Are cost advantages from a modern Indian hospital transferable to the United States?

BACKGROUND: Multiple modern Indian hospitals operate at very low cost while meeting US-equivalent quality accreditation standards. Though US hospitals face intensifying pressure to lower their cost, including proposals to extend Medicare payment rates to all admissions, the transferability of Indian hospitals' cost advantages to US peers remains unclear. METHODS: Using time-driven activity-based costing methods, we estimate the average cost of personnel and space for an elective coronary artery bypass graft (CABG) surgery at two American hospitals and one Indian hospital (NH). All three hospitals are Joint Commission accredited and have reputations for use of modern performance management methods. Our case study applies several analytic steps to distinguish transferable from non-transferable sources of NH's cost savings. RESULTS: After removing non-transferable sources of efficiency, NH's residual cost advantage primarily rests on shifting tasks to less-credentialed and/or less-experienced personnel who are supervised by highly-skilled personnel when perceived risk of complications is low. NH's high annual CABG volume facilitates such supervised work "downshifting." The study is subject to limitations inherent in case studies, does not account for the younger age of NH's patients, or capture savings attributable to NH's negligible frequency of re-admission or post-acute care facility placement. CONCLUSIONS: Most transferable bases for a modern Indian hospital's cost advantage would require more flexible American states' hospital and health professional licensing regulations, greater family participation in inpatient care, and stronger support by hospital executives and clinicians for substantially lowering the cost of care via regionalization of complex surgeries and weekend use of costly operating rooms.

Time to presentation with acute myocardial infarction in the elderly: associations with race, sex, and socioeconomic characteristics.

BACKGROUND: Although prompt treatment is a cornerstone of the management of acute myocardial infarction (AMI), prior studies have shown that one fourth of AMI patients arrive at the hospital >6 hours after symptom onset. It would be valuable to identify individuals at highest risk for late arrival, but predisposing factors have yet to be fully characterized. METHODS AND RESULTS: Data from the Cooperative Cardiovascular Project, involving Medicare beneficiaries aged >65 years hospitalized between January 1994 and February 1996 with confirmed AMI, were used to identify patients who presented "late" (>/=6 hours after symptom onset). Patient characteristics were tested for associations with late presentation by use of backward stepwise logistic regression. Among 102 339 subjects, 29.4% arrived late. Significant predictors of late arrival (odds ratio, 95% CI) included diabetes (1.11, 1.07 to 1.14) and a history of angina (1.32, 1.28 to 1.35), whereas prior MI (0.82, 0.79 to 0.85), prior angioplasty (0.80, 0.75 to 0.85), prior bypass surgery (0.93, 0.89 to 0.98), and cardiac arrest (0.52, 0.46 to 0. 58) predicted early presentation. Additionally, initial evaluation at an outpatient clinic (2.63, 2.51 to 2.75) and daytime presentation (1.67, 1.59 to 1.72) predicted late arrival. Finally, female sex, black race, and poverty, which were evaluated with an 8-level race-sex-socioeconomic status interaction term, were also risk factors for delay. CONCLUSIONS: Delayed hospital presentation is a common problem among Medicare beneficiaries with AMI. Factors associated with delay include not only clinical and logistical issues but also race, sex, and socioeconomic characteristics. Education efforts designed to hasten AMI treatment should be directed at individuals with risk factors for late arrival.

Race, sex, poverty, and the medical treatment of acute myocardial infarction in the elderly.

BACKGROUND: Race, sex, and poverty are associated with the use of diagnostic cardiac catheterization and coronary revascularization during treatment of acute myocardial infarction (AMI). However, the association of sociodemographic characteristics with the use of less costly, more readily available medical therapies remains poorly characterized. METHODS AND RESULTS: We evaluated 169 079 Medicare beneficiaries >/=65 years of age treated for AMI between January 1994 and February 1996 to determine the association of patient race, sex, and poverty with the use of medical therapy. Multivariable regression models were constructed to evaluate the unadjusted and adjusted influence of sociodemographic characteristics on the use of 2 admission (aspirin, reperfusion) and 2 discharge therapies (aspirin, beta-blockers) indicated during the treatment of AMI. Therapy use varied by patient race, sex, and poverty status. Black patients were less likely to undergo reperfusion (RR 0.84, 95% CI 0. 78, 0.91) or receive aspirin on admission (RR 0.97, 95% CI 0.96, 0. 99) and beta-blockers (RR 0.94, 95% CI 0.88, 1.00) at discharge. Female patients were less likely to receive aspirin on admission (RR 0.98, 95% CI 0.97, 0.99) and discharge (RR 0.98, 95% CI 0.96, 0.99). Poor patients were less likely to receive aspirin (RR 0.97, 95% CI 0. 96, 0.98) or reperfusion (RR 0.97, 95% CI 0.93, 1.00) on admission and aspirin (RR 0.98, 95% CI 0.96, 1.00), or beta-blockers (RR 0.95, 95% CI 0.91, 0.99) on discharge. CONCLUSIONS: Medical therapies are currently underused in the treatment of black, female, and poor patients with AMI.

Acute myocardial infarction complicated by atrial fibrillation in the elderly: prevalence and outcomes.

BACKGROUND: Although atrial fibrillation (AF) is a common complication of acute myocardial infarction (MI), patient characteristics and association with outcomes remain poorly defined in the elderly. METHODS AND RESULTS: We evaluated 106 780 Medicare beneficiaries > or =65 years of age from the Cooperative Cardiovascular Project treated for acute MI between January 1994 and February 1996 to determine the prevalence and prognostic significance of AF complicating acute MI in elderly patients. Patients were categorized on the basis of the presence of AF, and those with AF were further subdivided by time of AF (present on arrival versus developing during hospitalization). AF and non-AF patients were compared by univariate analysis, and logistic regression modeling was used to identify clinical predictors of AF. The influence of AF on outcomes was evaluated by unadjusted Kaplan-Meier survival curves and logistic regression models. AF was documented in 23 565 patients (22. 1%): 11 510 presented with AF and 12,055 developed AF during hospitalization. AF patients were older, had more advanced heart failure, and were more likely to have had a prior MI and undergone coronary revascularization. AF patients had poorer outcomes, including higher in-hospital (25.3% versus 16.0%), 30-day (29.3% versus 19.1%), and 1-year (48.3% versus 32.7%) mortality. AF remained an independent predictor of in-hospital (odds ratio [OR], 1. 21), 30-day (OR, 1.20), and 1-year (OR, 1.34) mortality after multivariate adjustment. Patients developing AF during hospitalization had a worse prognosis than patients who presented with AF. CONCLUSIONS: AF is a common complication of acute MI in elderly patients and independently influences mortality, particularly when it develops during hospitalization.

Access to bone marrow transplantation for leukemia and lymphoma: the role of sociodemographic factors.

PURPOSE: Use of bone marrow transplantation (BMT), a complex, costly treatment for many forms of cancers, has increased significantly in recent years. The increasingly competitive health care marketplace raises concerns about patient access to costly medical procedures such as BMT. We attempted to evaluate patient access to BMT for the treatment of leukemias and lymphomas. METHODS: We analyzed inpatient hospital discharge data from four states (California, Maryland, Massachusetts, and New York) for 2 years (1988 and 1991) to examine whether the use of BMT for patients with either leukemia or lymphoma varies by sociodemographic characteristics and insurance coverage. We developed a sorting algorithm to collapse the discharge data into patient level records. We used logistic regression to analyze the odds of receiving a BMT stratified by disease type (leukemia or lymphoma). RESULTS: After controlling for other factors, black patients with leukemia are 51% to 53% as likely as whites, while black patients with lymphoma are 34% to 45% as likely as white patients to undergo a BMT (P < .05). Medicaid, self-pay patients, and Health Maintenance Organization (HMO) enrollees with either leukemia or lymphoma are significantly less likely to undergo a BMT compared with patients with private insurance. Younger patients are significantly more predisposed to undergo a BMT than older patients. The odds of receiving a BMT have increased over time, but the rates of increase vary by state. Consistent with clinical expectations, the relative odds of BMT vary significantly by type of leukemia or lymphoma. CONCLUSION: Substantial variation exists in access to BMT for patients with either leukemia or lymphoma. Black patients, those enrolled in HMOs, those covered by Medicaid, and self-pay patients were less likely to receive a BMT when admitted for either leukemia or lymphoma. These findings raise concerns about access to cancer treatments for patients in the current health care system.

Effect of CD34(+) cell dose on resource utilization in patients after high-dose chemotherapy with peripheral-blood stem-cell support.

PURPOSE: The mean time to neutrophil and platelet recovery for patients receiving high-dose chemotherapy (HDC) supported with peripheral-blood stem cells (PBSCs) is related to the dose of CD34(+) cells infused. The effect of cell dose on resource utilization after transplantation has not been previously reported. MATERIALS AND METHODS: We assessed CD34(+) cell dose and resource utilization for 1,317 patients undergoing transplantation with PBSCs from April 1991 to June 1997. PBSCs were collected after mobilization with chemotherapy and recombinant human granulocyte colony-stimulating factor (rhG-CSF). Daily measurement of the CD34(+) content of the PBSC collection was performed by a central laboratory using a single CD34(+) analysis technique. Resource utilization included engraftment parameters, length of stay, and transfusion requirements for 100 days posttransplantation. Analysis included descriptive statistics and multiple regression. RESULTS: Mean patient age was 47 years, and 86% of patients were female. Median cell dose was 3.6 x 10(6)/kg and 13.2 x 10(6)/kg for patients receiving less than 5.0 x 10(6) CD34(+) cells/kg and 5.0 x 10(6) or more CD34(+) cells/kg, respectively. Patients receiving less than 5. 0 x 10(6) CD34(+) cells/kg were more likely to have metastatic breast cancer or non-Hodgkin's lymphoma and required more platelet and RBC transfusions, 3.3 more hospital days, and increased antibiotic and antifungal use. In univariate analysis, the cost of care was $41,516 (+/-$20,876 SD) and $32,382 (+/-$16,353 SD) for patients with less than 5.0 x 10(6) CD34(+) cells/kg and 5.0 x 10(6) or more CD34(+) cells/kg, respectively. In multivariate analysis, patients with less than 5.0 x 10(6) CD34(+) cells/kg had an increase in costs of $5,062 (+/- $1,262 SE). CONCLUSION: Infusion of more than 5.0 x 10(6) CD34(+) cells/kg was associated with a reduction in resource utilization. Achieving a target of 5.0 x 10(6) CD34(+) cells/kg should have important clinical and economic benefits for patients.

Randomized trial of filgrastim, sargramostim, or sequential sargramostim and filgrastim after myelosuppressive chemotherapy for the harvesting of peripheral-blood stem cells.

PURPOSE: The purpose of this study was to compare the effects of filgrastim, sargramostim, or sequential sargramostim and filgrastim on CD34(+) cell yields and morbidity after myelosuppressive mobilization chemotherapy (MC). PATIENTS AND METHODS: One hundred fifty-six patients were randomized to receive filgrastim (n = 51), sargramostim (n = 52), or sargramostim for 5 days followed by filgrastim (n = 53) after MC with either cyclophosphamide and etoposide (n = 75) or paclitaxel and cyclophosphamide (n = 81). RESULTS: Compared with those who received sargramostim, patients who received filgrastim had faster recovery of an absolute neutrophil count of 0.5 x 10(9)/L or greater (a median of 11 v 14 days; P =. 0001), with fewer patients requiring RBC transfusions (P =.008), fewer patients with fever (18% v 52%; P = 0.001), fewer hospital admissions (20% v 42%; P =.013), and less intravenous antibiotic therapy (24% v 69%; P =.001). Patients who received filgrastim yielded more CD34(+) cells (median, 7.1 v 2.0 x 10(6)/kg/apheresis; P =.0001), and a higher fraction achieved 2.5 x 10(6) (94% v 78%; P =.021) and 5 x 10(6) (88% v 53%; P =.001) or more CD34(+) cells/kg with fewer aphereses (median, 2 v 3; P =.002) and fewer days of growth-factor treatment (median, 12 v 14; P =.0001). There were no major differences in outcomes between the filgrastim alone and the sequential regimens. After high-dose chemotherapy, patients who had peripheral-blood stem cells (PBSCs) mobilized with filgrastim or the sequential regimen received higher numbers of CD34(+) cells and had faster platelet recovery (P =.015), with fewer patients (P =.014) receiving fewer platelet transfusions (P =.001) than patients receiving sargramostim-mobilized PBSCs. CONCLUSION: It was concluded that filgrastim alone or sequential sargramostim and filgrastim were superior to sargramostim alone for the mobilization of CD34(+) cells and reduction of toxicities after MC.

Impact of quality of life on patient expectations regarding phase I clinical trials.

PURPOSE: Quality of life (QOL) is increasingly recognized as a critical cancer-treatment outcome measure, but little is known about the impact of QOL on the patient decision-making process. A pilot study was conducted in an effort to (1) measure the expectations of patients, physicians, and research nurses regarding the potential benefits and toxicities from experimental and standard therapies, and (2) determine the relationship of QOL to patient perceptions regarding treatment options. METHODS: Thirty cancer patients enrolling in phase I clinical trials, their physicians, and their research nurses were administered questionnaires that assessed demographics, QOL, and treatment expectations. RESULTS: Compared with their physicians, patients overestimated potential benefits and toxicities from experimental therapy (mean expected benefit, 59.8% v 23.8%, P <.01; mean expected toxicity, 29.8% v 16.0%, P <.01). Patients estimated a greater potential for benefit (59.8% v 36.8%, P <.01) and less potential for toxicity (29.8% v 45.6%, P =.01) for experimental therapy, compared with standard therapy. Short Form-36 general health perception correlated with patient perception of potential benefit from experimental therapy (r =.48, P =.01). CONCLUSION: Participants in phase I clinical trial have high expectations regarding the success of experimental therapy and discount potential toxicity. Patient QOL may affect the expectation of benefit from experimental therapy and, ultimately, treatment choice. Understanding the interactions between QOL and patient expectations may guide the development of improved strategies to present appropriate information to patients considering early-phase clinical trials.

Physician resource use and willingness to participate in assisted suicide.

OBJECTIVE: To explore the relationship between general internists' tendency to conserve medical resources and their willingness to participate in physician-assisted suicide (PAS). DESIGN AND PARTICIPANTS: Survey of a random sample of general internists in 6 urban areas of the United States. MEASUREMENTS: We assessed the physicians' use of medical resources by constructing a scale based on 6 hypothetical clinical scenarios in which respondents were given a choice between resource-intensive and resource-conserving options. We then presented a scenario of a competent terminally ill patient with breast cancer making stable and persistent requests for PAS. RESULTS: Sixty-seven (33%) of the 206 respondents indicated that they would participate in the suicide of the depicted patient. In a multivariate model, physicians who were more conservative with resources were 6.4 times more likely than their resource-intensive counterparts to prescribe the requested drugs (P = .02); minority physicians were less willing than whites to participate in PAS (odds ratio, 0.34; P = .03). Physicians' number of years in practice, location, sex, reported percentage of fee-for-service patients, and self-reported strength and direction of financial incentives in the respondents' practices were not associated with willingness to prescribe drugs for PAS. CONCLUSIONS: Most general internists, especially minority physicians, are personally reluctant to participate in PAS. While the characteristics of their practices do not affect PAS, physicians who tend to practice resource-conserving medicine are significantly more likely than their resource-intensive counterparts to provide a lethal prescription at the request of a terminally ill patient.

Lying for patients: physician deception of third-party payers.

BACKGROUND: Some physicians may resort to deception to secure third-party payer approval for patient procedures. Related physician attitudes, including willingness to use deception, are not well understood. OBJECTIVE: To determine physician willingness to deceive a third-party payer and physician attitudes toward deception of third-party payers. METHODS: A cross-sectional mailed survey was used to evaluate physician willingness to use deception in 6 vignettes of varying clinical severity: coronary bypass surgery, arterial revascularization, intravenous pain medication and nutrition, screening mammography, emergent psychiatric referral, and cosmetic rhinoplasty. We evaluated 169 board-certified internists randomly selected from 4 high- and 4 low-managed care penetration metropolitan markets nationwide for willingness to use deception in each vignette. RESULTS: Physicians were willing to use deception in the coronary bypass surgery (57.7%), arterial revascularization (56.2%), intravenous pain medication and nutrition (47.5%), screening mammography (34.8%), and emergent psychiatric referral (32.1%) vignettes. There was little willingness to use deception for cosmetic rhinoplasty (2.5%). Rates were highest for physicians practicing in predominantly managed care markets, for clinically severe vignettes, and for physicians spending less time in clinical practice. Physician ratings of the justifiability of deception varied by perspective and vignette. CONCLUSIONS: Many physicians sanction the use of deception to secure third-party payers' approval of medically indicated care. Such deception may reflect a tension between the traditional ethic of patient advocacy and the new ethic of cost control that restricts patient and physician choice in the use of limited resources.

Determining an episode of care using claims data. Diabetic foot ulcer.

OBJECTIVE: Amid changes in the organization and financing of health care, health care decision makers are increasingly interested in episodes of care. We sought to determine an episode of care for diabetic foot ulcer using an administrative claims database. RESEARCH DESIGN AND METHODS: We used 1993-1995 claims data to assess resource utilization for privately insured patients with diabetic foot ulcers. Over a 26-week period, we determined the episode length by comparing differences in average daily charges and proportion of patients with charges before and after foot ulcer diagnosis. All 13 weeks before diagnosis were used to calculate baseline values. Significance was determined by CIs, which were calculated by a nonparametric bootstrap technique. Costs associated with the episode were also calculated. A sensitivity analysis using weeks with highest and lowest values as baseline was also conducted. RESULTS: Based on average daily charges, the episode of care for diabetic foot ulcer was 5 weeks. Using proportion of patients with charges, the episode was longer than 13 weeks. The cost for an episode of care ranged from $900 to $2,600. In the sensitivity analyses, episodes of care ranged from 1 to 13 weeks. CONCLUSIONS: Episodes of care can be defined by the period beginning with increased resource consumption and ending when resource consumption returns to baseline levels. With the growth of managed care and disease management programs, episode-of-care analysis may have an increasingly important role in health care provision and delivery.

Economic evaluation of high-dose chemotherapy and bone marrow transplantation for patients with breast cancer of stage II or IIIA with more than 10 positive lymph nodes at surgical excision.

Several issues related to the design of a multi-year study of the outcome of high-dose chemotherapy with bone marrow transplant for patients with metastatic breast cancer have been discussed. Important considerations include an understanding of the types of data that would be collected through the study period, an assessment of treatment benefits and complications, an understanding of the valuation of differences in resource consumption across treatment arms, and an informed definition of the study population to include in the economic assessment. The resulting analysis will help patients and policy makers better understand the impact of aggressive treatment options for breast cancer patients.

Introduction to clinical economics: assessment of cancer therapies.

Recent changes in the health care environment have led to the assessment of the costs and benefits of cancer treatment as criteria for the evaluation of new cancer therapies. The methodological framework for these assessments is provided by the field of clinical economics, a discipline that combines the techniques from clinical medicine, economics, epidemiology, and biostatistics. This article reviews the concepts of clinical economics and issues related to economic analysis of new therapies. This type of information is increasingly important to patients, clinicians, and health policy decision makers to help assure patients' access to effective cancer therapies.

Resource utilization control processes as indicators of quality in managed care organizations: a proposal.

Central to the development of a competitive market for managed care products is the ability to measure the quality of care provided by individual managed care organizations (MCOs). Several types of quality measures could be considered for this purpose: patient and provider satisfaction, a listing of specific services provided to patients, or clinical outcomes of such services. Although assessing quality with measures of the process of care is commonplace, we propose developing measures of the production processes that control utilization of health care resources within an organization. Evaluation of these production or resource utilization control processes, although not a substitute for health outcomes assessment, may improve our knowledge of the delivery of services within managed care organizations. We present a paradigm for evaluation of health care resource utilization control processes within managed care organizations based on our description of internal and external controls for health care resource management. This paradigm can serve as a framework for further research into the quality of care provided by these organizations and the processes of health care resource management in MCOs.

The effects of patient sex and race on medical students' ratings of quality of life.

PURPOSE: Although previous studies have examined race and sex differences in health care, few studies have investigated the possible role of physician bias. We evaluated the influence of race and sex on medical students' perceptions of patients' symptoms to determine if there are differences in these perceptions early in medical training. SUBJECTS AND METHODS: One-hundred sixty-four medical students were randomly assigned to view a video of a black female or white male actor portraying patients with identical symptoms of angina. We evaluated students' perceptions of the actors' health state (based on their assessment of quality of life) using a visual analog scale and a standard rating technique, as well as the type of chest pain diagnosis. RESULTS: Students assigned a lower value (indicating a less desirable health state) to the black woman than to the white man with identical symptoms [visual scale (mean +/- SD): 72 +/- 13 vs 67 +/- 12, P <0.02; standard gamble: 87 +/- 10 vs 80 +/- 15, P < 0.001). Nonminority students reported higher mean values for the white male patient (standard gamble: 89 +/- 8 vs 81 +/- 14 for the black female patient), whereas minority students' assessments did not differ by patient. Male students assigned a slightly lower value to the black female patient (standard gamble: 76 +/- 16 vs 87 +/- 10 for the white male patient). Students were less likely to characterize the black female patient's symptoms as angina (46% vs 74% for the white male patient, P = 0.001). CONCLUSIONS: The way that medical students perceive patient symptoms appears to be affected by nonmedical factors.

Effect of diabetes mellitus and insulin use on survival after acute myocardial infarction in the elderly (the Cooperative Cardiovascular Project).

Using data from a retrospective cohort study of Medicare beneficiaries hospitalized with an acute myocardial infarction (AMI), we evaluated the role of diabetes mellitus on 30-day and 1-year mortality. We classified subjects as nondiabetics, diabetics controlled with diet alone, diabetics receiving an oral hypoglycemic agent, and diabetics on insulin at time of admission. We compared baseline admission characteristics of subgroups using chi-square and Wilcoxon rank-sum tests and evaluated the effect of each diabetic state using sequential logistic models. We identified 80,832 nondiabetic patients, 9,862 diet-controlled diabetic patients, 14,664 diabetics receiving an oral hypoglycemic agent, and 12,241 diabetic patients on insulin therapy. Although mean age was similar among the groups, prevalence of hypertension, prior AMI, prior congestive heart failure, and prior revascularization were higher among diabetic patients, particularly those taking insulin. Diabetic patients, particularly those taking insulin, were less likely to receive aspirin and beta blockers and to undergo coronary revascularization. Diabetic patients had higher 30-day and 1-year mortality than nondiabetic patients. After adjustment for demographics, clinical and hospital characteristics, and treatment strategies, insulin-treated diabetics had the highest risk of mortality, followed by diabetics receiving oral hypoglycemic agents, followed by diet-controlled diabetics. Thus, diabetes is highly prevalent among elderly patients with an AMI. Mortality rates for these patients, particularly insulin-using diabetics, are higher than among their nondiabetic counterparts. Preventive and therapeutic strategies must be developed to ensure improved short- and long-term outcomes for elderly patients with diabetes and AMI.

Clinical economics in clinical trials: the measurement of cost and outcomes in the assessment of clinical services through clinical trials.

As the population ages and more expensive high-technology services become available, health care costs continue to spiral upward. Because the financial resources for health care are limited, economic analysis can help to evaluate expenditures and set priorities. Economic analysis of medical technology or medical care evaluates a medical service by comparing its dollar cost with its dollar benefit (cost-benefit), by measuring its dollar cost in relation to its outcomes (cost-effectiveness) as well as in relation to its utility or quality-adjusted outcomes (cost-utility), or simply by tabulating the costs involved (cost-identification). Direct costs are generated as services are provided. In addition, patients' productivity is affected, and these costs can be considered, especially in determining the benefit of a service that decreases morbidity or mortality. Intangible costs are those of pain, suffering, and grief. The point of view, or perspective, of the study determines the costs and benefits that will be measured in the analysis. Sensitivity analysis, which can evaluate the stability of the conclusions to the data used, is an important assessment within economic analysis. Economic analysis of new pharmaceutical therapies is increasingly being incorporated into clinical trials. Although there are some limitations of pharmacoeconomic information in clinical studies of drug safety and efficacy, these trials are often the only opportunity for economic data collection before adoption and reimbursement decisions are made. Validation after the drug has been introduced should complement economic information developed from clinical trials.

Mobilization of peripheral blood stem cells following myelosuppressive chemotherapy: a randomized comparison of filgrastim, sargramostim, or sequential sargramostim and filgrastim.

Myelosuppressive chemotherapy is frequently used for mobilization of autologous CD34(+) progenitor cells into the peripheral blood for subsequent collection and support of high-dose chemotherapy. The administration of myelosuppressive chemotherapy is typically followed by a myeloid growth factor and is associated with variable CD34 cell yields and morbidity. The two most commonly used myeloid growth factors for facilitation of CD34 cell harvests are granulocyte colony-stimulating factor (G-CSF) and granulocyte-macrophage colony-stimulating factor (GM-CSF). We performed a randomized phase III clinical trial comparing G-CSF, GM-CSF, and sequential administration of GM-CSF and G-CSF following administration of myelosuppressive chemotherapy. We evaluated CD34 yields, morbidity, and cost-effectiveness of the three cytokine schedules. One hundred and fifty-six patients with multiple myeloma, breast cancer, or lymphoma received cyclophosphamide with either paclitaxel or etoposide and were randomized to receive G-CSF 6 microg/kg/day s.c., GM-CSF 250 microg/m(2)/day s.c., or GM-CSF for 6 days followed by G-CSF until completion of the stem cell harvest. Compared with patients who received GM-CSF, patients who received G-CSF had faster recovery of absolute neutrophil count to 0.5 x 10(9) per liter (median of 11 vs14 days, P = 0.0001) with fewer patients requiring red blood cell transfusions (P= 0.008); fewer patients with fever (18% vs 52%, P = 0.001); fewer hospital admissions (20% vs 42%, P = 0.13); and less intravenous antibiotic therapy (24% vs 59%, P = 0.001). Patients who received G-CSF also yielded more CD34 cells (median 7.1 vs 2.0 x 10(6) kg per apheresis, P = 0.0001) and a higher percentage achieved 2.5 x 10(6) CD34 cells per kilogram (94% vs 78%, P = 0.21) and 5 x 10(6) CD34 cells per kilogram (88% vs 53%, P = 0.01) or more CD34 cells per kilogram with fewer aphereses (median 2 vs 3, P = 0.002) and fewer days of growth factor treatment (median 12 vs 14, P = 0.0001). There were no significant differences in outcomes between groups receiving G-CSF alone and the sequential regimen. After high-dose chemotherapy, patients who had peripheral blood stem cells mobilized with G-CSF or the sequential regimen received higher numbers of CD34 cells and had faster platelet recovery with fewer patients requiring platelet transfusions than patients receiving peripheral blood stem cells mobilized by GM-CSF. In summary, G-CSF alone is superior to GM-CSF alone for the mobilization of CD34(+) cells and reduction of toxicities following myelosuppressive chemotherapy. An economic analysis evaluating the cost-effectiveness of these three effective schedules is ongoing at the time of this writing.

Economic evaluation of filgrastim, sargramostim, and sequential sargramostim and filgrastim after myelosuppressive chemotherapy.

Filgrastim alone and sequential sargramostim and filgrastim have been shown to be more effective than sargramostim alone in the mobilization of CD34(+) cells after myelosuppressive chemotherapy (MC). We sought to compare costs and resource use associated with these regimens. Data were collected prospectively alongside a multicenter, randomized trial of filgrastim, sargramostim, and sequential sargramostim and filgrastim. Direct medical costs were calculated for inpatient and outpatient visits and procedures, including administration of growth factors and MC. We followed 156 patients for 30 days or until initiation of high-dose chemotherapy. The main outcome measures were resource use and costs of inpatient and outpatient visits, platelet and red blood cell transfusions, antibiotic use, and apheresis procedures. Hospital admissions, red blood cell transfusions, and use of i.v. antibiotics were significantly more common in the sargramostim group than in the other treatment arms. In univariate and multivariable analyses, total costs were higher for patients receiving sargramostim alone than for patients in the other groups. Mean costs in multivariable analysis for the filgrastim and sequential sargramostim and filgrastim arms were not significantly different. Filgrastim alone and sequential sargramostim and filgrastim are less costly than sargramostim alone after MC, as well as therapeutically more beneficial.

Prospective economic evaluation accompanying a trial of GM-CSF/IL-3 in patients undergoing autologous bone marrow transplantation for Hodgkin's and non-Hodgkin's lymphoma. IL-3 BMT Study Team.

Our objective was to assess the economic impact of a new cytokine therapy that was being compared to standard therapy as supportive care in patients receiving autologous bone marrow transplantation for treatment of lymphoma. We performed an economic study accompanying a multicenter, randomized, controlled clinical trial in academic medical centers. One hundred and fifteen patients consented to participate in a parallel economic study of a randomized controlled trial of sequential IL-3 followed by GM-CSF vs GM-CSF alone after autologous bone marrow transplantation. We measured costs and quality-adjusted survival over a 13-month follow-up period. For the 13-month study period, the total cost estimates were $79892 (95% CI $69343 to $90544) for patients receiving GM-CSF alone and $89651 (95% CI $79769 to $102114) for patients receiving IL-3/GM-CSF. The difference was not statistically significant. During the 13-month study period, the total number of quality-adjusted life-months in the GM-CSF arm was 6.67 (95% CI 5.75 to 7.56) months, while the total number of quality-adjusted life-months in the IL-3 arm was 6.26 (95% CI 5.34 to 7.15) months. The difference in quality-adjusted life-months between the two treatment arms was not statistically significant. We conclude that economic analysis of a phase III clinical trial of IL-3/GM-CSF compared with GM-CSF alone showed no significant effect of IL-3 on the costs of care for patients undergoing bone marrow transplantation for a period of up to 13 months after the procedure. This study demonstrates the feasibility of prospective economic evaluation within phase III trials of new cancer therapies. Data from this type of economic protocol could be used to help physicians, patients and managed care organizations understand the effect of new treatments from both a clinical and an economic perspective.