS

EDTPPX should be recommended whenever VTE incidence exceeds 2.39 %. When post-discharge estimated VTE risk is 0.88-2.39 %, patient preferences about self-injections and medication costs should be considered.

Anastomotic leak or organ space surgical site infection: What are we missing in our quality improvement programs?

BACKGROUND

Expert panels of colorectal surgeons consistently rank anastomotic leak as among the most important quality metrics for colectomies. Nonetheless, most administrative and clinical databases do not collect data on anastomotic leaks and rely on reported organ space surgical site infections (OSI) as a proxy for identifying anastomotic leaks. This study questions the validity of using OSI as a surrogate for anastomotic leak.

METHODS

The Upstate New York Surgical Quality Initiative (UNYSQI) is a collaboration of 12 hospitals that prospectively collects colectomy-specific metrics, including anastomotic leak, in addition to standard National Surgical Quality Improvement Program (NSQIP) data, including OSIs. Cases with an organ space infection and/or anastomotic leak were selected from the 2010-2011 UNYSQI database. Patient characteristics and outcomes were compared for cases with organ space infections and anastomotic leaks.

RESULTS

Overall, 3% of colectomies had a reported organ space infection and 4% had an anastomotic leak. Among cases having anastomotic leaks, only 25% were also coded as having an organ space infection, leaving 75% of anastomotic leaks not captured by the NSQIP database (κ = 0.272; P ≤ .001).

CONCLUSION

Organ space infection is a poor surrogate for anastomotic leak, resulting in grossly underestimated leak rates and seemingly represents different postoperative courses. Procedure-specific quality measures for colorectal surgery should include data collection on anastomotic leaks to provide accurate data for use in improving patient care.