Optimizing decision-making among childcare staff on fever and common infections: cluster randomized controlled trial.

BACKGROUND: Children 0-4 years attending childcare are more prone to acquire infections than home-cared children. Childcare illness absenteeism due to fever is mostly driven by fear towards fever in childcare staff and parents. This may cause high childcare absenteeism, healthcare service use, and work absenteeism in parents. This study evaluates a multicomponent intervention targeting determinants of decision-making among childcare staff on illness absenteeism due to fever and common infections. METHODS: The multicomponent intervention was developed based on the Intervention Mapping approach and consisted of (i) an educational session, (ii) a decision tool, (iii) an information booklet and (iv) an online video. The intervention was evaluated in a cluster randomized controlled trial in Southern Netherlands. Nine centres received the intervention and nine provided childcare-as-usual. Primary outcome measure was the percentage of illness absenteeism on cluster level, defined as number of childcare days absent due to illness on total of registered childcare contract days in a 12-week period. Secondary outcome measures included intended behaviour, attitude, risk perception, knowledge and self-efficacy of childcare staff. Outcomes were analyzed using linear mixed models accounting for clustering. Knowledge was descriptively analysed. RESULTS: Overall illness absenteeism was comparable in intervention (2.95%) and control group (2.52%). Secondary outcomes showed significant improvements in intervention group regarding intended behaviour, two of three attitude dimensions. Knowledge increased compared with control but no differences regarding self-efficacy. CONCLUSION: The intervention was not effective in reducing illness absenteeism. However, the intervention improved determinants of decision-making such as intended behaviour, attitude, and knowledge on fever. TRIAL REGISTRATION: NTR6402 (registered on 21 April 2017).

Optimising decision making on illness absenteeism due to fever and common infections within childcare centres: development of a multicomponent intervention and study protocol of a cluster randomised controlled trial.

BACKGROUND: Evidence has shown that children 0-4 year-old attending childcare are prone to acquire infections compared to children cared for at home, with fever being the most common symptom. Illness absenteeism due to fever and common infections is substantial and mostly driven by unrealistic concerns and negative attitude towards fever of both childcare staff and parents, resulting in illness absenteeism from childcare, work absenteeism among parents and healthcare service use. The objective of this study is to optimise decision making among childcare staff on illness absenteeism due to fever and common infections in childcare. Underlying determinants of behavioural change were targeted by means of a multicomponent intervention. METHODS: A multicomponent intervention was developed to improve decision making, using the stepwise approach of Intervention Mapping, and in close collaboration with stakeholders and experts. The intervention consisted of 1) a two-hour educational session on fever among childcare staff; 2) an online video for childcare staff and parents emphasising key information of the educational session; 3) a decision tool for childcare staff and parents in the format of a traffic light system to estimate the severity of illness and corresponding advices for childcare staff and parents; 4) an information booklet regarding childhood fever, common infections, and self-management strategies for childcare staff and parents. The multicomponent intervention will be evaluated in a cluster randomised trial with a 12-week follow-up period and absenteeism due to illness (defined as the percentage of childcare days absent due to illness on the total of childcare days during a 12-week period) as primary outcome measure. Secondary outcome measures are: incidence rate and duration of illness episodes, knowledge, attitude, self-efficacy, and risk perception on fever and common infections of childcare staff and parents, healthcare service use in general and paracetamol use, and work absenteeism of parents. DISCUSSION: This study aims to develop a multicomponent intervention and to evaluate to what extent illness absenteeism due to fever and common infections can be affected by implementing a multicomponent intervention addressing decision making and underlying determinants among childcare staff and parents of children attending daycare. TRIAL REGISTRATION: NTR6402 (registered on 21-apr-2017).

Complications on a general surgery service: incidence and reporting.

OBJECTIVES: To determine the incidence and nature of complications on a general surgery service and to compare these results with pre-existing institutional recording and reporting methods. DESIGN: A single observer prospectively monitored the presence and documentation of complications for all patients admitted to the general surgery service at the Wellesley Central Hospital over a 2-month period, through daily chart reviews, attendance at rounds and surgical operating rooms, frequent patient visits on the ward and interviews with the health care team. SETTING: The general surgery service of an urban, university-affiliated teaching hospital. PATIENTS: One hundred and ninety-two general surgery inpatients over 1277 patient-days from June 16, 1996, until Aug. 15, 1996. Same-day surgery patients were not included. RESULTS: Seventy-five (39%) of the 192 patients suffered a total of 144 complications. Two complications (1%) were fatal, 10 (7%) were life threatening, 90 (63%) were of moderate severity and 42 (29%) were trivial. Of these 144 complications, 26 (18%) were deemed potentially attributable to error. One hundred and twelve (78%) of the complications occurred during or after a surgical operation and were related directly or indirectly to it. Only 9 (6%) complications were not documented in the progress notes of the patients' charts. However, 115 (80%) were not presented at weekly morbidity and mortality rounds, and 95 (66%) were not documented on the face sheet of the patients' final medical records. CONCLUSIONS: Complications are common and are underreported by traditional methods. Since hospital funding and quality improvement efforts depend on accurate identification and recording of adverse events, strategies to improve the recording and reporting of complications must be developed.

Clinical and scientific importance of source control in abdominal infections: summary of a symposium.

In May 1997, a panel of surgeon-investigators met to discuss the clinical importance and research implications of controlling the source of abdominal infections. It was concluded that source control is critical to therapeutic success and that antimicrobial therapy and other adjunctive interventions will fail if the source of infection is not controlled by resection, exteriorization or other means. The panelists presented different definitions of source control, depending on the scientific purpose of the definition. All participants agreed that failure to consider the adequacy of source control of infection has limited the value of most clinical trials of therapeutic anti-infective agents. Besides recognizing source control as an essential goal of patient care, the panelists emphasized the need for further investigative work to define, record and stratify the adequacy of source control in clinical trials of therapeutic agents for abdominal infections.

Predicting infection in localized intraabdominal fluid collections: value of pH and pO2 measurements.

PURPOSE: To evaluate the use of pH, pO2, and the subjective opinion of the radiologist compared with bacterial culture in accurate diagnoses of bacterial infection in intraabdominal fluid collections. MATERIALS AND METHODS: Prospectively, 79 patients who were suspected of having an intraabdominal fluid collection underwent diagnostic fluid aspiration. The aspirate was cultured and measured for pH and pO2. A pH < or = 7.1 and a PO2 < or = 49 mm Hg were threshold values used to separate infected from sterile fluid collections. RESULTS: pH alone had a 92% sensitivity and 79% specificity, whereas PO2 alone had a 51% sensitivity and 79% specificity. pH or pO2 combined yielded a 92% sensitivity and 60% specificity. The radiologist's opinion produced a 83% sensitivity and 92% specificity. pH and the radiologist's opinion combined produced a 78% sensitivity and 96% specificity. pH or the radiologist's opinion combined had a 95% sensitivity and a 63% specificity. CONCLUSION: pH is the most sensitive indicator of infection and the radiologist's opinion is the most specific. We recommend proceeding to drainage if the radiologist believes the collection to be infected and performing pH analysis if not. If the pH < or = 7.04, proceed to drainage. If neither of the above criteria are met, drainage could be delayed, pending the results of culture.

The role of oral antimicrobials for the management of intra-abdominal infections.

BACKGROUND: Oral therapy for patients with complicated intra-abdominal infections has been very limited because those patients are frequently ill and need surgery. In addition, at the time of diagnosis and initial treatment, the infection is often accompanied by ileus, gastrointestinal tract function is frequently unknown, and many patients cannot tolerate oral intake. The use of oral antimicrobials in this setting is a recent advance resulting from the availability of agents with good tissue pharmacokinetics and potent aerobic gram-negative activity. This is the first prospective blinded study of oral therapy to provide data on the characteristics of patients eligible for oral treatment and the consequences of such treatment. STUDY DESIGN: In blinded fashion, patients with complicated intra-abdominal infections were randomized to either i.v. ciprofloxacin plus metronidazole or i.v. imipenem throughout their treatment course, or i.v. ciprofloxacin plus metronidazole and treatment with oral ciprofloxacin plus metronidazole when oral feeding was resumed (CIP/MTZ i.v./oral). Physicians could switch the patient to oral therapy between 3 and 8 days after the start of i.v. treatment. RESULTS: One hundred fifty-five of 330 (47%) patients were switched to active or placebo oral therapy. Patients who received i.v./oral therapy were treated, overall, for an average of 8.6 +/- 3.6 days, with an average of 4.0 +/- 3.0 days of oral treatment. Of 46 CIP/MTZ i.v./oral patients (active oral arm), treatment failure occurred in 2 patients (4%) compared with 41 patients (23%) who were not switched to oral agents. No patient or disease features, such as Acute Physiology and Chronic Health Evaluation II score, severity of illness at study entry, organ source of infection, or duration of treatment were identified as predictors of conversion to oral treatment. CONCLUSIONS: In this first prospective examination of sequential i.v./oral therapy for complicated intra-abdominal infections, conversion to oral therapy with ciprofloxacin plus metronidazole appears as effective as continued i.v. therapy for patients able to tolerate oral feedings. Patients who can tolerate oral intake may be treated with appropriate oral antimicrobials and are not at any significant increased risk for failure.

Predicting the need for reoperation for abdominal infection.

Abdominal infection complicating abdominal operation is a serious clinical problem that is subject to diagnostic delay, which is a risk factor for adverse outcomes. Clinical examination and laboratory and imaging modalities become more accurate at achieving a diagnosis once the patient becomes sicker from infection but cannot reliably predict the need for reoperation early in the postoperative course. The Abdominal Reoperation Predictive Index scoring system synthesizes common sense and objective measurements in an attempt to predict the need for reintervention before it is too late. We encourage other centers to test this predictor in their own patient populations.

Antibiotic therapy for abdominal infection.

Abdominal infections are treated by resuscitation, abdominal drainage, control of the source of infection, and antimicrobial agents. Ideally, antimicrobial therapy is active against expected pathogens, safe and effective in clinical trials, inexpensive, and unlikely to promote drug resistance. Numerous single-agent and combination-drug regimens have been efficacious in clinical trials, based on coverage of Escherichia coli and Bacteroides species, the predominant pathogens isolated. Whether expanded antimicrobial coverage is required, especially in hospital-acquired infections, is controversial. Candida infections should be treated with antifungal therapy in patients with recurrent abdominal infections, immunosuppressed patients, and those with candidal abscesses. Most agents have few serious adverse effects; aminoglycosides are the least expensive agents but cause nephro- and ototoxicity. There is little information on the promotion of drug resistance in this condition. Recent developments include the introduction of ticarcillin/clavulanic acid, ampicillin/ sulbactam, piperacillin/tazobactam, meropenem, aztreonam/clindamycin, and ciprofloxacin/metronidazole; success with once-daily aminoglycosides; evidence that antibiotics limit infectious complications of pancreatitis; controversy over the value of diagnostic cultures; the use of oral therapy; evidence in favor of shorter courses of treatment; and the introduction of pharmacoeconomic studies. Clinical investigators are challenged to improve drug trials by stratifying and controlling for the adequacy of surgical intervention.

Soft tissue infections and the diabetic foot.

Soft tissue infections are classified as local or spreading. Spreading soft tissue infections are potentially life-threatening conditions, requiring prompt diagnosis and treatment. The information presented is based on a literature review and the authors' clinical experience. Diagnosis of soft tissue infections is aimed at determining the level of infection (skin, fascia, muscle) and whether necrosis is present. The bacteriology of these infections is varied and is of secondary importance. Treatment of skin infections that have no dead tissue is with antibiotics alone. Infections at the fascial or muscle level and those with necrosis at any level require surgical debridement and adjuvant antibiotics. The feet of diabetic patients are prone to plantar forefoot ulcers associated with tissue destruction and infection. The vast majority are caused by mechanical factors. If local immune defenses are adequate, bacterial colonization occurs without infection. Most diabetic foot ulcers will respond to relief of pressure, which may require total contact casting. Antibiotics and debridement are required in infected or deep ulcers, or when the ulcer does not respond to total contact casting.

Surgical Infection Society position on vancomycin-resistant Enterococcus.

The risk of transfer of vancomycin resistance to staphylococci is a real possibility and has been achieved in the laboratory. Prolonged colonization occurs with vancomycin-resistant Enterococcus (VRE), and many more patients are colonized than infected. The failure to identify, isolate, and adhere to infection control measures when caring for VRE-colonized patients dooms to failure any means to control its spread. Control of vancomycin use alone is unlikely to greatly affect the number of patients at risk for VRE colonization. The global spread of VRE may be impossible to stop, but infection control measures are the most important line of defense inside hospitals.

Pharmacodynamics of antimicrobial therapy in surgery.

INTRODUCTION: "Pharmacodynamics" refers to the relationship of drug concentrations in serum or tissues to effects on biologic systems. Concepts used to describe antimicrobial pharmacodynamics include the minimal inhibitory concentration (MIC), the minimal bactericidal concentration (MBC), and serum bactericidal titers (SBT), as well as post-antibiotic effect. METHODS: Pertinent published literature was identified through a MEDLINE search. RESULTS: Aminoglycosides have a concentration-dependent effect on bacteria killing and possess a relatively long postantibiotic effect. Given these characteristics, single-daily dosing, where the total daily dose with a traditional aminoglycoside regimen is given as one dose, may be more efficacious compared with more frequent dosing. For beta-lactam antimicrobials, bacterial killing is related to the duration of time that the free drug concentration exceeds the bacterial MIC. Beta-lactam antimicrobials have been shown to have no, or a short postantibiotic effect. Beta-lactam antimicrobials may be more effective when administered as continuous intravenous infusions. CONCLUSIONS: Pharmacodynamic variation may result from differences in drug sensitivity among individuals and the nature of the interaction between antimicrobials and microorganisms. Proper use of pharmacokinetic and pharmacodynamic principles can result in more effective and less toxic antimicrobial regimens.

Results of a randomized trial comparing sequential intravenous/oral treatment with ciprofloxacin plus metronidazole to imipenem/cilastatin for intra-abdominal infections. The Intra-Abdominal Infection Study Group.

OBJECTIVE: In a randomized, double-blind, multicenter trial, ciprofloxacin/metronidazole was compared with imipenem/cilastatin for treatment of complicated intra-abdominal infections. A secondary objective was to demonstrate the ability to switch responding patients from intravenous (IV) to oral (PO) therapy. SUMMARY BACKGROUND DATA: Intra-abdominal infections result in substantial morbidity, mortality, and cost. Antimicrobial therapy often includes a 7- to 10-day intravenous course. The use of oral antimicrobials is a recent advance due to the availability of agents with good tissue pharmacokinetics and potent aerobic gram-negative activity. METHODS: Patients were randomized to either ciprofloxacin plus metronidazole intravenously (CIP/MTZ IV) or imipenem intravenously (IMI IV) throughout their treatment course, or ciprofloxacin plus metronidazole intravenously and treatment with oral ciprofloxacin plus metronidazole when oral feeding was resumed (CIP/MTZ IV/PO). RESULTS: Among 671 patients who constituted the intent-to-treat population, overall success rates were as follows: 82% for the group treated with CIP/MTZ IV; 84% for the CIP/MTZ IV/PO group; and 82% for the IMI IV group. For 330 valid patients, treatment success occurred in 84% of patients treated with CIP/MTZ IV, 86% of those treated with CIP/MTZ IV/PO, and 81% of the patients treated with IMI IV. Analysis of microbiology in the 30 patients undergoing intervention after treatment failure suggested that persistence of gram-negative organisms was more common in the IMI IV-treated patients who subsequently failed. Of 46 CIP/MTZ IV/PO patients (active oral arm), treatment success occurred in 96%, compared with 89% for those treated with CIP/MTZ IV and 89% for those receiving IMI IV. Patients who received intravenous/oral therapy were treated, overall, for an average of 8.6 +/- 3.6 days, with an average of 4.0 +/- 3.0 days of oral treatment. CONCLUSIONS: These results demonstrate statistical equivalence between CIP/MTZ IV and IMI IV in both the intent-to-treat and valid populations. Conversion to oral therapy with CIP/MTZ appears as effective as continued intravenous therapy in patients able to tolerate oral feedings.

Liposomal cefoxitin in a porcine model of intra-abdominal sepsis: hemodynamic changes.

The effects of free versus liposomal cefoxitin on various physiological parameters in a porcine model of Gram-negative intra-abdominal sepsis were evaluated. Four different doses of Escherichia coli inoculum mixed with sterile pig feces were used (10(8), 10(9), 10(10), and 10(11) cfu/animal), and the most consistent hemodynamic changes were observed with an inoculum of approximately 10(11) bacteria/20 kg animal. Two treatment groups were established as follows: free cefoxitin (n = 9) and liposomal cefoxitin (n = 9). All animals were maintained under anesthesia for the duration of the study, and then euthanized 24 h following intra-abdominal inoculation. The inoculated and nontreated animals showed increases in heart rate, mean pulmonary arterial pressure, systemic and pulmonary vascular resistance, and decreases in mean systemic arterial pressure and cardiac index. These changes were significant (p < .05) compared with a control group injected with normal saline. Liposomal cefoxitin-treated animals showed significantly lower decreases in mean systemic arterial pressure and increases in heart rate (p < .05) compared with both the inoculated nontreated and free cefoxitin-treated groups. Both liposomal and free cefoxitin significantly modulated the mean pulmonary arterial pressure compared with the inoculated nontreated animals (p < .05). Acidosis that developed during intra-abdominal infection diminished 6 h following the first dose of liposomal cefoxitin (p < .05). The results of these experiments demonstrate that liposomal cefoxitin exerts a beneficial modulation of some of the hemodynamic disturbances during intra-abdominal Gram-negative sepsis.

Duration of antibiotic treatment in surgical infections of the abdomen. Postoperative peritonitis.

Postoperative peritonitis is potentially lethal and is usually caused by leakage of gut contents. Successful management depends on early diagnosis and treatment which require clinical suspicion and aggressive diagnostic imaging. Treatment consists of fluid and nutritional resuscitation, peritoneal toilet, control of gut leakage and initiation of antimicrobial therapy. Since delay in diagnosis is common, antimicrobial treatment is usually begun when the infection has become well developed. Experimental evidence has shown that in some settings antimicrobial agents do not perform as well in later stages of infection but whether this applies to peritonitis is not known. The optimal duration of antimicrobial therapy has not been studied specifically in postoperative peritonitis. Arguments for prolonged treatment include the potential for greater killing of bacteria in patients who have severe infections. Arguments against prolongation of therapy include the lesser role of antibiotics compared with operative management, doubt about the value of antibiotics' ability to kill bacteria at later stages of infection and the significant number of infective and non-infective complications of drug therapy. Limitation of antimicrobial treatment to no more than seven days is advocated. Persistent clinical signs, fever, or leucocytosis should prompt a search for a drainable focus of infection in the abdomen or treatable site elsewhere.

Liposomal cefoxitin in a porcine model of intra-abdominal sepsis: bactericidal efficacy.

The bactericidal effect of free versus liposomal cefoxitin was evaluated in the major reticuloendothelial organs in a porcine model of intra-abdominal sepsis. Yorkshire Landrace pigs were inoculated with 3.2 x 10(10) (n = 5) or 1.4 x 10(11) (n = 7) cfu of Escherichia coli mixed in sterile feces/animal. Two treatment groups inoculated with 1.4 x 10(11) cfu were established: free cefoxitin (n = 9) and liposomal cefoxitin (n = 9). All animals were maintained under anesthesia and euthanized after 24 h. The number of E. coli recovered in the liver, lungs, and spleen was significantly affected by inoculum size (p < .05). The liver had significantly higher numbers of bacteria (p < .05) compared with the other organs, regardless of the inoculum size. The liver and the lung of the liposomal cefoxitin-treated group showed significantly lower numbers of E. coli (5.0 x 10(4) and 6.3 x 10(2), respectively) compared with the untreated (liver, 6.3 x 10(7); lung, 2.0 x 10(6)) and free cefoxitin (liver, 5.0 x 10(6); lung, 7.9 x 10(4))-treated groups (p < .05). At 2 h following the injection of free and liposomal cefoxitin, the decrease of E. coli in peritoneal fluid compared with the nontreated septic group was significant (p < .05). No growth was observed from blood cultures taken 24 h after sepsis induction. All control experiments yielded negative cultures. The results of these experiments demonstrated that liposomal cefoxitin exerts an enhanced bactericidal effect in liver and lungs during Gram-negative sepsis.

Definition of the role of enterococcus in intraabdominal infection: analysis of a prospective randomized trial.

BACKGROUND: The role of enterococcus in intraabdominal infection is controversial. This study examines the contribution of enterococcus to adverse outcome in a large intraabdominal infection trial. METHODS: A randomized prospective double-blind trial was performed to compare two different antimicrobial regimens in combination with surgical or percutaneous drainage in the treatment of complicated intraabdominal infections. A total of 330 valid patients was enrolled from 22 centers in North America. RESULTS: In 330 valid patients, 71 had enterococcus isolated from the initial drainage of an intraabdominal focus of infection. This finding was associated with a significantly higher treatment failure rate than that of patients without enterococcus (28% versus 14%, p < 0.01). In addition, only Acute Physiology and Chronic Health Evaluation II score and presence of enterococcus were significant independent predictors of treatment failure when stepwise logistic regression was performed (p < 0.01 and < 0.03). Risk factors for the presence of enterococcus include age, Acute Physiology and Chronic Health Evaluation II, preinfection hospital length of stay, postoperative infections, and anatomic source of infection. There was no difference between the clinical trial treatment regimens with regard to overall failure, failure associated with enterococcus, or frequency of enterococcal isolation. CONCLUSIONS: This study is the first to report enterococcus as a predictor of treatment failure in complicated intraabdominal infections. This trial also identifies several significant risk factors for the presence of enterococcus in such infections.

Unilamellar liposomes modulate secretion of tumor necrosis factor by lipopolysaccharide-stimulated macrophages.

Liposomal encapsulation of antimicrobial agents has been used to improve drug delivery, particularly against intracellular pathogens. The effect of unilamellar liposomes on macrophage activation in response to Escherichia coli lipopolysaccharide was examined. Liposomes caused a dose- and time-dependent inhibition of tumor necrosis factor release by lipopolysaccharide-treated cells. The accumulation of tumor necrosis factor mRNA transcripts was unaffected, suggesting a posttranscriptional mechanism for this effect. However, induction of macrophage procoagulant activity was unaffected by liposomes, indicating a selective rather than a global inhibition. These data suggest that liposomes used for drug delivery may modulate the host response to infection.

Deferoxamine induces hypotension in experimental gram-negative septicemia.

Multiple organ system failure may result from tissue damage caused by activated neutrophils or endotoxin. A significant part of this tissue damage is due to peroxidation induced by oxygen-free radicals and requires iron as a co-factor. Iron chelation has been shown to prevent tissue damage in some models. This experiment was carried out to determine whether iron chelation with deferoxamine (DFO) would prevent lung damage in a swine model of Gram-negative septicemia. Fifteen animals were randomized to control, Pseudomonas aeruginosa infusion at a rate of 2 x 10(7) colony forming units/20 kg/min (septic group), or Pseudomonas infusion combined with DFO pretreatment at a dose of 80 mg/kg/h (septic-treated group). Three of six septic-treated animals became severely hypotensive and died during the course of the experiment as opposed to none of six septic animals. Surviving septic-treated animals were significantly hypotensive (60 +/- 24 mmHg mean arterial pressure) compared to septic (122 +/- 9 mmHg) and control (109 +/- 8 mmHg) animals. DFO did not improve respiratory function (e.g., pO2) or morphology in septic animals. We conclude that iron-chelation therapy with DFO at the above dosage results in a significant deterioration in cardiovascular function in septic swine. Lung damage was not prevented.

Steroids, APACHE II score, and the outcome of abdominal infection.

OBJECTIVE: To compare the outcome of abdominal infection in patients with or without previous systemic glucocorticoid therapy and to determine the effect of steroid administration on the relationship between APACHE II (Acute Physiology and Chronic Health Evaluation) scores and mortality. HYPOTHESIS: Steroid therapy leads to greater mortality and relatively lower APACHE II scores. DESIGN: Prospective cohort study. SETTING: University hospital. PATIENTS: Two hundred ninety-seven consecutive adult patients with abdominal infection treated by surgical or percutaneous drainage. Treatment was at the clinician's discretion. Seventy-one patients received systemic steroid therapy. OUTCOME MEASURES: APACHE II score, clinical course, and death in hospital; relationship between APACHE II score and mortality in the steroid and no steroid groups. RESULTS: Thirty-three patients receiving steroid therapy (46%) died vs 55 (24%) of 226 patients not receiving steroid therapy. The APACHE II score (P < .0001) and steroid administration (P = .04) were each independently associated with death. Steroid-treated patients had the same probability of dying as "nonsteroid" patients with an APACHE II score a mean of 3.7 points higher (95% confidence limits, 0.03 and 7.7). The confidence that 2, 3, or 4 extra APACHE II points is the appropriate correction for steroid-treated patients is 80%, 60%, or 40%, respectively. CONCLUSIONS: Patients receiving steroid therapy appear to be at higher risk of dying of abdominal infection than predicted by APACHE II scores. The number of patients receiving cancer chemotherapy was too small to determine whether this was an additional risk factor. In the design of clinical trials stratified by APACHE II scores, steroid-treated patients should either be excluded or assigned two extra APACHE II points.