SPECIAL ARTICLE. Interhospital transfer delays emergency abdominal surgery and prolongs stay. Introduction. Alexandra M. Limmer and Michael B.

Transcription

1 SPECIAL ARTICLE ANZJSurg.com Interhospital transfer delays emergency abdominal surgery and prolongs stay Alexandra M. Limmer and Michael B. Edye Department of Surgery, Western Sydney University, Sydney, New South Wales, Australia Key words abdominal pain, length of stay, patient transfer, surgery, time to treatment. Correspondence Dr Alexandra M. Limmer, School of Medicine, Western Sydney University, Goldsmith Avenue, Campbelltown, NSW 2560, Australia. A. M. Limmer MBBS (Hons); M. B. Edye FRACS. Accepted for publication 22 September doi: /ans Abstract Background: Interhospital transfer of patients requiring emergency surgery is common practice. It has the potential to delay surgical intervention, increase rate of complications and thus length of hospital stay. Methods: A retrospective cohort study was conducted of adult patients who underwent emergency surgery for abdominal pain at a large metropolitan hospital in New South Wales (Hospital A) in The impact of interhospital transfer on time to surgical intervention, post-operative length of stay and overall length of stay was assessed. Results: Of the 910 adult patients who underwent emergency surgery for abdominal pain at Hospital A in 2013, 31.9% (n = 290) were transferred by road ambulance from a local district hospital (Hospital B). The leading surgical procedures performed were appendicectomy (n = 299, 32.9%), cholecystectomy (n = 174, 19.1%), gastrointestinal endoscopy (n = 95, 10.4%), cystoscopy (n = 86, 9.5%), hernia repair (n = 45, 4.9%), salpingectomy (n = 19, 2.1%) and oversewing of perforated peptic ulcer (n = 13, 1.4%). Overall, interhospital transfer (n = 290, 31.9%) was associated with increases in mean time to surgical intervention (14.2 h, P < 0.001), post-operative length of stay (1.1 days, P = 0.001) and overall length of stay (1.6 days, P < 0.001). Delayed surgical intervention was observed across all procedure types except surgery for perforated peptic ulcer, where transferred patients underwent surgery within a comparable timeframe to direct admissions. Conclusion: Interhospital transfer delays surgical intervention and increases length of hospital stay. This mandates attention due to the implications for patient outcomes and added burden to the healthcare system. The system did, however, show capability to appropriately expedite surgery for acutely life-threatening cases. Introduction Abdominal pain is a leading presenting complaint to Australian emergency departments (EDs) and general practices. It comprises approximately 31% of ED pain presentations 1 and is one of the most common indications for emergency surgery. 2 Many peripheral hospitals do not provide 24-h emergency surgical services and thus rely on interhospital transfer of acute surgical patients. Early recognition and transfer of these patients is vital in facilitating timely provision of emergency surgery. The literature examining the impact of interhospital transfer on time to surgical intervention and length of stay in an Australian context is modest and primarily in a rural setting. This study examined a year of emergency surgery for abdominal pain at a large metropolitan hospital in New South Wales, Hospital A. Hospital A provides 24-h emergency surgery in a diverse range of subspecialties including general surgery, urology, obstetrics and gynaecology, orthopaedics and cardiothoracic surgery. Approximately one-third of all admissions to Hospital A are transferred from a local district hospital, Hospital B, which is located approximately 20 min away by road and operates in close affiliation. Hospital B provides predominantly elective surgery in a limited range of subspecialties (general, orthopaedic and maxillofacial surgery) between 08:00 and 17:00 hours. Apart from rare circumstances where emergency surgical cases can be accommodated at Hospital B (less than 1% of cases), the vast majority of patients needing surgical intervention are transferred via road ambulance to Hospital A. Interhospital transfer is vulnerable to delay at multiple levels. These include administrative delays associated with organizing transport and transfer of care, the physical distance between the ANZ J Surg 87 (2017)

2 868 Limmer and Edye two hospitals, loss of continuity of care and the need for repeat triage and assessment of the patient at the receiving hospital. 3 5 Delayed surgical intervention has the potential to allow progression of intra-abdominal pathology; leading to higher incidence of complications. In acute appendicitis, for instance, rate of appendiceal perforation has been shown to increase from 28.8% for appendicectomy performed on the day of admission to 33.3% on day 2 and 78.8% on day 8. 6 Patients with perforated appendicitis have longer length of stay and higher total hospital charges. 6 Pathologies such as major intra-abdominal haemorrhage, perforated viscus or bowel ischaemia may also be rapidly fatal or cause significant morbidity without prompt surgical intervention. 7 Thus, delayed surgical intervention not only has implications for the individual patient but may prolong length of stay which places additional financial and material burden on the Australian healthcare system. With these considerations, this study aimed to assess the impact of interhospital transfer on time to surgical intervention, post-operative and overall length of stay. Methods Study design A retrospective, closed cohort study was conducted using electronic medical records (emr, Cerner Powerchart). The medical record numbers of all patients who underwent emergency diagnostic or therapeutic surgery at Hospital A in 2013 were obtained from an electronic surgical database. Inclusion criteria were: (1) Age 18 years or over. (2) Presentation to Hospital A ED; either directly or transferred from Hospital B ED. (3) Abdominal pain or tenderness; as a primary presenting complaint, associated complaint or elicited on routine physical examination. (4) Emergency surgery performed for the diagnosis or treatment of abdominal pain in calendar year Data collected were age, sex, presence of specific comorbidities (diabetes mellitus, hypertension and ischaemic heart disease), admission details (date and time of presentation, transfer and discharge) and procedure details (procedure type, date and time of procedure). Main outcome measures were time to surgical intervention, post-operative length of stay and overall length of stay. Patients were categorized into two groups based on admission status. The transferred patients group was defined as patients who presented initially to Hospital B ED and were subsequently transferred via road ambulance to Hospital A for surgical intervention. The direct admissions group included all patients who presented initially to Hospital A ED. Emergency surgical procedures were classified into eight groups for analysis: appendicectomy, cholecystectomy, gastrointestinal endoscopy, cystoscopy, hernia repair, salpingectomy, oversewing of perforated peptic ulcer and other. Appendicectomy, cholecystectomy, salpingectomy and oversewing of perforated peptic ulcer included both open and laparoscopic procedures. Gastrointestinal endoscopy was defined as any diagnostic or therapeutic gastroscopy, colonoscopy or sigmoidoscopy with or without biopsy or intervention. Cystoscopy was defined as any cystoscopic examination of the urinary tract with or without retrograde pyelogram, insertion of ureteric stents or other intervention. Time to surgical intervention was defined as the time between admission to either Hospital B or Hospital A ED and patient arrival in the operating theatre. Post-operative length of stay was defined as the time between departing the operating theatre and discharge from hospital. Overall length of stay was defined as the total time from admission to discharge. The null hypothesis was that interhospital transfer was not associated with increased time to surgical intervention, post-operative or overall length of stay. Statistical analysis Data were recorded and analysed using SPSS software, version 20 (SPSS Inc., Chicago, IL, USA). Descriptive statistics were computed including frequencies and percentages for categorical variables and means and standard deviations for continuous variables. Pearson s chi-squared tests were performed to compare patient characteristics and procedure type between transferred patients and direct admissions. Distribution of continuous variables was tested for normality and Mann Whitney U-tests were conducted to evaluate differences in mean time to surgical intervention, mean postoperative length of stay and mean overall length of stay for each procedure type based on admission status. Subgroup analysis was performed to independently assess the impact of admission status and patient comorbidities on length of stay. Two-tailed statistical significance was set at P < Ethics approval for the study was received from the Human Research Ethics Committee of the Western Sydney Local Health District, ethics approval ID HREC2014/3/6.4(3950)QA. Results Characteristics of study population A total of 910 adult patients underwent emergency surgery for abdominal pain at Hospital A in Of these, 620 (68.1%) were direct admissions and 290 (31.9%) were transferred from Hospital B by road ambulance. Females comprised 58.0% (n = 528) of the study population and mean age was 44 years (range 18 to 95 years). Appendicectomy was the leading procedure type (n = 299, 32.9%) followed by cholecystectomy (n = 174, 19.1%), gastrointestinal endoscopy (n = 95, 10.4%), cystoscopy (n = 86, 9.5%), hernia repair (n = 45, 4.9%), salpingectomy (n = 19, 2.1%) and oversewing of perforated peptic ulcer (n = 13, 1.4%). The remaining 19.7% of procedures (n = 179) were listed as other. Transferred patients had a higher prevalence of diabetes mellitus (17.9 versus 10.0%, P = 0.001) and hypertension (26.6 versus 19.4%, P = 0.01) compared with direct admissions. There were otherwise no significant differences in patient characteristics or procedure type between transferred patients and direct admissions (Table 1).

3 Interhospital transfer and surgical delay 869 Table 1 Characteristics of study population by admission status Characteristic Transferred patients (n = 290) Direct admissions (n = 620) χ 2 P value Female sex 163 (56.2) 365 (58.9) 0.45 Age (years): mean (SD) 45.6 (18.8) 43.3 (18.5) 0.06 Comorbidities Diabetes mellitus 52 (17.9) 62 (10.0) Hypertension 77 (26.6) 120 (19.4) 0.01 Ischaemic heart disease 18 (6.2) 28 (4.5) 0.28 Procedure type Appendicectomy 90 (31.0) 209 (33.7) 0.14 Cholecystectomy 52 (17.9) 122 (19.7) Cystoscopy 26 (9.0) 60 (9.7) Gastrointestinal endoscopy 42 (14.5) 53 (8.5) Hernia repair 14 (4.8) 31 (5.0) Oversewing of perforated peptic ulcer 6 (2.1) 7 (1.1) Salpingectomy 3 (1.0) 16 (2.6) Other 57 (19.7) 122 (19.7) Values are expressed as n (%). Mann Whitney U-test Time to surgical intervention Time to surgical intervention was a mean 46.9 h for transferred patients compared to 32.7 h for direct admissions (P < 0.001). Transferred patients spent a mean 8.1 h in Hospital B ED and a mean 28 min in transit to Hospital A. Mean time to surgical intervention for specific procedures by admission status is displayed in Table 2. Interhospital transfer was associated with a significant increase in time to surgical intervention for patients who underwent appendicectomy, cholecystectomy and cystoscopy. Surgery for perforated peptic ulcer was performed at an average of 10.3 h from presentation in transferred patients versus 11.7 h in direct admissions (P = 0.40). Length of stay Post-operative length of stay was 4.2 days for transferred patients versus 3.1 days for direct admissions (P = 0.001). Overall length Table 2 Outcomes by admission status Outcome: mean SD Transferred patients (n = 290) Direct admissions (n = 620) Mean difference P value Time to surgical intervention (hours) Appendicectomy <0.001 Cholecystectomy <0.001 Cystoscopy Gastrointestinal endoscopy Hernia repair Oversewing of perforated peptic ulcer Salpingectomy Other <0.17 Overall <0.001 Overall excluding other <0.001 Post-operative length of stay (days) Appendicectomy Cholecystectomy Cystoscopy Gastrointestinal endoscopy Hernia repair Oversewing of perforated peptic ulcer Salpingectomy Other Overall Overall excluding other Overall length of stay (days) Appendicectomy Cholecystectomy Cystoscopy Gastrointestinal endoscopy Hernia repair Oversewing of perforated peptic ulcer Salpingectomy Other Overall <0.001 Overall excluding other <0.001 Values are expressed as mean standard deviation. Mann Whitney U-test.

4 870 Limmer and Edye Table 3 Length of stay by sex and comorbidity status Outcome: mean SD Yes No Mean difference P value Post-operative length of stay (days) Female sex Diabetes <0.001 Hypertension <0.001 Ischaemic heart disease Overall length of stay (days) Female sex Diabetes <0.001 Hypertension <0.001 Ischaemic heart disease <0.001 Values are expressed as mean standard deviation. Mann Whitney U-test. of stay for these groups were 6.2 and 4.6 days, respectively (P < 0.001). Interhospital transfer was associated with increased mean post-operative and overall length of stay for all procedure types, to varying degrees of significance, except for salpingectomy and surgery for perforated peptic ulcer (Table 2). Mean postoperative and overall length of stay were also significantly longer for patients with diabetes mellitus, ischaemic heart disease and hypertension (Table 3). The increases in length of stay associated with interhospital transfer were observed independently of the difference in prevalence of diabetes mellitus and hypertension between transferred patients and direct admissions (Table 4). Discussion The principal findings of this study were that interhospital transfer for emergency surgery for abdominal pain was associated with a mean 14.2 h increase in time to surgical intervention, 1.1 day longer post-operative length of stay and 1.6 day longer overall length of stay. When examined separately, mean time to surgery was significantly longer for transferred patients who underwent appendicectomy, cholecystectomy and cystoscopy. Mean postoperative length of stay was also increased for these patient groups however did not reach statistical significance. Despite the small number of cases, the observation that surgery for perforated peptic ulcer was performed within a comparable timeframe in both transferred patients and direct admissions highlights the capacity for the system to respond appropriately and efficiently expedite surgery for acutely life-threatening cases. Few Australian studies have examined the delays associated with interhospital transfer. A 2015 study of general surgical patients at a rural Victorian hospital found that a median 9.3 h elapsed between decision and execution of interhospital transfer. 8 A primary reason for delay was lack of bed availability in the receiving hospital. 8 A 2007 Queensland study also found significant delays in decompressive craniotomy and increased mortality in patients with extradural haemorrhage who underwent interhospital transfer. 9 Similar results have been reported by several international studies. A 2013 Canadian retrospective review of surgical delays in patients with hip fractures reported median time to surgery of 93 h for transferred patients compared to 44 h for direct admissions (P < 0.001). 10 Median length of hospital stay was 20 and 13 days for transferred and direct admissions, respectively (P < 0.001). 10 There were no significant differences in patient characteristics including age, sex, fracture type and fixation method between groups. 8,10 A 2014 US study of paediatric burns patients also reported an association between interhospital transfer and increased length of stay independent of the percentage of total body surface area affected. 11 Although not directly examined in the present study, an important consideration is whether the surgical delay associated with interhospital transfer negatively impacts patient outcomes. The related increases in post-operative length of stay in this study suggest that patient morbidity or rate of post-operative complications Table 4 Length of stay by admission and comorbidity status Outcome: mean SD Transferred patients (n = 290) Direct admissions (n = 620) Mean difference P value Post-operative length of stay (days) Diabetes mellitus No diabetes mellitus Hypertension No hypertension Both diabetes mellitus and hypertension Neither diabetes mellitus nor hypertension Overall length of stay (days) Diabetes mellitus No diabetes mellitus <0.001 Hypertension No hypertension <0.001 Both diabetes mellitus and hypertension Neither diabetes mellitus nor hypertension <0.001 Values are expressed as mean standard deviation. Mann Whitney U-test.

5 Interhospital transfer and surgical delay 871 may have been greater, requiring additional or prolonged postoperative monitoring and care. Despite this, mean in-hospital delay to appendicectomy for both transferred patients and direct admissions was within 24 h (23.8 and 16.6 h, respectively), which metaanalyses show is not associated with increased rate of complex appendicitis. 12 Likewise, mean time to cholecystectomy for both groups (72.7 and 48.8 h, respectively) was also in accordance with the 72 h timeframe advocated for early cholecystectomy. 13,14 Further studies also suggest that delay in presentation to hospital has a greater impact on outcomes than in-hospital delay. 15 A 2014 US study of surgical quality metrics for academic medical centres found that interhospital transfer was associated with higher mortality rate, longer length of stay and higher consumption of resources including intensive care (53 versus 32% for direct admissions), respiratory therapy (53 versus 42%), imaging (92 versus 82%) and rehabilitation services (63 versus 49%). 16 Mean total costs were over 50% higher for transferred patients compared with direct admissions. Mean intensive care unit (ICU) stay was 2.5 days longer and mean ICU costs were $ and $11 887, respectively. 16 Conflicting results were reported by a 2014 US review of emergency general surgical cases derived from the NSQIP database. 5 Transferred patients, who comprised 7% of the cohort, were found to have a significantly higher rate of postoperative complications (56 versus 38% for direct admissions), higher mortality (10 versus 4%), 1.74 times longer post-operative length of stay and higher rate of readmission (13 versus 10%). 5 However, regression analysis revealed that the increased frequency of adverse outcomes amongst transferred patients occurred primarily due to confounding differences in patient characteristics and procedure type. 5 This confounding effect was excluded in the present study as procedure type was not significantly different between transferred patients and direct admissions. Furthermore, increased length of stay occurred independently of the differing prevalence of diabetes mellitus and hypertension between groups. The mean additional increases in post-operative and overall length of stay observed in this study (1.1 and 1.6 days, respectively) may be relatively insignificant for the individual patient but represent substantial added burden on the health system. Increased length of stay not only places greater burden on human and material resources and contributes to the bed block experienced in Australian hospitals, but also adds considerably to government health expenditure. Given that the average cost per bed day for acute patients in NSW hospitals in was $1400, a hospital with a similar volume of transferred surgical patients as occurred in this study could expect to generate over $ annually in additional healthcare costs. 17 Provision of 24 h emergency surgical services at peripheral hospitals would be costly and difficult in terms of staffing and resource requirements. Hence, system improvements are clearly essential to facilitate the process of arranging and conducting transfer and to expedite surgery on arrival at the receiving hospital. In rural Australia, supporting the training and practice of general practitioner surgeons may also reduce the need to transfer patients requiring simpler procedures such as appendicectomy or abscess drainage. 18 Another significant challenge is adherence to the 4 hour rule that was introduced by NSW Health in 2012 with the aim to reduce ED length of stay. 19 The mean 8.1 h length of stay in Hospital B ED for emergency surgical patients in this study fell considerably short of this standard. Potential strategies to address this shortcoming include development of transfer policies for the acute abdomen. Such policies have been shown to improve patient outcomes and reduce mortality for major trauma, severe burns injury and paediatric surgery. 20 A 2005 NSW study also suggested implementation of a direct door to operating theatre approach, by which anaesthetic consults and emergency theatre time are arranged prior to rather than after patient arrival at the receiving hospital. 4 Such considerations are particularly important in planning of new hospital infrastructure in order to avoid replication of the issues identified in this study. A limitation of this study is the use of post-operative length of stay as a surrogate for patient outcomes. Further study is required to determine whether rates of specific post-operative complications, readmission, morbidity or mortality are increased for transferred patients who undergo emergency surgery for abdominal pain. Investigation is also necessary to elucidate factors underlying the delay associated with interhospital transfer and identify any administrative or clinician barriers. Conclusion In conclusion, the significant delay in surgical intervention and increase in length of hospital stay associated with interhospital transfer demands attention. System modification is vital to enhance the efficiency of interhospital transfer and thus minimize the potential impact on patient outcomes and burden on the Australian healthcare system. Acknowledgements We extend our thanks to Dr Christina Abdel Shaheed, Associate Lecturer at Western Sydney University, for her assistance with statistical analysis. References 1. Emergency Demand Coordination Group. Hospital Admission Risk Program (HARP) Background Paper. Melbourne: Victorian Government Dept. of Human Services, Australian Institute of Health and Welfare. Australian Hospital Statistics Canberra: AIHW, Royal Australasian College of Surgeons. Position Statement on Emergency Surgery. Guidelines and Position Papers. FES_FES_2269_P, February Wong K, Levy R. Interhospital transfers of patients with surgical emergencies: areas for improvement. Aust. J. Rural Health 2005; 13: Lucas D, Ejaz A, Haut E, Spolverato G, Haider A, Pawlik T. Interhospital transfer and adverse outcomes after general surgery: implications for pay for performance. J. Am. Coll. Surg. 2014; 218: Papandria D, Goldstein S, Rhee D et al. Risk of perforation increases with delay in recognition and surgery for acute appendicitis. J. Surg. Res. 2013; 184:

6 872 Limmer and Edye 7. Leppäniemi A. What is acceptable delay in emergency abdominal surgery? Scand. J. Surg. 2013; 102: Dobson H, Ranasinghe W, Hong M et al. Waiting for definitive care: an analysis of elapsed time from decision to surgery or transfer in a rural centre. Aust. J. Rural Health 2015; 23: Deverill J, Aitken L. Treatment of extradural haemorrhage in Queensland: interhospital transfer, preoperative delay and clinical outcome. Emerg. Med. Australas. 2007; 19: Desai S, Patel J, Abdo H, Lawendy A, Sanders D. A comparison of surgical delays in directly admitted versus transferred patients with hip fractures: opportunities for improvement? Can. J. Surg. 2014; 57: Myers J, Smith M, Woods C, Espinosa C, Lehna C. The effect of transfers between health care facilities on costs and length of stay for pediatric burn patients. J. Burn Care Res. 2015; 36: United Kingdom National Surgical Research. Collaborative Safety of short, in-hospital delays before surgery for acute appendicitis: multicentre cohort study, systematic review, and meta-analysis. Ann. Surg. 2014; 259: Gurusamy K, Davidson C, Gluud C, Davidson B. Early versus delayed laparoscopic cholecystectomy for people with acute cholecystitis. Cochrane Database Syst. Rev. 2013; 6: CD Cao A, Eslick G, Cox M. Early cholecystectomy is superior to delayed cholecystectomy for acute cholecystitis: a meta-analysis. J. Gastrointest. Surg. 2015; 19: Gandy R, Truskett P, Wong S et al. Outcomes of appendicectomy in an acute care surgical model. Med. J. Aust. 2010; 193: Crippen C, Hughes S, Chen S, Behrns K. The impact of interhospital transfers on surgical quality metrics for academic medical centers. Am. Surg. 2014; 80: Hehir G, Auditor-General, NSW Health. New South Wales Auditor- General s Report Performance Audit. Sydney: Audit Office of New South Wales, Appendix 3: Cost per bed day for overnight admissions of ABF hospitals in for acute patients, excluding mental health, renal dialysis and hospital in the home episodes. 18. Abbott B, Laurence C, Elliot T. GP surgeons: what are they? An audit of GP surgeons in South Australia. Rural Remote Health 2014; 14: Emergency Care Institute. NEAT The Basics [Cited 26 Feb 2014.] Available from URL: NSW Department of Health. Emergency Surgery Guidelines. North Sydney, NSW; 23 June 2009.