To investigate surgical blood usage during the siege of Sarajevo.

Data on blood usage and pre-transfusion hematocrit (Hct) values from blood transfusion request forms in 250 wartime emergency surgical procedures during August through October 1992 (experimental group), and in 146 peacetime elective surgical procedures (control group) during April through June 1991 at the State Hospital of Sarajevo, were reviewed.

The mean number of blood units transfused per patient (blood usage rate) was 1.13 in the experimental group versus 2.56 in the control group (p <0.001). During the war, for blood conservation, normovolemic hemodilution was practiced widely. A significantly lower mean pre-transfusion Hct value of 0.21 was observed in the experimental group versus 0.27 in the control group (p <0.001).

Blood-usage rate was lower during emergency surgical procedures in war than during elective surgical procedures in peacetime without apparent adverse patient outcome. This decrease in blood-usage rate in the face of increased numbers of trauma victims was the result of a planned blood-conservation program which included: stringent blood-usage criteria, and widespread implementation of casualty resuscitation using normovolemic hemodilution with colloid and crystalloid plasma substitutes.

Anecdotal observations about prehospital emergency medical care in major natural and human-made disasters, such as earthquakes, have suggested that some injured victims survive the initial impact, but eventually die because of a delay in the application of life-saving medical therapy.

A multidisciplinary, retrospective structured interview methodology to investigate injury risk factors, and causes and circumstances of prehospital death after major disasters was developed. In this study, a team of United States researchers and Costa Rican health officials conducted a survey of lay survivors and health care professionals who participated in the emergency medical response to the earthquake in Costa Rica on 22 April 1991.

Fifty-four deaths occurred prior to hospitatization (crude death rate = 0.4/1,000 population). Seventeen percent of these deaths (9/54) were of casualties who survived the initial impact but died at the scene or during transport. Twenty-two percent (2/9) were judged preventable if earlier emergency medical care had been available. Most injuries and deaths occurred in victims who were inside wooden buildings (p <.O1) as opposed to other building types or were pinned by rubble from building collapse. Autopsies performed on a sample of victims showed crush injury to be the predominant cause of death.

A substantial proportion of earthquake mortality in Costa Rica was protracted. Crush injury was the principal mechanism of injury and cause of death. The rapid institution of enhanced prehospital emergency medical services may be associated with a significant life- saving potential in these events.

To study hospital disaster operations following a major United States disaster.

Researchers interviewed all 51 hospital administrators and 49 of 51 emergency department (ED) charge nurses and emergency physicians who were on duty at the study hospitals during the 13-hour period immediately following the 1989 Loma Prieta earthquake.

The 51 acute-care hospitals in the six northern California counties most affected by the Loma Prieta earthquake.

Questionnaires and inperson interviews.

The most frequently noted problem was lack of communications within and among organizations. Hospitals received inadequate information about the disaster from local governmental agencies. Forty-three percent of hospitals had inadequate back-up power configurations, and five hospitals sustained total back-up generator failures. Twenty hospitals performed partial evacuations.

The Loma Prieta earthquake did not cause total disruption of hospital services. Hospitals need to work with local governmental agencies and internal hospital departments to improve disaster communications.

Motor vehicle injuries are a major public health problem. They are a primary cause of: 1) death and injury in the United States; and 2) result in a substantial loss of productive life. These injuries and fatalities have serious social and economic consequences for the injured individual, their families, and society. This report focuses on the portion of health care expense borne by the public and the tax revenue implications of these injuries and fatalities.

The relationship between motor vehicle injuries and fatalities, health care costs, and income taxes was analyzed for four situations: 1) 1990 baseline; 2) achievement of modest goals for safety improvements; 3) population growth with constant injury and fatality rates; and 4) the effect of higher injury and fatality rates. Total health care costs, publicly funded health care costs, lost income tax revenue, and increased public assistance were estimated at the [U.S.] federal level, and at the state and local level.

Study of these relationships indicate that: 1) the lifetime economic cost of motor vehicle injuries, fatalities, and property damage that occurred in 1990 is $137.5 billion. American taxpayers will pay $11.4 billion of that total to cover publicly funded health care ($3.7 billion), reduced income tax revenue ($6.1 billion), and increased public assistance expenses ($1.6 billion); 2) the lifetime economic cost of alcohol-related, motor vehicle injuries, fatalities, and property damage that occurred in 1990 was $46.1 billion. Of this, the American taxpayer will pay $1.4 billion to cover publicly funded health care and $3.8 billion to cover reduced income tax revenue and increased public assistance; 3) reducing the percentage of the alcohol-related portion of these fatalities from 45% to 43% (1,200 lives saved), and alcohol-related injuries by a proportionate amount, would save American taxpayers $73 million in publicly funded health care and $208 million in income taxes and public assistance; 4) by increasing observed safety-belt usage in passenger cars from 62% to 75%, (1,700 lives saved plus a proportionate reduction in injuries), publicly funded health care costs would be reduced by $180 million, and $328 million would be saved in the combination of increased income tax revenues and reduced public assistance; 5) Further reductions in publicly funded health care, increases in income tax revenues, and reductions in public assistance are possible as a result of reasonable gains in other areas, such as increased safety-belt usage in light trucks, increased usage of motorcycle helmets, increased correct usage of child safety seats, and reducing the number of speeding drivers; 6) if injury and fatality rates remain at the 1992 level, population increases alone would result in 3,300 more fatalities in the year 2000. Economic costs from these fatalities and a proportionate increase in injuries would increase by an estimated $7.4 billion, including a $277 million increase in publicly funded health care costs, and $573 million in reduced income tax revenue and increased public assistance; and 7) if injury and fatality rates increase from the 1992 level, injuries, fatalities, and costs will increase. In one scenario, with 5,800 more fatalities than the population growth scenario, economic costs would increase by $13 billion, including a $350 million increase in publicly funded health care, and an additional $1 billion in taxes to cover lost income tax revenue and increased public assistance.

It is obvious that inaction is a costly alternative and that anticipated population gains will require further reductions in injury and fatality rates just to maintain current injury and fatality rates. Fortunately, countermeasures are to be available that can accomplish this. Lack of vigilance that would result in deterioration of safety levels would be even more costly.

Decisions to send an ambulance with or without lights and siren are made every day. While travel with lights and siren is presumed to have relatively more risk associated with it than travel without, little epidemiologic analysis has been conducted to compare the two modes of travel or to characterize collisions in general.

To characterize ambulance collisions and assess the risk of traveling with lights and siren in an urban 9-1-1 environment.

Retrospective analysis of all consecutive ambulance collisions of the Paramedic Division of the San Francisco Department of Public Health during a 27-month period.

The overall collision rate for lights and siren (LS) travel was higher than that for non-lights and siren travel, although the difference was not statistically significant (45.9 collisions per 100,000 LS patient travels, 95% confidence limits 29.7, 62.1, versus 27.0/100,000 for non-LS travel, 95% confidence limits 18.3, 35.7). However, the rates of resulting injuries displayed a statistically significant difference (22.2 injuries per 100,00 LS patient travel, 95% confidence limits 11.0, 33.5, versus 1.5/100,000 for non-LS travel, 95% confidence limits −0.6, 3.5). While the majority of collisions (60.0%) occurred during patient-related travel, 35.6% occurred while the ambulance was available awaiting assignment, and 4.4% in a hospital parking lot. The majority of collisions were due to inattention, failure of on-coming traffic to yield, or unsafe parking; unsafe speed was an infrequent cause. Most crashes occurred during daylight, in dry weather, and involved another vehicle.

There is some elevated risk for collision and added injury during lights and siren travel compared to travel without LS. The causes for these collisions suggest that interventions designed to improve driver skills and increase citizen awareness of an approaching ambulance could help reduce the number of collisions.

The safe and timely provision of blood is of crucial importance in the prevention and mitigation of morbidity and mortality due to trauma. The use of blood in the treatment of war casualties, soldiers as well as civilians, was analyzed retrospectively and the impact of massive blood transfusion on blood banking services and reserves of blood during the war in Sarajevo was assessed.

A retrospective analysis of 3,215 war casualties (1,815 civilians plus 1,400 military) who arrived to the casualty reception center of the State Hospital of Sarajevo during the period 11 May through 31 October 1992 was performed. Blood usage was reviewed in three stages: within 24 hours (h) of admission, after seven days of hospitalization, and after 30 days of hospitalization. The types of injury, survival rate, and blood-usage rate in a sample of 37 war casualties who required massive blood transfusions (MBT) during the period 11 May through 31 December 1992 was examined.

The civilian casualty rate in this series of patients was 56.5%. A total of 1,217/3,215 (37.9%) casualties were hospitalized. In this study, 16% (504/3,215) of total number of persons wounded received blood transfusion. Of these patients, 504/1,217 (41.4%) were transfused. A total of 971.1 liters of blood were transfused through 31 October 1992; 68% within 24 h of admission, 91% within the first seven days, and 100% within the first 30 days. From a total of 37 MBT recipients, 36 (97%) were injured by firearms. Survival rate among MBT patients was 30%. The MBT recipients comprised 2% of total hospitalized patients and 6% of total number of patients transfused. The amount of blood needed during episodes of MBT was 15% of total blood used through 31 December 1992.

Based on these data, prospective requirements for blood usage should take into account casualty triage, as follows: for each casualty transported to the hospital, hospitalized, or transfused, 0.302, 0.796, and 1.912 liters of blood respectively, will be needed for the first 30 days of treatment. Recipients of massive blood transfusions are a significant drain on blood reserves in war. This experience can be utilized in the development of revised guidelines for blood usage for an entire population affected by war.

The aim of this study was to compare the patient care measures provided by paramedics according to standing orders versus measures ordered by direct [on-line] medical command in order to determine the types and frequency of medical command orders.

Prospective identification of patient care measures done as part of a prehospital quality assurance program.

An urban paramedic service in the northeast United States with direct medical command from three local hospitals.

One thousand eight paramedic reports from October 1992 through March 1993.

All patient care interventions recorded as done by standing orders or by direct medical command orders. Errors in patient care were determined by the same criteria as in the prior two studies of the same system.

Direct medical command gave orders in 143/1,008 (14.2%) cases. Paramedics performed 2,453/2,624 (93.5%) of the total patient care interventions using standing orders. In 61 cases (6.1 %), medical command ordered a potentially beneficial intervention not specified by standing orders or not done by the paramedic. 21/171 (12.3%) command orders were for additional doses of epinephrine or atropine in cardiac arrest cases (where the initial doses had been given under standing orders), and 59/171 (34.5%) were for interventions already mandated or permitted by standing orders. The paramedic error rate was 0.6%, and the medical command error rate was 1.8% (unchanged form the prior study of the same standing-orders system).

Direct medical command gave orders in 14% of cases in this standing-orders system, but 35% of command orders only reiterated the standing orders. More selective and reduced uses of on-line command could be done in this system with no change in the types or numbers of patient care interventions performed.

To analyze the characteristics of fatal ambulance crashes to assist emergency medical services (EMS) directors in objectively developing their EMS system's policy governing ambulance operations.

No difference exists between the characteristics of fatal ambulance crashes during emergency and nonemergency use.

Retrospective, cross-sectional, comparative analysis of ambulance crashes resulting in fatalities reported to the Fatal Accident Reporting System (FARS) from 1987 to 1990.

Twenty variables, representing characteristics of fatal ambulance crashes, were selected from the National Highway Traffic Administration FARS Codebook and were evaluated using tests of significance for categorical data grouped by emergency use and nonemergency use. Crash variable categories examined included demographics, accident configuration, accident severity, vehicle description, and ambulance operator action.

During the four-year study period, 109 fatal ambulance crashes occurred producing 126 deaths. Four states, New York, Michigan, California, and North Carolina, accounted for 37.5% of all fatal crashes. Seventy-five fatal crashes (69%) occurred during emergency use (EU) and 34 fatal crashes (31%) occurred during nonemergency use (NEU). The total number of fatal crashes varied in a downward trend (1987:32; 1988:24; 1989:28; 1990:25). The number of fatal EU crashes also varied in a downward trend (1987:28; 1988:16; 1989:19; 1990:12), while the number of fatal NEU crashes increased each year [1987:4; 1988:8; 1989:9; 1990:13](p = .016). Most EU fatal crashes occurred between 1200 h and 1800 h (p = .009). Most NEU fatal crashes occurred during times when light conditions were poor (p = .003). When a violation was charged to the ambulance driver (17 cited), the vehicle was more likely to be in EU (p = .056). No statistically significant differences between EU and NEU were identified by: 1) day of week; 2) season; 3) atmospheric conditions; 4) roadway surface type; 5) roadway surface condition; 6) speed limit; 7) roadway alignment; 8) relationship to junction; 9) manner of collision; 10) year manufactured; 11) vehicle role; 12) vehicle maneuver; 13) manner leaving scene; 14) extent of deformation; 15) violations charged; or 16) number of persons killed in accident.

Few characteristics differentiate between fatal ambulance crashes during EU and NEU. The difference between EU and NEU were statistically significant in only three out of the 20 variables examined: 1) year occurred; 2) time of day; and 3) light condition. These data provide few objective measures that may be used to develop ambulance operation policies to decrease fatal ambulance crashes.

Supplying an adequate amount of drinking water to a population is a complex problem that becomes an extremely difficult task in war conditions. In this paper, several simple methods for obtaining individual supplies of drinking water by filtration of atmospheric water with common household items are reported.

Samples of atmospheric water (rain and snow) were collected, filtered, and analyzed for bacteriological and chemical content. The ability of commonly available household materials (newspaper, filter paper, gauze, cotton, and white cotton cloth) to filter water from the environmental sources was compared.

According to chemical and biological analysis, the best results were obtained by filtering melted snow from the ground through white cotton cloth.

Atmospheric water collected during war or in extreme shortage conditions can be purified with simple improvised filtering techniques and, if chlorinated, used as an emergency potable water source.

Use of an oxygen-powered demand-valve to ventilate through an endotracheal tube is considered inappropriate due to concern regarding excessive airway pressure.

It was hypothesized that ventilation through an endotracheal tube using a bag-valve (BV) device and the recently modified demand-valve (DV) would produce similar tidal volumes (Vt), minute ventilation (MV), and peak airway pressures (PAP).

This is a prospective, randomized vitro experimental model. Subjects were blinded to volume and pressure gauges. Thirty-nine EMTs (mean age 27 years with mean experience five years) volunteered to ventilate a mechanical test lung through an endotracheal tube for 10 minutes. Each subject was randomized to BV or DV and to either normal (0.1 L/cm H2O) or poor (0.04 L/cm H2O) lung compliance. This DV delivers set flow of 40 L/min at maximum 50±5 cm H2O. Subjects were instructed to use their “usual” technique for an average size adult in respiratory arrest with normal heart rate and blood pressure. The Vt and PAP were recorded for each breat; the MV and maximum PAP (PAP-max) for each minute was noted. Data were analyzed using repeated measures ANOVA and Tuke multiple comparisons with alpha set at 0.05.

Overall average tidal volumes and minute ventilations were acceptable with both ventilalory devices at both normal and poor compliance for the first, fifth, and 10th minute of continuous ventilation. Average airway pressures and peak airway pressures during the first, fifth, and 10th minute of ventilation all were significantly higher with those of the bag-valve than with the use of the demandvalve at both normal and poor compliance.

In this model, ventilation with bag-valve and demand-valve both provided more than adequate Vt and MV; values wer similar except for higher Vt with BV at normal compliance. However, DV yielded significantly lower PAP and PAPmax at both poor and normal compliance. These findings need corrobration in an in vivo model, but suggest that with proper training, demand-valve ventilation through an endotracheal tube may be preferable.

In war, abdominal injuries constitute a significant proportion of the total injuries. These injuries are associated with high mortality and their treatment poses dianostic, surgical, and therapeutic dilemmas. This article presents the epidemiology of abdominal war injuries during the siege of Sarajevo, and briefly describes the surgical techniques and therapeutic practices used in their treatment.

A retrospective medical record review was performed of 273 war casualties with trauma to visceral and vascular structures in the abdomen inflicted during a 7.5 month period in 1992.

Most patients underwent exploratory laparotomy. Six percent had negative laparotomies, and there were no deaths in this group. In 18.3%, injuries were limited to one organ system, while 81.7% sustained combined injuries to multiple-organ systems. The crude mortality rate was 26.0%. Mortality rate excluding deaths within 24 hours of injury was 10.3%. Injuries were caused by metal fragments from artillery shrapnel, mortar and contact mines, or hand grenades. Because of a shortage of colostomy bags, resections of the colon with primary end-to-end anastomoses rather than colostomy were performed in 72% of the cases.

Mortality was highest in those victims with four or more injured organ systems (81.3%) or with major vascular injuries (64.7%). The primary cause of death within the first 24 hours was prolonged hemorrhagic shock.