Impact of a multidisciplinary standardized clinical pathway on perioperative outcomes in patients with oesophageal cancer

Authors


  • Presented to the Annual Meeting of the Digestive Disorders Federation, Liverpool, UK, June 2012

Abstract

Background:

Defined clinical pathways can contribute to improved outcomes in patients undergoing oesophageal cancer surgery. A standardized oesophagectomy clinical pathway (SOCP) established at the Virginia Mason Medical Center (VMMC) in Seattle, Washington, USA was introduced into the Royal Surrey County Hospital (RSCH), Guildford, UK in 2011. The aim of this study was to see whether transfer and implementation of an oesophagectomy care pathway could change postoperative outcomes significantly.

Methods:

Three consecutively accrued study groups were examined at the RSCH: patients operated on immediately before the introduction of the SOCP (group 1), patients operated on after the introduction of the SOCP but not included in the pathway (group 2), and patients managed according to the SOCP (group 3). Outcomes were compared with those of patients who had surgery at the VMMC between 2009 and 2011 using the SOCP (group 4).

Results:

There were 12 patients in each of the first three groups and 74 in group 4. All groups were similar with respect to body mass index, medical co-morbidities and clinical stage. The median age of patients in group 3 was significantly lower than that in group 1, and median American Society of Anesthesiologists score was significantly better in group 3 compared with group 4. Following initiation of the SOCP there was an increase in immediate extubation (8 of 12 in group 1 versus 12 of 12 in group 3) and first-day mobilization (1 of 12 versus 12 of 12 respectively), and a reduction in complications (9 of 12 versus 4 of 12), length of critical care stay (4 (range 2–20) days in group 1 versus 3 (1–5) days in group 3) and length of hospital stay (17 (12–30) to 7 (6–37) days respectively). Patients not on the pathway but who had surgery during the same interval experienced small but non-significant improvements in length of critical care and hospital stay, and in first-day mobilization.

Conclusion:

The study demonstrated improvement in short-term outcomes after oesophagectomy following the adoption of an established multidisciplinary standardized postoperative pathway. Copyright © 2012 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd.

Introduction

Oesophagectomy remains one of the most demanding surgical procedures, and is associated with significant morbidity and mortality. Recent studies examining the volume–outcome relationship for oesophageal resection suggest that high-volume institutions with a larger caseload and appropriate infrastructure are better prepared to deliver consistently high-quality outcomes1–3. One reason why high-volume institutions have better outcomes may be consistency in many aspects of care associated with high patient throughput. The introduction of formal clinical care pathways has facilitated a systematic approach, particularly for postoperative recovery after major surgery. These pathways provide a template for all medical personnel interacting with these patients, and can outline a goal-directed recovery for each patient. A multidisciplinary approach to the formulation and implementation of a standardized oesophagectomy clinical pathway (SOCP) is important to facilitate success. SOCPs are usually of multifaceted design in order to optimize every aspect of a patient's treatment and postoperative recovery. These pathways can provide a framework for quality improvement, leading to better postoperative outcomes and reducing costs4–6. Clinical care pathways should ideally be initiated at the time of a patient's initial referral and give a goal-directed approach to the physiological work-up, staging, treatment and recovery. Description of specific goals of the SOCP at the initial consultation will help to establish patients' and relatives' expectations during investigations and treatment.

No information exists to indicate whether multidisciplinary pathways can be transferred successfully between different cancer programmes and health systems to produce a change in short-term outcomes. The aim of this study was to determine the effect of the implementation of a validated multidisciplinary oesophagectomy care pathway on postoperative outcomes when transferred to a different healthcare system.

Methods

The Virginia Mason Medical Center (VMMC) in Seattle, Washington, USA introduced a multidisciplinary standardized oesophagectomy clinical care pathway to set specific goals to guide all aspects of surgical management in 19917. This SOCP underwent five revisions up to 2011 (Table 1). The evolution of the pathway involved representatives of all services and subspecialties involved in the patients' staging, preoperative, intraoperative and postoperative management. During this time, utilizing the pathway while prospectively documenting outcomes in an Institutional Review Board (IRB)-approved database, progressive improvement in perioperative results was noted. In March 2011, a multidisciplinary team from another high-volume oesophagogastric centre, the Royal Surrey County Hospital (RSCH) in Guildford, UK, visited VMMC to observe the application of this pathway, and in April 2011 initiated a similar multidisciplinary pathway at the RSCH (Table 2). The group included an oesophagogastric consultant surgeon, anaesthesia/intensive care unit consultant, physiotherapist, nutritionist, cancer specialist nurse and hospital manager.

Table 1. Clinical care pathway at Virginia Mason Medical Center, 2011
Postoperative dayAim or goal
  1. MAP, mean arterial pressure; PPI, proton pump inhibitor, HDU, high-dependency unit; ICU, intensive care unit.

Evening of day ofSits up in bed
 surgeryMAP > 70 mmHg
 PPI initiated
 Physiotherapy visit and introduction to incentive spirometry
Day 1Walks in the corridor in the morning before discharge from HDU, then walks of 100–200 feet 3 times on day 1
 Discharge from or step down from ICU
 Start jejunal tube feeding
 Remove apical chest drain if no air leak
Day 2 onwardsWalks 3–4 times per day ± physiotherapy consult
 Titrate epidural to facilitate mobilization and maintain MAP > 70 mmHg
Day 3, 4 or 5Chest drain 2 removed—depending on whether chest or cervical incision
 Jejunal tube feeds increased to goal feeding
 Gastrografin® (Bracco Diagnostics, Princetown, New Jersey, USA)/barium study to assess anastomosis and gastric emptying
Day 4 or 5Nasogastric tube removed
Day 5 or 6Discontinue epidural
 Switch to oral or jejunostomy tube analgesics—all medications must be crushed or given as liquid down jejunostomy tube, and PPI given orally
 Jejunal tube feeds moved to nocturnal feeding
 Dietary and home health consultation
 Limited oral liquid intake on discharge: 1 cup every 2 h
 No attempt to initiate solid diet before discharge
 Dietary team provides specific directions regarding advancing oral intake over subsequent 3–4 weeks
Day 6 or 7Planned discharge
Table 2. Clinical care pathway at Royal Surrey County Hospital, 2011
Postoperative dayAim or goal
  • *

    LiDCO Rapid (LiDCO, Cambridge, UK) provides real-time continuous assessment of a patient's haemodynamic status using a bolus indicator dilution method. Cardiac output is estimated by injection of a small dose of lithium chloride via a central or peripheral venous line; the resulting arterial lithium concentration–time curve is recorded by withdrawing blood past a lithium sensor attached the patient's existing arterial line. LiDCO Rapid facilitates stroke volume optimization by displaying the parameters including: mean arterial pressure, heart rate, stroke volume, cardiac output, pulse pressure variation and stroke volume variation in a continuous fashion. HDU, high-dependency unit.

Evening of day of surgeryGoal-directed fluid therapy with LiDCO Rapid* for 6 h to maximize fluid status
 Epidural analgesia ± additional intravenous analgesia
 Sits up in bed for 4 h
Day 1Mobilizes twice in the day as far as the patient is able
 Sits up in chair/bed for 4 h
 Start jejunal feeding (30 ml/h)
 Start 30 ml/h water orally
 Commence prophylactic anticoagulation
Day 2Mobilizes three times in the day
 Increase jejunal feeding to 50 ml/h continuously
 Remove nasogastric aspirate if < 300 ml per 24 h
 Transfer to surgical HDU
Day 3Mobilizes four times in the day
 Basal chest drain may be removed if serous output < 250 ml per 24 h
Day 4Mobilizes five times in the day
 When flatus passed, increased jejunal feeds to 80 ml/h, and oral intake to 50 ml/h including cups of tea/coffee
Day 5Mobilizes six times in the day, increasing distance of walks
 Consider discontinuation of epidural following pain team review
 Central line removal
 If bowels opened, discontinue urinary catheter
 Convert intravenous proton pump inhibitor to oral formulation
Day 6Commence free fluids
 Remove apical chest drain if clear fluid is draining
 Receive pump training for jejunostomy
Day 7Commence puréed diet and free (non-fizzy) fluids
 Dietitian education on postoperative diet and enteral feeding regimen
 Planned discharge, following physiotherapy assessment and instructions

Following adoption of the SOCP at the RSCH, consecutive patients who underwent oesophageal resection by a single surgeon as part of this pathway from April to October 2011 were allocated to one patient group (group 3). Information from these patients was collected and recorded in an IRB-approved database. Information from a further group of consecutive patients (group 2), who were operated on between April and October 2011 by other surgeons at the RSCH but not included in the pathway, were also recorded in this database. Consecutive patients who had undergone oesophageal resection at the RSCH immediately before implementation of the SOCP (group 1) were used for comparison, along with consecutive patients from VMMC who had oesophagectomy between 2009 and 2011 (group 4).

Patients were compared with respect to preoperative demographics including medical co-morbidities, clinical stage and use of neoadjuvant therapies. Perioperative data were recorded, including intraoperative blood loss, fluid administration and duration of surgery. Postoperative outcomes included length of hospital and critical care stay, and postoperative complications.

Statistical analysis

Dichotomous data were analysed with the χ2 or Fisher's exact test. Continuous variables were analysed using Student's t test or Mann–Whitney U test. All P values reported were two-tailed, and P < 0·050 was considered to be statistically significant. Univariable and multivariable regression analyses were performed to determine factors associated with morbidity (complications) and length of hospital stay. Independent variables included in univariable analysis were: treatment group; patient age; body mass image (BMI); medical co-morbidities including arrhythmia, ischaemic heart disease, diabetes, hypertension, renal insufficiency, asthma, chronic obstructive pulmonary disease and peripheral vascular disease; American Society of Anesthesiologists (ASA) score; Charlson co-morbidity index score, and SOCP. Factors with P < 0·100 in univariable analysis were then used in multivariable analysis. SPSS® version 18 (IBM, Armonk, New York, USA) was used for statistical analysis.

Results

The RSCH multidisciplinary team identified specific areas for change: immediate postoperative extubation, improved preoperative patient and family education regarding pathway goals, early mobilization protocols, and modification of in-hospital dietary goals. Conservative intraoperative fluid utilization protocols and regional anaesthesia pain protocols were already established at the RSCH and were not amended.

There were 12 patients in groups 1–3, and 74 patients in group 4 (from VMMC).

Comparisons among groups 1–3

Presenting demographics were similar in all groups with respect to BMI and medical co-morbidities (Table 3). Median age was significantly lower in group 3 than in group 1: 65 (range 47–80) versus 74·5 (58–80) years respectively (P = 0·040). Clinical stage and use of neoadjuvant chemotherapy were similar across all groups.

Table 3. Preoperative demographic data
     P
 Pre-pathway (group 1) (n = 12)Non-pathway (group 2) (n = 12)Pathway (group 3) (n = 12)VMMC (group 4) (n = 74)Group 3 versus 4Group 1 versus 3Group 1 versus 2
  • *

    Values are median (range).

  • According to American Joint Committee on Cancer/International Union Against Cancer tumour node metastasis staging system (sixth edition). VMMC, Virginia Mason Medical Center; MI, myocardial infarction; CABG, coronary artery bypass graft surgery; COPD, chronic obstructive pulmonary disease; ASA, American Society of Anesthesiologists.

  • Fisher's exact test, except

  • §

    Mann–Whitney U test and

  • χ2 test.

Age (years)*74·5 (58–80)66 (58–81)65 (47–80)65 (42–84)0·627§0·040§0·105§
Body mass index (kg/m2)*24 (18–29)27 (21–35)26 (23–33)27 (19–41)0·494§0·097§0·063§
Co-morbidities       
 Arrhythmia111120·6830·9990·999
 MI/coronary artery disease03170·9990·9990·217
 CABG01020·9990·999
 Diabetes042130·9990·4780·093
 Hypertension787200·7550·9990·999
 Renal insufficiency00110·2610·999
 Asthma01280·6250·4780·999
 COPD21060·5890·4780·999
 Deep vein thrombosis10050·9990·9990·999
 Peripheral vascular disease01030·9990·999
Smoker       
 Previous70650   
 Current30110   
ASA score*2 (2–2)2·5 (2–3)2 (1–3)3 (2–3)0·041§0·186§0·082§
Clinical stage       
 I131110·6010·9990·590
 II213310·5210·9990·999
 III988260·0310·9990·999
Neoadjuvant therapy       
 Chemotherapy9896< 0·0010·9990·999
 Chemoradiotherapy00039< 0·001

The operative approach was also similar for all groups, with laparoscopic gastric mobilization in 26 of the 36 patients. Following initiation of the SOCP, the rate of immediate extubation increased from 8 of 12 patients in group 1 to 12 of 12 patients in group 3 (P = 0·028) (Table 4). Median length of operation was longer in the pathway group compared with that in the pre-pathway group (429 (range 385–530) versus 365 (300–460) min respectively; P = 0·004). Median estimated intraoperative blood loss and intraoperative intravenous fluid administration were similar across all three groups (Table 4).

Table 4. Management strategies
     P
 Pre-pathway (group 1) (n = 12)Non-pathway (group 2) (n = 12)Pathway (group 3) (n = 12)VMMC (group 4) (n = 74)Group 3 versus 4Group 1 versus 3Group 1 versus 2
  • *

    Values are median (range). VMMC, Virginia Mason Medical Center; i.v., intravenous.

  • Student's t test, except

  • Mann–Whitney U test,

  • §

    Fisher's exact test and

  • χ2 test.

Surgical approach       
 Laparoscopic or open81170   
 Open, two-phase41535   
 Left thoracoabdominal00034   
Duration of surgery (min)*365 (300–460)435 (360–660)429 (385–530)416 (244–664)0·3400·0040·006
Blood loss (ml)*403 (213–1000)600 (250–1100)422 (100–940)150 (50–400)0·0040·5640·156
Intraoperative i.v. fluids (ml)*2475 (2000–4750)3000 (1750–5500)2375 (1500–4500)2700 (1250–7900)0·1990·4660·212
Immediate extubation8812740·9990·0280·999
First-day mobilization1512690·999< 0·001§0·155§

Fewer patients experienced postoperative complications following introduction of the clinical pathway (9 of 12 in group 1 versus 4 of 12 in group 3; P = 0·041), although no individual complication was significantly different between the groups (Table 5). Multivariable logistic regression analysis indicated that the SOCP was the only variable significantly associated with a reduced incidence of postoperative complications (β = 0·391, 95 per cent confidence interval (c.i.) 0·003 to 0·490; P = 0·025).

Table 5. Postoperative outcomes
     P
 Pre-pathway (group 1) (n = 12)Non-pathway (group 2) (n = 12)Pathway (group 3) (n = 12)VMMC (group 4) (n = 74)Group 3 versus 4Group 1 versus 3Group 1 versus 2
  • *

    Values are median (range). VMMC, Virginia Mason Medical Center.

  • Fisher's exact test, except

  • Mann–Whitney U test and

  • §

    χ2 test.

Complications994350·534§0·041§0·999§
 Anastomotic leak10130·4580·9990·999
 Chyle leak31010·9990·2170·590
 Reoperation10110·2610·9990·999
 Arrhythmia133230·6700·5900·590
 Myocardial infarction0000
 Pneumonia451100·6180·3170·999
 Pleural effusion31140·5370·5900·590
 Pneumothorax02140·5370·9990·478
 Reintubation12030·9990·9990·999
 Renal failure00120·3670·999
 Stroke00010·999
 Deep vein thrombosis00030·999
 Pulmonary embolus00030·999
 Wound infection00020·999
 Delirium00030·999
Critical care stay (days)*4 (2–20)3 (2–9)3 (1–5)1 (0–22)< 0·0010·0220·087

Significant improvements seen following implementation of the clinical pathway included the proportion of patients mobilizing on postoperative day 1 (1 of 12 in group 1 versus 12 of 12 in group 3; P < 0·001) (Table 4), median length of critical care stay (4 (range 2–20) versus 3 (1–5) days respectively; P < 0·001) and median length of hospital stay (17 (12–30) versus 7 (6–37) days; P = 0·022) (Table 5). Multivariable linear regression analysis confirmed the SOCP to be the only variable significantly associated with a reduced length of hospital stay (β = − 7·171, 95 per cent c.i. − 13·241 to − 1·092; P = 0·002).

For patients who had surgery after the introduction of the SOCP but were not included in the pathway (group 2), there were smaller improvements, which failed to reach statistical significance, in median length of critical care stay (3 (range 2–9) days versus 4 (2–20) days in group 1; P = 0·129), hospital stay (13 (8–22) versus 17 (12–30) days respectively; P = 0·087) and first-day mobilization (5 of 12 versus 1 of 12; P = 0·155) (Tables 4 and 5).

Comparison between groups 3 and 4

Age was similar in these two groups. Median ASA score was significantly lower in group 3 (2 (range 1–3) versus 3 (2–3) in group 4; P = 0·041). The RSCH pathway group contained a greater proportion of stage III oesophageal cancers (8 of 12 versus 26 of 74 respectively; P = 0·031). Chemotherapy alone was used preferentially at RSCH (9 of 12 patients) and chemoradiotherapy preferentially at VMMC (39 of 74 patients) (Table 3).

Laparoscopic gastric mobilization was employed in 7 of 12 patients at the RSCH, whereas only open thoracoabdominal operations were performed at VMMC. Median estimated blood loss at RSCH was greater in group 3 (422 (range 100–940) ml versus 150 (50–400) ml in group 4; P = 0·004) (Table 4). The two groups had similar postoperative outcomes in terms of complications, first-day mobilization and length of hospital stay (Tables 4 and 5). Median length of critical care stay was significantly shorter at VMMC (1 (range 0–22) versus 3 (1–5) in group 3; P < 0·001). No patient in any group died in hospital or within 30 days of surgery.

Discussion

Hospitals and healthcare systems are increasingly encouraged to optimize both use of resources and clinical outcomes. Despite recent improvements in postoperative mortality following oesophagectomy, morbidity and length of hospital stay remain highly variable8–10. Even after minimally invasive oesophagectomy, the reported median length of hospital stay is 15 (range 12–22) days in the UK11. Standardized clinical care pathways provide a goal-directed approach for patients, typically involving initial assessment, procedure selection, intraoperative management and postoperative care. Previous reports have demonstrated that standardized clinical care pathways can reduce length of stay and costs, not only in patients undergoing oesophagectomy5–7, 12–14 but also in those having a wide variety of major operations15–17.

Over the 6-month interval covered by the present study, changes in management included an increase in the rate of immediate extubation and in the proportion of patients mobilizing on the first postoperative day. There were subsequent significant reductions in postoperative complication rates, as well as shorter critical care and overall hospital stays following implementation of the multidisciplinary standardized clinical pathway.

An initial impression may suggest that the improvements observed in this study are the result of a short-term concerted effort and surgery in a selected group of patients. However, measures that involve a change from a third of patients remaining intubated overnight to all patients mobilizing on the morning of the first postoperative day demonstrate how much processes are based on historical perceptions of immediate postoperative management. Improvement in length of hospital stay from 17 to 7 days, with no increase in mortality, complications or readmissions over a 6-month period, shows that pathways that took 20 years to evolve in one centre can be implemented in a short time period in another high-volume centre, as long as there is an institutional commitment to changing process, perceptions and infrastructure. This commitment must exist in both clinical and administrative personnel, and be accepted by all of the allied health personnel involved in the postoperative care of patients.

The main limitation of this study is the relatively small number of patients included, owing to the short period of time studied in order to demonstrate a rapid improvement in outcomes. There were some differences between patients in the pathway group and those in the pre-pathway group, and it should be borne in mind that the results pertain only to short-term outcomes. It remains to be seen whether this improvement in perioperative outcomes can be sustained, as has been seen at the VMMC7.

Historically, changes in surgical practice have been initiated after visits by lead surgeons to other units, and have involved modifications in operative technique and factors directly under the control of the surgeon. The selected members of the visiting team illustrate the importance of a multidisciplinary approach with the involvement of the entire healthcare team in the design, implementation and evaluation of the pathway. Each member of the team was selected for the visit because they were key stakeholders in critical areas that would require significant modification to standard processes. This approach enables each member to act as a lead within their own department to facilitate the implementation of a standardized clinical care pathway within a global change of approach to perioperative care. Previous studies5–7, 12–14 have used terms such as ‘fast-track’ and ‘enhanced recovery’, which potentially fail to describe completely the infrastructure and multidisciplinary approach taken in the design, implementation and evolution of standardized clinical pathways in the treatment of oesophageal cancer.

After the visit of the multidisciplinary team to VMMC, several changes were targeted for immediate implementation within each department at RSCH. This included a change that enabled operating theatre staff to begin work at the earlier time of 06.00 hours. With a typical duration of surgery for oesophagectomy of 6–7 h, this allowed operations to be completed in the middle of the afternoon when staffing was at normal levels. This probably facilitated the assessment and monitoring of the immediately extubated patient. Although early extubation has been demonstrated previously to reduce mortality and morbidity rates18, it had the additional benefit of reducing the need for postoperative intensive care unit admission, instead allowing patients to be managed in a high-dependency setting and reducing overall cost. Beds that can be adjusted to take the patient from a horizontal to a sitting position and lightweight chest drainage units were purchased in order to facilitate early mobilization—a financial outlay that was offset by reduced length of hospital stay.

The VMMC visit provided a change in staff perceptions regarding the potential for an expedited recovery following oesophagectomy. Good preoperative education regarding the elements of the SOCP is necessary to guide patients and relatives about what is expected in the postoperative period. Nutrition is a key issue, and a dietetics service plays an important role in this communication. Resumption of an oral diet may take several weeks, and routine discharge with supplementary feeding via a jejunostomy tube facilitates discharge and avoids many of the problems with attempts to consume a full oral diet shortly after oesophagectomy.

An interesting finding in the present study was the improvement in outcomes seen in patients not taking part in the pathway but operated on at the time the pathway was in place at the RSCH. First-day mobilization, length of critical care stay and length of overall hospital stay were all better than historical outcomes before the SOCP was introduced. This indicates that the introduction of a SOCP may lead to improved hospital processes and modification in infrastructure, which can benefit other patients.

Improvements in perioperative outcomes after oesophagectomy are multifactorial. Meticulous surgical technique remains critical, but improvements in hospital processes and infrastructure are probably the areas most likely to lead to better short-term outcomes. SOCPs provide a template for all medical personnel interacting with these patients, and can outline a goal-directed recovery for each patient. Involving an entire healthcare team in the design, implementation, evaluation and revision of these pathways ensures that all team members are committed to achieving specific goals that enhance recovery after oesophagectomy.

Contributors

Other important contributors to this manuscript, who did not fulfil the criteria for authorship, include: Dr M. Carraretto (Department of Anaesthesia and Intensive Care, RSCH), G. Pinn (Cancer Nurse Co-ordinator, RSCH), S. Gregory (Department of Physiotherapy, RSCH), F. Macharg (Nutritionist, RSCH), N. Waring-Edkins (Department of Surgery, RSCH), S. Herson (Intensive Care Unit, RSCH), and Dr D. Nickinovich (Health Science Matrix, Medical Outcomes Research Design and Analysis, Seattle, Washington, USA), who carried out the statistical analysis for this study.

Acknowledgements

The authors wish to acknowledge the Ryan Hill Research Foundation for their support for this work.

Disclosure: The authors declare no conflict of interest.

Ancillary