The Costs of Conducting Clinical Research By Ezekiel J. Emanuel, Lowell E. Schnipper, Deborah Y. Kamin, Jenifer Levinson, and Allen S. Lichter Purpose: Physicians frequently receive payment for enrolling subjects onto clinical trials. Some view these payments as conflicts of interest. Others contend that these payments are necessary reimbursements for conducting clinical research. We evaluated the clinical and nonclinical hours and costs associated with conducting a mock phase III clinical research trial. Methods: We collected data from representatives of 21 clinical sites, on the numbers of hours associated with 13 activities necessary to the conduct of clinical research. The hours were based on enrolling 20 patients in a 12-month randomized placebo-controlled trial of a new chemotherapeutic agent. The outcome measures were disease progression and quality-of-life reports. These costs were evaluated for both government and pharmaceutical industry–sponsored trials. Results: On average, 4,012 hours (range, 1,512 to 13,319 hours) were required for a government-sponsored

trial, and 3,998 hours (range: 1735 to 15,699) were required for a pharmaceutical industry–sponsored trial involving 20 subjects with 17 office visits, or approximately 200 hours per subject. Thirty-two percent of the hours were devoted to nonclinical activities, such as institutional review board submission and completion of clinical reporting forms. On average, excluding overhead expenses, it cost slightly more than $6,094 (range, $2,098 to $19,285) per enrolled subject for an industry-sponsored trial, including $1,999 devoted to nonclinical costs. Conclusion: Based on the results of our mock trial, the time required for nontreatment trial activities is considerable, and the associated costs are substantial. J Clin Oncol 21:4145-4150. © 2003 by American Society of Clinical Oncology.

N GENERAL, in the United States, there is significant support for clinical research and the therapeutic benefits it makes possible. Nevertheless, there are serious concerns about the conduct of clinical research, particularly as it relates to clinical researchers’ potential conflicts of interest. Indeed, the concern prompted the Department of Health and Human Services to issue new conflict-of-interest guidelines.1 One aspect of these conflicts of interest relate to payments that physicians and investigators receive for enrolling subjects onto clinical trials. This longstanding practice has often been criticized.2-5 Some view these payments as conflicts of interest because they “pose a temptation for the investigator to enter a patient in an experimental study”6 that may not be in the patient’s interest. Some claim researchers are “enticing and cajoling patients” to clinical studies in order to obtain payment.6 The New York Times ran two prominent articles highlighting the practice of physicians receiving payments, and even listed the fees of some pharmaceutical industry– sponsored drug trials, including a diabetes trial paying $4,410 per subject, a migraine study paying $3,600 per subject, and a study of an arthritis drug paying $1,600 per subject.6 Indeed, one report claimed that physicians could reap up to $225,000 for enrolling subjects in clinical research trials.4 Furthermore, critics contend that physicians should inform subjects about the payments they receive, but they almost never do.2-4 Responses to these criticisms generally center on the argument that enrolling subjects onto clinical research trials is not free; there is time and effort related to obtaining institutional review board (IRB) approval for research studies, explaining the research to prospective subjects, obtaining signed informed consent, completing clinical reporting forms, maintaining records for potential audits, and other related tasks. The payments for recruiting subjects cover such added expenses. Viewed in this

light, these payments do not constitute a conflict-of-interest and need not be a part of the informed-consent process. Furthermore, given physicians’ tremendous time and financial constraints, these payments are essential to facilitating enrollment, and without them, clinical progress would be retarded. Without information pertaining to the actual cost of enrolling and reporting the outcomes of subjects in clinical research studies, it is impossible to know whether current payments to physicians are excessive and create a conflict of interest for physicians, or are adequate to cover basic expenses related to conducting clinical research. To our knowledge, there is no published accounting of the time and costs incurred in conducting a clinical research trial. To address this gap in data, we assessed the costs of enrolling patients in a mock, randomized controlled trial of a chemotherapy agent.

I

From the Department of Clinical Bioethics, Warren G. Magnuson Clinical Center, National Institutes of Health, Bethesda, MD; Division of Hematology and Oncology, Beth Israel Deaconess Medical Center, Boston, MA; Department of Public Policy and Practice, American Society of Clinical Oncology, Alexandria, VA; The Lewin Group, Falls Church, VA (current address: Health Strategies Consultancy Washington, DC); and the University of Michigan School of Medicine, Ann Arbor, MI. Submitted August 23, 2002; accepted August 22, 2003. Supported by the American Society of Clinical Oncology, Alexandria, VA. Authors’ disclosures of potential conflicts of interest are found at the end of this article. Address reprint requests to Allen S. Lichter, MD, Office of the Dean, 4101 Medical Science Building I, 1301 Catherine Rd, University of Michigan School of Medicine, Ann Arbor, MI 48109-0624; e-mail: [email protected]. © 2003 by American Society of Clinical Oncology. 0732-183X/03/2122-4145/$20.00

Journal of Clinical Oncology, Vol 21, No 22 (November 15), 2003: pp 4145-4150 DOI: 10.1200/JCO.2003.08.156

Downloaded from ascopubs.org by 37.44.207.106 on January 28, 2017 from 037.044.207.106 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

4145

4146

EMANUEL ET AL Table 1.

Mock Protocol Visit Schedule for a Phase III Placebo-Controlled Trial of a Chemotherapeutic Agent Week No.

Month No.

Required Studies

1

2

3

4

5

6

7

8

9

10

11

12

13

3

6

9

12

History Physical exam Clinical laboratory analyses PSA Chest x-ray CT scan Bone scan Pain medicine stabilization Pain medicine diary* BPI RRFLS KPS FACT-G Progression Study drug Hydrocortisone LHRH antagonist PK sampling Adverse events

X X X X X X X X X X X X X

X X X X

X X X X

X X X X

X X X X

X X X X

X X X X

X X X X

X X X X

X X X X

X X X X

X X X X

X X X X X X

X X X X X X

X X X X X X

X X X X X X

X X X X X X

X X X X X X X

X X X X X X X

X X X X X X X

X X X X X X X

X X X X X X X

X X X X X X X

X X X X X X X X X X X X X X X

X X X X X X X X X X X X X X

X X X X X X X X X X X X X X

X X X X X X X X X X X X X X

X X X X X X X X X X X X X X

X X

X X X X

X X X X X

3 3 3 X X

X

X X

X

X

X

X

X X

X

X

X X

X

X X

X X

NOTE. “X” indicates the occurence of a given study. Arrows indicate the continuation of a study. Abbreviations: PSA, prostate-specific antigen; CT, computed tomography; BPI, Brief Pain Inventory; RRFLS, Revised Rand Functional Limitations Scale; KPS, Karnofsky Performance Status Scale; FACT-G, Functional Assessment of Cancer Therapy—General; LHRH, luteinizing hormone-releasing hormone; PK, pharmacokinetic. *Begun 2 weeks before treatment, and recorded daily.

METHODS

Site Recruitment Three distinct types of sites were identified: academic medical centers, including comprehensive cancer centers; private oncology or multispecialty practices; and group model managed-care organizations. The United States was divided into the four census regions, and within each region, potential participants of each type were identified and contacted regarding their participation. Participating sites were administered both written and inperson surveys.

Mock Protocol We developed a mock phase III randomized research protocol evaluating an experimental chemotherapeutic agent for hormonally refractory prostate cancer against placebo. The protocol entailed a 2-week period to stabilize pain medications, followed by stratification by prostate-specific antigen level and randomization to either placebo, or the experimental drug administered intravenously over 1 hour (Table 1). Five doses would be administered during the first week, two each would be given in the second and third weeks, and then the drug would be administered weekly for a total of 11 weeks of treatment. After completing treatment, the subjects would be observed for 12 additional months with four follow-up visits. Overall, there would be 17 office visits—13 for treatment and four for follow-up. End points to be measured included disease progression, formal pain assessments, performance measures, and quality-of-life assessments. The protocol assumed that each site enrolled 20 subjects.

Activities Related to Clinical Research We delineated 13 separate activities related to the successful management of a clinical trial: (1) pre-IRB submission activities; (2) study initiation; (3) subject recruitment; (4) informed consent; (5) randomization; (6) supply management and dispensing of an investigational drug; (7) office visits for administering the drug; (8) office visits for follow-up care; (9) management, including office visits, related to adverse events; (10) data collection, management, analysis, and reporting; (11) audits and monitoring visits; (12) communication with the sponsor; and (13) poststudy investigator meetings.

Office visits for administering the drug, and follow-up and management of adverse events (activities 7 and 8) were considered clinical care. The other activities were deemed nonclinical.

Data Collection and Validation The Lewin Group, a nationally known medical consulting organization, adapted a methodology previously utilized to determine costs and expenses associated with health care services, to determine the expenses of clinical research (reports by The Lewin Group for the American Academy of Neurology, The Neurology and Psychiatry Practice Expense Study, September 15, 1997; The North American Society of Pacing and Electrophysiology Practice Expense Study, November 25, 1997; The American College of Emergency Physicians Practice Expense Study, August 17, 1998). The methodology includes both a written and an in-person survey that were administered by the Lewin Group. Written survey. Each potential site was mailed a written survey to collect background data on practice characteristics; history of clinical trials participation; general and staff costs, including salaries and fringe benefits for 1999; and experience in receiving third-party reimbursement for clinical trials participation. For staff involved in clinical trials, additional data were obtained for numbers of hours worked per year, percentage of salary paid by the practice, and the percentage of time devoted to clinical trials. In-person survey. After the written survey, researchers from the Lewin Group conducted 1-day in-person meetings with practice setting representatives to delineate the specific staff time devoted to each one of the 13 clinical research–related activities. Representatives were asked to specify the number of hours for each activity, the specific staff involved in each activity, and who paid the staff member in 1999. To standardize data, the responses were related to the mock placebo-controlled phase III research protocol. Because time devoted to particular activities could vary depending on the sponsor, this process was conducted twice— once as though the clinical research trial were sponsored by the National Institutes of Health (NIH), and once as though it were a trial sponsored by the pharmaceutical industry. Data from the written and in-person surveys were combined and validated with each site. Outlier data points were investigated and either corrected or substantiated. In addition, researchers from the Lewin Group “shadowed” the principal investigators, nurse coordinators, chemotherapy nurses, data man-

Downloaded from ascopubs.org by 37.44.207.106 on January 28, 2017 from 037.044.207.106 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

4147

COSTS OF CONDUCTING CLINICAL RESEARCH Table 2.

Characteristics of the Study Sites

Time for Clinical Research Activities

Characteristic

No. of Sites

Type of site Academic medical center comprehensive cancer center Managed-care organization Oncology group practice Geographic region (United States) Northeast Midwest South West Practice size (No. of physician-investigators) ⬍5 5-10 ⬎ 10

2 3 16 3 7 4 7 6 11 4

agers, and others at the four study sites to assess whether the reported time was comparable to the actual observed time. This was not a formal time-motion study, but rather an informal assessment of time devoted to specific activities, such as preparing the IRB submission, explaining the clinical research study and obtaining informed consent, and completing clinical reporting forms.

Converting Time Allocations to Costs The hours spent on each activity by specified staff were converted into costs using the actual 1999 salary and fringe benefit rates provided by each site on the written survey. In less than 15% of cases, salary was not available, usually because there were too few employees performing the activity, and sites were concerned about breaching personnel confidentiality. In these cases, national salary rates (Bureau of Labor Statistics) and an average fringe-benefit rate of 20% (the average of the fringe benefit rates among the sample sites) were used. The 1999 average hourly rates without fringe benefits for the 1997 data inflated to 1999 dollars were: oncology nurse, $29.82; registered nurse, $19.91; medical/laboratory technician, $13.31; pharmacist, $27.88; and physician, $48.52.

RESULTS

Characteristics of the Participating Sites Table 2 provides characteristics of the 21 participating sites.

Table 3.

On average, slightly less than 4,000 cumulative hours (range, 1,512 to 13,319 hours) were required to conduct a randomized, controlled clinical trial of 20 subjects, including time for the clinical care of the enrolled subjects (Table 3). For all clinical research activities, from protocol development and study initiation to administration of the study drug, recording the clinical data, and reporting the data to the sponsor, approximately 200 hours (range, 87 to 785 hours) of all staff time per subject were required. There was no significant difference in the total hours between the NIH- and industry-sponsored clinical studies (Table 3). The distribution of hours by the three major types of professionals involved in clinical research showed that physicians contribute slightly less than 9% of the hours, nurses contribute 32%, and data managers, 36% (Table 3). Other professionals, such as pharmacists and administrative staff, contribute the remaining time. Dividing the time devoted to clinical research trials into clinical and nonclinical care hours showed that 32% and 28% of all hours involved nonclinical activities for industry- and NIHsponsored clinical research trials, respectively. At academic centers, there were 135 nonclinical care hours per patient; at managed-care organizations, there were 113 nonclinical care hours; and at oncology group practices, there were 46 nonclinical care hours. As illustrated by the differences in nonclinical care time, the different types of sites have very different amounts of time devoted to each type of research-related activity (Table 4). Academic centers, including comprehensive cancer centers, utilize approximately three times as many hours per subject as oncology group practices, with managed-care organizations in between. This differential is more substantial for industrythan NIH-sponsored research trials, with academic centers requiring an excess of 310 hours per patient for industrysponsored clinical research trials and 249 hours per subject for NIH-sponsored studies.

Average Time Required for Conducting Randomized, Controlled Clinical Research Study Involving a Drug for 20 Subjects Average No. of Hours for an Industry-Sponsored Clinical Research Trial*

No. of Hours to Enroll and Treat 20 Subjects NIH-Sponsored Activity

Pre-IRB submission and initiation of study Subject recruitment and informed consent Randomization and dispensing drugs Office visits Management of adverse events Data management and analysis Audits and communications with the sponsor Post study follow-up and meetings Total hours

Industry-Sponsored

Average

Range

Average

Range

PhysicianInvestigator

Registered Nurse

Data Manager

All Other Staff

57 201 113 2,846 69 640 65 21 4,012

5-177 63-453 10-295 1,047-10,809 1.5-225 104-2,494 2-156 2-120 1,512-13,319

117 190 156 2,691 72 652 112 48 3,998

6-223 63-223 17-487 988-11,611 14-225 40-3,301 12-235 1-155 1,735-15,699

27 28 3 241 15 12 9 9 344

34 82 56 922 22 133 36 10 1,295

36 51 25 788 30 445 59 21 1,455

20 29 72 740 5 62 8 8 944

NOTE. This is the average number of hours for all 21 sites for a randomized controlled trial of a drug for hormonal-resistant prostate cancer involving 20 subjects, 13 weeks of treatment, and 12 months of follow-up with a total of 17 out-patient visits. Abbreviations: NIH, National Institutes of Health; IRB, institutional review board. *This is the average number of hours across all 21 sites for an industry-sponsored trial.

Downloaded from ascopubs.org by 37.44.207.106 on January 28, 2017 from 037.044.207.106 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

4148

EMANUEL ET AL Table 4.

Average Time Required for Conducting Clinical Trials per Subject Stratified by Type of Site Industry-Sponsored Total (hours)

Activity

Pre-IRB submission and initiation of study Subject recruitment and informed consent Randomization and dispensing drugs Office visits Management of adverse events Data management and analysis Audits and communications with the sponsor Poststudy follow-up and meetings Total

Academic Medical and Cancer Center Total (hours)

Managed-Care Organizations Total (hours)

Oncology Group Practices Total (hours)

Average

Range

Average

Range

Average

Range

Average

Range

6 10 8 135 4 33 6 2 200

0.3-11 3-23 0.9-24 49-581 0.7-11 2-165 0.6-12 0-8 87-785

7 13 8 319 6 94 7 3 454

4-10 13-13 5-11 57-581 6-6 24-165 3-11 1-6 123-785

10 18 6 194 3 67 7 3 307

9-11 9-23 1-13 163-220 2-5 16-147 7-8 2-4 250-391

5 7 8 98 4 17 5 2 144

0.3-9 3-16 1-24 49-158 1-11 2-37 2-37 0-8 87-261

NOTE. This is the average number of hours per subject for a randomized controlled trial of a drug for hormonal-resistant prostate cancer subjects involving 20 subjects, 13 weeks of treatment, and 12 months of follow-up with a total of 17 out-patient visits. Abbreviation: IRB, institutional review board.

Costs for Clinical Research Activities Using 1999 salary and fringe benefit rates, the hours devoted to clinical research trials can be converted into costs (Table 5). On average, it costs slightly more than $6,000 (range, $2,098 to $19,285) per oncology subject to participate in this randomized controlled trial of a chemotherapeutic agent, excluding costs for diagnostic tests, blood work, and overhead. The average clinical costs amount to slightly more that $4,000, with the nonclinical costs constituting $1,999 for an industry-sponsored clinical research trial. The costs are different for different types of sites, with academic medical centers reporting almost three times the total cost of oncology group practices. The nonclinical care cost for academic medical centers amounted to $3,735 per subject; for managed-care organizations, $3,917 per subject; and for oncology group practices, $1,384 per subject. Qualitative Observations of Time for Clinical Research Activities Direct observation of clinical research staff indicated that conducting clinical research is labor intensive, and that representatives tend to underestimate the time required for each activity. For instance, one research nurse indicated that because of the need for precise tests on a fixed schedule dictated by the

Table 5.

research protocol, extensive time was needed to schedule and coordinate the various tests into a subject’s scheduled visit. To ensure that subjects receive all tests, research nurses frequently accompanied individual subjects to each scheduled test. Similarly, physicians complained that there is a significant “hassle factor” with industry-sponsored trials because of frequent protocol revisions that require additional time for informing the IRB, receiving its approval, informing other members of the research team, and changing procedures. In addition, the pharmaceutical company, or more likely, the clinical research organization contracted to conduct the research, frequently calls to inquire about accrual, to monitor completion of clinical reporting forms, to verify outcome values, and to collect additional data. This additional time is consistent with the greater time reported for pre-IRB submission, audits, communication with sponsor, and poststudy follow-up of industry-sponsored clinical research. Finally, in observing the informed consent process, the provision of chemotherapy to patients, and the completion of clinical reporting and other forms, the research team found that the time specified in written surveys by the study site participants was conservative. In each of these steps, hours spent, especially by research nurses, tended to be more than that which was actually reported by the representatives of the

Average Costs Required for Conducting Clinical Trials per Subject Stratified by Type of Site Industry-Sponsored Total Costs ($)

Activity

Pre-IRB submission and initiation of study Subject recruitment and informed consent Randomization and dispensing drugs Office visits Management of adverse events Data management and analysis Audits and communications with the sponsor Poststudy follow-up and meetings Total

Academic Medical and Cancer Center Total Costs ($)

Managed-Care Organizations Total Costs ($)

Oncology Group Practices Total Costs ($)

Average

Range

Average

Range

Average

Range

Average

Range

287 372 203 4,095 184 741 176 97 6,094

7-1,108 63-959 22-469 1,369-14,151 23-632 41-4,003 12-774 1-435 2,098-19,285

285 498 237 8,497 269 2,167 215 198 12,233

113-457 433-561 144-329 2,844-14,151 269-269 432-3,903 100-331 24-373 5,181-19,285

686 713 179 7,125 167 1,816 245 109 11,041

339-1,108 405-959 22-440 5,772-8,505 95-297 341-4,003 197-291 59-195 8,893-13,650

208 287 203 2,902 181 307 156 80 4,285

7-626 63-847 25-496 1,369-4,941 23-632 23-632 12-774 1-435 2,098-8,608

NOTE. These are the average costs per subject for a randomized controlled trial of a drug for hormonal-resistant prostate cancer subjects involving 20 subjects, 13 weeks of treatment, and 12 months of follow-up with a total of 17 out-patient visits. Abbreviation: IRB, institutional review board.

Downloaded from ascopubs.org by 37.44.207.106 on January 28, 2017 from 037.044.207.106 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

4149

COSTS OF CONDUCTING CLINICAL RESEARCH

study sites. In no case were the hours observed fewer than those reported by study site representatives. DISCUSSION

This study systematically determined the actual time and nonclinical costs required to initiate and manage successfully a clinical research trial. The data indicate that enrolling a subject onto a clinical research trial, providing the experimental intervention, monitoring the subject, and recording and processing useable data are time consuming and costly, at least for oncology research trials. Conducting an oncology clinical trial requires at least 13 separate activities. According to this study, these activities require a substantial amount of time, approximately 200 hours for each enrolled subject. Slightly less than one-third of these hours—in excess of 50 hours per subject—are devoted to nonclinical activities related to research, ranging from assembling the IRB submission to managing experimental medications to completing clinical research forms and responding to queries from the sponsor. These data indicate that on average, the nonclinical costs associated with the research-related activities are approximately $2,000 per subject for an oncology research trial. Indeed, at academic and cancer centers, the nonclinical costs of research are slightly less than $4,000 per subject. Currently, the National Cancer Institute (NCI) provides cooperative oncology groups (such as the Cancer and Leukemia Group B [CALGB], Eastern Cooperative Oncology Group, and National Surgical Adjuvant Breast and Bowel Project) with $1,725 per enrolled subject, which is intended to cover both the direct costs of data management as well as administrative and indirect costs; how much of these funds is actually disbursed from the cooperative oncology groups to individual investigators or their institutions depends on the practices of the group and its relationship to investigators (Richard Schilsky, CALGB, personal communication, May, 2002). (The NCI also provides up to an additional $400 for enrolling subjects in “ancillary studies” and an additional $25 for completion of each long-term follow-up form.) This study suggests that such payments do not actually cover the full costs associated with conducting clinical research. Indeed, enrolling subjects in NCI-sponsored research may actually cost physician practices and medical centers. The higher levels of payment to physicians by the pharmaceutical industry may in part be because of the greater “paperwork” demands associated with completing clinical reporting forms, monitoring, and auditing. Furthermore, the higher industry payments may have the inadvertent effect of off-setting and cross-subsidizing the lower payments made by the NCI; profits from industry payments may permit medical centers and practices to conduct NCI-sponsored clinical research without an overall revenue loss. That oncologists and cancer centers continue to participate in clinical research trials even though they may lose money in the process is not surprising for two reasons. First, data from an unpublished survey of 3,063 American Society of Clinical Oncology members indicates that only 22.2% of oncologists, or their practice colleagues, have actually calculated the costs of participating in clinical research (E.J. Emanuel, personal com-

munication, May, 2002). Thus, the vast majority of oncologists have no factual basis for determining whether $1,500 for enrolling a subject in a clinical trial is a net financial gain or loss. Furthermore, at least for oncologists, there is the belief that clinical research trials represent the best oncologic care available for patients. This may mean that to be able to offer the best care for patients, oncologists and cancer centers may be willing to absorb some nonclinical costs. That physicians might make hundreds of thousands of dollars in profit by enrolling many subjects onto clinical research trials seems unlikely given these data. Even enrolling subjects onto pharmaceutical company–sponsored trials, in which payments probably exceed costs, could result in substantial profits for physicians only if the condition being studied is so common that a practice or medical center could enroll many patients who fulfill the eligibility requirements, and only if the trial is relatively short, simple, and with few substantial risks, so that the effort necessary to conduct the trial is modest. These conditions are likely to be met for trials in some common conditions and diseases, but not for a disease such as cancer. There are substantial differences in the overall time and costs required to conduct clinical research trials between group practices and academic and cancer centers. While in general, academic and cancer centers were much more expensive than group practices, for some activities such as subject recruitment, dispensing drugs, management of adverse events, and communication with sponsors, particular group practices had the highest costs. Physicians and nurses at academic centers tended to spend considerably more time during each office visit with subjects than those in group practices, probably reflecting the higher volume of patient care in private practice settings. In group practices, staff tends not to be dedicated to clinical research; the nonclinical research-related activities are often “added” on to existing clinical responsibilities. Conversely, academic centers tend to use personnel dedicated exclusively to clinical research. Further, academic and cancer centers tended to have several layers of data managers and processes for quality assurance, enhancing the data quality with a concomitant financial cost. In addition, academic and cancer centers have offices of research administration involved in study initiation and dispensing of drugs. In addition, in providing clinical care, academic and cancer centers have more professionals interacting with patients for more time than in private practices. They also tended to have higher-level research nurses and physicians performing more of the clinical trial functions, adding to costs per unit of time. This study has several important limitations. First, the small sample size, especially for academic medical and comprehensive cancer centers, limits the generalizability of the data. Substantial efforts were made to recruit additional academic sites, but these efforts were unsuccessful. Clearly with more sites participating the confidence intervals on time and cost estimates would be smaller and the generalizability enhanced. Similarly, small size precludes statistical comparisons. Second, factors associated with lower time and costs of the nonclinical aspects of clinical research, such as whether greater volumes of research produce efficiencies and lower costs and whether computerized systems

Downloaded from ascopubs.org by 37.44.207.106 on January 28, 2017 from 037.044.207.106 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

4150

EMANUEL ET AL

devoted to clinical research lower costs, could not be evaluated. Third, these costs exclude overhead costs, such as those associated with legal advice, operating the IRB, telephone, postage and mailing, and office rental and operations. Therefore, the total costs may be even higher. Fourth, representatives at the sites estimated the time allocations. They may have over- or underestimated the time required for each activity. Our in-person monitoring at four sites suggests that if anything, representatives tend to underestimate the time required for activities, rather than overestimating them. It would be helpful to validate these data with detailed time-motion studies, though this is likely to be complicated and costly given the length of time during which clinical trials are conducted and the variety of staff involved. Finally, these data apply to a mock randomized, placebocontrolled trial of a chemotherapeutic agent measuring disease progression as well as many quality-of-life end points. The precise time required and the cost estimates for other types of

oncologic trials may be significantly different. More importantly, the precise time and costs required for relatively simpler nononcologic trials, for example, of drugs for hypertension or migraine headaches, may be substantially different. There are 13 separate activities involved in completing an oncology clinical research trial. Based on the mock trial we conducted, it can be concluded that nontreatment activities require substantial time (on average, 61 hours per subject enrolled), and the associated costs are substantial. Given the key role clinical trials play in developing new therapies and improving existing ones, the cost of conducting various types of clinical trials is an area that requires further study. AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The authors indicated no potential conflicts of interest.

REFERENCES 1. National Institutes of Health Office of Extramural Research: http:// grants.nih.gov/grants/policy/emprograms/overview/ep-coi.htm 2. Spiro HM: Mammon and medicine: The rewards of clinical trials. JAMA 255:1174-1175, 1986 3. Lind SE: Finder’s fees for research subjects. N Engl J Med 323:192195, 1990 4. Shimm DS, Spece RG: Industry reimbursement for entering patients into clinical trials: Legal and ethical issues. Ann Intern Med 115:148-151, 1991

5. Lind SE: Financial issues and incentives related to clinical research and innovative therapies, in HY Vanderpool (ed): The Ethics of Research Involving Human Subjects: Facing the 21st Century. Frederick, MD, University Publishing Group, 1996, pp185-202 6. Eichenwald K, Kolata G: Drug trials hide conflicts for doctors. New York Times, May 16, 1999, A1, A28

Downloaded from ascopubs.org by 37.44.207.106 on January 28, 2017 from 037.044.207.106 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.