Cost-effectiveness of tissue plasminogen activator for acute ischemic stroke. NINDS rt-PA Stroke Study Group

Tissue plasminogen activator (tPA) has been shown to improve 3-month outcome in stroke patients treated within 3 hours of symptom onset. The costs associated with this new treatment will be a factor in determining the extent of its utilization. Data from the NINDS rt-PA Stroke Trial and the medical literature were used to estimate the health and economic outcomes associated with using tPA in acute stroke patients. A Markov model was developed to estimate the costs per 1,000 patients eligible for treatment with tPA compared with the costs per 1,000 untreated patients. One-way and multiway sensitivity analyses (using Monte Carlo simulation) were performed to estimate the overall uncertainty of the model results. In the NINDS rt-PA Stroke Trial, the average length of stay was significantly shorter in tPA-treated patients than in placebo-treated patients (10.9 versus 12.4 days; p = 0.02) and more tPA patients were discharged to home than to inpatient rehabilitation or a nursing home (48% versus 36%; p = 0.002). The Markov model estimated an increase in hospitalization costs of $1.7 million and a decrease in rehabilitation costs of $1.4 million and nursing home cost of $4.8 million per 1,000 eligible treated patients for a health care system that includes acute through long-term care facilities. Multiway sensitivity analysis revealed a greater than 90% probability of cost savings. The estimated impact on long-term health outcomes was 564 (3 to 850) quality-adjusted life-years saved over 30 years of the model per 1,000 patients. Treating acute ischemic stroke patients with tPA within 3 hours of symptom onset improves functional outcome at 3 months and is likely to result in a net cost savings to the health care system.     

Plasma levels of activated protein C in healthy subjects and patients with previous venous thromboembolism: relationships with plasma homocysteine levels

BACKGROUND/PURPOSE: In the pediatric population, appendicitis remains the most common surgical emergency encountered. The purpose of this study was to determine the impact of an evidence-based clinical pathway for acute appendicitis on patient care as well as hospital and home care costs at the authors' pediatric institution. METHODS: A prospective evaluation was conducted of an appendicitis clinical pathway (June 1996 through November 1996) compared with historical control patients (June 1994 through November 1994) not cared for by the pathway. RESULTS: Data (average +/- SD) for 120 pathway (P) patients were compared with 122 control (C) patients. Age (11.5 +/- 3.6 years for C v 11.2 +/- 3.9 years for P), rates of negative appendectomy (12.3% for C v 9.2% for P) and perforation (26.2% for C v 18.3% for P) were similar. Pathway patients with nonperforated appendicitis were more often discharged from the hospital within 24 hours (48% for C v 67% for P; P = .014) with lower hospital costs ($4,095 +/- $1,280 for C v $3,638 +/- $1,633 for P; P = .001). Pathway patients with perforated appendicitis had shorter hospitalization (185.2 +/- 59 hours for C v 113 +/- 44 hours for P; P = .0001) and lower hospital costs ($11,175 +/- $3,893 for C v $7,823 +/- $2,366 for P; P = .0001). CONCLUSION: An evidence-based appendicitis pathway decreased duration of hospitalization and cost without adversely affecting diagnosis or therapy. Clinical pathways for surgical diagnoses may prove useful as a means to minimize costs without compromising patient care.     

Research without billing data. Econometric estimation of patient-specific costs

OBJECTIVES: This article describes a method for computing the cost of care provided to individual patients in health care systems that do not routinely generate billing data, but gather information on patient utilization and total facility costs. METHODS: Aggregate data on cost and utilization were used to estimate how costs vary with characteristics of patients and facilities of the US Department of Veterans Affairs. A set of cost functions was estimated, taking advantage of the department-level organization of the data. Casemix measures were used to determine the costs of acute hospital and long-term care. RESULTS: Hospitalization for medical conditions cost an average of $5,642 per US Health Care Financing Administration diagnosis-related group weight; surgical hospitalizations cost $11,836. Nursing home care cost $197.33 per day, intermediate care cost $280.66 per day, psychiatric care cost $307.33 per day, and domiciliary care cost $111.84 per day. Outpatient visits cost an average of $90.36. These estimates include the cost of physician services. CONCLUSIONS: The econometric method presented here accounts for variation in resource use caused by casemix that is not reflected in length of stay and for the effects of medical education, research, facility size, and wage rates. Data on non-Veteran's Affairs hospital stays suggest that the method accounts for 40% of the variation in acute hospital care costs and is superior to cost estimates based on length of stay or diagnosis-related group weight alone.     

Use of antipsychotic drugs in a long-term care institution. Experiences with implementing clinical guidelines]

Anti-psychotic drugs (neuroleptics) are useful for treating psychoses. However, non-psychotic patients, particularly patients with a deviant behaviour pattern, are often also treated with anti-psychotic drugs. The drugs may induce serious side-effects and should only be used on strict indications and at the lowest possible dosage. In a nursing home for deaf people with additional handicaps we introduced clinical guidelines for the use of anti-psychotic drugs and recorded their use during a two-year period. We found there was a reduction in the number of patients taking anti-psychotic drugs (from 32/54 to 26/54, p = 0.03), as well as a decrease in dosage per user (from median 2.4 mg to 1.7 mg equivalents of haloperidol, p = 0.05). Both the number of patients who were given depot injections and the number of different anti-psychotic drugs per patient were reduced. We conclude that it is possible to reduce the use of anti-psychotic drugs in institutions where long-term care is provided for disabled people.     

How access to long-term care affects home health transfers

This study examines the determinants of home health use after hospitalization for acute illness for eleven diagnosis-related groups (DRGs) in 1985, drawing on data from four primary sources: Medicare hospital bills, Medicare home health bills, the Medicare and Medicaid Automated Certification System files, and the American Hospital Association Survey. Separate Tobit models are estimated for each DRG. The analysis shows that transfers to home health care are heavily influenced by the hospital's long-term care arrangement and by conditions in local nursing home and home health care markets. Especially important is whether a hospital has its own long-term care unit, swing beds, or both, and whether nursing home beds are available in the local area. Patients discharged from hospitals are more likely to use home health care in areas with a low supply of nursing home beds and low Medicaid reimbursement levels for skilled nursing facilities. The results of this study have implications for proposals to extend Medicare's Prospective Payment System for hospital services to include postacute care. Proponents of a "bundled payment" that encompasses both acute and postacute services argue that the current system leads to inefficiencies and inequities. This analysis points to systematic relationships between home health and nursing home services, which should be factored into the development of a bundled payment policy.     

Acute and long-term outcomes of extremely injured blunt trauma victims

Long-term outcomes after blunt trauma remain poorly defined. The purpose of this study was to document such outcomes in extremely injured adults (Injury Severity Score > or = 50). From April 1990 to June 1993, 76 patients (5% of all trauma victims) had an ISS > or = 50 at a single trauma center. Thirty-five (46%) survived to hospital discharge. The mean duration of hospital stay was longer for survivors than for nonsurvivors (92 days versus 16 days, p < 0.001). Of the 35 survivors, 26% were discharged directly home, 60% to a rehabilitation hospital, 8% to a chronic care facility, and 6% to an acute care hospital. After a mean follow-up of 27 months, 6% had died, 9% refused participation, and the remaining 30 patients (91% of long-term survivors) demonstrated significant residual disabilities in physical, emotional, and mental health status. We suggest that extremely injured patients comprise a small proportion of blunt trauma victims, consume substantial acute care hospital resources, often survive, and yet frequently have residual disability. A reduction in this long-term disability may represent the greatest challenge in modern trauma care.     

Appendicitis in children in the managed care era

Acute appendicitis is the most common condition requiring emergency operation in children. Late appendicitis is still a major source of morbidity and potential mortality. It has been suggested that managed care programs are responsible for a delay in surgical referral and consequently an increased risk of morbidity and mortality. In light of the increasing use of managed care, the authors reviewed their experience with pediatric acute appendicitis in managed care and indemnity insurance patients. The charts of all pediatric appendectomy patients treated between January 1990 and March 1995 were reviewed. Payor status, surgical and pathological findings, hospital course, and follow-up findings were documented. If the operative note or the pathology report described the appendix as gangrenous or perforated, the case was considered to be late appendicitis. Group I patients had traditional indemnity insurance; group II patients were in our institution's managed care plan. One hundred two patients were identified (28 in group 1, 74 in group II). Late appendicits was found less often in the managed care group (21.6% v 42.9%; P < .01). This resulted in a lower rate of major complications (1.4% v 3.6%) and a lower overall complication rate (2.7% v 7.1%). Group II also had a shorter hospital stay (2.6 days v 4.5 days; (P < .01) and lower average hospital charges ($6,507 v $8,754 (P < .01). These results do not demonstrate any adverse affect on outcome for children with acute appendicitis who have a managed care plan. In fact, the incidence of late appendicitis among these patients was half of that of the indemnity-insured patients. The lower risk of late appendicitis resulted in a shorter length of stay and lower hospital charges. These results suggest that managed care programs can provide quality care along with a significant reduction in costs; no delay in appropriate surgical referral was demonstrated.     

A comparison between emergency diagnostic and treatment unit and inpatient care in the management of acute asthma

BACKGROUND: Emergency diagnostic and treatment units (EDTUs) may provide an alternative to hospitalization for patients with reversible diseases, such as asthma, who fail to adequately respond to emergency department therapy. OBJECTIVE: To evaluate the medical and cost-effectiveness, patient satisfaction, and quality of life of patients receiving EDTU care for acute asthma compared with inpatient care. METHODS: A prospective, randomized clinical trial performed at 2 urban public hospitals enrolled patients with acute asthma (age range, 18-55 years) not meeting discharge criteria after 3 hours of emergency department therapy. Patients were treated with inhaled adrenergic agonists and steroids in an EDTU for up to 9 hours after randomization or with routine therapy in a hospital ward. Patients were followed up for 8 weeks. MAIN OUTCOME MEASURES: Discharge rate from the EDTU, length of stay, relapse rates, days missed from work or school, days incapacitated during waking hours, symptom-free days and nights, nocturnal awakenings, direct medical costs, patients satisfaction, and patient quality of life. RESULTS: The study consisted of 222 patients with asthma. Sixty-five patients (59%) treated in an EDTU were discharged home; the remainder were admitted to the hospital. There were no differences during the follow-up period in relapse rates (P = .74) or in any other morbidities between the EDTU and inpatient groups. There were significant differences in the length of stay, patient satisfaction, and quality of life favoring EDTU care. The mean (+/-SD) cost per patient in the EDTU group was $1202.79 +/- $1343.96, compared with $2247.32 +/- $1110.18 for the control group (P < .001). CONCLUSIONS: Treatment of selected patients with asthma in an EDTU results in the safe discharge of most such patients. This study suggests that quality gains and cost-effective measures can be achieved by the use of such units.     

Can home care services achieve cost savings in long-term care for older people?

OBJECTIVE: To determine whether efficient allocation of home care services can produce net long-term care cost savings. METHODS: Hazard function analysis and nonlinear mathematical programming. RESULTS: Optimal allocation of home care services resulted in a 10% net reduction in overall long-term care costs for the frail older population served by the National Long-Term Care (Channeling) Demonstration, in contrast to the 12% net cost increase produced by the demonstration intervention itself. DISCUSSION: Our findings suggest that the long-sought goal of overall cost-neutrality or even cost-savings through reducing nursing home use sufficiently to more than offset home care costs is technically feasible, but requires tighter targeting of services and a more medically oriented service mix than major home care demonstrations have implemented to date.     

Risk factors for higher cost in congenital heart operations

BACKGROUND: For many congenital heart defects, hospital mortality is no longer a sensitive parameter by which to measure outcome. Although hospital survival rates are now excellent for a wide variety of lesions, many patients require expensive and extensive hospital-based services during the perioperative period to enable their convalescence. These services can substantially increase the cost of care delivery. In today's managed care environment, it would be useful if risk factors for higher cost could be identified preoperatively so that appropriate resources could be made available for the care of these patients. The focus of this retrospective investigation is to determine if risk factors for high cost for repair of congenital heart defects can be identified. METHODS: We assessed financial risk by tracking actual hospital costs (not charges) for 144 patients undergoing repair of atrial septal defect (58 patients), ventricular septal defect (48 patients), atrioventricular canals (14 patients), or tetralogy of Fallot (24 patients) at Duke University Medical Center between July 1, 1992, and September 15, 1995. Furthermore, we were able to identify where the costs occurred within the hospital. Financial risk was defined as a large (> 60% of mean costs) standard deviation, which indicated unpredictability and variability in the treatment for a group of patients. RESULTS: Cost for atrial septal defect repair was predictably consistent (low standard deviation) and was related to hospital length of stay. There were factors, however, for ventricular septal defect, atrioventricular canal, and tetralogy of Fallot repair that are identifiable preoperatively that predict low- and high-risk groups using cost as an outcome parameter. Patients undergoing ventricular septal defect repair who were younger than 6 months of age at the time of repair, who required preoperative hospital stays of longer than 7 days before surgical repair, or who had Down's syndrome had a less predictable cost picture than patients undergoing ventricular septal defect repair who were older than 2 years, who had short (< 4 days) preoperative hospitalization, or who did not have Down's syndrome ($48,252 +/- $42,539 versus $15,819 +/- $7,219; p = 0.008). Patients with atrioventricular canals who had long preoperative hospitalization (> 7 days), usually due to pneumonia (respiratory syncytial virus) with preoperative mechanical ventilation had significantly higher cost than patients with atrioventricular canals who underwent elective repair with short preoperative hospitalization ($83,324 +/- $60,138 versus $26,904 +/- $5,384; p = 0.05). Patients with tetralogy of Fallot had higher costs if they had multiple congenital anomalies, previous palliation (combining costs of both surgical procedures and hospital stays), or severe "tet" spells at the time of presentation for operation compared with patients without these risk factors ($114,202 +/- $88,524 versus $22,241 +/- $7,071; p = 0.0005). One patient (with tetralogy of Fallot) with multiple congenital anomalies died 42 days after tetralogy of Fallot repair of sepsis after a gastrointestinal operation. Otherwise, hospital mortality was 0% for all groups. CONCLUSIONS: Low mortality and good long-term outcome for surgical correction of congenital heart defects is now commonplace, but can be expensive as some patients with complex problems receive the care necessary to survive. This study demonstrates that it is possible to identify factors preoperatively that predict financial risk. This knowledge may facilitate implementation of risk adjustments for managed care contracting and for strategic resource allocation.     

The physician decision-making process in transferring nursing home patients to the hospital

BACKGROUND: Each year more than 25% of nursing home patients are transferred to the emergency department or hospital for evaluation and treatment of infection. These transfers may have an adverse impact on the quality and cost of patient care. This study examined physician assessment and management of acute infections in the nursing home. METHODS: A cross-sectional study was conducted of all acute urinary tract infections and lower respiratory tract infections occurring from February through June 1991 in eight randomly selected urban nursing homes. The numbers of transfers to the emergency department of hospital were recorded along with identification of the clinical, psychosocial, and institutional factors that influenced the physician's decision to transfer. RESULTS: Three hundred fifty-nine patients had 258 urinary tract infections and 219 respiratory tract infections. Eighty-one (17%) of these events resulted in transfer to a hospital for evaluation (16/81) and/or admission (65/81). Less than one third (30.4%) of the events caused the patient to be examined in the nursing home by a physician before the decision to transfer to the hospital. The mean time between the staff notification of an acute event and physician response by telephone was 5.12 hours. Independent mobility (P < or = .05), a transfer to the hospital during the previous 6 months (P < or = .01), and fewer nursing home laboratory tests and treatments (P < or = .01) were all associated with hospital transfer. CONCLUSIONS: In this sample of acutely ill nursing home patients, physicians collected limited clinical data before the decision to transfer. Although some transfers may be appropriate, a reduction in the transfer rate may reduce health care costs and limit the risk of iatrogenesis, thus improving the outcome of acute illnesses occurring in the nursing home.     

Chemotherapeutics for treatment of influenza and other viral respiratory tract infections in children

Regardless of the primary care model used in the long-term care facility, each of the three approaches offers quality care improvement and greater consistency for residents at reduced costs. Of the three, an all licensed nursing staff model could best meet the higher acuity levels of residents and the disintegrating availability of qualified nursing assistants. If nurses are unable to "sell" this model to administration, it may be helpful to pilot the concept one one unit for a period of time and compare resident, family, and staff satisfaction with that of a similar unit. Also, it is critical to compare the financial implications, including cost per resident per day and rate of staff turnover, to weigh the model's effectiveness. This small sampling of five facilities indicates the average cost per resident per day is $10 less when using either the primary team or all licensed staff models than in facilities of comparable size. Hospitals have already passed the time when they have had to work smarter, leaner, and more efficiently. Can long-term care facilities afford not to do the same?     

Opportunities for control of hospital cost for total knee arthroplasty

The hospital financial records of 120 consecutive patients who underwent unilateral knee replacement surgery at one hospital during 1995 were reviewed to determine opportunities for control of hospital cost for total knee arthroplasty. The average hospital length of stay for these patients was 4.27 days (range, 3-10 days). The average hospital cost was $10,231. All 120 patients were classified under Diagnosis Related Group 209, principle procedure 81.54 primary total knee arthroplasty. Medicare paid for 70% of the patients. All payers were profitable except Medicaid and one managed care organization. When hospital cost for total knee arthroplasty was allocated to hospital service centers, 78% of the cost was attributed to the operating room, nursing units, recovery room, and pharmacy. When hospital cost for total knee arthroplasty was allocated to hospital days, 80% of the hospital cost occurred during the first 48 hours of hospitalization. Hospital reimbursement for total knee arthroplasty is primarily a prospective case price payment system. After initial cost containment efforts reduce the hospital length of stay for total knee arthroplasty to 4 to 6 days, additional control of hospital cost should focus on these areas of opportunity.     

Health care utilization in multiple sclerosis: a population-based study in Olmsted County, MN

OBJECTIVE: The authors sought to determine acute ambulatory- and hospital-billed charges for the Olmsted County, Minnesota Multiple Sclerosis (MS) Disability Prevalence Cohort and compare them to those incurred by the general population. METHODS: Billed charges for 155 people with clinically definite or laboratory-supported MS were compared with those of age- and gender-matched non-MS controls. Billing data, including all inpatient and outpatient acute and rehabilitative medical care charges over a 5-year period surrounding a December 1, 1991 prevalence date, were analyzed. Data were correlated with level of disability using the Minimal Record of Disability for MS. RESULTS: Median total annual billed charges for most individuals with MS, including those with less severe ($1,277) and relapsing-remitting illness ($1,348), did not differ from those for controls ($1,327, p=0.075). Only those with severe MS (22.6%) had median annual medical charges higher than controls ($5,440, p < 0.001). Male patients with MS had higher median annual total charges ($2,353) than male controls ($762, p=0.003). Total charges for female patients with MS ($1,440) were not different from those for female controls ($1469). Median annual outpatient charges were 15% more for the MS group ($1,418) than for controls ($1,231). Patients with MS had a mean of 0.2 hospital admissions annually compared with 0.1 annual admissions per control patient. Among variables collected on persons with MS, the Expanded Disability Status Scale was the strongest predictor of level of charges (p < 0.001). CONCLUSION: Acute ambulatory- and hospital-billed charges for most patients with MS do not differ from those of the general population.     

Evaluation of cost minimization strategies of anaesthetic drugs in a tertiary care hospital

A survey was undertaken to compare anaesthetic drug expenditures over a three-year period, to evaluate the impact of strategies offered to curtain continuously rising drug costs. Suggestions to control rising expenditures were based primarily on education of staff and residents regarding drug costs, emphasizing rational use of the more expensive drugs, and minimizing drug wastage. To assess the impact of these measures, a review of annual hospital budgets, global pharmacy expenditures, and anaesthetic drug expenditures was conducted for the period 1991 to 1993. Both absolute and proportional costs of anaesthetic drugs were compared, by year, according to six major classes: opioid analgesics (OA), muscle relaxants (MR), inhalational anaesthetic drugs (INH), intravenous anaesthetic drugs (i.v.), local anaesthetic drugs (LA) and a category labelled other drugs (OTH). In addition, the utilization patterns and unit price changes were compared for each drug for the periods 1991-92, and 1992-93. Total hospital drug costs increased from $7.1 M to $8.5M over the three years. During the same period, the cost of anaesthetic drugs decreased from $379K to $361K, despite an augmentation in annual case load from 12,507 to 13,076 surgical procedures. For the entire survey period, the mean cumulative anaesthetic drug cost was 4.6% of the pharmacy budget, or 0.24% of the hospital budget. Analysis by drug class revealed a $51K decrease in expenditures on OA. due to decreased utilization of fentanyl and alfentanil, and a decrease in the price of fentanyl. The increased expenditure on INH drugs was primarily due to an increase in acquisition costs.(ABSTRACT TRUNCATED AT 250 WORDS)     

Improving functional outcomes in older patients: lessons from an acute care for elders unit

BACKGROUND: Hospitalization often marks the beginning, and may be partially responsible for, a downward trajectory characterized by declining function, worsening quality of life, placement in a long term care facility, and death. At the University Hospitals of Cleveland, an Acute Care for Elders (ACE) unit that reengineered the process of caring for older patients (> or = 70 years of age) to improve functional outcomes was established in September 1990. DESCRIPTION OF INTERVENTION: The general principles of ACE included an approach to care guided by the biopsychosocial model and recognition of the importance of fitting the hospital environment to the patient's needs. The design of the intervention was consistent with principles of comprehensive geriatric assessment and continuous quality improvement. Care, which focused on maintaining function, was directed by an interdisciplinary team that considered the patient's needs both at home and in the hospital. The major components of the ACE Unit intervention included patient-centered nursing care (daily assessment of functional needs by nursing, nursing-based protocols to improve outcomes, daily rounds by a multidisciplinary team), a prepared environment, planning for discharge, and medical care review. RESULTS: In a randomized trial comparing ACE with usual care, patients receiving ACE had improved functional outcomes at discharge. The costs to the hospital for ACE unit care were less than for usual care. The functional status of ACE and usual care patients was similar 90 days after discharge. FUTURE DIRECTIONS: The ACE unit intervention is being expanded to preserve the improvements observed during the hospitalization in the outpatient setting. In addition, needs other than function which are critical to patients' long-term quality of life are being considered.     

Arthroscopic or open Bankart procedures: what are the costs?

The cost implications and resource utilization of arthroscopic and open Bankart procedures were evaluated to determine if differences exist between these procedures when performed in a community setting. Billing and hospital records of consecutive patients who underwent either open or arthroscopic Bankart procedures at the three facilities in our city during an 18-month period were analyzed. Procedure type (open or arthroscopic), location (hospital or surgicenter), operation time, operating room time, postanesthesia care unit time, step-down area time, charges for each of these, and anesthesiologist charges were analyzed for 11 open and 13 arthroscopic Bankart procedures. Open procedures took longer and required more operating room time than arthroscopic procedures regardless of location (P < .01). Open procedures required longer postanesthesia care unit time than arthroscopic procedures (P < .01). Facility made no difference. Anesthesia fees were less for arthroscopic ($882) than open Bankarts ($1,075) (P = .002). Total facility and anesthesia fees were less for arthroscopic ($4,747) than for open Bankarts ($6,062) (P = .05). The arthroscopic Bankart repair was performed more quickly than the open Bankart procedure, regardless of facility choice, and resulted in lower total charges. A surgicenter is less expensive whether these procedures are performed arthroscopically or open.     

Bone scintigraphy equipped with a pinhole collimator for diagnosis of avascular necrosis of the femoral head

The hypothesis that direct nursing hours correlate with the cost of a patient stay in intensive care was tested. One hundred and thirty-nine patients were studied and the data collected included: (a) direct nursing hours applied to each patient; (b) a daily TISS score: (c) a detailed costing of each patient (all costs are shown in N.Z.$). There was a strong correlation between the direct nursing hours and the total cost per patient (r2 = 0.98) (total cost = 54 x direct nursing hours + 344). Also a strong correlation existed between the total TISS scores and the total costs per patient (r2 = 0.96) (total cost = 67.13 x TISS). Direct nursing hours offer a relatively simple and logical method of allocating costs per patient.     

Attracting nurses to long-term care

1. A major issue facing nursing, and society as a whole, is the longstanding and continuing nursing shortage in long-term care. 2. The demand for RNs in nursing homes will continue to intensify into the next century as the population of older adults with complex care needs continues to grow. 3. The findings of this study suggest that RN recruitment needs in long-term care include having enough qualified and dedicated staff, supportive and competent administration; competitive salaries and benefits; functional, attractive facilities; improved professional and public image; a caring, supportive environment; realistic regulations; decreased paperwork; progressive nurse practice models; and opportunities for educational advancement and career growth.     

Safety and cost of rapid i.v. injection of famotidine in critically ill patients

The safety and cost of famotidine in intensive care patients given the drug by rapid i.v. injection or slow i.v. infusion were studied. All patients admitted to the medical-coronary care and surgical intensive care units (ICUs) at a university teaching hospital over a two-month period who had orders for at least one dose of famotidine injection for any indication were randomly assigned to receive the drug by rapid i.v. injection or slow i.v. infusion via volumetric chamber. Data on patient demographics, drug administration time, adverse effects, cardiovascular variables, and costs (including drug acquisition, supply, and nursing personnel costs) were collected prospectively. Fifty-three patients received famotidine by i.v. injection (a total of 1041 doses) and 52 by i.v. infusion (1006 doses). The mean +/- S.D. duration of famotidine administration was 44 +/- 12 seconds in the i.v.-injection group and 19 +/- 5 minutes in the i.v.-infusion group. Adverse effects possibly related to famotidine occurred in three injection-group patients and two infusion-group patients. No significant difference between the groups in cardiovascular variables (mean arterial pressure, heart rate, and respiratory rate) was noted. Cost savings for the injection group relative to the infusion group totaled $2886 for the two-month study period. Half of the savings came from reduced supply costs and half from reduced personnel costs. The annualized savings to the institution would be about $17,300. Rapid i.v. injection of famotidine appeared to be as safe in ICU patients as giving the drug by slow i.v. infusion and was less costly.