PubMed Health⌕ Search

Biomedical subjects

E P Sloan

Publications and source records attributed to E P Sloan.

At least 19 recordsLinked to original sources
PubMed2000

Setting-based practice variation in the management of simple febrile seizure.

OBJECTIVE: To identify provider-based differences in the ED assessment and management of children presenting with uncomplicated, first-time febrile seizures. METHODS: Multicenter, retrospective cohort study of seven EDs in-the Chicago area: two tertiary academic pediatric EDs (PEDs) and five community-based general EDs (GEDs). The visits of all patients with a discharge diagnosis including the term "seizure" were identified from a 30-month period. Records of patients who met criteria for simple, first-time febrile seizure were reviewed (age 6-60 months; temperature > or =38.0 degrees C; single, generalized, tonic-clonic seizure <20 minutes; "alert" or "arousable" on presentation; absence of known neurologic disease). RESULTS: Four hundred fifty-five records were included: 330 and 125 patients presenting to GEDs and PEDs, respectively. The two groups did not differ in mean age, vital signs, reported duration of seizure, or prior antibiotic use. Lumbar puncture (LP) was performed more often in the GED group (33% vs 22%). No patients were found to have bacterial meningitis. The patients in the GED group were more likely to receive parenteral antibiotics in the ED (56% vs 22%) and to be admitted or transferred (18% vs 4%). In a logistic regression model incorporating age, temperature, seizure duration, seizure in the ED, prior antibiotic use, primary care, and insurance status, the GED patients remained more likely to have an LP (OR 1.5), receive parenteral antibiotics (OR 2.5), and be admitted or transferred (OR 2.5). CONCLUSIONS: There were significant setting-based differences in the evaluation and management of children with simple febrile seizures presenting to GEDs and PEDs.

Adult↗
PubMed1999

Diaspirin cross-linked hemoglobin (DCLHb) in the treatment of severe traumatic hemorrhagic shock: a randomized controlled efficacy trial.

CONTEXT: Severe, uncompensated, traumatic hemorrhagic shock causes significant morbidity and mortality, but resuscitation with an oxygen-carrying fluid might improve patient outcomes. OBJECTIVE: To determine if the infusion of up to 1000 mL of diaspirin cross-linked hemoglobin (DCLHb) during the initial hospital resuscitation could reduce 28-day mortality in traumatic hemorrhagic shock patients. DESIGN AND SETTING: Multicenter, randomized, controlled, single-blinded efficacy trial conducted between February 1997 and January 1998 at 18 US trauma centers selected for their high volume of critically injured trauma patients, but 1 did not enroll patients. PATIENTS: A total of 112 patients with traumatic hemorrhagic shock and unstable vital signs or a critical base deficit, who had a mean (SD) patient age of 39 (20) years. Of the infused patients, 79% were male and 56% were white. An exception to informed consent was used when necessary. INTERVENTION: All patients were to be infused with 500 mL of DCLHb or saline solution. Critically ill patients who still met entry criteria could have received up to an additional 500 mL during the 1-hour infusion period. MAIN OUTCOME MEASURES: Twenty-eight day mortality, 28-day morbidity, 48-hour mortality, and 24-hour lactate levels. RESULTS: Of the 112 patients, 98 (88%) were infused with DCLHb or saline solution. At 28 days, 24 (46%) of the 52 patients infused with DCLHb died, and 8 (17%) of the 46 patients infused with the saline solution died (P = .003). At 48 hours, 20 (38%) of the 52 patients infused with DCLHb died and 7 (15%) of the 46 patients infused with the saline solution died (P = .01). The 28-day morbidity rate, as measured by the multiple organ dysfunction score, was 72% higher in the DCLHb group (P = .03). There was no difference in adverse event rates or the 24-hour lactate levels. CONCLUSIONS: Mortality was higher for patients treated with DCLHb. Although further analysis should investigate whether the mortality difference was solely due to a direct treatment effect or to other factors, DCLHb does not appear to be an effective resuscitation fluid.

Adult↗
PubMed1999

Nocturnal and daytime panic attacks--comparison of sleep architecture, heart rate variability, and response to sodium lactate challenge.

BACKGROUND: The purpose of this study was to determine if nocturnal panic patients have greater autonomic dysregulation than patients with daytime panic. METHODS: Three groups were studied: patients who suffer from panic attacks during sleep (n = 12), those who suffer from daytime panic attacks only (n = 12), and control subjects (n = 12). Each subject underwent 24-hour holter monitoring for heart rate variability (HRV), an overnight sleep recording, and sodium lactate challenge during wakefulness. RESULTS: There was a marked subjective response to the sodium lactate challenge in the panic disorder (PD) patients but not in control subjects. Each group showed changes in HRV in response to sodium lactate challenge. The decrease in HRV measures was more marked in PD patients as a whole than in control subjects. During non-rapid eye movement (REM) sleep the value for total power (TP) was significantly higher in the nocturnal panic patients. The PD patients as a whole had higher values for TP and low-frequency (LF) power during REM sleep than control subjects. There were no significant differences between the two PD groups in sleep architecture. The PD patients as a whole had lower sleep efficiency and less stage 4 sleep than control subjects. CONCLUSIONS: These findings indicate that there are substantial differences between PD and control subjects in autonomic regulation and that there are small differences between patients with daytime panic attacks and those with sleep-related panic attacks.

Adult↗
PubMed1999

The informed consent process and the use of the exception to informed consent in the clinical trial of diaspirin cross-linked hemoglobin (DCLHb) in severe traumatic hemorrhagic shock. DCLHb Traumatic Hemorrhagic Shock study group.

UNLABELLED: In the clinical trial of diaspirin cross-linked hemoglobin (DCLHb), optimal therapy required the immediate enrollment of patients with severe, uncompensated, traumatic hemorrhagic shock. When it was not feasible to obtain prospective consent, an exception to informed consent was used according to FDA regulation 21 CFR 50.24. OBJECTIVES: To examine the informed consent process and the use of the consent exception and consent to continue (CTC), and to describe the patients for whom this process was used. METHODS: This was a multicenter, randomized, controlled, single-blinded efficacy trial of DCLHb as an adjunct to standard therapy in the treatment of severe, traumatic hemorrhagic shock. Patients with unstable vital signs or a critical base deficit were treated, with a primary study endpoint of 28-day mortality. RESULTS: During the 11-month study period, 112 patients were randomized in 18 U.S. trauma centers, and data from 98 of the infused patients were analyzed. Prospective consent was obtained from two patients, three family members, and one legally authorized representative (LAR) (6%). Consent to continue was requested for 89 patients (89%), and full participation was granted for 87 of these patients (98%). Consent to continue was provided by 54 (98%) of the 55 patients approached. The mean number of days for family/LAR CTC was 1.1 +/-3.8 days, and 50% of the time it was obtained on the day of study enrollment. Patient CTC was obtained in an average of 13 +/- 23 days, with a median of four days. Patients treated in this protocol were more likely to have sustained penetrating trauma than the overall trauma patient population treated in these trauma centers (44% vs 21%, p = 0.002). CONCLUSIONS: Informed consent in this study of an emergent therapy most often involved the use of the consent exception and consent to continue, the latter of which occurred in a timely manner. Nearly all of those who were approached for CTC approved full participation in the study, suggesting acceptance of the process outlined in the new regulations. Patients treated in a hemorrhagic shock clinical trial may differ from the general trauma patient population.

Adult↗
PubMed1999

A proposed consent process in studies that use an exception to informed consent.

Federal regulations allow an exception to informed consent when it is not feasible to obtain informed consent in certain emergency research circumstances. A multicenter, randomized, single-blinded, normal saline procedure-controlled efficacy trial of diaspirin cross-linked hemoglobin (DCLHb) in acute traumatic hemorrhagic shock was conducted. The study intended to include 850 of the most severely injured trauma patients with hemorrhage and persistent hypoperfusion as demonstrated by vital signs suggestive of vascular collapse or a base deficit that signified prolonged hypoperfusion. It was anticipated that some patients would be unable to provide informed consent, and that identification and availability of some patients' legally authorized representatives (LARs) would be unlikely within the therapeutic window of the intervention. Each participating institution therefore developed a process to implement exception to informed consent. Each hospital's proposed process was reviewed by the institutional review board, the sponsor, the FDA, and the study's data monitoring committee chair. The goal was the development of local implementation processes by which the best interests of patients and their families could be fulfilled using prospective informed consent, the exception to informed consent, and consent to continue in emergency research, as appropriate for each individual patient. This paper describes the proposed implementation method developed for Cook County Hospital. It includes several important features, 1) prospective informed consent by the patient, when feasible; 2) the ability of the patient to decline participation, even when deemed incompetent to provide prospective informed consent; 3) prospective consent by the family/LAR, when feasible; 4) the use of a scripted abbreviated consent by the patient family/ LAR in life-threatening situations when it is possible only to briefly discuss the research being conducted; 5) independent approval for the use of the consent exception by a second physician immediately prior to patient enrollment; 6) the repeated use of consent to continue (both for the family/LAR and by the patient) when an exception to consent has been utilized; and 7) ongoing review of the informed consent process on a case-by-case basis by the institution's scientific review committee. The authors believe this proposed informed consent process maximizes the communication between investigators, patients and their proxies, and the institution's scientific review committee. Multiple mechanisms exist that allow for consent to be provided or declined, both prior to and after enrollment in the research protocol. The ongoing immediate review of the process allows for process enhancements to be made as needed.

Aspirin↗
PubMed1998

ST segment elevation and the prediction of hospital life-threatening complications: the role of right ventricular and posterior leads.

UNLABELLED: Accurate prognosis in suspected acute myocardial infarction (AMI) is essential for appropriate use of thrombolytic therapy and primary angioplasty. However, previous models may be limited because the 12-lead electrocardiogram (ECG) does not examine the right ventricular (RV) and posterior myocardium. We evaluated ST segment elevation (STSE) in posterior (V7-V9) and RV (V4R-V6R) leads to determine their predictive value for hospital life-threatening complications (HLTCs). METHOD AND RESULTS: This prospective trial of seven Midwestern hospital emergency departments (EDs) had inclusion criteria of age 35 years, chest pain suggestive of ischemia, and coronary care unit (CCU) admission. ECG leads were test positive if STSE was > 0.1 mV. Patients were positive for HLTCs if ED or inpatient hospital course included: ventricular fibrillation or tachycardia, second- or third-degree block, shock, arrest, or death. Univariate and multivariate analyses were performed to test each lead's association with HLTCs. Of 533 patients, 64.7% (345/533) had AMI and 15.8% (85/533) had HLTCs. The sensitivity of 18 leads for HLTCS was increased by 5.8%, but specificity decreased by 8.2%. ECG subgroups by STSE were associated with the following HLTC rates: inferior/+RV (32.4%); anterior (29.5%), lateral (23.1%), inferior RV (17.9%), and posterior (16.2%). V1 (odds = 3.2) and V6R (odds = 3.1) were statistically significant independent predictors. CONCLUSION: Posterior and RV leads did not increase the ECG's overall prognostic value, but in the presence of inferior STSE, were associated with low and high complication rates, respectively. Right and left precordial leads were the best predictors of HTLCs.

Adult↗
PubMed1997

Value of posterior and right ventricular leads in comparison to the standard 12-lead electrocardiogram in evaluation of ST-segment elevation in suspected acute myocardial infarction.

In this multicenter prospective trial, we studied posterior (V7 to V9) and right ventricular (V4R to V6R) leads to assess their accuracy compared with standard 12-lead electrocardiograms (ECGs) for the diagnosis of acute myocardial infarction (AMI). Patients aged >34 years with suspected AMI received posterior and right ventricular leads immediately after the initial 12-lead ECG. ST elevation of 0.1 mV in 2 leads was blindly determined and inter-rater reliability estimated. AMI was diagnosed by World Health Organization criteria. The diagnostic value of nonstandard leads was determined when 12-lead ST elevation was absent and present and multivariate stepwise regression analysis was also performed. Of 533 study patients, 64.7% (345 of 533) had AMI and 24.8% received thrombolytic therapy. Posterior and right ventricular leads increased sensitivity for AMI by 8.4% (p = 0.03) but decreased specificity by 7.0% (p = 0.06). The likelihood ratios of a positive test for 12, 12 + posterior, and 12 + right ventricular ECGs were 6.4, 5.6, and 4.5, respectively. Increased AMI rates (positive predictive values) were found when ST elevation was present on 6 nonstandard leads (69.1%), on 12 leads only (88.4%), and on both 6 and 12 leads (96.8%; p <0.001). Treatment rates with thrombolytic therapy increased in parallel with this electrocardiographic gradient. Logistic regression analysis showed that 4 leads were independently predictive of AMI (p <0.001): leads I, II, V3, V5R; V9 approached statistical significance (p = 0.055). The standard ECG is not optimal for detecting ST-segment elevation in AMI, but its accuracy is only modestly improved by the addition of posterior and right ventricular leads.

Aged↗
PubMed1996

The emergency department electrocardiogram and hospital complications in myocardial infarction patients.

OBJECTIVE: To determine whether acute myocardial infarction (AMI) patients who have negative ECGs on presentation have significantly lower complication rates than do those AMI patients who have positive ECGs on presentation. METHODS: Retrospective, cohort analysis comparing rates of hospital complications (ventricular fibrillation or tachycardia, shock, atrial arrhythmia or bradyarrhythmia with systolic blood pressure < or = 90 mm Hg, pulmonary edema) or interventions among patients with a final hospital diagnosis of AMI and an initially negative vs positive ECG. A negative ECG was normal or had nonspecific ST-segment and/or T-wave abnormalities (upright, flattened T waves; an isolated inverted T wave; ST depression < 0.1 mV; tall T waves with J-point elevation) or minor nonischemic abnormalities. Sample size was adequate to detect a 30% between-group difference in complication rates [alpha = 0.05, 1 - beta (power) = 0.80]. RESULTS: The 27 negative-ECG AMI patients differed from the 38 control patients in (mean +/- SD) age [57 +/- 12 vs 66 +/- 12 years, p < 0.01] but not in gender or history of AMI. The negative- and positive-ECG groups had similar rates of hospital complications [30% (95% CI: 13-47%) vs 42% (95% CI: 26-58%), p = 0.44] and intensive procedures [19% (95% CI: 4-34%) vs 29% (95% CI: 15-43%), p = 0.50], respectively. The negative-ECG patients with hospital complications had ECG evolution precede the event in 83% (95% CI: 69-97%) of cases; persistently negative-ECG patients had no complications [(95% CI: 0-33%), p = 0.06]. CONCLUSIONS: Negative- and positive-ECG AMI patients do not have moderate or large differences in the rates of in-hospital complications. Most negative-ECG patients who suffer complications evolve ECG changes prior to the event and such changes indicate the potential need for a higher level of care.

Adult↗
PubMed1996

Circadian rhythms and psychiatric disorders in the elderly.

This article reviews changes in circadian rhythms that have been reported to occur in the elderly psychiatric population. Data relating to circadian changes in normal aging are included where relevant. Information was obtained from: (1) a computerized MEDLINE search from 1975 to May 1996; (2) a review of bibliographies of papers obtained through the computerized search; and (3) texts on chronobiology. We could not locate any information relating to circadian rhythms and mania, anxiety, or paranoid disorders in old age. Disruption to the sleep/wake cycle, temperature, melatonin, and motor activity rhythms have been reported in dementia and depression, and disruption to some neuroendocrine and cardiovascular rhythms are reported in dementia. Disruption to circadian rhythmicity has implications for the management of dementia patients: for example, the sleep/wake and behavioral disturbances, and for the long-term management of mood disorders. A number of circadian markers have not been studied and several patient groups have received no research attention to date.

Aged↗
PubMed1995

Obstructive sleep apnea in a consecutive series of obese women.

The prevalence of obstructive sleep apnea (OSA) in women who were being seen for obesity assessment, rather than for assessment of sleep disturbance, was assessed. A consecutive series of referrals to an eating disorders clinic (n = 40) were studied in the sleep laboratory. Their body mass index (BMI) ranged from 29.4 to 66.9 kg/m2. Overnight polysomnograms were carried out, with respiration and oxygen saturation being monitored. Four women (10%), 1 postmenopausal and 3 premenopausal, had significant OSA. They tended to have a higher percentage of body fat and higher BMIs than the nonapnea group. These findings indicate that the prevalence of OSA is higher in the general population of obese women than is generally thought. They highlight the need for professionals who work with obese women to consider the possibility that some of the complaints of fatigue and tiredness may be a result of OSA.

Adult↗
PubMed1995

The relationship between hospital per diem billing and DRG reimbursement for urban trauma patients.

STUDY OBJECTIVE: To study the relationship between a trauma center per diem charges and medicare DRG reimbursement. DESIGN: Retrospective comparison of charges ($630/day, $1500/ICU day) and hypothetical DRG reimbursement using medical records ICD-9 N and P codes and version 5.0 of grouper. SETTING: An urban level I trauma center that participates in a trauma system that serves a population of three million people. PATIENT POPULATION: Trauma patients > or = 16 years old (mean age of 32 years) admitted and discharged between 1/1/88 and 9/30/88. The group was 86% male, 75% black, with a blunt mechanism of injury in 64%. The mean ICU stay was 0.9 days, and the mean total length of stay was 5.0 days. RESULTS: Total per diem charges were $8,652,159, and DRG reimbursement was $8,636,505, causing a net loss of $15,654, or 0.2% of charges. Mean charges and reimbursement did not differ for the entire group. The mean loss per patient was $8. Mean charges and reimbursement differed in penetrating trauma patients (mean loss = $138), as well as those with different lengths of stay. The correlation between charges and reimbursement was 0.42; for penetrating trauma patients, the correlation was 0.58. (p < .001) CONCLUSION: If DRG reimbursement were provided for all admitted trauma patients, the amount would equal per diem rates. Trauma centers with similar patients and lengths of stay can use these per diem rates to estimate DRG reimbursement.

Adolescent↗
PubMed1995

The relationship between hospital per diem billing and DRG reimbursement for urban trauma patients.

STUDY OBJECTIVE: To study the relationship between a trauma center per diem charges and medicare DRG reimbursement. DESIGN: Retrospective comparison of charges ($630/day, $1500/ICU day) and hypothetical DRG reimbursement using medical records ICD-9 N and P codes and version 5.0 of grouper. SETTING: An urban level I trauma center that participates in a trauma system that serves a population of 3 million people. PATIENT POPULATION: Trauma patients > or = 16 years old (mean age of 32 years) admitted and discharged between 1/1/88 and 9/30/88. The group was 85% male, 75% black, with a blunt mechanism of injury in 64%. The mean ICU stay was 0.9 days, and the mean total length of stay was 5.0 days. RESULTS: Total per diem charges were $8,652.159, and DRG reimbursement was $8,636,505, causing a net loss of $15,654, or 0.2% of charges. Mean charges and reimbursement did not differ for the entire group. The mean loss per patient was "8. Mean charges and reimbursement differed in penetrating trauma patients (mean loss = $138), as well as those with different lengths of stay. The correlation between charges and reimbursement was 0.42; for penetrating trauma patients, the correlation was 0.58 (p < .001). CONCLUSION: If DRG reimbursement were provided for all admitted trauma patients, the amount would equal per diem rates. Trauma centers with similar patients and lengths of stay can use these per diem rates to estimate DRG reimbursement.

Adolescent↗
PubMed1995

Electroencephalography and single photon emission computed tomography in dementia: a comparative study.

A series of elderly patients with dementia of Alzheimer type (AD), multi-infarct dementia (MID) and functional (non-organic) psychiatric illness (major depressive disorder) were selected according to DSM-III-R criteria and received: a battery of cognitive tests, EEG and Single Photon Emission Computed Tomography (SPECT) using 99mTc HMPAO. The EEG and SPECT scans were examined independently of the clinical data. The former were divided into two abnormal categories, those showing AD type change and vascular change respectively, and a normal group. The SPECT scans were classified as follows: a SPECT rCBF pattern showing bilateral temporoparietal perfusion deficits (AD type); those showing single focal perfusion deficits or multiple areas of low perfusion in the cerebral cortex suggestive of ischaemic change (vascular type SPECT picture); a mixed AD/MID pattern; and those with normal scan findings. There were significant associations between clinical diagnosis, EEG rating and SPECT rCBF pattern, approximately three-quarters of AD and MID patients having the predicted EEG and scan changes. Normal EEG recordings were more common in the MID patients. The two tests agreed in about two-thirds of cases, with no consistent pattern apparent in the cases with divergent findings. Each test misclassified a minority of dementia patients, but in only one patient were both investigations normal. Almost half of the so called 'functionally ill' patients had abnormal rCBF changes, showing mainly vascular changes while one-fifth had abnormal EEGs.

Aged↗
PubMed1995

Skin and mucous membrane contacts with blood during surgical procedures: risk and prevention.

OBJECTIVE: To study the epidemiology and preventability of blood contact with skin and mucous membranes during surgical procedures. DESIGN: Observers present at 1,382 surgical procedures recorded information about the procedure, the personnel present, and the contacts that occurred. SETTING: Four US teaching hospitals during 1990. PARTICIPANTS: Operating room personnel in five surgical specialties. MAIN OUTCOME MEASURES: Numbers and circumstances of contact between the patient's blood (or other infective fluids) and surgical personnel's mucous membranes (mucous membrane contacts) or skin (skin contacts, excluding percutaneous injuries). RESULTS: A total of 1,069 skin (including 620 hand, 258 body, and 172 face) and 32 mucous membrane (all affecting eyes) contacts were observed. Surgeons sustained most contacts (19% had > or = 1 skin contact and 0.5% had > or = 1 mucous membrane-eye contact). Hand contacts were 72% lower among surgeons who double gloved, and face contacts were prevented reliably by face shields. Mucous membrane-eye contacts were significantly less frequent in surgeons wearing eyeglasses and were absent in surgeons wearing goggles or face shields. Among surgeons, risk factors for skin contact depended on the area of contact: hand contacts were associated most closely with procedure duration (adjusted odds ratio [OR], 9.4; > or = 4 versus < 1 hour); body contacts (arms, legs, and torso) with estimated blood losses (adjusted OR, 8.4; > or = 1,000 versus < 100 mL); and face contacts, with orthopedic service (adjusted OR, 7.5 compared with general surgery). CONCLUSION: Skin and mucous membrane contacts are preventable by appropriate barrier precautions, yet occur commonly during surgery. Surgeons who perform procedures similar to those included in this study should strongly consider double gloving, changing gloves routinely during surgery, or both.

Adult↗
PubMed1995

Human immunodeficiency virus and hepatitis B virus seroprevalence in an urban trauma population.

OBJECTIVE: To determine the seroprevalence of the human immunodeficiency virus (HIV) and the hepatitis B virus (HBV) in patients of an urban level I trauma center. DESIGN: Prospective, blinded point prevalence study of serum HIV and HBV antibody and antigen. SETTING: An urban level I trauma center that participates in a trauma system serving three million people. PATIENTS: The study included 994 (94.8%) of 1049 consecutive trauma service patients treated between June 6, 1988 and September 22, 1988. The patients were 82.2% male and 73.1% black, with a mean age of 28.8 +/- 12.3 years. Blunt trauma was seen in 65.4% of patients, 5.2% were in shock, and 96.2% survived their trauma. MAIN OUTCOME MEASURES: HIV and HBV seroprevalence, using both antibody and antigen testing. RESULTS: HIV infection was seen in 43 patients (4.3%); 41 (95.3%) were HIV Ab+ and two (4.7%) were HIV Ab-/HIV Ag+. Infection with the HBsAg was seen in 31 patients (3.1%). Infection with either virus was seen in 70 patients (7%); four patients (0.4%) were infectious for both viruses. Infection was related to age 20 to 49 years, i.v. drug use, a hepatitis or sexually transmitted disease history, prior HIV testing, shock, and death (p < 0.05). Penetrating trauma was not predictive of infection. In a logistic regression model, IV drug use was the single significant predictor of infection (p < 0.05). CONCLUSIONS: Young urban trauma patients, because of drug-related intentional violence, are 15.3 to 17.6 times more likely to be HIV infected and 3.9 to 7.9 times more likely to be infectious for HIV or HBV than the trauma population overall. The 12 to 21% infection rates in critically injured patients who require shock resuscitation and/or die reinforces the need for mandated universal precautions and for clear policies which govern the performance of procedures by physicians in training. Primary HIV infection in critically injured patients may worsen their outcome and may adversely affect the exposed health care worker. Emergency departments and trauma units should develop a referral system to HIV primary care services (HIV counselling and testing) for high risk patients and for adversely exposed health care workers.

AIDS Serodiagnosis↗
PubMed1995

The incidence and outcome of penetrating and blunt trauma in central Bosnia: the Nova Bila Hospital for War Wounded.

OBJECTIVES: Since February 1992, the civil war in the former Yugoslavia has left over 140,000 people dead. This study describes the injuries and outcome of patients treated at a provisional war hospital in Bosnia and compares mortality rates with war hospitals from prior armed conflicts. DESIGN: This is a retrospective review of 1,703 trauma patients treated between March 1993 and October 1993 at the Nova Bila War Hospital. RESULTS: In Bosnia, the overall mortality was 10.3% (5.0% dead on arrival and 5.3% inhospital deaths). Head injuries, seen in 19.2% of patients, had a mortality of 23.8%. The odds of sustaining a gunshot wound were 2.8 times greater in Bosnia than in Vietnam (p < 0.05). The odds of sustaining a head injury were 1.1 to 1.6 times greater in Bosnia than in Lebanon and Afghanistan (p < 0.05). The overall mortality odds in Bosnia were 3.1 times greater than in Vietnam (p < 0.001), but were equal to those in Lebanon and Afghanistan. CONCLUSIONS: Despite limited personnel and supplies, the inhospital mortality rate was comparable with those found in other war hospitals. The Nova Bila Hospital represents a unique response to the great medical need brought about by the continued fighting in the former Yugoslavia.

Adolescent↗