Routine ultrasound and limited computed tomography for the diagnosis of acute appendicitis.

BACKGROUND: Acute appendicitis continues to be a challenging diagnosis. Preoperative radiological imaging using ultrasound (US) or computed tomography (CT) has gained popularity as it may offer a more accurate diagnosis than classic clinical evaluation. The optimal implementation of these diagnostic modalities has yet to be established. The aim of the present study was to investigate a diagnostic pathway that uses routine US, limited CT, and clinical re-evaluation for patients with acute appendicitis. METHODS: A prospective analysis was performed of all patients presenting with acute abdominal pain at the emergency department from June 2005 until July 2006 using a structured diagnosis and management flowchart. Daily practice was mimicked, while ensuring a valid assessment of clinical and radiological diagnostic accuracies and the effect they had on patient management. RESULTS: A total of 802 patients were included in this analysis. Additional radiological imaging was performed in 96.3% of patients with suspected appendicitis (n = 164). Use of CT was kept to a minimum (17.9%), with a US:CT ratio of approximately 6:1. Positive and negative predictive values for the clinical diagnosis of appendicitis were 63 and 98%, respectively; for US 94 and 97%, respectively; and for CT 100 and 100%, respectively. The negative appendicitis rate was 3.3%, the perforation rate was 23.5%, and the missed perforated appendicitis rate was 3.4%. No (diagnostic) laparoscopies were performed. CONCLUSIONS: A diagnostic pathway using routine US, limited CT, and clinical re-evaluation for patients with acute abdominal pain can provide excellent results for the diagnosis and treatment of appendicitis.

Patient and procedure-related risk factors for adverse events after infrainguinal bypass.

BACKGROUND: Current medical practice urges individual health care facilities and medical professionals to obtain and provide detailed insight in quality of care with the possibility of comparing data between institutions. Adverse event (AE) analysis serves as a mainstay in quality assessment in vascular surgery, but the comparison of AE data between facilities can be complex. The aim of the present study was to assess independent risk factors for AE occurrence: patient, disease and operation characteristics besides general differences between health care facilities. METHODS: All AEs after infrainguinal bypass graft procedures (BGPs) in three health care facilities in the Netherlands were evaluated. AEs were defined identically in the facilities. RESULTS: Of 601 BGPs performed, 662 AEs were registered. Independent predictors of AEs were female gender (odds ratio [OR], 2.13; 95% confidence interval [CI], 1.39-3.26; P < .01), age >or=60 years (OR, 0.57; 95% CI, 0.34-0.95; P = .03), American Society of Anesthesiologists classification 3-4 (OR, 1.79; 95% CI, 1.01-3.17; P = .05), comorbidities of pulmonary disease (OR, 2.99; 95% CI, 1.67-5.34; P < .01) and diabetes mellitus (OR, 2.49; 95% CI, 1.58-3.94; P < .01), distal anastomosis level at below knee femoropopliteal BGP (OR, 2.01; 95% CI, 1.26-3.22; P < .01), femorotibial BGP (OR, 2.40; 95% CI, 1.37-4.19; P < .01), and popliteopedal BGP (OR, 92.39; 95% CI, 11.13-766.98; P < .01). One health care facility had significantly fewer AEs than the other two (OR, 0.21; 95% CI, 0.13-0.35; P < .01). CONCLUSION: Age, gender, comorbidity, and type of surgery are all independent predictors of AE occurrence in vascular surgery. After adjustment for differences in these factors, one health care facility still had lower AE occurrence, which needs to be examined further.

Standard outpatient re-evaluation for patients not admitted to the hospital after emergency department evaluation for acute abdominal pain.

BACKGROUND: The aim of the present study was to investigate the efficacy and safety of standard outpatient re-evaluation for patients who are not admitted to the hospital after emergency department surgical consultation for acute abdominal pain. METHODS: All patients seen at the emergency department between June 2005 and July 2006 for acute abdominal pain were included in a prospective study using a structured diagnosis and management flowchart. Patients not admitted to the hospital were given appointments for re-evaluation at the outpatient clinic within 24 h. All clinical parameters, radiological results, diagnostic considerations, and management proposals were scored prospectively. RESULTS: Five-hundred patients were included in this analysis. For 148 patients (30%), the final diagnosis was different from the diagnosis after initial evaluation. Eighty-five patients (17%) had a change in management after re-evaluation, and 20 of them (4%) were admitted to the hospital for an operation. Only 6 patients (1.2%) had a delay in diagnosis and treatment, which did not cause extra morbidity. CONCLUSIONS: Standard outpatient re-evaluation is a safe and effective means of improving diagnostic accuracy and helps to adapt management for patients that are not admitted to the hospital after surgical consultation for acute abdominal pain at the emergency department.

Assessment of incidence, cause, and consequences of pressure ulcers to evaluate quality of provided care.

BACKGROUND: Pressure ulcers are one of the most frequently registered complications in general surgery. OBJECTIVE: To obtain insight into the incidence, cause, and consequences of pressure ulcers and to evaluate the value of pressure ulcer registration to assess quality of care. RESULTS: During the 9-year study period, 275 pressure ulcers were registered (5.8% of total registered complications). Age and female sex were independent risk factors for pressure ulcer development. Pressure ulcer classification was as follows: mild (53.3%), moderate (35.6%), severe (9.5%), and irreversible damage (1.5%). Patients undergoing hip surgery and major limb amputation were at risk for pressure ulcer development (10.4% and 8.8%, respectively). In most patients (89.5%), pressure ulcers had no consequences other than local wound therapy; in 12 patients (4.4%), pressure ulceration led to alteration in medication; in 15 patients (5.5%), length of hospital stay was prolonged; and four patients (0.4%) suffered from irreversible damage. CONCLUSION: The incidence of pressure ulcers is strongly correlated to sex, age, and indication of admittance. Most ulcers were classified as mild and had no consequences. The insight obtained into incidence, cause, and consequences of pressure ulcers can be used as an indicator of quality of provided care if adjusted for case mix and indication of operation.

Wound complications at the groin after peripheral arterial surgery sparing the lymphatic tissue: a double-blind randomized clinical trial.

BACKGROUND: The groin incision after arterial reconstructive surgery is most likely at risk for infectious or lymphatic wound complications. Theoretically; sparing lymphatic tissue by a lateral approach to the femoral artery should minimize these. The aim of this study was to assess the incidence of wound complications after the lateral versus the direct approach of the common femoral artery. METHODS: The study population included all patients who underwent an exploration of the common femoral artery between May 2002 and December 2005. RESULTS: After 6 weeks, no statistical differences in the occurrence of wound complications could be shown. A wound infection was present after 6 weeks in 6.1% in the direct group versus 6.0% in the lateral group. Lymphorrhea was persistent in 3.1% in the direct group versus 5.0% in the lateral group. CONCLUSIONS: Using a lateral vertical incision for the approach of the common femoral artery did not decrease the incidence of postoperative wound complications.

A prospective study of incidence of saphenous nerve injury after total great saphenous vein stripping.

BACKGROUND AND OBJECTIVES: Total stripping of the great saphenous vein (GSV) is a validated surgical strategy of treating patients with primary varicose veins (PVV). An often cited, but not well documented and studied, complication of total stripping is postoperative damage of the saphenous nerve (SN). OBJECTIVE: The objective was to evaluate the incidence of SN damage and to assess the therapeutic efficacy after total stripping of the GSV. MATERIALS AND METHODS: Patients undergoing total stripping of the GSV because of PVV in the entire lower limb were enrolled. Pre- and postoperative neurologic examination was performed to identify potential sensory neurologic deficits. RESULTS: Total stripping of the GSV in 69 limbs occurred because of pain (9%) or a tired feeling in the limbs (77%) or for cosmetic reasons (14%). The overall incidence of postoperative sensory neurologic deficits was 7 and 6%, respectively, after 6-week follow-up and both 3% after 3-month follow-up. In 99% of the patients, total stripping of the GSV resulted in reduction of the primary signs and symptoms. CONCLUSION: The incidence of SN damage after total stripping of the GSV is low. Thus, total stripping of the GSV resulted in improvement of the primary complaint in almost all patients. Total stripping of the GSV is an effective surgical strategy in treating PVV.

The effect of implementation of an optimized care protocol on the outcome of arteriovenous hemodialysis access surgery.

BACKGROUND: The long-term patency of arteriovenous fistulas (AVFs) and arteriovenous grafts (AVGs) suffers from a high incidence of primary failure due to early thrombosis, myointimal hyperplasia at the venous access site, or failure to mature. A multidisciplinary meeting in vascular access surgery was initiated to optimize the timing, indication, type of intervention, and the logistics of AVFs/AVGs during the preoperative and postoperative period. This study evaluated the influence of the new optimized care protocol on the incidence of revisions (surgical and endovascular) and patency rates. METHODS: This protocol for vascular access surgery of AVFs/AVGs for hemodialysis was introduced in January 2004. It was initiated with the presence of the vascular surgeons, nephrologists, interventional radiologists, dialysis nurses, and the ultrasound technicians. Every patient who needed an AVF/AVG because of long-term treatment of chronic renal failure or awaiting kidney transplantation, or who needed a revision of an AVF/AVG, was discussed. Two groups were compared. Group I patients were treated with an AVF/AVG before the introduction of the new protocol (2001 and 2002). Group II patients were treated with an AVF/AVG after the introduction of the new optimized care protocol (2004 and 2005). Both groups were followed up after 12 months. RESULTS: During the study period, 146 AVFs/AVGs were attempted, and 111 postoperative revisions were performed to restore primary and secondary patency: 63 in group I (60 surgical, 3 radiology) and 48 in group II (23 surgical, 25 radiology). Significantly more segmental access replacements (P < 0.027) occurred in group I than in group II. Significantly fewer surgical revisions (P < 0.019) and more endovascular balloon angioplasties (P < 0.001) occurred in group II. Significantly higher cumulative primary and secondary patency rates of all AVFs/AVGs (P < 0.001), radial-cephalic direct wrist AVFs (P < 0.001), and brachial-cephalic forearm looped transposition AVGs (P < 0.001) were achieved in group II after follow-up. CONCLUSION: The new protocol outlined in a bimonthly multidisciplinary meeting for vascular access surgery of AVFs/AVGs for hemodialysis resulted in more effective logistics according to preoperative diagnostics and operation. More importantly, a significant increase in endovascular balloon angioplasties and a significant decrease in surgical revisions was observed, resulting in less patient morbidity. Also, higher primary and secondary patency was achieved after the introduction of the new optimized care protocol.

The incidence of unplanned returns to the operating room after peripheral arterial bypass surgery and its value as indicator of quality of care.

BACKGROUND: In recent years, a growing need has arisen to define possible indicators of quality of care. METHODS: To examine whether unplanned return to the operating room within 30 days after the initial procedure could serve as an indicator to assess quality of care in peripheral arterial bypass surgery, all bypass procedures performed between January 1996 and January 2004 were evaluated. Data were obtained from a prospectively kept hospital registration system. RESULTS: A total of 607 consecutive procedures were performed in 468 patients. The overall unexpected return to the operating room rate was 11.2%. Patients requiring peripheral arterial bypass surgery for critical ischemia with gangrene were significantly more at risk for an unplanned reoperation (20.2%) than patients with disabling claudication (2.1%) (P < .0001). Patients requiring femorocrural bypass surgery (24.2%) were also more at risk than patients with a suprageniculate bypass procedure (5.2%) (P < .0001). CONCLUSIONS: Unplanned return to the operating room within 30 days after the initial operation can be a useful indicator of quality of care after peripheral arterial bypass surgery. However, a prospective, well-defined registration system to collect all data is essential. Furthermore, the severity of peripheral arterial disease and the type of procedure performed should be taken into account.

Can the incidence of unplanned reoperations be used as an indicator of quality of care in surgery?

The incidence of unplanned reoperations could potentially be used as an indicator of quality of care. This study provided insight into the incidence of unplanned reoperations in a surgical department and added to the discussion of the value of unplanned reoperations as an indicator of quality of care. Between January 1996 and December 2003, all unplanned reoperations were entered prospectively into a complication registration system. The number of unplanned reoperations was 447 (1.7%). Unplanned reoperations occurred frequently after vascular (6.5%) and colon surgery (5.7%) and were caused predominantly by errors in surgical technique (70%) and patients' comorbidities (21%). Mortality for patients requiring unplanned reoperations was significantly higher than for patients who did not require reoperations (10.3% versus 4.0%). Unplanned reoperation rates can be an indicator of quality of care. However, a prospective, well-defined registry is essential to ensure an accurate assessment of the quality of care provided.

Can pyloromyotomy for infantile hypertrophic pyloric stenosis be performed in any hospital? Results from two teaching hospitals.

In order to document the incidence of perioperative complications in patients with infantile hypertrophic pyloric stenosis, a descriptive cohort study was performed in two teaching hospitals in the Netherlands. One hospital specialized in pediatric surgery and the other was a general surgery teaching hospital. All consecutive infants who underwent pyloromyotomy for the diagnosis hypertrophic pyloric stenosis in both hospitals between 1998 and 2002 were included. The children were diagnosed and treated according to a standard protocol. From all charts, complications durante- and post-operationem were recorded. A total of 256 pyloromyotomies were performed. Registered perioperative complications were duodenal mucosal perforation (n=6; 2%). Perioperatively unrecognized duodenal mucosal perforation occurred four times (1%). One re-operation was performed for an incomplete pyloromyotomy (0.3%). Persistent vomiting after the operation occurred in 18 children (7%). A large majority of postoperative complications were wound infections (n=16; 6%), 12 after right upper quadrant incision and 4 after umbilical incision; most of them were treated with antibiotics and/or incision for drainage of an abscess. An incisional hernia occurred four times. Prolonged vomiting was the only postoperative complication that differed significantly between the two teaching hospitals. The overall percentages of complications were equal to complication rates in literature, and since there were no extensive differences in major complications between the two teaching hospitals in this study, we can conclude that pyloromyotomy can be performed safely in specialized centers and in general centers provided with a multidisciplinary team.

Computed tomography severity index is an early prognostic tool for acute pancreatitis.

BACKGROUND: Acute pancreatitis is a severe disease with unpredictable course and outcomes. It is especially hard to identify early those patients who will have a fulminant course. In a prospective observational study, we tested the hypothesis that the CT Severity Index (CTSI), established within 48hours after admission, is prognostic for morbidity and mortality and can predict the necessity for admission to an ICU. STUDY DESIGN: From January 1994 to October 2002, all patients with the diagnosis of first time acute pancreatitis underwent spiral CT with intravenous contrast within 48hours of admission. The extent of inflammation and necrosis was assessed to define the CTSI. Patients were initially managed in an ICU in a standardized fashion. Complications and mortality were registered in a systematic manner. RESULTS: Seventy-nine patients were admitted with acute pancreatitis. The overall complication rate was 57%; mortality was 9%. In patients with a CTSI of 0 to 3, these rates were 42% and 2%, respectively; in those with CTSI of 4 to 6, 81% and 19%, respectively; and in those with CTSI of 7 to 10, 100% and 33%, respectively. Outcomes of subsequent CT scans did not alter the initial prognosis. Early CTSI correlated well with the incidence of complications, sepsis, mortality, and necessity for ICU admission. CONCLUSIONS: Acute pancreatitis is associated with marked morbidity and mortality. Initial admission to an ICU and standardized conservative treatment are justified for all patients. Early establishment of the CTSI is an excellent prognostic tool for complications and mortality. Patients with a CTSI of 0 to 3 can safely be discharged from the ICU.

Short versus long-term postoperative drainage of the axilla after axillary lymph node dissection. A prospective randomized study.

BACKGROUND: Axillary lymph node dissection (ALND) is a standard procedure in the treatment of breast cancer. Current practice following ALND involves several days of drainage of the axilla to reduce the formation of seroma. The aim of this study is to investigate the feasibility of 24 h drainage. STUDY DESIGN: A prospective randomized trial was performed comparing 24 h drainage to long-term drainage. The primary outcome measure was duration of hospital stay. Formation of seroma and wound related complications were secondary outcome measures. RESULTS: Fifty patients were randomised to the 24 h drainage group and 50 patients to the long-term drainage group. 24 h drainage was associated with a shorter hospital stay (2.5 versus 4.6 days, p < 0.001). Seroma aspiration was required in 76% of the patients after 24 h drainage and in 64% after long-term drainage (p = 0.19). The number of wound related complications was higher after long-term drainage (13 versus 9, p = 0.33). Infectious complications were seen in 11 patients after long-term drainage versus 6 after 24 h drainage (p = 0.18). CONCLUSION: These results indicate that 24 h drainage following ALND is feasible and facilitates early hospital discharge. Furthermore, 24 h drainage is not associated with excess wound related complications compared to long-term drainage.

Quality control of colorectal surgery with an extensive complication registration system.

BACKGROUND/AIMS: Since 1996, a standard registration is used to obtain a clear understanding of the complications that occur after colorectal surgery. In our registration system the cause of a complication and the detrimental effect to the patient's health are coded. METHODS: The treatment of colorectal diseases was evaluated to analyze the quality of medical care. RESULTS: From 1996 to 2000, a total of 169 complications were documented in 108 of the 353 patients operated on. Leakage of the anastomosis occurred in 22 cases (6%) and wound infection occurred in 11 cases (3%). Most of the complications required no or little medical attention (n = 101, 59%). 51% of the complications could be attributed to the physical condition of the patient. Surgical complications were the cause in 31% of the cases and management problems in 16% of the cases. CONCLUSION: A registration system provides good insight into the frequency and severity of the complications after colorectal surgery. Extensive registration is mandatory to provide reliable information, comparing the results year by year. This provides the basis for continuous improvement of medical protocols on the surgical ward.

Mortality after peripheral bypass surgery: value of a mortality scoring system in evaluating the quality of care.

The objective of this study was to determine the role of a mortality registration in the quality control of patients who died after peripheral bypass surgery. We developed a mortality registration to classify causes of death, to evaluate shortcomings in treatment, and to determine the extent of agreement between pre- and postmortem findings. In a 10-year period, 28 of the 1,022 patients (2.7%) who underwent peripheral arterial reconstruction died. Fifty-three percent of the patients died owing to postoperative complications, most frequently a myocardial infarction. A shortcoming in the medical treatment was observed in only one patient. Forty-three percent of the relatives gave permission for an autopsy. In only two cases, the autopsy report revealed a myocardial infarction that had remained unnoticed during the clinical course. In this selected group of patients undergoing a peripheral bypass operation, the causes of death and the shortcomings in medical care could usually be identified without the help of autopsy data.

The need for a prophylactic gastrojejunostomy for unresectable periampullary cancer: a prospective randomized multicenter trial with special focus on assessment of quality of life.

OBJECTIVE: To evaluate the effect of a prophylactic gastrojejunostomy on the development of gastric outlet obstruction and quality of life in patients with unresectable periampullary cancer found during explorative laparotomy. SUMMARY BACKGROUND DATA: Several studies, including one randomized trial, propagate to perform a prophylactic gastrojejunostomy routinely in patients with periampullary cancer found to be unresectable during laparotomy. Others suggest an increase of postoperative complications. Controversy still exists in general surgical practice if a double bypass should be performed routinely in these patients. METHODS: Between December 1998 and March 2002, patients with a periampullary carcinoma who were found to be unresectable during exploration were randomized to receive a double bypass (hepaticojejunostomy and a retrocolic gastrojejunostomy) or a single bypass (hepaticojejunostomy). Randomization was stratified for center and presence of metastases. Patients with gastrointestinal obstruction and patients treated endoscopically for more than 3 months were excluded. Primary endpoints were development of clinical gastric outlet obstruction and surgical intervention for gastric outlet obstruction. Secondary endpoints were mortality, morbidity, hospital stay, survival, and quality of life, measured prospectively by the EORTC-C30 and Pan26 questionnaires. It was decided to perform an interim analysis after inclusion of 50% of the patients (n = 70). RESULTS: Five of the 70 patients randomized were lost to follow-up. From the remaining 65 patients, 36 patients underwent a double and 29 a single bypass. There were no differences in patient demographics, preoperative symptoms, and surgical findings between the groups. Clinical symptoms of gastric outlet obstruction were found in 2 of the 36 patients (5.5%) with a double bypass, and in 12 of the 29 patients (41.4%) with a single bypass (P = 0.001). In the double bypass group, one patient (2.8%) and in the single bypass group 6 patients (20.7%) required (re-)gastrojejunostomy during follow-up (P = 0.04). The absolute risk reduction for reoperation in the double bypass group was 18%, and the numbers needed to treat was 6. Postoperative morbidity rates, including delayed gastric emptying, were 31% in the double versus 28% in the single bypass group (P = 0.12). Median postoperative length of stay was 11 days (range 4-76 days) in the double versus 9 days (range 6-20 days) in the single bypass group (P = 0.06); median survival was 7.2 months in the double versus 8.4 months in the single bypass group (P = 0.15). No differences were found in the quality of life between both groups. After surgery most quality of life scores deteriorated temporarily and were restored to their baseline score (t = -1) within 4 months. CONCLUSIONS: Prophylactic gastrojejunostomy significantly decreases the incidence of gastric outlet obstruction without increasing complication rates. There were no differences in quality of life between the two groups. Together with the previous randomized trial from the Hopkins group, this study provides sufficient evidence to state that a double bypass consisting of a hepaticojejunostomy and a prophylactic gastrojejunostomy is preferable to a single bypass consisting of only a hepaticojejunostomy in patients undergoing surgical palliation for unresectable periampullary carcinoma. Therefore, the trial was stopped earlier than planned.

Vein versus polytetrafluoroethylene in above-knee femoropopliteal bypass grafting: five-year results of a randomized controlled trial.

OBJECTIVE: Controversy still exists whether polytetrafluoroethylene is equivalent to vein as bypass graft material for the above-knee femoropopliteal bypass. Therefore, a prospective randomized trial was performed to compare vein with polytetrafluoroethylene for femoropopliteal bypasses with the distal anastomosis above the knee. METHODS: Between January 1993 and December 1996, 151 above-knee femoropopliteal bypasses were performed. The indications for operation were severe claudication in 120 cases, rest pain in 20 cases, and ulceration in 11 cases. After randomization, 75 reversed saphenous venous bypasses and 76 polytetrafluoroethylene bypasses were performed. RESULTS: No perioperative mortality was seen, and 5% of the patients had minor infections of the wound, not resulting in loss of the bypass, the limb, or life. After 5 years, 38% of the patients had died and 7% were lost to follow-up. Only once was the saphenous vein necessary for coronary artery bypass grafting. Primary patency rates after 5 years were 75.6% for venous bypass grafts and 51.9% for polytetrafluoroethylene grafts (P =.035). Secondary patency rates were 79.7% for vein and 57.2% for polytetrafluoroethylene bypasses (P =.036). In the venous group, 14 bypasses failed, leading to five new bypasses. In the polytetrafluoroethylene group, 29 bypasses failed, leading to 16 reinterventions. For these 16 new bypasses, in four cases, the ipsilateral preserved saphenous vein was used. In both groups, one above-knee amputation and one below-knee amputation had to be performed. CONCLUSION: We conclude after 5 years of follow-up of this randomized controlled trial that a bypass with saphenous vein has better patency rates at all intervals and needs fewer reoperations. Saphenous vein should be the graft material of choice for above-knee femoropopliteal bypasses and should not be preserved for reinterventions. Polytetrafluoroethylene is an acceptable alternative if the saphenous vein is not available.

Factors associated with perioperative complications and long-term results after pulmonary resection for primary carcinoma of the lung.

OBJECTIVE: The aim of this study was to describe perioperative morbidity and mortality of patients presenting with resectable lung cancer and to investigate the long-term survival. METHODS: We reviewed the records of 344 patients who underwent lung resection for bronchogenic carcinoma. Follow-up information was obtained from visits to the outpatient clinic. RESULTS: Between January 1991 and December 1995 there were 263 males and 81 females included with a mean age of 65.7 years. One hundred and eight (31%) patients underwent a pneumonectomy, 159 (46%) a lobectomy, 43 (13%) a bilobectomy, four (1%) a segmental resection and 30 (9%) an explorative thoracotomy. A total of 341 complications occurred. The 30 day mortality rate was 7.9% (27 patients). Patients with a low FEV1% and older patients have a higher risk of mortality within 30 days. Postoperative myocardial infarction and pneumonia were associated with an increase in 30 day mortality. The median survival was 3.6 years for stage I, 1.9 years for stage II, 1.0 years for stage IIIa, 0.9 years for stage IIIb and 0.9 years for stage IV. Prognostic factors for the long-term survival included stage, pneumonectomy, percentage FEV1 <70, and large cell carcinoma. CONCLUSIONS: Pulmonary resection can be performed at an acceptable risk. Critical reviewing of our results made it possible to make recommendations for improvements.

Value of keeping records of mortality.

OBJECTIVE: To evaluate treatment and complications which is essential for good medical practice. DESIGN: Prospective audit. SETTING: City hospital, The Netherlands. SUBJECTS: All the patients who died on the surgical ward between 1994 and 1998 and were classified according to four categories of mortality recording. INTERVENTIONS: The causes of death, inaccuracies in treatment and the extent of agreement between premortem and postmortem findings were documented. MAIN OUTCOME MEASURES: Morbidity and mortality. RESULTS: Of the 11,195 patients admitted, 420 (4%) deceased during their hospital stay. Most patients died of the disease with which they presented at admission (n = 176, 42%) or of complications (n = 167, 40%). In 20% (n = 83) of the cases a shortcoming in the clinical course was found. 251 of the 420 patients who died (60%) had a necropsy. 53 of the 251 reports (21%) gave information that could have had an effect on the treatment or the clinical course. CONCLUSIONS: Recording mortality is a way of testing the diagnostic and therapeutic accuracy in our quest for a high quality of care.