Contract research organizations in oncology clinical research: Challenges and opportunities
Abstract
Contract research organizations (CROs) represent a multibillion dollar industry that is firmly embedded in the contemporary clinical trial process. Over the past 30 years, and especially within the last decade, the reach of CROs has extended to service all phases of drug trials in an increasingly global research environment. The presence of CROs is particularly noticeable in medical oncology because of the large number of investigational compounds developed to treat cancer that are currently undergoing testing in human subjects. Although limited data are available with which to objectively define the effects that CROs have had on the clinical trial process, with the expansion of these organizations, several reports have called into question whether ethical and professional standards in research conduct are at times secondary to economic considerations. CROs can add considerable value to the clinical trial process, but difficulty communicating with CRO representatives and time spent answering trivial data queries generated by CROs are current obstacles for study site personnel interacting with CROs. Further study of the effect of the CRO industry on the clinical trial process is needed to ensure efficient data collection and patient safety while collaboratively developing novel therapies in an expedited fashion. Cancer 2016;122:1476–82. © 2016 American Cancer Society.
INTRODUCTION
Contract research organizations (CROs) are private corporations that perform a wide variety of clinical research-related duties and functions on behalf of biotechnology, pharmaceutical, and medical device companies sponsoring studies in human subjects.1 Because of economic and regulatory factors, CROs are playing an increasingly visible role in clinical research and the management of clinical trials, such that interactions with CRO personnel and CRO-originated queries regarding study participants have become a ubiquitous presence in the professional lives of clinical investigators, especially in fields such as medical oncology, in which there are currently a large number of investigational compounds and ongoing interventional trials.
High-quality CROs have the potential to add substantial benefit to the clinical trial process, including improving the quality of data collection and trial standardization. After more than 30 years of CRO activity, the concept appears to be here to stay. However, interactions between study site teams and CROs are not always efficient or productive. CROs are widely perceived as adding bureaucratic burden to trial conduct, increasing costs, and contributing to delays in protocol conduct, without a demonstrable benefit in the quality of data collection or improvements in patient safety.2-4 Investigators also have expressed concern regarding a creeping process of “research fundamentalism” inhibiting cancer research, to which the micromanagement culture of some CROs contributes.4, 5 In this article, we survey the activities of CROs in the contemporary clinical trial landscape, describe the benefits that CROs offer and the risks that they can pose to safe and ethical research conduct, highlight the burden that some CROs currently impose on clinical investigators, and outline potential opportunities for improvement.
Origin and Expansion of CROs
The birth of the CRO can be traced to 1982. In that year, Quintiles was founded by a biostatistician from the University of North Carolina to provide fee-for-service consultative data support for clinical trials, and a newly formed company, Theradex Oncology Experts, received the clinical trials monitoring service contract from the National Cancer Institute.6, 7 In the 1980s, both the National Cancer Institute and large pharmaceutical companies were searching for ways to address rising research and development costs, including increasingly detailed requirements for New Drug Application submission to regulatory agencies such as the US Food and Drug Administration (FDA) and the various national precursors of the European Medicines Agency,8 at a time when available capital for research was perceived as increasingly limited.9 CROs were envisioned as a form of outsourcing, a broader corporate trend of subcontracting that originated during the same period (the term “outsourcing,” a contraction of “outside resourcing,” dates to 1981).10 CRO outsourcing provided “spillover capacity” for data management and biostatistical analysis during peak activity periods, when the pharmaceutical companies’ own employees had insufficient capacity to complete all necessary tasks.
By the 1990s, the scope of CRO activities had expanded to service all aspects of Phase 2 to 4 trials, including interfacing with investigators, collecting samples for laboratory analysis and conducting analyses, and assisting with dealer management systems and the preparation of regulatory submissions. The CRO market has grown from an estimated $2 billion in 1993 to $10 billion in 2005 and $25.3 billion in 2013, a substantial percentage of the estimated annual clinical research and development expenditures of $130 billion made by the pharmaceutical industry in that latter year.11-13 The increasingly prominent role of CROs in medical research also is evidenced by the shift of sponsors from using “preferred providers,” longitudinal relationships with a single CRO, to choosing “strategic partnerships” with CROs on a rotating basis, with contract terms of up to 3 to 5 years.11
Recent estimates indicate that by 2020, greater than 70% of clinical trials will be conducted by CROs, with the largest 4 CROs, Quintiles, Parexel, Pharmaceutical Product Development LLC, and INC Research, currently associated with the lion's share of trials.14 Other smaller CROs, including Covance Inc, Medpace Inc, PRA Health Sciences, inVentiv Health, and Chiltern International Ltd, were responsible for approximately 20% of trials in a 2014 survey.15 A 2015 survey predicted the CRO compound annual growth rate at 9.83% during the next 5 years.
Despite this high market penetrance and continued growth, details regarding the activities of CROs and their influence on research conduct have been poorly studied to date. For example, to our knowledge, there currently are no mechanisms with which to comparatively evaluate the quality, efficiency, competence and expertise, accuracy of data collection, and cost-effectiveness of CROs involved in cancer research.16
Several major forces are driving the continued increased market share of CROs, including the globalization of clinical trials, the growth in narrowly used and costly specialty drugs, and the increasing challenge and cost to pharmaceutical companies of hiring skilled benefits-eligible permanent workers. In 2004, approximately 60% of CRO revenue came from the United States, whereas approximately 40% was international; in 2013, these numbers were nearly reversed at 43% and 57%, with a growing role of CROs noted in emerging clinical trial markets such as India, China, and Central and Eastern Europe.9, 11 “Specialty drugs” requiring expertise in delivery routes, rare diseases, or genomic and other biomarkers are projected to account for 42% of drug costs in 2016, up from 23% in 2011, and CROs are increasingly offering extended services related to approved specialty drugs. These postapproval commitments represent an important source of ongoing growth for the CRO industry.17
The explosive growth of CROs in the last 20 years rests primarily on economic grounds. Conversely, increasing product development safety and improving clinical trial scientific methodology have not yet been key drivers of trends in CRO use. The lengthy time to new drug approval and the high cost of conducting all phases of clinical trials are major challenges for sponsors; according to a review by the Tufts Center for the Study of Drug Development, drug development times are faster with “high CRO usage” sponsors.18 However, despite the growth of CROs, the cost of drug development still substantially increased between 2003 to 2014, from a reported median of $802 million to $2.6 billion per new approval,19 although these figures have been widely challenged.20-22
Potential Risks of CROs With Respect to Study Conduct
To the best of our knowledge, it remains unclear how the conduct and integrity of clinical trials might be affected by the increased use of CROs. CROs could potentially increase the quality of collected data and rate of trial progress. However, although high-quality data are lacking, observational reports and investigative journalism have uncovered several concerns that merit attention from the clinical research community.
Existing data regarding CROs
Few studies to date have compared traditional academic organization-pharmaceutical sponsor partnership research structures with the CRO model. In one analysis of CROs, semistructured interviews were conducted with employees at 25 US private-sector trial sites in the Southwest United States between October 2003 and September 2004.23 The Southwest was chosen because of a CenterWatch database indicating a “preponderance” of private-sector clinical trials in the area.23 The interviews were conducted until “saturation was reached,” meaning that the investigators found that no new useful insights were emerging from further discussion. The lead investigator, who was trained in qualitative sociologic methods, assessed interview transcripts and provided all interviewees with the full transcript of their interviews and offered an opportunity for them to edit their transcripts. The 62 informants included 12 principal investigators (PIs), 18 study coordinators, 3 patient recruiters, 7 administrators, and 22 research participants and representatives of the pharmaceutical companies. The physicians interviewed were representative of the national demographics of industry-funded investigators. Data analysis was based on grounded theory, with data analysis involving coding field notes and interviews for emerging categories and themes.
A chief finding from these interviews was that: “The physicians perceived their roles in terms of business rather than science…physicians saw their primary ethical responsibility as providing accurate data to the companies that hired them and did not explicitly refer to their ethical responsibility to trial participants.”23 However, the research model and incentives may differ depending on the setting of the trial. For example, salaried academic investigators might view clinical trials as a bridge to career advancement, whereas financial incentives might be a more important motivation to private-sector physicians.23
Although to our knowledge there are no data available regarding publication bias as it relates specifically to CROs, multiple analyses have indicated a strong association between pharmaceutical funding and the preferential publishing of results that favor the interests of those companies.24-26 Given their principal source of funding, the priorities of CROs are more likely to be aligned with the interests of sponsors than with enrolled subjects or enrolling physicians. Contractual agreements reached by CROs with pharmaceutical companies have raised concerns regarding the inappropriate control of data by industry sponsors, prompting the International Committee of Medical Journal Editors to express concern regarding the conduct of CROs with regard to trial design, access to raw data, and publication rights.27
Communication and bureaucratic hurdles
Medical professionals and administrative staff with experience interacting with CROs often report frustrating logistical and communication issues. To understand these concerns better, we interviewed a convenience sample of nonphysician study personnel, including regulatory coordinators, Institutional Review Boards, clinical research coordinators, research nurses, and data managers at 4 academic institutions (Dana-Farber Cancer Institute/Brigham and Women's Hospital, Mayo Clinic, Massachusetts General Hospital, and The University of Texas MD Anderson Cancer Center). Although the interviewees stated that some CROs had been highly responsive to queries, the most commonly cited concern was the tedious process of adverse event (AE) reporting, which requires clinicians to attribute the likelihood of an AE to an effect of a study drug or a combination of drugs about which little is known, the distribution of AE descriptions to all sites and Institutional Review Boards involved in the trial, and possibly multiple follow-up filings detailing the subsequent clinical course of the patient.5 Numerous queries regarding patient medical history and concomitant medications serve to increase data points collected without necessarily providing increased clarity regarding the effectiveness or AE profile of the drug.
One senior investigator interviewed by Duke University social scientist Jill Fisher suggested that the query process may be due to CROs needing to impress pharmaceutical companies: “CROs, it seems, always have to prove to the company that they're doing a nice job, and they look at minutia so much so that in some ways (it) can kill you.”28 In contrast to CROs, the investigator stated that with regard to pharmaceutical sponsors, moving toward drug approval is the highest priority and, “petty things are not on their agenda.”28
Trial staff at sites in which research is conducted also often expressed frustration arising from on-site coordinators' interaction with monitors, those employees assigned by the CRO to serve as a liaison with clinical facilities enrolling research subjects. Monitors occasionally have limited clinical experience, yet are tasked with coordinating complex trials and collecting data from multiple sites. Monitors also frequently are under pressure from their CRO to fulfill an internal set of metrics and generate a certain number of queries to be fully compensated, because many CROs contract via a unit-based payment model in which CROs are paid by the sponsor for completing specific tasks, including the resolution of queries.29
Situations have arisen in which monitors have undermined on-site coordinators by e-mailing on-site physicians stating that an on-site study coordinator has violated the contract because clinically irrelevant metrics related to these queries have not been met.30 At the same time, on-site coordinators reported a frequent inability to obtain timely responses from monitors regarding clinically pressing questions related to trial conduct, which has resulted in treatment delays that occasionally have affected patient care. Site personnel describe instances in which they have had to contact the sponsor directly to have questions answered, bypassing an assigned but unresponsive CRO. In several cases, monitors assigned by CROs to oversee studies conducted in the United States did not speak English proficiently and were unable to communicate effectively with study staff.
The frequent high turnover rate noted for monitors (due to their relatively low pay and often-cited desire to be hired by a pharmaceutical company because such companies usually provide greater upward mobility and employee benefits than CROs) compounds this process for the on-site study coordinators and clinical investigators. Fisher described the hiring process and transient nature of the CRO monitor position.28 The majority of monitors are hired with little prior experience and usually are pending or recent college graduates. CRO monitors often aspire to and leave for pharmaceutical positions, whereas others move within the CRO industry to other monitoring positions with more favorable benefits. Subsequently, according to Fisher, “Most drug studies with a duration of more than six months will have at least one change in monitor over the course of the study, and it is not uncommon to hear about four or more monitors assigned to studies that last more than a year…Roughly 60 percent of all monitors change companies within three years, and nearly 80 percent of all monitors are no longer in monitoring positions at all after five years.” Because of the high turnover rate of monitors and their relative inexperience with cancer generally or the specific tumor type under study, some monitors are perceived to hound researchers, as stated by one oncology PI we interviewed, “with the zeal of Inspector Javert (from Les Miserables) and the competence of Inspector Clouseau (from The Pink Panther).”
CRO personnel who are responsible for prestudy site selection visits also often have little, if any, clinical experience, a finding backed by Fisher's investigations.28 Numerous study staff and PIs describe being asked the same questions on multiple CRO-led “feasibility visits,” such as whether their institute has a -70 °C freezer or an on-site investigational pharmacy, even when the study site is a major comprehensive cancer center that has other trials concurrently open and accruing patients with the same CRO. In 2015, several major centers, including Weill Cornell Medical Center, decided to institute a policy of no longer conducting in-person preselection feasibility visits because these visits were believed to be inefficient (G. Roboz, personal communication, November 17, 2015). Similarly, site initiation visit personnel frequently read from prepared slides and are reported to be unable to answer basic questions regarding study conduct.
Ethical concerns
Ethical concerns regarding certain CRO practices have arisen, although these may be exceptional cases. One defunct CRO, SFBC International, converted a 675-bed Holiday Inn hotel in Miami, Florida, into a clinic in which undocumented immigrants were paid to receive a study drug in a trial overseen by an unlicensed physician. The resulting scandal, reported by Bloomberg Markets in December 2005 (http://www.journalism.columbia.edu/system/documents/523/original/2006_Evans_Big_Pharma_s_Shameful_Secret_MAG.pdf), was investigated by a committee of the United States Senate and resulted in the firm changing its name to PharmaNet Inc, which was later acquired by inVentiv Health Clinical.31 The hotel was subsequently demolished for fire and safety concerns.32 At another SFBC International site in Montreal, Quebec, Canada, a patient with active tuberculosis was allowed to participate in a trial without respiratory isolation, leading to 9 other participants testing positive for tuberculosis at a later date.33 Other newsworthy events include the CRO Parexel hiring an undertrained physician to administer an investigational monoclonal antibody without after-hours clinical coverage; 6 subjects enrolled by the physician developed organ failure.33
Although these examples are egregious, because monitors from CROs often lack clinical experience, they may not notice genuine ethical concerns. For example, one clinical investigator from the Southern California Research Institute made up patients, failed to conduct physical examinations on some patients and falsified results on others, and substituted employee information for patient data. In September 1998, this investigator was found guilty of fraud, sentenced to 15 months in prison, banned from research for 20 years, and ordered to repay $800,000 to several pharmaceutical companies. During this fraudulent activity, dozens of monitors assigned to the trials were blind to the actions of the Southern California Research Institute's investigator.28
For monitors, their technical training is usually not enough to allow them to accurately detect fraud. Instead, there is a large emphasis on intuition. Several of the monitors interviewed by Fisher described a strong emphasis on intuition in their work: “You sort of get a feel for-when you see patients really go in and out of the site-you get a feel for whether the site is a legit site or not.”28
Outsourcing of clinical trials to the developing world is a potential risk for CROs.34 Medical ethicists have cited the limited ability of regulatory bodies and Institutional Review Boards to monitor the integrity of trial participants and the methods of data collection in far-flung sites.35 In 2008, the US Department of Health and Human Services (DHHS) and the Office of the Inspector General (OIG) released a report on the FDA's capability of monitoring foreign clinical trials. The DHHS and the OIG found that although approximately 80% of marketing applications contained data from clinical trials with sites located outside of the United States, and greater than one-half of all subjects in pivotal trials leading to US drug approvals come from outside the United States, the FDA inspected far fewer sites in foreign countries compared with those with domestic locations.35 The DHHS/OIG report also found that the FDA might be unaware of some ongoing, early-phase clinical trials because they are being conducted without an Investigational New Drug application. Other reasons for the oversight limitations of the FDA involving sites outside of the United States include data submission occurring in nonstandard formats and sponsors failing to provide details regarding site locations and subject enrollment in clinical study reports.35
Many CROs have joined international regulatory bodies and been found to be in compliance with standards set by the Organisation for Economic Co-operation and Development.9 Nevertheless, concerns remain that poor economic conditions in foreign markets may lead to patients being enrolled on CRO-coordinated clinical trials due to a lack of other care options, often without knowledge that they are entering onto a trial.36
Potential Benefits
CROs provide an additional labor pool that is rapidly available during times of increased sponsor clinical activity, and certain aspects of CRO activities have the potential to be beneficial for drug development, including faster development times. Independent monitors who are well trained can provide benefit with respect to the integrity of trials by adding a layer of professional oversight to data collection. The insight of experienced monitors into trial conduct can be helpful to site coordinators, who also are often clinically inexperienced. Central laboratory analysis provided by CROs helps to standardize study results across multiple study sites. Although irrelevant clinical data also are routinely collected, CROs do ensure that key data necessary for regulatory submission are collected.
However, the concerns outlined above do call into question whether CROs as currently constituted add an unnecessary layer of complexity to a clinical trial process already burdened by bureaucracy. In addition, since the emergence of the CRO industry, there clearly is a potential risk to scientific integrity and ethical standards for sponsors conducting research through outside entities. This highlights the importance of ensuring clear and thorough contract negotiations between all parties to provide a strong framework with which to maintain scientific and ethical integrity.
It is important to note that the majority of physicians who work for CROs, and the CRO industry itself, appear to strive to produce high-quality research to benefit society. As one group pointed out in a recent essay, there are many positive aspects to nonacademic research efforts, including rigorous rules for data collection, the absence of the high-pressure environment of academia, and the opportunity to work in a variety of clinical areas.8 The academic setting contains numerous detrimental bureaucratic hurdles to research as well, including the “passing on” of projects from grant-obtaining senior investigators to subordinates, resulting in researchers with less experience performing the studies. A better relationship and cooperation is needed between academic and nonacademic research spheres: “2 is better than 1.”8
Opportunities for Improvement
To the best of our knowledge, there has been only limited scientific study of CROs. In the absence of objective data, the Association of American Medical Colleges has issued 2 reports with suggestions and recommendations to maintain the integrity of clinical trials, including proposed safeguards for registering and reporting trials and guidelines to ensure that study methodology, data analysis, and research reporting maintain integrity.37 The FDA and European Medicines Agency have suggested that it is desirable that data monitoring committees for clinical trials prepare additional statistical reports by independent statisticians, such as an academic institution or government agency, that are not associated with the commercial sponsor or CRO. However, to address the many challenges brought about by the privatization of clinical trials with CROs, it most likely will take more than loose recommendations; stronger government regulatory action mandating these safeguards may be required, as well as clarity regarding what those regulations do and do not require. CROs may collect trivial data because of confusion or uncertainty regarding just what data are strictly necessary to comply with regulations. Modifying policy will require serious political effort as well as time.
More data concerning CRO-driven research is needed regarding publication bias, fraud rates compared with non-CRO trials, and adherence to regulations before drawing specific conclusions regarding the objective effects of this trend in the clinical trials process. In the meantime, we believe there are a few practical measures that, if instituted, could address many of the aforementioned concerns.
To ensure smooth communication between research sites and CROs, monitors should be required to undergo specific training regarding the clinical application of the study drug. They also should be required to speak the language of the site conducting the trial to a basic level of competence. Monitors should be held to metrics important for study conduct and patient safety, such as response time to inquiries from clinical staff, rather than more easily measurable but potentially abused metrics such as the number of data queries generated and resolved. In the event that CRO monitors are unresponsive to site coordinators, CROs should always provide a direct path with which to contact the sponsor. Prestudy site feasibility visits are often unnecessary, especially if the study site is well established with the specific CRO from prior trials. Personnel coming to sites for study site initiation visits should be knowledgeable regarding the trial design and the investigational product and be able to answer questions concerning study conduct.
Reports that CRO reimbursement models are beginning to change from unit-based or task-based pricing to a “desired outputs” model may represent a change toward rewarding clinically meaningful quality standards, and away from a list of checkboxes that result in many of the communication problems mentioned above.29 However, to our knowledge, these quality standards are as yet ill-defined. Should these changes become widely implemented in contracts between CROs and their sponsors, we strongly encourage representation and input from trial site leaders in both academic institutions and community practices to ensure further efficiency for data collection and patient safety.
Finally, although the value of such indices is uncertain, a “CRO rating index” could be tested. This index could include the transparent disclosure of metrics such as efficiency, monitor turnover frequency, the percentage of CRO-managed studies that resulted in peer-reviewed publication, experience with specific cancer types, and cost. Such an index could help sponsors to identify those CROs that are not only the least expensive but deliver the best services. Contracts between investigators and CROs could include language imposing a surcharge to CROs for excessive queries or too many changes in CRO personnel, or other controls.
Conclusions
CROs are playing an increasingly important role in clinical trial conduct in oncology and have the potential to add benefit to the research enterprise, but the potential effects of CRO use on study safety, ethical conduct, and efficiency are understudied. Concerns raised by study personnel and occasional reports of misconduct suggest that the current CRO model is flawed and that new standards for CRO monitors are needed. Further analysis should aid in modifying interactions between CROs and clinical sites and achieve universally valued goals, including the rapid development of safe and effective novel therapies.
FUNDING SUPPORT
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
David P. Steensma has received personal fees from Celgene for registry consulting and personal fees from Amgen and Takeda for participation in Data Safety Monitoring Committees, and has acted as a trial design consultant for Novartis for work performed outside of the current study.
AUTHOR CONTRIBUTIONS
Daniel A. Roberts: Conceptualization, methodology, investigation, resources, writing–original draft, writing–review and editing, visualization, and project administration. Hagop M. Kantarjian: Conceptualization, investigation, writing–original draft, and writing–review and editing. David P. Steensma: Conceptualization, methodology, writing–original draft, writing–review and editing, and supervision.
