Clinical research
How can we decrease mortality in surgery units?

Introduction
Mortality in general surgery units in hospitals in Poland oscillates from few tenths of a per cent to a few per cent. It depends on numerous factors. The most important ones include: the unit profile, severity of the admitted patients’ clinical condition, skills, knowledge and practice of the unit staff, the unit sanitary conditions [1-3]. In the opinion of the authors of this study, the model of postoperative care and the principles of the cooperation with the intensive care unit are also of great importance. In the year 2006, there functioned 10 public health care institutions in the area of Lodz, having in their structure surgical unit classified as general surgery unit (code: 4500). They were three university teaching, three provincial, three county and one departmental hospitals. The lowest percentage of mortality was noted in the surgical unit of University Teaching Hospital No. 5 in Lodz (UH No. 1). The authors decided to analyse the causes of such low mortality in this hospital. Two remaining university teaching hospitals: University Teaching Hospital No. 1 in Lodz (UH No. 1) and University Teaching Hospital No. 2 in Lodz (UH No. 2) were selected for comparative analysis. Postoperative care at the University Teaching Hospital No 5 is performed by the staff of the Department of Anaesthesiology and Intensive Care Unit. The surgeons are consultants. Such a solution is extremely rare not only in Poland but also all over the world. In the two remaining university hospitals postoperative care is conducted in a traditional way by surgeons in recovery rooms. The authors decided to analyse the factors affecting mortality in selected general surgery units with particular consideration of the functioning model of postoperative care.
Material and methods
The study was a retrospective analysis of mortality in general surgery units located at three university hospitals: N. Barlicki University Hospital No. 1 in Lodz, WAM University Hospital No. 2 in Lodz and B. Szarecki University Hospital No. 5. The selection of the hospitals was dictated by a few reasons. The Medical University of Lodz is the founding body of all the hospitals subjected to analysis. These hospitals are only few kilometres away from each other. The units have similar number of beds, and well-educated medical and nursing staff. Heads of the hospital departments have all been awarded professorships. Health benefits are provided on the basis of the same list of benefits as part of contracts with the same payer – Lodz Provincial Branch of National Health Fund (LPB NHF). The study comprised 26,020 patients treated in these units from 01.01.2003 to 31.12.2006. The available statistical material was analysed. In the first stage the statistical data were analysed of the Provincial Centre of Public Health in Lodz. The available statistical material was analysed. In the first stage the statistical data were analysed by the Provincial Centre of Public Health in Lodz (PCPH) (data from centralized public health). The obtained information concerned the number of treated patients, the number of patients transferred, discharged or dead, the number of man-days of treatment, mean bed use, mean hospitalisation time, mean number of patients per bed and mortality. Authors used also data included in the questionnaires, filled by the heads of surgical departments in Lodz province, including among others the general information about the department, about the number and qualifications of the medical staff and data about the performed surgeries. In the second stage the structure of the analysed units and the structure of the selected groups of diagnoses were compared. A relative structure similarity index (Pw) was used to compare the structure of hospitalised patients in the analysed units: m Smin(uk1,uk2) k=1 Pw=––––––––––––––– m Smax(uk1,uk2) k=1 The third stage was focused on explaining the reasons of significantly lower mortality among patients hospitalised in surgical unit of UH No. 5. The treatment of postoperative patients in recovery rooms in the conditions similar or identical to those in the intensive therapy unit is extremely rare. The general surgery unit of B. Szarecki University Teaching Hospital in Lodz provides such conditions. An anaesthesiologist and 2 anaesthesiological nurses serve permanent medical care (4-bedded rooms). When needed, a surgeon is asked for consultation. All the operated-on patients are moved to recovery rooms where their life functions are monitored. Most frequently the patients stay there for 24 hours after the surgical procedure. The scope of monitoring depends on the kind and range of surgery. Standard monitoring includes: ECG, arterial blood pressure by indirect method, plethysmography, respiratory rate, body temperature, fluid supply, and diuresis. In the case of severe surgery, monitoring is significantly broadened and includes central venous pressure in patients with central venous access, arterial blood pressure by direct method and capnography in ventilated patients. All these allow early detection and treatment of any occurring complications. The application of high quality analgesic treatment with the use of phentanyl infusion in combination with nonsteroidal anti-inflammatory drugs is of importance. Effective analgesia not only improves the patient’s comfort but prevents the occurrence of numerous complications as well. In the conditions similar to those in intensive care units, such treatment is completely safe. Wide availability of infusion pumps allows for application of drug infusion (e.g. catecholamines, nitroglycerin, beta-blockers, hypotensive drugs, etc.) and fluids. Perioperative procedures applied in B. Szarecki University Hospital No. 5 make it easier to transfer the patients immediately after severe surgery into the Intensive Care Unit, which in our opinion is a prevention against the development of serious, life-threatening symptoms. The patients who were transferred to the ICU also meet the requirements given by LPB NHF, that is, obtaining the adequate number of points on the TISS 28 scale (25 points). In the general surgery units of N. Barlicki University Hospital No. 1 in Lodz and WAM University Hospital No. 2 in Lodz, postoperative care is provided by surgeons and surgical nurses. In the conditions of surgical units, there is not always a possibility of application of adequate monitoring and rapid reaction to life-threatening conditions. Very often, analgesic therapy is limited to application of nonsteroidal anti-inflammatory drugs on the patient’s demand. A relatively insignificant percentage of postoperative patients, due to different organizational, medical or other problems, land in the Intensive Care Units at N. Barlicki University Teaching Hospital No. 1 in Lodz and WAM University Hospital No. 2 in Lodz. These are most frequently patients in very severe conditions. It seems that earlier transfer of patients with high postoperative risk could prevent such conditions. The statistical analysis was performed with Microsoft Excel. The tests for two means for independent samples (HO: µ1=µ2; H1: µ1≠µ2) were used for the statistical analysis.
Results
Structure of hospitalization in general surgery units in selected hospitals according to basic disease (A00-A99; …; Z00-Z99) in the years 2003 to 2006 is presented in Table I. The relative structure similarity indices of the hospitalized patients in general surgery units in the selected hospitals acc. to the basic disease (A00-A99; …; Z00-Z99) in the years 2003-2006 were respectively: UH No. 1/UH No. 2 – 0.511; UH No. 1/UH No. 5 – 0.549; UH No. 2/UH No. 5 – 0.637 and 0.418 together for UH No. 1/UH No. 2/UH No. 5. To compare objectively the mortality, the most numerous groups of patients: with diagnosed C00-C97; D00-D89 and K00-K93 were selected for further analysis. The mortality in group of patients with diagnosed C00-C97 was in UH No. 1 15.21%, in UH No. 2 8.12% and in UH No. 5 1.46% (UH No. 1/UH No. 5, p<0.001 and UH No. 2/UH No. 5, p<0.001). Relative structure similarity indices were in the range of C00-C97 diagnoses respectively: UH No. 1/UH No. 2 – 0.335; UH No. 1/UH No. 5 – 0.358; UH No. 2/UH No. 5 – 0.271 and 0.207 together for UH No. 1/UH No. 2/UH No. 5. After taking into account these indices (appropriate modification of the structure and number of deaths) the mortality was respectively: in UH No. 1 14.52%, in UH No. 2 6.21% and in UH No. 5 2.42% (UH No. 1/UH No. 5, p<0.001 and UH No. 2/UH No. 5, p<0.02). The mortality in group of patients with diagnosed D00-D89 was in UH No. 1 5.32%, in UH No. 2 0.59% and in UH No. 5 0.12% (UH No. 1/UH No. 5, p<0.001 and UH No. 2/UH No. 5 NS). Relative structure similarity indices were in the range of D00-D89 diagnoses respectively: UH No. 1/UH No. 2 – 0.236; UH No. 1/UH No. 5 – 0.553; UH No. 2/UH No. 5 – 0.178 and 0.147 together for UH No. 1/UH No. 2/UH No. 5. After taking into account these indices (appropriate modification of the structure and number of deaths) the mortality was respectively: in UH No. 1 5.38%, in UH No. 2 0.43% and in UH No. 5 0.08% (UH No. 1/UH No. 5, p<0.05 and UH No. 2/UH No. 5 NS). In the most numerous group (K00-K93) the mortality was in UH No. 1 2.49%, in UH No. 2 0.75% and in UH No. 5 0.08% (UH No. 1/UH No. 5, p<0.001 and UH No. 2/UH No. 5, p<0.001). Relative structure similarity indices were in the range of K00-K93 diagnoses respectively: UH No. 1/UH No. 2 – 0.668; UH No. 1/UH No. 5 – 0.486; UH No. 2/UH No. 5 – 0.512 and 0.416 together for UH No. 1/UH No. 2/UH No. 5. After taking into account these indices (appropriate modification of the structure and number of deaths) the mortality was respectively: in UH No. 1 1.30%, in UH No. 2 0.62% and in UH No. 5 0.10% (UH No. 1/UH No. 5, p<0.001 and UH No. 2/UH No. 5, p<0.03). Mortality and statistical significance in the group of hospitalized patients with the diagnosis C00; …; C97; D00; …; D89 and K00; …; K93 according to the basic disease in general surgery units in selected hospitals in the years 2003-2006 after modification of the structure and number of deaths are presented in Table II. The age structure of patients and the mode of hospital admission with the diagnosis C00; …; C97; D00; …; D89 and K00; …; K93 for the selected hospitals between 2003-2006 year are presented on Figures 1-6. To assess the effect of the quality of postoperative care and the current model of postoperative care on mortality, the mortality of patients with the same diagnoses was subjected to analysis according to the basic disease in the groups C00; …; C97; D00; …; D89 and K00; …; K93 (after modification of the structure and the number of deaths). The results are demonstrated in Table III. The selected data from the questionnaires filled annually by heads of surgical departments are presented on Table IV.
Discussion
The available in literature reports comparing mortality in surgical units concern first of all the mortality of patients subjected to a specific surgical procedure. The comparison of mortality in surgical units, comprising both the patients operated on and those treated conservatively, is very rare due to methodological difficulties faced by researchers as well as numerous factors hard to objectivize, affecting mortality in the analysed units. The structure of the admitted patients, their clinical condition, the kind of performed procedures, skills and experience of the operating surgeons, kind of postoperative care, cooperation with other units including intensive care unit are the most essential factors affecting mortality in surgical units. Postoperative care performed by surgeons in the conditions of a recovery room or intensive (surgical) care room is widespread solution also in other countries. Sometimes an anesthesiologist is included into the surgical team, however his task is first of all anaesthetization of patients admitted to this unit, less frequently he takes part in the process of postoperative treatment. Among well educated physicians it does not matter if you are a surgeon or an anesthesiologist but the conditions to care the patients. However, experience of anesthesiologists in the management of patients with respiratory, circulatory or urinary dysfunction is in Poland significantly greater than that of surgeons. It results first of all from the binding model of education and from the organization of national health care system. Similar situation is found in a lot of countries all over the world. From this point of view lower mortality in a surgical unit, where postoperative care is conducted by anesthesiologists, should not be a surprise. In the opinion of the authors of this study, the higher is the per cent of patients operated on in severe condition or with significant burden, the more noticeable should be the difference in mortality in the case of the change in postoperative care organization. High-intensity ICUs have been associated with improved outcomes. Angus et al. defined an ICU as “high intensity” if ł80% of patients were cared for by a critical care physician (intensivist) and defined an ICU as compliant with Leapfrog if it was both high-intensity and providing some form of in-house physician coverage during all hours [1-3]. In case of UH No. 5 all patients subjected to surgery are under the intensive care for all day (and for at least 1 day), whereas in the University Hospitals No. 1 and 2 the intensive care finishes in the moment when patients comes round from the anesthetic. High-intensity vs. low-intensity ICU physician staffing is associated with reduced hospital and ICU mortality and hospital and ICU [4]. The authors analysed in detail the patients with the diagnoses C00; …; C97; D00; …; D89 and K00; ...; K93 according to the basic disease, from the years 2003-2006. The patients with these diagnoses, the number of whom was at least 15 in each hospital (after modification of the structure) were qualified for further analysis. These diagnoses had the following statistical numbers: C16 (stomach cancer), C18 (colon cancer), C20 (malignant neoplasm of the rectum), D37 (oral cavity and digestive system neoplasms of uncertain behaviour and/or unspecified nature), K29 (gastritis and duodenitis), K31 (other diseases of stomach and duodenum), K35 (acute appendicitis), K40 (inguinal hernia), K42 (umbilical hernia), K43 (abdominal hernia), K56 (paralytic ileus), K60 (fissure and fistula of anus and rectum), K62 (other diseases of anus and rectum), K63 (other intestinal diseases i.e. abscess, perforation, fistula), K80 (cholelithiasis), K81 (cholecystitis), K83 (other disorders of biliary tract), K85 (acute pancreatitis), K86 (other diseases of pancreas) and K92 (other diseases of digestive system). Among these diagnoses: K35 (acute appendicitis), K40 (inguinal hernia), K42 (umbilical hernia), K43 (abdominal hernia), K60 (fissure and fistula of anus and rectum), K62 (other diseases of anus and rectum) and K80 (cholelithiasis) are the ones in which mortality should be low and it should not differ statistically. If such statistically significant differences occurred, they would manifest inadequate qualifications of the operating surgeons and/or improper sanitary rigour. The diagnoses: C16 (malignant neoplasm of the rectum), C18 (colon cancer), K56 (paralytic ileus), K85 (acute pancreatitis) are the ones with the increased risk of complications and death, in which proper postoperative care or in cases of need – the treatment in intensive care unit may be of importance to decrease mortality. The decrease of mortality in these groups of patients in any of the analysed hospitals at insignificant difference in mortality of patients with the diagnoses of: K35, K40, K42, K43, K60, K62 or K80 would manifest, according to the authors of this study, the advantage of the postoperative care system in this hospital over the other system. The carried out statistical analysis demonstrated very low mortality in the groups of patients with the diagnoses of: K35, K40, K42, K43, K60, K62 and K80. The mortality in these groups of patients did not differ statistically among the analysed hospitals. In the opinion of the authors it proves that the skills of the operating team similar (high) in all the analysed hospitals and typical complications for these diagnoses e.g. infections, were not a distinct problem from the point of view of mortality in these groups of patients. It also points indirectly to the sufficient sanitary rigor. The presented in Table III number of hospitalizations, mortality and statistical significance concern the group of patients after modification of the structure and number of deaths. It is obvious that this procedure is necessary in the case of the comparison of aggregated mortality. However, it is not necessary in the comparison of mortality within the same basic diagnosis. Then, mortality is the same before and after modification and only the number of hospitalizations and deaths changes. These, of course affect the level of statistical significance in the carried out analyses. For instance, before modification of the structure and the number of deaths, more statistically significant difference in mortality was observed in the diagnosis of C16 (UH No. 1/UH No. 5, p<0.02), C18 (UH No. 1/UH No. 5, p<0.001) and K56 (UH No. 1/UH No. 5, p<0.01). The influence of the sample quantity is obvious in comparison to the data presented in Table III. However, the sample size did not affect the statistical significance in the case of the following diagnoses: K35, K40, K42, K43, K60, K62 and K80. It is also worth paying attention to the fact that mortality in the group of patients with diagnosed C00-C97 was respectively: in UH No. 1 14.52%, in UH No. 2 6.21% and in UH No. 5 2.42% (UH No. 1/UH No. 5, p<0.001 and UH No. 2/UH No. 5, p<0.02) and in the group of patients with diagnosed K00-K93 the mortality was respectively: in UH No. 1 1.30%, in UH No. 2 0.62% and in UH No. 5 0.10% (UH No. 1/UH No. 5, p<0.001 and UH No. 2/UH No. 5, p<0.03) (after appropriate modification of the structure and number of deaths). In the opinion of the authors, on one hand the results demonstrate that the mortality in the analysed groups of patients is significantly lower in UH No. 1 than in the other hospitals, on the other hand they consider in the analysis the patients with increased risk of complications and death in whom proper postoperative care may be of importance to decrease mortality. Information that the Department of General and Colorectal Surgery at the University Hospital No. 5 has been functioning since January 2003 deserves particular attention. Until 31 December, 2002 majority of physicians and nurses from this team (including the Head of the Department) worked in the surgical unit of the Ministry of Internal Affairs and Administration Hospital in Lodz. The system of postoperative care in surgical unit of that hospital was organized in similar way as in University Hospital No. 1 and No. 2 and mortality was on the average 1.55% in the years 2000-2002. The range of surgical procedures was significantly smaller than those performed in University Hospital No. 5, both in objective (ICD-10) and subjective (operating surgeons) evaluation. Analyzing the questionnaires handed every year by the heads of the surgical units to the provincial consultant in surgery, the authors did not find any other significant causes explaining the differences in mortality of patients in the analysed surgical units of the selected hospitals. A detailed analysis of the diagnoses and of mortality allows to presume that early postoperative intensive care in severely ill patients and in cases of need immediate admission to ICU significantly decrease mortality. The latter is of great importance in the case of explanation of the difference in mortality between surgical units of UH No. 1 and UH No. 2. The Intensive Care Unit at UH No. 2 can admit more patients from other hospital units due to underestimation of the value of contract with National Health Fund and thus fewer patients than in ICU at UH No. 1. The authors are aware of the fact that with continuous shortage of places in intensive care units for first of all the treatment of traumatic cases and acute cardio-pulmonological states, rapid change of the system of treatment of postoperative patients is impossible. Nevertheless, the significance of the numbers quoted in this study seems to justify the usefulness of considering such activities. In conclusions the change of the system of postoperative care consisting in: taking over postoperative care by physicians and anaesthesiological nurses, intensive monitoring of postoperative patients, immediate transfer of patients with life hazard to Intensive Care Unit, decreases significantly mortality in a surgical unit.
References
1. Angus DC, Shorr AF, White A, Dremsizov T, Schmitz R, Kelley M. Critical care delivery in the United States: distribution of services and compliance with Leapfrog recommendations. Crit Care Med 2006; 34: 1016-24. 2. Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002; 346: 1128-37. 3. Piechota M, Banach M. Changes of the system of postoperative care decreases mortality in a surgical unit. Crit Care 2007; 11 (Suppl 2): P480. 4. Pronovost PJ, Angus DC, Dorman T, Robinson KA, Dremsizov TT, Young TL. Physician staffing patterns and clinical outcomes in critically ill patients: a systematic review. JAMA 2002; 288: 2151-62.