PURPOSE: To evaluate the cost-effectiveness of recombinant human activated protein C (rhAPC) compared with usual therapy for patients with severe sepsis, and also to determine the influence that severity of illness exerts on cost-effectiveness. MATERIALS AND METHODS: We use a Markov model-based cost-effectiveness analysis of treatment strategies for patients with severe sepsis. Therapy includes treatment with either rhAPC and usual therapy, or usual therapy alone. Probabilities for clinical outcomes were obtained from a large randomized clinical trial comparing the use of rhAPC with placebo (PROWESS study) and from outcomes literature for patients with severe sepsis and its complications. Cost estimates were based on Medicare reimbursement rates, Health Care Financing Administration information and the literature. Outcome measures include life-years, quality-adjusted life-years (QALYs), costs, and incremental cost-effectiveness. RESULTS: Compared with usual therapy alone, rhAPC treatment for patients with very severe sepsis (APACHE II score > or = 25) was associated with an incremental cost-effectiveness ratio of $13 493/QALY. Treatment of patients with less severe sepsis with rhAPC (APACHE II score 25) had an incremental cost-effectiveness ratio of $403,000/QALY. For patients with very severe sepsis the incremental cost-effectiveness ratio for treatment with rhAPC remained under $30,000/QALY, over a broad range of variables, including costs of rhAPC, costs of acute care and costs and probabilities of complications of treatment. For patients with less severe sepsis, drug costs would need to fall well below current market price before achieving cost-effectiveness. A probabilistic sensitivity analysis comparing rhAPC treatment with usual therapy for patients with very severe sepsis showed that 1% of Monte Carlo simulations had incremental cost-effectiveness ratios > $50,000/QALY. CONCLUSIONS: The use of rhAPC for the treatment of patients with very severe sepsis, as determined by APACHE II score > or = 25, appears cost-effective, while treatment of patients with APACHE II score 25 is not cost-effective.

PURPOSE: To conduct an empirical analysis of self-referred whole-body computed tomography (CT) and develop a profile of the geographic and demographic distribution of centers, types of services and modalities, costs, and procedures for reporting results. MATERIALS AND METHODS: An analysis was conducted of Web sites for imaging centers accepting self-referred patients identified by two widely used Internet search engines with large indexes. These Web sites were analyzed for geographic location, type of screening center, services, costs, and procedures for managing imaging results. Demographic data were extrapolated for analysis on the basis of center location. Descriptive statistics, such as frequencies, means, SDs, ranges, and CIs, were generated to describe the characteristics of the samples. Data were compared with national norms by using a distribution-free method for calculating a 95% CI (P .05) for the median. RESULTS: Eighty-eight centers identified with the search methods were widely distributed across the United States, with a concentration on both coasts. Demographic analysis further situated them in areas of the country characterized by a population that consisted largely of European Americans (P .05) and individuals of higher education (P .05) and socioeconomic status (P .05). Forty-seven centers offered whole-body screening; heart and lung examinations were most frequently offered. Procedures for reporting results were highly variable. CONCLUSION: The geographic distribution of the centers suggests target populations of educated health-conscious consumers who can assume high out-of-pocket costs. Guidelines developed from within the profession and further research are needed to ensure that benefits of these services outweigh risks to individuals and the health care system.

OBJECTIVE: Positron emission tomography (PET) is a high-cost imaging tool primarily used in oncology, cardiology, and neuropsychiatry. Accurate estimates of the cost of PET are needed to assess its cost effectiveness and determine the appropriate role for this modality in clinical applications. We performed a survey-based cost analysis of PET with FDG by estimating direct, indirect, and capital costs from eight PET centers. A breakdown of the operational budget of PET centers and FDG-compounding facilities is presented along with the costs per scan. Differences in costs between sites that purchase FDG and those that manufacture FDG are also examined. MATERIALS AND METHODS: We sent surveys to managers of eight Veterans Affairs and two non-Veterans Affairs PET scanning and FDG-compounding facilities. The survey included questions about service volume and the direct costs of equipment, personnel, space, supplies, and repairs needed for FDG compounding and PET scanning and interpretation. We estimated the indirect costs associated with FDG compounding, PET scanning, and PET interpretation. RESULTS: Of the eight sites that responded to our survey, three sites manufacture FDG on-site, three sites purchase FDG, and two sites do both. The total mean cost per scan using manufactured FDG is 1885 US dollars, and it is 1898 US dollars using purchased FDG. CONCLUSION: PET is expensive. The cost is similar when FDG is manufactured or purchased. Because both PET and cyclotron facilities have high fixed costs, increasing the number of scans obtained and the number of FDG doses manufactured may lead to a decrease in unit costs.

Background: In recent years, there has been significant controversy over how institutional review board (IRB) procedures are structured, administered, and balanced against other concerns (for example, the feasibility of multicenter research). Formal economic analysis has been largely absent in these debates. With hospitals and research facilities under budgetary pressures, it is essential that cost-effective IRB systems be developed.

Objective: To estimate the cost of IRB procedures in an eight-site observational substance abuse treatment study.

Methods and Findings: We estimated the cost of each IRB "action" (initial reviews, continuing reviews, amendments, and adverse event reports) at each project site using a method that has been described in detail elsewhere. In brief, a national administrative database on IRBs was consulted to determine the estimated total cost of operating IRBs of various sizes. These total costs comprised personnel, space, supplies, and training. Dividing these total costs by the total number of actions conducted per year by an IRB yielded a cost per action. When these data were used as a reference point, each of the eight IRBs in our study was estimated to have a cost per action equal to that of IRBs of a comparable size in the national database.

All study procedures for the substance abuse treatment study were approved by our local university IRB. Nevertheless, the IRBs at each of the eight participating medical centers elected to conduct an independent, full-board review of the research protocol. The focus of our economic study was the costs of these "supplemental" IRB actions. In other words, how much money was expended on further reviewing the study after the home IRB had approved the project?

Expenditures for the supplemental IRB activities were estimated at $56 191 in 2001 dollars. This consumed 16.8% of the total research grant budget for all activities (recruiting sites, enrolling and interviewing participants, analyzing data) over the same period. Most of these costs ($29 824) reflected the IRBs' expenditures for 70 supplemental actions. Coordinating center personnel, space, and supply costs made up the second largest expense, at $16 951. The additional effort for the investigators at the eight participating sites totaled $9416.

Discussion: The high costs of IRBs were driven by several factors. First, our home IRB and committees at local sites frequently disagreed on the formatting of consent procedures and patient survey forms, and all expected final say on study implementation. This resulted in consent forms and patient survey instruments being reviewed and re-reviewed many times. Second, participating institutions required review both by hospital and affiliated university IRBs. We believe this redundancy was due to fear generated by recent high-profile cases in which research institutions were penalized for human subjects violations. Third, our coordinating center was often not allowed to directly contact IRB staff at the participating sites, nor would the nine IRBs involved (including our home IRB) communicate with each other. Coordinating center staff thus spent long periods translating and retransmitting information to IRB staff and local investigators scattered throughout the country, which slowed and sometimes garbled communication.

Conclusion: These extensive IRB procedures had no discernible impact on human subjects protection. The essential procedures of the study never changed substantially, despite exchanges of over 15 000 pages of material among the nine sites. In addition, as an empirical matter, no adverse incidents occurred among the participants. Researchers and policymakers should weigh the cost and benefits of multisite review for minimal-risk studies, such as observational health services research. The National Cancer Institute's centralized IRB for multisite phase III trials is an intriguing model. Other systems exist, although none has explicitly discussed minimal-risk studies. If costs such as those identified in this project continue to grow without consideration of the benefits of facilitating scientific projects, we worry that future research may be unjustly affected.

We compared results of safety climate survey questions from health care respondents with those from naval aviation, a high-reliability organization. Separate surveys containing a subset of 23 similar questions were conducted among employees from 15 hospitals and from naval aviators from 226 squadrons. For each question a "problematic response" was defined that suggested an absence of a safety climate. Overall, the problematic response rate was 5.6% for naval aviators versus 17.5% for hospital personnel (p0.0001). The problematic response was 20.9% in high-hazard domains such as emergency departments and operating room. Problematic response among hospital workers was up to 12 times greater than that among aviators on certain questions.

CONCLUSION: Although further research on safety climate in health care is warranted, hospitals may need to make substantial changes to achieve a safety climate consistent with the status of high-reliability organizations.

Employment-based health insurance is failing. Costs are out of control. Employers have no effective strategy to deal with this. They must think strategically about fundamental change. This analysis explains how employers' purchasing policies contribute to rising costs and block growth of economical care. Single-source managed care is ineffective, and effective managed care cannot be a single source. Employers should create exchanges through which they can offer employees wide, responsible, individual, multiple choices among health care delivery systems and create serious competition based on value for money. Recently introduced technology can assist this process.

Context The Internet has attracted considerable attention as a means to improve health and health care delivery, but it is not clear how prevalent Internet use for health care really is or what impact it has on health care utilization. Available estimates of use and impact vary widely. Without accurate estimates of use and effects, it is difficult to focus policy discussions or design appropriate policy activities.

Objectives To measure the extent of Internet use for health care among a representative sample of the US population, to examine the prevalence of e-mail use for health care, and to examine the effects that Internet and e-mail use has on users' knowledge about health care matters and their use of the health care system.

Design, Setting, and Participants Survey conducted in December 2001 and January 2002 among a sample drawn from a research panel of more than 60,000 US households developed and maintained by Knowledge Networks. Responses were analyzed from 4764 individuals aged 21 years or older who were self-reported Internet users.

Main Outcome Measures Self-reported rates in the past year of Internet and e-mail use to obtain information related to health, contact health care professionals, and obtain prescriptions; perceived effects of Internet and e-mail use on health care use.

Results Approximately 40% of respondents with Internet access reported using the Internet to look for advice or information about health or health care in 2001. Six percent reported using e-mail to contact a physician or other health care professional. About one third of those using the Internet for health reported that using the Internet affected a decision about health or their health care, but very few reported impacts on measurable health care utilization; 94% said that Internet use had no effect on the number of physician visits they had and 93% said it had no effect on the number of telephone contacts. Five percent or less reported use of the Internet to obtain prescriptions or purchase pharmaceutical products.

Conclusions Although many people use the Internet for health information, use is not as common as is sometimes reported. Effects on actual health care utilization are also less substantial than some have claimed. Discussions of the role of the Internet in health care and the development of policies that might influence this role should not presume that use of the Internet for health information is universal or that the Internet strongly influences health care utilization.