Abstract

Background

Emergency laparotomy is a common intra-abdominal procedure. Outcomes are generally recognized to be poor, but there is a paucity of hard UK data, and reports have mainly been confined to single-centre studies.

Methods

Clinicians were invited to join an ‘Emergency Laparotomy Network’ and to collect prospective non-risk-adjusted outcome data from a large number of NHS Trusts providing emergency surgical care. Data concerning what were considered to be key aspects of perioperative care, including thirty-day mortality, were collected over a 3 month period.

Results

Data from 1853 patients were collected from 35 NHS hospitals. The unadjusted 30 day mortality was 14.9% for all patients and 24.4% in patients aged 80 or over. There was a wide variation between units in terms of the proportion of cases subject to key interventions that may affect outcomes. The presence of a consultant surgeon in theatre varied between 40.6% and 100% of cases, while a consultant anaesthetist was present in theatre for 25–100% of cases. Goal-directed fluid management was used in 0–63% of cases. Between 0% and 68.9% of the patients returned to the ward (level one) after surgery, and between 9.7% and 87.5% were admitted to intensive care (level three). Mortality rates varied from 3.6% to 41.7%.

Conclusions

This study confirms that emergency laparotomy in the UK carries a high mortality. The variation in clinical management and outcomes indicates the need for a national quality improvement programme.

Editor's key points

  • This is an important first report of the Emergency Laparotomy Network in the UK.

  • The report provides evidence for high mortality in these patients.

  • Also, crucially, the report points to the variability in care provided to this patient group.

  • Areas of concern, and where improvements are required, include pre-optimization, consultant presence, and postoperative high dependency care.

Despite being one of the most common urgent surgical procedures in the UK,1 there is a scarcity of outcome data concerning postoperative mortality rates after emergency laparotomy. A recent single-centre study demonstrated an overall mortality rate of ∼19%, with a considerably higher mortality rate of 42% in octogenarians.2 The finding of a high mortality rate is supported by outcome data from a larger analysis of patients undergoing surgery in 94 UK acute Trusts,3 which demonstrated that a relatively small proportion of the surgical population categorized as ‘high-risk’ accounted for over 80% of the postoperative deaths. Of the high-risk procedures, nearly 90% were emergencies, many of which are likely to have been an emergency laparotomy.

The term ‘emergency laparotomy’ describes an exploratory procedure for which the clinical presentation, underlying pathology, anatomical site of surgery, and perioperative management vary considerably. The total number of surgical procedures that can be coded within this emergency laparotomy population exceeds 400, reflecting the diverse nature of this surgical cohort.4 The variation in surgical pathology, coupled with the limited time period in which to optimize co-morbidities, is likely to contribute significantly to postoperative morbidity and mortality.

In 2007, the Association of Surgeons of Great Britain and Ireland acknowledged that standards of care for emergency surgical admissions were often unsatisfactory. There was a failure to prioritize patients, inadequate senior input, and unsatisfactory allocation of infrastructure and manpower;5 and they characterized emergency surgery as the ‘Cinderella Service’. Patients admitted for high-risk elective surgery can expect to be provided with an extensive package of perioperative care; yet many emergency surgery patients may not receive an equal measure of resources,1 despite being older and sicker.4

The Emergency Laparotomy Network5 was established in January 2010 within the framework of NHS Networks. This free resource is dedicated to the sharing of ideas and problems within the NHS, with the aim of improving the service and care provided to patients. The current ELN membership consists of over 200 clinicians from more than 40 NHS hospitals in the UK. The ELN's broad aims are to bring together clinicians from relevant specialities in order to improve the outcome of patients undergoing emergency laparotomy. In order to establish a baseline and to provide a foundation for the Network, the Steering Committee designed and conducted a multicentre audit to measure non-risk-adjusted outcome after emergency laparotomy. The analysis of the pooled data has been returned to contributors, allowing them to identify and reflect upon their own outcomes. This paper presents a preliminary analysis of the results.

Methods

Ethical guidance was sought from the local ethics committee of South Tees Hospitals NHS Foundation Trust, who confirmed that the project fulfilled the criteria of a clinical audit as defined in the National Research Ethics Service document ‘Defining Research’,6 negating the requirement for the ethics committee approval.

Members of the UK Emergency Laparotomy Network were invited to submit prospective data on consecutive patients undergoing emergency laparotomy. Patients who had an emergency laparotomy between September 2010 and April 2011 were eligible for inclusion. Clinicians were asked to collect prospective anonymized data for a continuous period of 3 months within this time frame. The data set included a description of patient characteristics, timing of surgery, grade of clinical personnel in theatre, anatomical site of surgery, operative procedure, postoperative destination, length of stay, and 30 day mortality. Inclusion and exclusion criteria are shown in Supplementary Appendix 1. The design of the data set was a pragmatic attempt to focus on key issues that are likely to influence outcomes, and was intended to have a minimal impact upon the demands of the clinicians in the operating theatre. Thus, extensive detail about the anaesthetic care, surgical management, and the physiological derangement and sickness-severity of the patient was not sought. Likewise, detailed information about the timing of perioperative interventions was deliberately excluded because it was considered difficult to interpret without a large amount of supporting data. Full details of the data set are shown in Supplementary Appendix 2 and are available from the NHS Networks website of the Emergency Laparotomy Network.5

Data collection was carried out using a Microsoft Excel 2007 spreadsheet and returned to one of the authors (D.M.), who ensured that data were anonymized. Each hospital was designated by a unique two-letter code between AA and BK. To facilitate accurate and consistent data entry, participants were asked to select answers from a series of dropdown menus for each data field. Any apparent errors or inconsistencies in the data received were clarified with the submitting unit, and if no additional clarification was forthcoming, that data point was deleted from analysis. Hospitals were excluded from the analysis if they were unable to provide outcome data for more than 90% of the patients undergoing emergency laparotomy.

To avoid bias, and for the purposes of this report, patient characteristics and mortality calculations pertaining to repeat laparotomies have been excluded: the results reported relate to the primary emergency laparotomy for each patient.

Simple statistical analysis was performed within the Microsoft Excel program; the Fisher's exact test was performed using GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA, USA); and the funnel plot was generated using methodology as described by Spiegelhalter,7 with the standardized mortality ratio (SMR) defined as the ratio of observed to expected numbers of deaths in each hospital after adjusting for age group and gender variation.

Results

Data were returned from 37 NHS hospitals. The results from two hospitals were excluded because outcome data were available for fewer than 50% of the patients. The remaining 35 hospitals reported outcome data for more than 90% of their patients and were included in the analysis. One thousand eight hundred and fifty-three patients underwent 1941 separate emergency laparotomies, with a median [inter-quartile range (IQR) (range)] of 46 [30.5–68.8 (8–184)] operations per hospital. Thirty-day mortality data were available for 1819 patients, and was 14.9%. Baseline and outcome characteristics of laparotomy patients are shown in Table 1.

Table 1

Baseline characteristics and mortality data. *Characteristic-specific denominator figures for mortality calculations not shown: proportions refers to the number of cases with both characteristic and mortality data available. Mortality for that particular characteristic. Recent surgery defined as within 30 days (elective or emergency laparotomy).

Characteristic n (%) 30 day mortality* (%) 
All patients 1845 (100) 271 (14.9) 
Age band 
 Total patients with data (n1845 1819 
 <20 25 (1.4) 0 (0) 
 20–29 101 (5.5) 0 (0) 
 30–39 95 (5.1) 4 (4.3) 
 40–49 160 (8.7) 14 (9.0) 
 50–59 256 (13.9) 24 (9.4) 
 60–69 400 (21.7) 56 (14.0) 
 70–79 466 (25.3) 92 (20.0) 
 80–89 305 (16.5) 70 (23.6) 
 90–99 37 (2.0) 11 (31.4) 
Gender 
 Total patients with data 1845 1819 
 Male 879 (47.6) 139 (16.0) 
 Female 966 (52.4) 132 (13.9) 
Admitting speciality 
 Total patients with data 1845 1819 
 General surgery 1541 (83.5) 206 (13.6) 
 Medicine 213 (11.5) 43 (20.4) 
 Gynaecology 28 (1.5) 2 (7.4) 
 Vascular 14 (0.8) 9 (64.3) 
 Orthopaedics 11 (0.6) 5 (45.5) 
 Urology 22 (1.2) 2 (9.1) 
 Other 16 (0.9) 4 (26.7) 
Urgency 
 Total patients with data 1807 1782 
 Expedited (days) 230 (12.7) 19 (8.3) 
 Urgent (h) 1352 (74.8) 184 (13.8) 
 Immediate (min) 225 (12.5) 63 (28.3) 
Laparotomy, a complication of recent surgery, patients with data=1528 238 (15.6) 34 (14.5) 
ASA physical status 
 Total patients with data 1705 1680 
 I 113 (6.6) 0 (0) 
 II 565 (33.1) 23 (4.1) 
 III 643 (37.7) 85 (13.4) 
 IV 332 (19.5) 111 (33.6) 
 V 52 (3.0) 36 (69.2) 
Consultant staff present in theatre 
 Surgeon, n=1840 1359 (73.9) 210 (15.6) 
 Anaesthetist, n=1835 1209 (65.9) 187 (15.6) 
Goal-directed fluid therapy 
 Total patients with data 1626 1599 
 Used 235 (14.5) 44 (18.9) 
 Not used 1391 (85.6) 199 (14.6) 
Postoperative placement, total n=1789 
 Total patients with data 1789 1765 
 Level 1 (ward) 700 (39.1) 46 (6.7) 
 Level 2 (HDU) 523 (29.2) 52 (10.1) 
 Level 3 (ICU) 527 (29.5) 160 (30.7) 
 Extended recovery (PACU) 39 (2.2) 5 (13.5) 
Characteristic n (%) 30 day mortality* (%) 
All patients 1845 (100) 271 (14.9) 
Age band 
 Total patients with data (n1845 1819 
 <20 25 (1.4) 0 (0) 
 20–29 101 (5.5) 0 (0) 
 30–39 95 (5.1) 4 (4.3) 
 40–49 160 (8.7) 14 (9.0) 
 50–59 256 (13.9) 24 (9.4) 
 60–69 400 (21.7) 56 (14.0) 
 70–79 466 (25.3) 92 (20.0) 
 80–89 305 (16.5) 70 (23.6) 
 90–99 37 (2.0) 11 (31.4) 
Gender 
 Total patients with data 1845 1819 
 Male 879 (47.6) 139 (16.0) 
 Female 966 (52.4) 132 (13.9) 
Admitting speciality 
 Total patients with data 1845 1819 
 General surgery 1541 (83.5) 206 (13.6) 
 Medicine 213 (11.5) 43 (20.4) 
 Gynaecology 28 (1.5) 2 (7.4) 
 Vascular 14 (0.8) 9 (64.3) 
 Orthopaedics 11 (0.6) 5 (45.5) 
 Urology 22 (1.2) 2 (9.1) 
 Other 16 (0.9) 4 (26.7) 
Urgency 
 Total patients with data 1807 1782 
 Expedited (days) 230 (12.7) 19 (8.3) 
 Urgent (h) 1352 (74.8) 184 (13.8) 
 Immediate (min) 225 (12.5) 63 (28.3) 
Laparotomy, a complication of recent surgery, patients with data=1528 238 (15.6) 34 (14.5) 
ASA physical status 
 Total patients with data 1705 1680 
 I 113 (6.6) 0 (0) 
 II 565 (33.1) 23 (4.1) 
 III 643 (37.7) 85 (13.4) 
 IV 332 (19.5) 111 (33.6) 
 V 52 (3.0) 36 (69.2) 
Consultant staff present in theatre 
 Surgeon, n=1840 1359 (73.9) 210 (15.6) 
 Anaesthetist, n=1835 1209 (65.9) 187 (15.6) 
Goal-directed fluid therapy 
 Total patients with data 1626 1599 
 Used 235 (14.5) 44 (18.9) 
 Not used 1391 (85.6) 199 (14.6) 
Postoperative placement, total n=1789 
 Total patients with data 1789 1765 
 Level 1 (ward) 700 (39.1) 46 (6.7) 
 Level 2 (HDU) 523 (29.2) 52 (10.1) 
 Level 3 (ICU) 527 (29.5) 160 (30.7) 
 Extended recovery (PACU) 39 (2.2) 5 (13.5) 

There appeared to be a direct relationship between increasing age of the patient and 30 day mortality; from a mortality of just under 10% for a patient in their 50s, mortality increased by ∼4% for each additional 10 yr of age. For patients aged 80 and over, the mortality was 24.4%.

Increased mortality was also associated with higher ASA physical status classification, with increased urgency of surgery, and with a need for a greater intensity of immediate postoperative care. More than 60% of the patients were aged 60 or over and had a mortality of >10%. Likewise, more than 60% of the patients were of ASA III or greater, and these patients also had a mortality in excess of 10%.

The majority of patients were admitted directly to general surgery teams; 11.5% of the patients were admitted to a medical team and had a significantly higher mortality (Fisher's exact test of medicine vs general surgery; two-tailed P=0.0115). Specialities other than general surgery or medicine represented only a small proportion of cases but had a high mortality. No adjustment for age, gender, surgery, or ASA was made in these mortality calculations.

Table 2 shows a breakdown of start times of cases and the proportion of cases where the senior anaesthetist and the surgeon present were at a consultant level, together with 30 day mortality. Senior clinicians do not necessarily attend together: in patients aged 60 or over, only 53.3% of the cases were undertaken by both a consultant surgeon and a consultant anaesthetist, with a corresponding figure for patients who were ASA III or greater of 56.0%.

Table 2

Time of day, seniority of medical staff, and 30 day mortality. *Time of anaesthetic induction

Time of day* n Consultant anaesthetist present (%) Consultant surgeon present (%) 30 day mortality (%) 
08:00–17:59 1044 75.2 80.8 14.2 
18:00–23:59 442 54.8 67.7 17.8 
00:00–07:59 152 40.8 61.8 20.3 
Time of day* n Consultant anaesthetist present (%) Consultant surgeon present (%) 30 day mortality (%) 
08:00–17:59 1044 75.2 80.8 14.2 
18:00–23:59 442 54.8 67.7 17.8 
00:00–07:59 152 40.8 61.8 20.3 

There was little evidence of a disparity between the availability of critical care resources after emergency laparotomy and the perceived clinical need. Of the patients who were felt to need intensive [intensive care unit (ICU)] care immediately after surgery, 99% were transferred to a level 3 bed. Similarly, 89% of those who were judged to require a high-dependency [high-dependency unit (HDU)] bed received this level of care, with a further 4% receiving level 2 care in an ICU bed. Mortality in patients returning to the ward (level 1) was 6.7%, HDU 10.1%, and ICU 30.7%. 2.2% of patients were cared for after operation in an extended recovery area [post-anaesthesia care unit (PACU)], and this group had a mortality of 13.5%. For the group of patients aged 60 or greater, and of ASA III or more (∼50% of all patients), 22% returned to the general ward after operation and had a mortality of 17.8%.

Postoperative length-of-stay data were available for 1736 patients. The median [IQR (range)] postoperative length of stay for all patients was 11 days [6–21 (0–216)], with a median stay of 12 days for patients returning to a general ward and 19 days for patients who initially received level 3 support.

With increasing urgency of the case, there was a decrease in elapsed times from admission to booking and from booking to the start of anaesthesia. The median [IQR (range)] admission to booking times in days for expedited, urgent, and immediate cases were 3 [1–8 (0–113)], 1 [0–4 (0–299)], and 1 [0–3 (0–137)], respectively; and the median [IQR (range)] booking to start times in hours for expedited, urgent, and immediate cases were 21.6 [15.0–30.6 (0–177.5)], 4.0 [2.0–11.2 (0–243.9)], and 1.3 [0.7–2.3 (0–30.1)], respectively.

Table 3 shows a variation between hospitals in the use of goal-directed fluid management; the location of immediate postoperative care; and the seniority of anaesthetic and surgical clinical staff in theatre at the time of surgery, together with non-adjusted 30 day mortality.

Table 3

Process and mortality variations between hospitals. GDFT, goal-directed fluid therapy; IQR, inter-quartile range. *No data supplied

Hospital Number (n=1853) Proportion consultant surg. (%) (n=1840) Proportion consultant anaes (%) (n=1835) GDFT (%) (n=1626) Postoperative placement (%) (n=1789)
 
30 day mortality (%) (n=1819) 
Level 1 Level 2 Level 3 PACU 
Median 46 73.9 64.1 9.1 43.9 22.6 30.3 0.0 14.9 
Range 8–184 40.6–100 25–100 0–63.0 0–68.9 0–74.5 9.7–87.5 0–35.0 3.6–41.7 
IQR 30.5–68.8 60–5–81.9 55.8–76.3 0.6–21.7 27.4–55.2 11.4–33.4 22.5–37.9 0–0 11.0–22.2 
AA 32 46.9 84.4 9.4 12.9 35.5 51.6 0.0 12.5 
AC 23 60.9 60.9 9.1 52.2 8.7 34.8 4.3 9.1 
AD 36 97.2 77.8 0.0 41.7 30.6 27.8 0.0 22.2 
AE 106 74.5 66.0 21.7 61.3 9.4 29.2 0.0 10.4 
AG 50 48.0 47.9 12.0 44.0 22.0 34.0 0.0 14.0 
AH 17 76.5 35.3 11.8 23.5 52.9 23.5 0.0 23.5 
AI 63 40.6 46.9 0.0 43.8 28.1 28.1 0.0 12.7 
AJ 60 60.0 28.3 41.7 25.4 42.4 32.2 0.0 21.7 
AK 94 88.3 62.8 7.4 19.1 41.5 38.3 1.1 19.0 
AL 35 54.3 55.9 0.0 28.6 40.0 31.4 0.0 8.6 
AM 45 80.0 84.4 2.3 60.0 22.2 15.6 2.2 6.7 
AN 60 78.3 50.0 3.6 11.7 10.0 43.3 35.0 23.3 
AO 66 69.7 75.8 28.8 54.5 4.5 28.8 12.1 12.3 
AP 24 70.8 25.0 0.0 66.7 16.7 16.7 0.0 16.7 
AQ 25 96.0 100.0 37.5 26.1 4.3 69.6 0.0 41.7 
AR 71 58.0 88.4 11.8 67.1 22.9 10.0 0.0 11.3 
AS 11 72.7 63.6 9.1 63.6 0.0 36.4 0.0 18.2 
AT 38 73.7 71.1 0.0 55.3 7.9 36.8 0.0 28.9 
AU 85 95.3 71.8 27.1 12.9 51.8 35.3 0.0 12.9 
AV 69 85.1 74.2 7.4 40.3 25.4 34.3 0.0 12.3 
AW 51 56.0 60.8 21.6 38.0 26.0 36.0 0.0 25.5 
AX 136 52.9 56.6 34.8 61.8 18.0 19.1 1.1 7.4 
AY 100.0 62.5 37.5 0.0 12.5 87.5 0.0 25.0 
AZ 27 44.4 55.6 63.0 29.6 44.4 25.9 0.0 22.2 
BA 38 73.7 65.8 0.0 44.7 13.2 42.1 0.0 22.9 
BB 45 68.9 93.3 4.5 68.9 20.0 11.1 0.0 4.4 
BC 37 78.4 86.5 0.0 40.5 2.7 56.8 0.0 27.8 
BD 28 67.9 67.9 0.0 53.6 25.0 21.4 0.0 3.6 
BE 52 94.2 98.1 9.6 28.8 32.7 38.5 0.0 15.7 
BF 184 97.8 82.1 24.5 7.6 74.5 17.9 0.0 7.6 
BG 26 80.8 73.9 60.0 28.0 12.0 0.0 23.1 
BH 77 74.0 33.8 0.0 57.1 16.9 26.0 0.0 27.3 
BI 45 91.1 57.8 11.1 35.6 26.7 37.8 0.0 13.3 
BJ 58 65.3 50.0 4.1 49.0 10.2 28.6 12.2 13.3 
BK 31 77.4 64.5 6.5 45.2 45.2 9.7 0.0 9.7 
Hospital Number (n=1853) Proportion consultant surg. (%) (n=1840) Proportion consultant anaes (%) (n=1835) GDFT (%) (n=1626) Postoperative placement (%) (n=1789)
 
30 day mortality (%) (n=1819) 
Level 1 Level 2 Level 3 PACU 
Median 46 73.9 64.1 9.1 43.9 22.6 30.3 0.0 14.9 
Range 8–184 40.6–100 25–100 0–63.0 0–68.9 0–74.5 9.7–87.5 0–35.0 3.6–41.7 
IQR 30.5–68.8 60–5–81.9 55.8–76.3 0.6–21.7 27.4–55.2 11.4–33.4 22.5–37.9 0–0 11.0–22.2 
AA 32 46.9 84.4 9.4 12.9 35.5 51.6 0.0 12.5 
AC 23 60.9 60.9 9.1 52.2 8.7 34.8 4.3 9.1 
AD 36 97.2 77.8 0.0 41.7 30.6 27.8 0.0 22.2 
AE 106 74.5 66.0 21.7 61.3 9.4 29.2 0.0 10.4 
AG 50 48.0 47.9 12.0 44.0 22.0 34.0 0.0 14.0 
AH 17 76.5 35.3 11.8 23.5 52.9 23.5 0.0 23.5 
AI 63 40.6 46.9 0.0 43.8 28.1 28.1 0.0 12.7 
AJ 60 60.0 28.3 41.7 25.4 42.4 32.2 0.0 21.7 
AK 94 88.3 62.8 7.4 19.1 41.5 38.3 1.1 19.0 
AL 35 54.3 55.9 0.0 28.6 40.0 31.4 0.0 8.6 
AM 45 80.0 84.4 2.3 60.0 22.2 15.6 2.2 6.7 
AN 60 78.3 50.0 3.6 11.7 10.0 43.3 35.0 23.3 
AO 66 69.7 75.8 28.8 54.5 4.5 28.8 12.1 12.3 
AP 24 70.8 25.0 0.0 66.7 16.7 16.7 0.0 16.7 
AQ 25 96.0 100.0 37.5 26.1 4.3 69.6 0.0 41.7 
AR 71 58.0 88.4 11.8 67.1 22.9 10.0 0.0 11.3 
AS 11 72.7 63.6 9.1 63.6 0.0 36.4 0.0 18.2 
AT 38 73.7 71.1 0.0 55.3 7.9 36.8 0.0 28.9 
AU 85 95.3 71.8 27.1 12.9 51.8 35.3 0.0 12.9 
AV 69 85.1 74.2 7.4 40.3 25.4 34.3 0.0 12.3 
AW 51 56.0 60.8 21.6 38.0 26.0 36.0 0.0 25.5 
AX 136 52.9 56.6 34.8 61.8 18.0 19.1 1.1 7.4 
AY 100.0 62.5 37.5 0.0 12.5 87.5 0.0 25.0 
AZ 27 44.4 55.6 63.0 29.6 44.4 25.9 0.0 22.2 
BA 38 73.7 65.8 0.0 44.7 13.2 42.1 0.0 22.9 
BB 45 68.9 93.3 4.5 68.9 20.0 11.1 0.0 4.4 
BC 37 78.4 86.5 0.0 40.5 2.7 56.8 0.0 27.8 
BD 28 67.9 67.9 0.0 53.6 25.0 21.4 0.0 3.6 
BE 52 94.2 98.1 9.6 28.8 32.7 38.5 0.0 15.7 
BF 184 97.8 82.1 24.5 7.6 74.5 17.9 0.0 7.6 
BG 26 80.8 73.9 60.0 28.0 12.0 0.0 23.1 
BH 77 74.0 33.8 0.0 57.1 16.9 26.0 0.0 27.3 
BI 45 91.1 57.8 11.1 35.6 26.7 37.8 0.0 13.3 
BJ 58 65.3 50.0 4.1 49.0 10.2 28.6 12.2 13.3 
BK 31 77.4 64.5 6.5 45.2 45.2 9.7 0.0 9.7 

The funnel plot (Fig. 1) shows an age-adjusted mortality ratio for each unit plotted against the number of cases, together with ±2.0 and 3.0 standard deviations (sd).

Fig 1

SMR for emergency laparotomy.

Fig 1

SMR for emergency laparotomy.

Discussion

This is the first report of the UK National Emergency Laparotomy Network and includes data from 35 NHS hospitals representing ∼25% of the UK hospitals that take unselected general surgical admissions.8

There are more than 400 OPCS codes describing surgery that could come under the umbrella term of ‘emergency laparotomy’, making problematic the use of hospital coding data to study this group of patients.4 We grouped all cases that met a pragmatic and clinical definition of emergency laparotomy, enabling us to compare the variation in key aspects of care and outcomes between different NHS hospitals. Previous publications have confirmed the high mortality rates within this group of patients, but extrapolation to a wider NHS population is potentially flawed: either because the published data are from a single centre,2 from overseas,9 or retrospective in nature.3 This prospective study sought to highlight variations in outcomes within the UK.

Our data confirm a high mortality from emergency laparotomy in the UK. The overall figure of 14.9% represents a heterogeneous group of patients and includes cases of varying complexity and sickness-severity. Higher mortalities are evident in subgroups, including those of the high ASA status and the elderly. Nonetheless, this is a mortality figure that would be unthinkable for common major elective general surgery [colorectal resection (2.7%),10 oesophagectomy (3.1%),11 gastrectomy (4.2%),11 and liver metastasis resection (1%)].12

Our mortality findings are in line with the results from the NCEPOD report from 2011.13 Mortality among patients classified as ‘high-risk’ by anaesthetists undergoing non-elective intra-abdominal surgery was 13.2%, but patients undergoing colorectal resection without primary anastomosis had the highest mortality (20.4%) of all patient groups in the study.

The association between increasing mortality and rising age is, while perhaps intuitive, striking in its clarity. Almost one-third of those aged 90 or above died by day 30 after operation, which may be lower than many would predict. In contrast, patients in their 30s had a mortality of 4.2%, although this still represents a significant risk. The decision to operate on an elderly patient with significant co-morbidities is always difficult; in the 2010 report of the Scottish Audit of Surgical Mortality, the most common cause of an assessor's concern was that in retrospect, the operation should not have been performed.14

In collaboration with other professional groups and the Department of Health (DoH), The Royal College of Surgeons (RCS) have recently published recommendations on standards of care for the emergency patient, and these highlight the need to identify ‘high risk’ within clinical pathways.15,16 The recommendations equate high risk with a predicted mortality of ≥10%, and our study shows that this group includes patients aged over 60, or those assessed as ASA Class III or higher. Seventy-eight per cent of the patients with data for 30 day mortality met either one or both of these criteria.

Length of stay varied greatly and increased with postoperative dependency in this study. The detail about the eventual discharge destination of the survivors of emergency laparotomy was beyond the scope of our audit, but it is likely that a significant proportion of patients required enhanced levels of support in the community. Prolonged lengths of stay are likely to reflect significant morbidity and a slow recovery after surgery. The cost of emergency laparotomies is high, both financially for the hospital and physically and emotionally for the patient and their family.

Unlike elective surgery, many risk factors are not amenable to modification at the time a patient requires emergency laparotomy. Age, co-morbidity, and the underlying pathology cannot be altered, but the way the process of care is provided may be varied according to the needs of the patient. Ensuring prompt assessment, early resuscitation, and timely access to theatre, with senior staff involvement and appropriate levels of postoperative care are all potentially modifiable factors. The 2011 NCEPOD report found that 20% of non-elective cases incurred delays in getting to the operating theatre.13 Detailed analysis of contributing factors leading to delay was beyond the scope of the audit, but we were able to demonstrate a wide range of time periods between booking for theatre and induction of anaesthesia. The RCS recommendations emphasize that definitive surgery is undertaken with an urgency that is escalated according to the degree of illness, particularly the severity of sepsis.15 Surgery should be commenced as soon as possible after adequate preparation in order to reduce suffering, morbidity, mortality, and cost.

The NCEPOD reports from 1991 to 199217 and 1995 to 199618 made specific recommendations to increase the involvement of consultants during urgent or emergency colorectal resection. In the most recent NCEPOD report,13 the assessors judged that in only 1% of all (elective and non-elective) cases was the level of the surgeon inappropriate, but that this rose to 4.7% for anaesthetists. In our study, the majority of patients underwent surgery with the direct supervision of a consultant surgeon (73.8% of cases), an anaesthetist (65.8% of cases), but in only 50% of all cases did patients undergo surgery in the presence of both a consultant surgeon and a consultant anaesthetist. Furthermore, the proportion of cases receiving consultant-delivered care decreased for those operations beginning after 6 p.m., with a further reduction for those cases starting after midnight, despite the fact that it seems reasonable to assume that many patients in the out-of-hours group were sicker than those for whom surgery was deferred. Best practice is that a consultant surgeon and a consultant anaesthetist are present for all cases with a predicted mortality of more than 10%,15,16 and the time of day should not affect the standard of care. Hospitals need to adjust their processes to ensure that delays in assessment, imaging, and preparedness do not increase the likelihood of high-risk procedures being undertaken out-of-hours. A comparison has been made with outcomes in the USA, which may be better than the UK for similar major surgical procedures.19 There are many differences between the two health-care systems making direct comparisons difficult, but recent payment changes in the USA ensure that the named qualified surgeon must be present at operation to receive payment for the case.

It is reassuring that a critical care bed was forthcoming in almost all occasions when it was requested for postoperative care. The variation between hospitals in the rate of admission to HDU or ICU needs to be better understood, but by and large the apparent demand for postoperative care in HDU or ICU was met. As expected, the mortality was highest among those patients who required ICU support after operation, and considerably lower among those sent to HDU. In the absence of an agreed clinical standard for admission to level 2 and 3 care after emergency laparotomy, it is difficult to interpret the variation between hospitals, but the rate of referral to critical care in this study appears inappropriately low when compared with the observed mortality of 6.7% for those discharged to a general ward. Surgical mortality is influenced by the ability of a hospital to recognize and ‘rescue’ surgical patients who have developed a complication,20 and early recognition and management of complications is likely to be better in a critical care unit than on a normal ward.21 The significant mortality rate in the group of patients sent to a general ward after operation does raise the possibility that the team caring for the patient at the time of surgery did not appreciate the risks of serious complications arising after discharge from recovery.

Goal-directed fluid therapy was used in fewer than 15% of emergency laparotomies. There is now substantial evidence to support its use in elective colorectal surgery,22–25 and it is recommended that patients undergoing non-elective major abdominal surgery should receive fluid to achieve an optimal stroke volume during and for the first 8 h after surgery.22 An elevated lactate is known to predict a poor outcome in high-risk patients,26 and the use of flow-based technologies to optimize tissue perfusion in the perioperative period may be indicated in the majority of the patients undergoing emergency laparotomy.

Our results demonstrate a wide variation in both the number of cases undertaken and the non-adjusted mortality rates between the 35 hospitals reporting data. It is unreasonable to compare different units simply in terms of mortality, as a significant difficulty with our study design was the limited ability to stratify risk in any detail according to sickness-severity. In the 3 month period during which clinicians were asked to collect data, it is possible that some units undertook an unusual proportion of cases with greater complexity or co-morbidity than normal, introducing unfavourable bias into their outcome data. Furthermore, some hospitals may have had difficulties accurately identifying which patients died within 30 days: indeed, the data from two hospitals were excluded from the mortality comparisons as the proportion of cases with outcomes recorded was too small.

Age is a well-recognized risk factor for postoperative mortality. After age and gender adjustment, all units have an SMR that is within 2 sds from the mean; while two units appear to have exceptionally good outcomes, with SMRs of almost 3 sds from the mean. There is an impression that units undertaking a greater number of cases have better outcomes, and there may be a benefit from examining the processes of care provided by these hospitals in order to identify why their outcomes are particularly good.

In conclusion, we have presented the first report of the Emergency Laparotomy Network and shown in a national, multicentre, and prospective audit that non-risk-adjusted mortality from emergency laparotomy is 14.9%, rising to 24.4% in those aged 80 or over. We have demonstrated a wide variation between units in the seniority of surgical and anaesthetic staff present during surgery, the use of critical care immediately after operation, and raw mortality. The weaknesses of our data collection include the limited ability to stratify risk for individual patients, and not attempting to independently verify that participating clinicians identified all eligible cases of emergency laparotomy during the 3 month data collection.

Standards of care for unscheduled surgical patients have been published by the RCS in collaboration with the DoH and other professional groups, and these should be used to drive service improvement for those patients undergoing emergency laparotomy. Our findings warrant further effort to collect risk-stratified data from a wider proportion of NHS hospitals undertaking emergency abdominal surgery, and quality improvement programmes aimed at reducing postoperative morbidity are likely to be cost effective.

Declaration of interest

All five authors are members of the steering group of the Emergency Laparotomy Network.

Funding

No financial funding was provided from any source for this work.

Acknowledgements

None of the data presented here would be available without the hard work of Emergency Laparotomy Network members, and we would like to say a special thank you to those who took a lead role in each hospital collecting, verifying, and submitting data. We would also like to thank Mr Tony Roberts and Mr Simon McInerney from the North East Quality Observatory for their statistical advice and assistance with the funnel plot.

References

1
Emergency General Surgery: The Future, A Consensus Statement
 
2
Clarke
A
Murdoch
H
Thomas
MJ
Cook
TM
Peden
CJ
Mortality and postoperative care after emergency laparotomy
Eur J Anaesthesiol
 , 
2011
, vol. 
28
 (pg. 
16
-
9
)
3
Pearse
RM
Harrison
DA
James
P
, et al.  . 
Identification and characterisation of the high-risk surgical population in the United Kingdom
Crit Care
 , 
2006
, vol. 
10
 pg. 
R81
 
4
Peden
CJ
Emergency surgery in the elderly patient: a quality improvement approach
Anaesthesia
 , 
2011
, vol. 
66
 (pg. 
440
-
5
)
5
Emergency Laparotomy Network
 
6
NPSA
National Research Ethics Service—facilitating ethical research
 
. Available from http://www.nres.nhs.uk/EasySiteWeb/GatewayLink.aspx?alId=355 (accessed 15th June 2012)
7
Spiegelhalter
DJ
Funnel plots for comparing institutional performance
Stat Med
 , 
2005
, vol. 
24
 (pg. 
1185
-
202
)
8
Faiz
O
Warusavitarne
J
Bottle
A
, et al.  . 
Nonelective excisional colorectal surgery in English National Health Service Trusts: a study of outcomes from Hospital Episode Statistics Data between 1996 and 2007
J Am Coll Surg
 , 
2010
, vol. 
210
 (pg. 
390
-
401
)
9
Ingraham
AM
Cohen
ME
Bilimoria
KY
, et al.  . 
Comparison of 30-day outcomes after emergency general surgery procedures: potential for targeted improvement
Surgery
 , 
2010
, vol. 
148
 (pg. 
217
-
38
)
11
2009_NOGCA_report.pdf
 
. Available from http://augis.org/clinical_audits/clinical_audits_og_cancer.htm (accessed 15th June 2012)
12
Lochan
R
White
SA
Manas
DM
Liver resection for colorectal liver metastasis
Surg Oncol
 , 
2007
, vol. 
16
 (pg. 
33
-
45
)
13
NCEPOD—POC: Knowing the Risk Report (2011)
 
. Available from http://www.ncepod.org.uk/2011poc.htm (accessed 15th June 2012)
14
The Scottish Audit of Surgical Mortality Annual Report 2010 (reporting on 2009)
 
. Available from http://www.sasm.org.uk/Publications/SASM_Annual_Report_2010.pdf (accessed 15th June 2012)
15
The Royal College of Surgeons of England
Emergency Surgery: standards for unscheduled care
 
16
The Royal College of Surgeons of England/DoH
The Higher Risk General Surgical Patient: towards improved care for a forgotten group
 
17
NCEPOD
Perioperative Deaths—The 1991/2 Report of NCEPOD
 
. 1993. Available from http://www.ncepod.org.uk/1991_2.htm (accessed 15th June 2012)
18
NCEPOD
Perioperative Deaths—The 1995/6 Report of NCEPOD
 
. 1997. Available from http://www.ncepod.org.uk/1995_6.htm (accessed 15th June 2012)
19
Bennett-Guerrero
E
Hyam
JA
Shaefi
S
, et al.  . 
Comparison of P-POSSUM risk-adjusted mortality rates after surgery between patients in the USA and the UK
Br J Surg
 , 
2003
, vol. 
90
 (pg. 
1593
-
8
)
20
Ghaferi
AA
Birkmeyer
JD
Dimick
JB
Variation in hospital mortality associated with inpatient surgery
N Engl J Med
 , 
2009
, vol. 
361
 (pg. 
1368
-
75
)
21
Goldhill
DR
Down
JF
Are we operating as well as we can? Critical care to minimise postoperative mortality and morbidity
Anaesthesia
 , 
2008
, vol. 
63
 (pg. 
689
-
92
)
22
Powell-Tuck
J
Gosling
P
Lobo
DN
British Consensus Guidelines on Intravenous Fluid Therapy for Adult Surgical Patients GIFTASUP.
 
7 March 2011. Available from www.bapen.org.uk/pdfs/bapen_pubs/giftasup.pdf (accessed 15th June 2012)
23
NICE
CardioQ-ODM (oesophageal Doppler monitor) to guide intravenous fluid management in patients undergoing surgery, or in critical care: equality impact assessment
 
24
Department of Health
Innovation, health and wealth
 
25
Giglio
MT
Marucci
M
Testini
M
Brienza
N
Goal-directed haemodynamic therapy and gastrointestinal complications in major surgery: a meta-analysis of randomized controlled trials
Br J Anaesth
 , 
2009
, vol. 
103
 (pg. 
637
-
46
)
26
Lamme
B
Mahler
CW
van Ruler
O
Gouma
DJ
Reitsma
JB
Boermeester
MA
Clinical predictors of ongoing infection in secondary peritonitis: systematic review
World J Surg
 , 
2006
, vol. 
30
 (pg. 
2170
-
81
)

Supplementary data

Comments

5 Comments
Challenges with improving outcomes after emergency surgery
3 September 2012
Alexander Davey

Editor - confirmation of significant mortality following emergency laparotomy in the first report from the UK Emergency Laparotomy Network [1] highlights some important issues with research into, or clinical audit of, the care of emergency surgical patients and our efforts to improve perioperative outcomes in this population. Observational studies can be used to identify patients at high risk of adverse outcomes and characterise them by demographics, co-morbidity and underlying pathology. Knowledge gained from such studies is fundamental to quality improvement. Clinical audit which does not collect new data or test a hypothesis but requires a comparison of outcomes against a predetermined standard is also an important and possibly more efficient means of developing care quality. Although this work shares a number of characteristics with the process of clinical audit there does not seem to be an explicit standard of care for comparison. Grey areas such as this have undoubtedly hampered many an enthusiastic trainee interested in undertaking work to assess and improve the quality of perioperative care.

Furthermore, the authors highlighted the potential for suboptimal care of high risk patients on general wards and recommended quality improvement programmes based on further efforts to collect risk stratified data from a wider population. In this report, authors did not attempt analysis of care quality in the pre, peri or post operative periods because of the cumbersome and complex nature of handling the data required. However, if outcomes such as 30 day mortality and postoperative morbidity are to be improved it is vital qualitative work of this nature is carried out to assess standard of care and identify areas for intervention. Confidential enquiries use expert case note review and multidisciplinary group consensus [2] to make subjective decisions about care quality but application of such methods at a local level is not practical. Further work should therefore focus on assessing care quality or developing a reliable generic definition of satisfactory quality immediate care that can be applied to the case note review process to facilitate more detailed audit of emergency surgical care quality at both a local and national level. Finally, it is exciting as a junior trainee to see research on the quality of emergency surgical care being developed. Ultimately, it is largely by focusing more attention on optimising the actual quality of immediate care that organisations will contribute to further improvements in outcomes for acutely or critically ill surgical patients.

References

1. Saunders DI, Murray D, Pichel AC, Varley, Peden CJ. Variations in mortality after emergency laparotomy: the first report of the UK Emergency Laparotomy Network. Br J Anaesth 2012; 109: 368 - 375

2. NCEPOD - POC: Knowing the risk report (2011). Available from http://www.ncepod.org.uk/2011report2/downloads/POC_fullreport.pdf (accessed 28th August 2012)

Conflict of Interest:

None declared

Submitted on 03/09/2012 8:00 PM GMT
Emergency laparotomy: time to assess risk, but not according to time.
1 October 2012
Andrew J. Beamish (with David S. Y. Chan)

Dear Editor,

Saunders et al. are to be commended for their important national multicentre prospective study on emergency laparotomy (1).

As clinicians we have long recognised the patent value of physiological parameters in predicting mortality in acute illness; a greater number and extent of physiological abnormalities conferring greater mortality (2). However, of particular interest in this study is the apparent relationship of such simple variables as age and ASA grade with 30-day mortality following emergency laparotomy, with a study design permitting transferability to the UK population. In an era of increasing engagement with and reliance upon risk stratification, many of the devices used to quantify operative risk are often infeasible or redundant in emergency major surgery. Cardiopulmonary exercise testing, echocardiography and POSSUM scores (3) are examples.

We are far from advocating an ageist or prejudiced approach and neither the over 60 year old nor the ASA grade >II patient should be denied a potentially beneficial laparotomy based on the high-risk (mortality >10%) status of their respective age or comorbid groups. Rather, the same decision making process using principles of individualised risk assessment and process adjustment in recognition and accommodation of risk status, already being championed in many areas of elective surgery, should be allowed to permeate the relatively neglected arena of emergency surgery.

It is also important to consider the heterogeneity between subgroups when interpreting the results relating to the time of day (Table 2) that laparotomy was performed. The work of NCEPOD has rightly led to firm guidance toward solely life, limb and organ-saving surgery being performed during the night (4,5). This is mentioned only in passing and with relation to consultant presence rather than case mix in this paper. We can assume, as the authors have, that these time of day groups are heterogenous, with a greater proportion of unwell patients among the cohort of 152 patients operated overnight than in the cohort of 1044 operated during the day. Therefore, we cannot effectively interpret the results in Table 2. With this important confounding variable in mind, it would be interesting to see the breakdown of the aforementioned simple measures of age and ASA grade within these time subgroups.

References:

1. Saunders DI, Murray D, Pichel AC, Varley S, Peden CJ, on behalf of the members of the UK Emergency Laparotomy Network. Variations in mortality after emergency laparotomy: the first report of the UK Emergency Laparotomy Network. British Journal of Anaesthesia 2012; 109(3): 368-75.

2. Goldhill DR, McNarry AF. Physiological abnormalities in early warning scores are related to mortality in adult inpatients. British Journal of Anaesthesia. 2004; 92: 882-4.

3. Copeland GP, Jones D, Walters M. POSSUM: a scoring system for surgical audit. British Journal of Surgery. 1991; 78(3): 355-60.

4. National Confidential Enquiry into Patient Outcome And Death. The NCEPOD Classification of Intervention. 2004. Available from http://www.ncepod.org.uk/pdf/NCEPODClassification.pdf (accessed 18th September 2012)

5. The Royal College of Surgeons of England. Emergency Surgery: standards for unscheduled care. 2011. Available from http://www.rcseng.ac.uk/publications/docs/emergency-surgery-standardsfor- unscheduled-care (accessed 18th September 2012)

Conflict of Interest:

None declared

Submitted on 01/10/2012 8:00 PM GMT
Response to eLetters regarding our paper
18 October 2012
David I Saunders (with Dave Murray, Carol Peden, Adam Pichel, and Simon Varley)

Dear Sir

We would like to acknowledge the interest in our paper and appreciate the opportunity to respond.

We agree with Beamish and Chan that risk stratification is an essential component of peri-operative management in patients undergoing an emergency laparotomy. Unlike the elective scenario, time pressures limit the degree of optimisation achievable, but the clinical management pathways we offer patients can be amended to take account of the likelihood of serious peri- and post-operative complications occurring.

The key to pragmatic risk stratification is utility. ASA, age and urgency of surgery are simple and repeatable variables, and in relation to our findings, offer a good starting point for helping to appreciate the level of risk in groups of patients. Our paper demonstrated increased mortality amongst those who underwent surgery overnight, and as might be predicted, further analysis of our data confirms the increased degree of urgency and higher ASA grades for cases undertaken after midnight. The proportion of cases of ASA 3,4 or 5 who had induction of anaesthesia between 0800 and 1759, 1800 and 2359, and 0000 and 0759 were 59.4%, 63.5% and 72.3% respectively. The proportion of cases deemed urgent or requiring immediate surgery increased across the same time periods as follows: 81.3%, 94.1% and 97.5%.

Some patients who undergo "emergency laparotomy" have no signs of sepsis, and do not meet criteria which would warrant anything other than day time operating, so an increased proportion of daytime patients of ASA 2 is expected, but the confirmation that overnight operating often takes place in sicker patients confirms our suggestion that there is room to improve our pathways of care for these high risk patients.

We disagree with Beamish and Chan that P-POSSUM is not practicable in the emergency patient. Most data for required fields is readily available from simple laboratory and bedside testing, and a reasonable pre-operative estimation made of likely blood loss, magnitude of surgery and peritoneal contamination. The calculation of predicted morbidity and mortality is easily performed using online software or using specific smart phone apps (1)(2). Repeating the calculation at the end of surgery when more definitive information is available (perhaps as part of an end of surgery "bundle") acts as a crosscheck of appropriateness of the planned care pathway and postoperative location(3).

Davey states that there remains a paucity of standards of care for emergency laparotomy patients. The Royal College of Surgeons and the Department of Health have both published recent guidance for the standards of care for the patient undergoing emergency surgery(3)(4). Risk stratification, consultant involvement and post-operative critical care placement are all modifiable elements of the per-operative pathway. The guidance is unambiguous, and should form the basis of locally defined pathways of care for emergency laparotomy. Certainly these standards will form the basis of standards against which the National Emergency Laparotomy Audit (NELA) will compare data, and from 2013, NELA will become one of the mandatory audits to be included in Trust Quality Accounts(5).

References

1. Surgical Risk for iPhone 3GS, iPhone 4, iPhone 4S, iPhone 5, iPod touch (3rd generation), iPod touch (4th generation), iPod touch (5th generation) and iPad on the iTunes App Store [cited 2012 Oct 9] Available from: http://itunes.apple.com/gb/app/surgical-risk/id502404932?mt=8

2. P-POSSUM Scoring - www.riskprediction.org.uk; [cited 2012 Oct 9] Available from: http://www.riskprediction.org.uk/pp-index.php

3. The Higher Risk General Surgical Patient: towards improved care for a forgotten group; [cited 2012 Oct 11] Available from: http://www.rcseng.ac.uk/publications/docs/higher-risk-surgical-patient/

4. Emergency Surgery: Standards for unscheduled care; [cited 2012 Oct 11] Available from: http://www.rcseng.ac.uk/publications/docs/emergency-surgery-standards-for- unscheduled-care

5. National Clinical Audits for inclusion in Quality Accounts; [cited 2012 Oct 11] Available from: http://www.hqip.org.uk/national- clinical-audits-for-inclusion-in-quality-accounts/

Conflict of Interest:

The authors of this letter are members of the Steering Group of the Emergency Laparotomy Network. Dave Murray is the Clinical Lead for the National Emergency Laparotomy Audit (NELA).

Submitted on 18/10/2012 8:00 PM GMT
Methodology of Emergency Laparotomy Network audit
4 November 2012
Alexander Davey

Editor - it appears the authors have misinterpreted my comments on guidelines for perioperative care in high risk patients, notable in their response to my correspondence on the publication of the Emergency Laparotomy Networks first report1. My comment on a lack of guidelines relates to the classification of the project as a clinical audit despite the lack of specific guidelines and a standard for comparison in the methodology section of their report. Ethically, for this and other reasons outlined below, this project should have been treated as an observational study with appropriate research ethics committee approval. Firstly, the stated purpose of the project was to establish a baseline and a data set was defined "de novo" for prospective data collection which could essentially be considered generalisable new knowledge. In addition, ethical guidance was only sought from one hospital for a project that involved collection of sensitive and controversial data in 40 institutions. There is some evidence in their discussion of comparison with predetermined standards but no evidence from the methods section that this was planned in advance when designing the project. Therefore, this work should be regarded as project that involves generation of new and potentially sensitive data across multiple sites without explicit comparison to a predetermined standard and as such would probably warrant ethical review2. The remainder of my letter highlighted the complexity of identifying issues influencing key outcomes such as 30 day mortality and the importance of work investigating processes or quality of care rather than outcomes. If such key issues are to be identified, data collection should relate to quality of care in all three phases of perioperative care and not be excluded because of the cumbersome nature of supporting data required. Work of the Emergency Laparotomy Network is an exciting development but more consideration of the ethical implications and detail required to meaningfully identify key issues in perioperative care of high risk patients is required. References 1. Saunders DI, Murray D, Pichel AC, Varley, Peden CJ. Variations in mortality after emergency laparotomy: the first report of the UK Emergency Laparotomy Network. Br J Anaesth 2012; 109: 368 - 375 2. NPSA. National Research Ethics Service--facilitating ethical research. Available from http://www.nres.nhs.uk/EasySiteWeb/GatewayLink.aspx?alId=355 (last accessed 30th October 2012).

Conflict of Interest:

None declared

Submitted on 04/11/2012 7:00 PM GMT
Further response
26 November 2012
David I Saunders (with Adam Pichel, Dave Murray, Simon Varley)

Thank you for a further opportunity to respond. Our principle goals were to report crude mortality and some simple process and outcome measures such as seniority of clinician, utilisation of critical care, and length of stay. This data would form the basis for developing and monitoring future quality improvement within the nascent Emergency Laparotomy Network, and based on the level of participation achieved, we believe we struck the right balance between complexity and utility. We agree with Dr Davey that the ethical implications of publishing quality improvement work are not straightforward. Our study involved no interventions; no hypothesis was tested; and there were neither randomisation nor protocol allocation. Our work was a service evaluation(1) and it is an oversimplification to assume that creating new "generalisable knowledge" means that the work should be classified as "research." The issues surrounding ethical approval for work such as ours have been discussed elsewhere, both in respect of the editorial and publication policy of medical journals(2) and in respect of the potential ethical considerations of quality improvement work(3). It is for this reason that we sought advice from a local Research Ethics Committee. Finally, we are pleased that reports such as ours have added to the growing debate over the standards of care provided for patients undergoing emergency laparotomy. References: 1. NPSA. Is your project research? [accessed 19th November 2012] Available from: http://www.nres.nhs.uk/applications/is-your-project- research/ 2. Yentis SM. Research, audit and journal policies. Anaesthesia 2011 Mar;66(3):155-156. 3. Dixon N. Research, audit and journal policies. Anaesthesia 2011 Sep;66(9):847-847.

Conflict of Interest:

None declared

Submitted on 26/11/2012 7:00 PM GMT