Department of Health and Human Services

OFFICE OF 

INSPECTOR GENERAL 


THE FOOD AND DRUG 

ADMINISTRATION’S OVERSIGHT 

OF CLINICAL TRIALS


Daniel R. Levinson

Inspector General 

September 2007

OEI-01-06-00160


Office of Inspector General

http://oig.hhs.gov

The mission of the Office of Inspector General (OIG), as mandated by Public Law 95-452,
as amended, is to protect the integrity of the Department of Health and Human Services
(HHS) programs, as well as the health and welfare of beneficiaries served by those
programs. This statutory mission is carried out through a nationwide network of audits,
investigations, and inspections conducted by the following operating components:

Office of Audit Services
The Office of Audit Services (OAS) provides all auditing services for HHS, either by
conducting audits with its own audit resources or by overseeing audit work done by others.
Audits examine the performance of HHS programs and/or its grantees and contractors
in carrying out their respective responsibilities and are intended to provide independent
assessments of HHS programs and operations. These assessments help reduce waste,
abuse, and mismanagement and promote economy and efficiency throughout HHS.

Office of Evaluation and Inspections
The Office of Evaluation and Inspections (OEI) conducts national evaluations to provide
HHS, Congress, and the public with timely, useful, and reliable information on
significant issues. Specifically, these evaluations focus on preventing fraud, waste, or
abuse and promoting economy, efficiency, and effectiveness in departmental programs.
To promote impact, the reports also present practical recommendations for improving
program operations.

Office of Investigations
The Office of Investigations (OI) conducts criminal, civil, and administrative
investigations of allegations of wrongdoing in HHS programs or to HHS beneficiaries
and of unjust enrichment by providers. The investigative efforts of OI lead to criminal
convictions, administrative sanctions, or civil monetary penalties.

Office of Counsel to the Inspector General
The Office of Counsel to the Inspector General (OCIG) provides general legal services to
OIG, rendering advice and opinions on HHS programs and operations and providing all
legal support in OIG's internal operations. OCIG imposes program exclusions and civil
monetary penalties on health care providers and litigates those actions within HHS.
OCIG also represents OIG in the global settlement of cases arising under the Civil False
Claims Act, develops and monitors corporate integrity agreements, develops compliance
program guidances, renders advisory opinions on OIG sanctions to the health care
community, and issues fraud alerts and other industry guidance.

I N T R O D

Δ

U C T

I O N

E X E C U T I V E

S U M M A R Y

OBJECTIVE
1.	 To determine the extent to which the Food and Drug Administration
(FDA) conducted inspections of clinical trials from fiscal year
(FY) 2000 to FY 2005.
2.	 To assess FDA’s processes for inspecting clinical trials.

BACKGROUND
The Federal Food, Drug, and Cosmetic Act generally requires all new
drugs and medical devices (hereinafter referred to as investigational
products) to undergo clinical trials on human subjects to demonstrate
the safety and efficacy of these products before they are approved for
sale in the United States. The sponsors, clinical investigators, and
institutional review boards (IRBs) that conduct and oversee these trials
must comply with FDA regulations designed to protect the human
subjects participating in them. The Office of Inspector General (OIG)
received a congressional request to review FDA oversight of clinical
trials after a series of news articles highlighted vulnerabilities.
Three FDA centers regulate medical investigational products for human
use: the Center for Drug Evaluation and Research, the Center for
Biologics Evaluation and Research, and the Center for Devices and
Radiological Health. FDA’s Office of Regulatory Affairs (ORA) conducts
onsite bioresearch monitoring (BiMo) inspections of sponsors, clinical
investigators, and IRBs as assigned by the centers.
BiMo inspections are not required by Federal regulations; the centers
decide when to assign them. Inspections can result in one of three
classifications: no action indicated (NAI), voluntary action indicated
(VAI), or official action indicated (OAI). After the inspection, BiMo
investigators recommend a classification for the inspection, and the
assigning center reviews the inspection report and assigns a final
classification. Although FDA takes no additional action for inspections
classified as NAI, it may take additional action for VAI and OAI
inspections.
We used seven data sources for this study: BiMo inspections data; file
reviews of all inspections that BiMo investigators or a center classified
as OAI; an e-mail survey of BiMo investigators; interviews with FDA
officials; observations of BiMo inspections; analysis of the National

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

i

E X E C U T I V E

S U

M M A R Y

Institutes of Health (NIH) clinical trial registry; and reviews of FDA
policies, procedures, and guidance documents.

FINDINGS
Data limitations inhibit FDA’s ability to effectively manage the BiMo
program. Because FDA does not maintain a clinical trial registry, it is
unable to identify all ongoing clinical trials and their associated trial sites.
Further, because FDA does not maintain an IRB registry, it is unable to
identify all IRBs. Even though FDA maintains six databases to track
BiMo inspections, none includes complete information needed to track all
such inspections. For example, ORA’s database does not include complete
information for inspection events that occur after it completes its
inspection. The center databases do not consistently track inspection
information.
Other factors hinder FDA’s ability to effectively manage the BiMo
program. Centers and ORA inconsistently classify OAI and NAI
inspections. FDA relies on voluntary compliance to correct violations of
regulatory significance. Uncertainty of timing and lack of coordination
impede FDA’s ability to conduct BiMo inspections. In addition, FDA
guidance and regulations do not reflect current clinical trials practices.
We estimate that FDA inspected 1 percent of clinical trial sites
during the fiscal year 2000–2005 period. FDA conducted 2,856 BiMo
inspections that required a clinical trial site visit during the
FY 2000–2005 period. Because FDA cannot identify the total number of
clinical trial sites, we used the NIH clinical trial registry to estimate the
proportion of clinical trial sites the BiMo inspections reached. The
centers conduct more inspections that verify clinical trial data than
inspections that focus on human subject protections. Seventy-five
percent of the BiMo inspections during the FY 2000–2005 period were
surveillance inspections, which generally target previously completed
trials and often focus on verifying the quality of clinical trial data. Also,
FDA inspected few IRBs, which offer considerable oversight of human
subject protections.

RECOMMENDATIONS
To improve the BiMo information systems and processes, FDA should
take the following actions:

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

ii

E X E C U T I V E

S U

M M A R Y

Improve information systems and processes.
Develop a clinical trial database that includes all clinical trials. FDA should
develop a comprehensive internal database of clinical trials to more
effectively identify and target ongoing clinical trials for inspection.
Create an IRB registry. This registry would give FDA basic information
about IRBs that it now lacks. By identifying all IRBs overseeing clinical
trials, FDA could target IRBs more effectively for inspection.
Create a cross-center database that allows complete tracking of BiMo
inspections. A database that includes timely and complete information
about all BiMo inspections would help FDA better coordinate and track
inspections.
Establish a mechanism to provide feedback to BiMo investigators on their
inspection reports and findings. Improved feedback between the centers and
BiMo investigators could lead to a common understanding of the
regulations and guidelines that govern BiMo inspections.
Seek legal authority to provide oversight that reflects current clinical trial
practices. FDA should consider seeking additional authority that covers
all of the stakeholders in the management and conduct of clinical trials.
In particular, FDA could seek to expand its authority to include the
colleagues and subordinates of a clinical investigator if they participate
in the conduct of a clinical trial.

AGENCY COMMENTS AND OFFICE OF INSPECTOR GENERAL
RESPONSE
FDA concurred with four of our five recommendations listed above.
FDA did not address our recommendation to establish a mechanism to
provide feedback to BiMo investigators on their inspection reports and
findings.
FDA pointed out that BiMo inspections make up only one part of its
efforts to ensure human subject protections, noting that it views its
protocol review before a clinical trial commences as the most important
step in protecting human subjects. We recognize the important role
that FDA’s protocol review plays in protecting human subjects. We
note, however, that this report addresses another important part of the
system for protecting human subjects: oversight of the trials once they
are actually underway.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

iii

E X E C U T I V E

S U

M M A R Y

The agency also highlighted the efforts of its Human Subject
Protection/Bioresearch Monitoring Council and emphasized the
importance of risk-based approaches to BiMo inspections.
Where appropriate, we made changes to the report based on FDA’s
technical comments.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

iv

Δ

T A B L E

O F

C O N T E N T S

EXECUTIVE SUMMARY .....................................i


INTRODUCTION ........................................... 1


F I N D I N G S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 


Data limitations inhibit FDA’s ability to manage effectively . . . . 10 

Other factors hinder FDA’s ability to manage effectively . . . . . . . 13 

We estimate that FDA inspected 1 percent of clinical trials . . . . . 18 

R E C O M M E N D A T I O N S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 


Agency comments and Office of Inspector General response . . . . 24 

A P P E N D I X E S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 


A: Methodology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 

B: Agency comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 


A C K N O W L E D G M E N T S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 


I N T R O D

Δ

U C T

I O N

I N T R O D U C T I O N

OBJECTIVE
1.	 To determine the extent to which the Food and Drug Administration
(FDA) conducted inspections of clinical trials from fiscal year
(FY) 2000 to FY 2005.
2.	 To assess FDA’s processes for inspecting clinical trials.

BACKGROUND
The Federal Food, Drug, and Cosmetic Act generally requires all new
drugs and medical devices (hereinafter referred to as investigational
products) to undergo clinical trials on human subjects to demonstrate
the safety and efficacy of these products before they are approved for
sale in the United States.1 2 FDA has promulgated regulations to
protect the rights, safety, and well-being of the human subjects who
participate in these trials. These regulations apply to the sponsors,
clinical investigators, and institutional review boards (IRBs) that
conduct or oversee clinical trials for investigational products. FDA
inspects clinical trials to determine whether each of these groups
complies with the relevant regulations.
The Office of Inspector General (OIG) received a congressional request
to review FDA’s oversight of clinical trials after a series of news articles
highlighted vulnerabilities.3 The series identified problems with FDA’s
oversight of clinical trials, including insufficient informed consent
procedures, inadequate training and certification requirements for
IRBs, limited Federal regulations, and FDA’s failure to enforce existing
regulations.
Clinical Trials
Sponsors. The person or entity responsible for developing and testing
an investigational product is the product’s sponsor. Sponsors of drug
and biological products must file an investigational new drug (IND)
application with FDA before they can begin clinical trials; device
1 In some instances devices do not go through a formal approval process but go to market

through a premarket notification and determination of substantial equivalence by FDA. In
these instances, clinical trials may not be necessary. See Federal Food, Drug, and Cosmetic
Act of 1938, P. L. No. 75-717, 52 Stat.1040 (1938) (amended 2004).
21 U.S.C. § 360(o).
2 21 U.S.C. §§ 355(i), 360(j).
3 David Evans, Mike Smith, Liz Willen, “Drug Industry Human Testing Masks Death,

Injury, Compliant FDA,” Bloomberg News, November 2, 2005.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

1

I N T R O D

U C T

I O N

sponsors submit investigational device exemptions (IDE).4 INDs and
IDEs provide FDA with information on the study protocol, the
qualifications of trial personnel, and assurances that trials will protect
human subjects’ welfare, among other details.5 A sponsor may begin its
clinical trial 30 days after FDA receives an IND or IDE, provided that
the agency does not place the study on clinical hold.6
FDA regulations require sponsors to select qualified clinical
investigators to conduct the trials needed to bring an investigational
product to the market.7 A clinical trial generally involves many clinical
investigators working in multiple trial sites, including sites outside the
United States.
Sponsors must also ensure that proper monitoring occurs throughout
the clinical trial.8 In part, this means that sponsors must ensure that
clinical investigators comply with the relevant FDA regulations. Key
FDA regulations pertaining to clinical trials address informed consent
procedures, data management practices, and clinical trial oversight
processes.9
If the clinical trials demonstrate the investigational product to be safe
and effective, sponsors that wish to market a product in the United
States must submit a new product application to FDA.10 11 Provisions
within the Prescription Drug User Fee Act of 1992 (PDUFA) and the
Medical Device User Fee and Modernization Act of 2002 (MDUFMA)
require FDA to expedite the process for the review of product
applications.12 FDA reviews the clinical trial data and may choose to
inspect facilities as part of the application review process.13 If FDA
4 21 CFR § 312.20 (drugs and biologics); 21 CFR § 812.20 (medical devices). 

5 21 CFR § 312.23.


6 21 CFR § 312.40 (drugs and biologics); 21 CFR § 812.30 (medical devices). 

7 21 CFR § 312.50 (drugs and biologics); 21 CFR § 812.40 (medical devices). 

8 Ibid. 

9 Ibid. See also FDA, “Compliance Program Guidance,” Chapter 48: Sponsors, Contract 

Research Organizations and Monitors Part 1—Background (February 2001). Available
online at http://www.fda.gov/ora/compliance_ref/bimo/7348_810/default.htm. Last accessed
on January 16, 2006.
10 The drug development and approval process may take several years. For more

information, see Congressional Research Service, “The U.S. Drug Approval Process: A
Primer,” June 2001.
11 21 CFR § 314.50 (drugs); 21 CFR § 601 (biologics); 21 CFR § 814.20 (devices). 

12 21 U.S.C. §§ 301, 379f(j). 

13 21 CFR § 314 (drugs); 21 CFR § 600.21 (biologics); 21 CFR § 814.44 (devices). 


OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

2

I N T R O D

U C T

I O N

approves a new product, sponsors can market and sell it in the United
States.
Clinical investigators. Clinical investigators’ responsibilities include
recruiting subjects, supervising clinical studies of the investigational
product, and reporting study results to the sponsor.14 Before they begin
drug and biological clinical trials, clinical investigators must sign an
FDA Form-1572 (Statement of Investigator).15 By signing this form,
investigators agree to follow a research protocol that the sponsor
provides to FDA, report any unexpected adverse outcomes to sponsors
and IRBs, obtain informed consent from all subjects participating in the
research, and comply with all relevant FDA regulations.16
Institutional Review Boards. IRBs are committees that an institution
designates to oversee clinical investigators and their research.17 IRBs
are often affiliated with hospitals and academic medical centers, but
they also exist in managed care organizations and Government agencies
and as for-profit entities that are independent of the institution in
which the research takes place. IRBs generally oversee many and
varied clinical trials.18
IRBs are intended to ensure that clinical investigators take appropriate
steps to protect the rights and welfare of human subjects. All clinical
trial research involving human subjects must be approved by an IRB.19
Once an IRB approves a clinical trial, Federal regulations require it to
reevaluate the trial at least once a year.20
Food and Drug Administration Oversight of Clinical Trials
Three centers within FDA individually regulate different types of
medical investigational products for human use. These centers are the
Center for Drug Evaluation and Research (CDER), the Center for
Biologics Evaluation and Research (CBER), and the Center for Devices

14 21 CFR § 312.60 (drugs and biologics); 21 CFR § 812.100 (devices). 

15 Although the Center for Devices and Radiological Health does not require a Form FDA 


1572, it does require an investigator agreement. See 21 CFR § 812.43(c) (devices);
21 CFR § 312.53 (drugs and biologics).
16 21 CFR § 50.312.
17 21 CFR § 56.102.
18 Larger IRBs may oversee more than 500 trials in a given year. See Institutional

Review Boards Registration Requirements, 69 Fed. Reg. 40,556 (July 6, 2004).
19 21 CFR § 56.109(a).
20 21 CFR § 56.109(f).

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

3

I N T R O D

U C T

I O N

and Radiological Health (CDRH).21 CDER and CBER share regulations
for the development of investigational products; CDRH has its own
regulations.22
The centers use onsite inspections to ensure that clinical investigators,
sponsors, and IRBs comply with FDA regulations while developing
investigational products. In 1977, FDA established the Bioresearch
Monitoring Program (BiMo) to develop cross-center guidelines for these
inspections of clinical investigators, sponsors, and IRBs (hereinafter
referred to as BiMo inspections).23 BiMo’s objective, as it relates to
clinical trials, is to “assure the quality and integrity of data submitted
to FDA to demonstrate the safety and efficacy of regulated products,
and to determine that human rights and the welfare of human and
animal research subjects are adequately protected.”24 The “Compliance
Program Guidance Manual” contains the current guidelines for BiMo
inspections.
FDA’s Office of Regulatory Affairs (ORA) conducts all onsite inspection
activities for the agency. ORA employs about 1,300 investigators who
conduct a total of 22,000 inspections annually. About 200 investigators
(hereinafter referred to as BiMo investigators) have received specialized
training to conduct BiMo inspections and have conducted these
inspections in recent years.25 BiMo investigators work out of all 19 of
ORA’s district offices.
Types of Bioresearch Monitoring inspections. The centers assign two
general types of BiMo inspections: surveillance inspections and directed
inspections. Most surveillance inspections target concluded clinical
trials and retrospectively review compliance with FDA regulations.
Many directed inspections target trials that are still treating human
subjects (ongoing trials). The centers can assign surveillance and
directed inspections to ORA at any point during the development of an

21 FDA’s Center for Veterinary Medicine and Center for Food Safety and Applied
Nutrition oversee investigational products as well, but are outside the scope of our
evaluation.
22 21 CFR § 312 (drugs and biologics); 21 CFR § 812 (devices). 

23 21 U.S.C. §§ 355, 360(i). 

24 FDA, “Bioresearch Monitoring Program Coordination Background.” Available online 


at: http://www.fda.gov/ora/compliance_ref/bimo/background.html. Last accessed on
February 5, 2007.
25 Although some investigators conduct BiMo inspections exclusively, most investigators
also conduct other types of inspections for FDA.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

4

I N T R O D

U C T

I O N

investigational product. Although FDA has the authority to conduct
surveillance and directed BiMo inspections, it is not required to do so.
Surveillance inspections are generally routine inspections that the
centers assign to ORA to verify that a clinical investigator, a sponsor, or
an IRB complied with FDA regulations during clinical trials for an
investigational product. Data verification inspections are a principal
type of surveillance inspection. Data verification inspections usually
occur 2 to 3 years after the clinical trials for an investigational product
conclude and focus on verifying the clinical trial data that a sponsor
submitted to support a new product application.
The centers assign directed inspections to ORA for a variety of reasons,
including complaints about the conduct of a clinical trial from an
interested party, problems noted during previous inspections, or
problems identified at similar sites. Other factors may also lead the
centers to assign a directed inspection of a clinical trial, such as an
unusually high number of subjects enrolled at one site under one
clinical investigator or a treatment that may be considered higher risk.
Directed inspections that target ongoing clinical trials allow the centers
and BiMo investigators to review the conduct of a clinical trial and
advise where corrective action may be needed while the trial is still
ongoing, thereby offering further human subject protections.
Bioresearch Monitoring inspection process. ORA’s BiMo investigators
conduct BiMo inspections according to the “Compliance Program
Guidance Manual.” Also, staff at the center that assigned a BiMo
inspection may issue additional instructions for the BiMo investigator.
The manual’s instructions vary based on the type of inspection
(i.e., clinical investigator, sponsor, or IRB).26 If a BiMo investigator
identifies violations of FDA regulations, he/she records them on a Form
FDA-483 (Inspectional Observations Report). After the inspection, the
BiMo investigator sends the assigning center a copy of the FDA-483 and
an Establishment Inspection Report (EIR), which includes background

26 FDA, “Compliance Program Guidance Manual,” Chapter 48.809: Institutional Review

Board (September 1997); Chapter 48.810: Bioresearch Monitoring—Sponsors, Contract
Research Organizations, and Monitors (February 2001); Chapter 48.111: Clinical
Investigators (September 2000).

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

5

I N T R O D

U C T

I O N

information and copies of all inspection documents. In the EIR, the
BiMo investigator recommends a classification for the inspection.27
Inspection classifications and actions taken. Staff at the assigning center
classify each inspection based on the EIR and other information they
have about the investigational product. BiMo inspections can result in
one of three classifications: no action indicated (NAI), voluntary action
indicated (VAI), or official action indicated (OAI). An NAI classification
signifies few or no violations of FDA regulations. Generally, FDA takes
no action in these cases. An inspection receives a VAI classification and
an untitled letter when the violations are serious enough to record, but
the center does not believe the violations cross “the threshold of
regulatory significance.”28 FDA does not require the clinical
investigators, sponsors, or IRBs that receive VAI classifications to
address the violations found during the inspection.
Centers classify inspections as OAI and issue warning letters to the
inspected entity when investigators find violations of “regulatory
significance.”29 FDA considers warning letters an important tool in
ensuring voluntary compliance. If the inspection subject does not take
corrective action to address violations cited in the warning letter, FDA
may take enforcement actions, such as disqualifying data or barring a
clinical investigator from conducting research. Warning letters and
enforcement actions are publicly available on FDA’s Web site.
FDA requests that the inspected entity respond to the violations listed
in a warning letter by sending a justification or a corrective action plan
to the assigning center. FDA’s “Regulatory Procedure Manual” states,
“If necessary to ensure that corrections have been implemented,
follow-up inspections should be conducted promptly after the agreed
upon date of completion of the promised corrections.”30
Recent Food and Drug Administration Initiatives
Recently, FDA has taken steps to improve its BiMo inspection
processes. In 2004, it created the BiMo Steering Committee to review
27 Although most recommended classifications made by BiMo investigators are classified
as no action indicated, voluntary action indicated, or official action indicated, BiMo
investigators may also choose not to recommend a classification.
28 Ibid., p. 4-27.
29 The centers are required to get the concurrence of FDA’s Office of Chief Counsel for all

warning and untitled letters. See FDA, “Regulatory Procedures Manual,” Chapter 4:
Advisory Actions, p. 4-26, March 2007.
30 Ibid., p. 4-11.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

6

I N T R O D

U C T

I O N

BiMo processes and make recommendations for improving them.
During the course of our study, FDA renamed that committee the
Human Subject Protection/Bioresearch Monitoring Council. In 2006,
the committee issued new guidance on clinical trial conduct and policy.31
Prior Office of Inspector General Reports
Previous OIG reports documented weaknesses in the oversight that
FDA and IRBs provide for clinical trials. In 1998, a series of reports
concluded that IRBs lacked the time and the expertise to sufficiently
monitor the research taking place under their jurisdiction.32 The
reports found that IRBs often conducted minimal review of studies after
the initial approval of the research. Additionally, IRBs provided little
training for investigators and board members.

A 2000 report documented weaknesses in clinical trial oversight. The
report found that data integrity concerns, more than human subject
protection, drove FDA’s oversight of clinical investigators and that the
BiMo program lacked clear and specific guidelines.33

METHODOLOGY
Scope
Our study focused on BiMo inspections of sponsors, clinical
investigators, and IRBs in clinical trials from FY 2000 through FY 2005.
We included all of the BiMo inspections conducted for CBER, CDER,
and CDRH that are listed in FDA’s inspections databases for this 6-year
period.
Data Sources and Analysis
We used seven data sources for this study. See Appendix A for a
detailed methodology.

31 In 2006, FDA issued the following new guidance: Draft Guidance: Process for

Handling Referrals to FDA Under 21 CFR § 50.54; Additional Safeguards for Children in
Clinical Investigations, May 2006; Guidance for Industry: Using a Centralized IRB Process
in Multicenter Clinical Trials, March 2006; Guidance for Clinical Trial Sponsors:
Establishment and Operation of Clinical Trial Data Monitoring Committees, March 2006;
Information Sheet Guidances for IRBs, Clinical Investigators, and Sponsors, January 2006.
32 Department of Health and Human Services, OIG, “Institutional Review Boards: A

Time for Reform,” OEI-01-97-00193, June 1998; OIG, “Low-Volume Institutional Review
Boards,” OEI-01-97-00194, October 1998; “Institutional Review Boards: Their Role in
Reviewing Approved Research,” OEI-01-97-00190, June 1998.
33 Department of Health and Human Services, OIG, “FDA Oversight of Clinical
Investigators,” OEI-05-99-00350, June 2000.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

7

I N T R O D

U C T

I O N

Food and Drug Administration’s Bioresearch Monitoring inspections data.
We analyzed FDA’s BiMo inspections data to evaluate how FDA tracks
and manages BiMo inspections. We also calculated the number and
type of inspections FDA conducts. We used data from six FDA
databases for BiMo inspections. The three centers maintain separate
databases to track BiMo inspections: two databases each for CBER and
CDER and one for CDRH. ORA maintains the Field Accomplishment
and Compliance Tracking System (FACTS).
From each database, we analyzed the records for BiMo inspections of
sponsors, clinical investigators, and IRBs that BiMo investigators
conducted for the three centers during the FY 2000–2005 period. We
excluded laboratory and bioequivalence inspections from our analysis.
File review. Using the FDA databases, we identified and reviewed all
BiMo inspections for which the BiMo investigator recommended OAI or
a center classified as OAI in FY 2000–2005. We reviewed 668 files,
including 99 CBER inspection files, 248 CDER inspection files, and
321 CDRH inspection files.
E-mail survey of Bioresearch Monitoring investigators. We surveyed BiMo
investigators about their experiences with FDA guidance, classification
recommendations, inspection followup with the assigning centers, and
challenges conducting BiMo inspections. We received completed
surveys from 76 percent of the BiMo investigators (170 out of 223).
Interviews with Food and Drug Administration officials. We conducted
telephone interviews with all 19 ORA district directors and supervisors
of investigation branches (hereinafter referred to as supervisors). The
interviews focused on the supervisors’ experiences with the supervisory
role, interactions with the BiMo investigators they supervise,
interactions with the assigning centers, classification recommendations,
and challenges in supervising BiMo inspections. We also interviewed
senior FDA officials in the agency’s headquarters and at each center.
Observation of Bioresearch Monitoring inspections. We observed the
process that BiMo investigators followed, as well as the interactions
between the investigator(s) and the inspection subject(s). We observed
two clinical investigator inspections and one IRB inspection.
Clinical trials registry site estimate. From the population of trials on the
National Institutes of Health’s (NIH) clinical trials registry as of
February 2007, we selected a stratified random sample of 150 trials.
Because the registry is separated by phase of the trial, the three strata
OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

8

I N T R O D

U C T

I O N

used for the sample were Phase I, Phase II, and Phase III of the clinical
trial. We excluded 43 of the sampled trials and the final sample size to
estimate that the number of trial sites was 107. (See Appendix A, Table
7, for our sample design.)
Review of Food and Drug Administration regulations, guidance, policies,
and procedures documents. We obtained and reviewed all relevant FDA
regulations, guidance, policies, and procedures documents for
conducting BiMo inspections.
Standards
We conducted this review in accordance with the “Quality Standards for
Inspections” issued by the President’s Council on Integrity and
Efficiency and the Executive Council on Integrity and Efficiency.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

9

Δ

F I N D I N G S

FDA relies on BiMo inspections as
a principal mechanism for
overseeing clinical trials after it
reviews the protocol. BiMo inspections help to ensure the quality and
integrity of clinical trial data, and to determine whether sponsors,
clinical investigators, and IRBs are complying with regulations that
protect human subjects while they are enrolled in clinical trials.

Data limitations inhibit FDA’s ability to
effectively manage the BiMo program

FDA lacks a complete clinical trial registry
Because FDA does not maintain a clinical trial registry, it is unable to
indentify all ongoing clinical trials and their associated trial sites.
Limited clinical trials that are regulated by FDA can be found on a
clinical trial registry run by NIH.34 Included within this data bank is
information on FDA-regulated studies to treat serious or
life-threatening diseases and conditions. Sponsors are not required to
submit information on any other type of clinical trials that FDA
regulates.
FDA lacks an IRB registry
Because FDA does not maintain an IRB registry, it is unable to identify
all IRBs. IRBs are important because their primary purpose is to
ensure that clinical investigators take appropriate steps to protect the
rights and welfare of human subjects. BiMo inspections of IRBs offer
the centers considerable oversight of human subject protections.
Because IRBs usually oversee many and varied clinical trials, the
centers can review numerous trials across the centers with each BiMo
inspection of IRBs.

The Office of Human Research Protections (OHRP) oversees all
federally sponsored research that includes human subjects.35 OHRP
maintains a database consisting of all IRBs that oversee this research.
Although the OHRP database likely contains many IRBs that oversee
some FDA-regulated clinical trials, it does not contain IRBs that oversee
clinical trials regulated exclusively by FDA.

34 Section 113 of the 1997 Food and Drug Administration Modernization Act directs the
Secretary of the Department of Health and Human Services to act through NIH to establish
and operate a databank of information on clinical trials for drugs to treat serious or lifethreatening diseases or conditions. 42 U.S.C. § 282(j)(3)(A).
35 OHRP oversees research that is conducted or supported by HHS. See 45 CFR Part 46.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

10

F

I N D I N G

S

FDA has acknowledged the importance of creating an IRB registry and
published a proposed rule in July 2004 for the creation of one.36
According to FDA, the benefits of such a registry would include being
able to identify IRBs that review clinical trials regulated by FDA,
having a complete list of IRBs for educational and outreach purposes,
and being able to easily identify IRBs for inspection.37
No FDA database includes complete information for all BiMo inspections,
hindering FDA’s ability to track BiMo inspections
FDA maintains six databases to track BiMo inspections. Although
ORA’s database, FACTS, includes information on all BiMo inspections,
it does not consistently include information about inspection-related
events that occur after an investigator submits an inspection report to
the centers. For example, 48 percent of BiMo inspection records in
FACTS (2,557 out of 5,312 records during the FY 2000–2005 period)
lack the centers’ final classifications. Because of these omissions, FDA
cannot use FACTS to track inspections from assignment through a
center’s final action.

Tracking BiMo inspections using data from the centers is challenging
because of the number of databases involved and their inconsistencies.
The three centers used five separate databases to track BiMo inspection
data in the FY 2000–2005 period.38
The center databases do not uniformly track inspections. First, each
database contains different information about BiMo inspections.
Second, when the centers track similar information, they do not do so in
the same way. For example, each database uses different categories to
identify the reason a center initiated an inspection.39 (See Table 1 on
the next page.) These categories are difficult to translate across
databases. For example, FACTS and one of CBER’s databases indicate
whether an inspection was based on a complaint. However, the other
databases combine complaint-based inspections with other types of
inspections, which means that FDA cannot identify which inspections
were prompted by complaints for all centers. Also, the databases define
different types of inspections as surveillance and as directed. FDA

36 Institutional Review Boards Registration Requirements, 69 FR 40,556 (July 6, 2004). 

37 Ibid., 40,557. 

38 CBER and CDER maintain two databases each to track inspections. 

39 None of the databases indicates whether an inspected trial was completed or ongoing. 


OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

11

F

I N D I N G

S

cannot use these databases to track inspections across centers unless it
controls for all of these discrepancies.
Table 1: Categories Center and ORA Databases Use To Track Reason for Inspection
Database

ORA
FACTS

CBER
IRB
CBER
CISM1

CDER
3
COMIS

CDER
DSI4

Inspections Included

Category

All BiMo inspections ORA
conducts

• Complaint (for cause, based on a complaint)
• Compliance (followup to previous violations)
• Surveillance (routine inspections that are not based on
preexisting information about a firm; includes data integrity
2
inspections)

CBER's IRB inspections

• Text field (most entries in this field do not indicate reason for
inspection)

CBER's clinical
investigator and sponsor 

inspections

• Complaint 

• Surveillance (site selected because of high risk; includes blood,
2
cell, gene, vaccine, and other)

Completed inspections of
trials for drug products

• Data audit (most are based on new product application)
• For cause (based on prior problems complaint)

CDER inspections
starting in FY 2004 or
later

•
•
•
•

For cause (based on previous problems or complaint)
Routine data audit (based on a new product application)
Information gathering
2
Surveillance

All CDRH BiMo
inspections of IRBs, 

clinical investigators, and 

sponsors 


•
•
•
•
•
•
•
•
•
•

Routine 

For cause 

Directed 

Vulnerable population
Probability sampling
Field initiated
Center initiated
Expedited review
2
Surveillance
Followup inspection after OAI inspection

CDRH
database

4

1

CISM is clinical investigator, sponsor, and monitor.

The databases define surveillance differently. 

COMIS is the Center Office Management Information System. 

4
DSI is the Division of Scientific Investigations. In the CDER DSI database, inspections can be more than one type. 

2

3

Source: OIG analysis of ORA, CBER, CDER, and CDRH data, 2006.

The five center databases do not uniformly track other important
inspection-related data, such as (1) whether the inspection targets an
ongoing or completed clinical trial, (2) ORA’s recommended
classification, and (3) follow-up inspection information.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

12

F

I N D I N G

S

Other factors hinder FDA’s ability to effectively
manage the BiMo program
Centers and ORA inconsistently classify OAI and NAI inspections
According to FY 2000–2005 CDER and CDRH data, these centers often
disagreed when ORA recommended an OAI classification, the most
serious classification category.40 (See Table 2.) During this 6-year
period, CDER revised 68 percent of ORA’s OAI recommendations and
CDRH revised 23 percent of ORA’s OAI recommendations. The centers
reclassified most of these inspections to VAI, a classification that
neither requires the inspected entity to formally address violations the
center identified nor makes the classification publicly available.

The centers do not systematically track their reasons for revising ORA’s
recommended classifications. However, our review of 668 FDA
inspection files identified two key reasons for changing OAI
classifications to VAI. The most common reasons were that the center,
in conjunction with an Office of Chief Counsel review, determined that
the violations were not serious enough to warrant a warning letter and
therefore an OAI classification, and that the inspection subject promised
corrective actions.

Table 2: CDER and CDRH Agreement With ORA Recommended Classifications, FY 2000–2005

Center

Percentage of All
Recommendations
Center Revised

Percentage of OAI
Recommendations Center
Revised

Percentage of VAI
Recommendations Center
Revised

Percentage of NAI
Recommendations
Center Revised

CDER

14% (303 of 2,090)

68% (133 of 195)

2% (25 of 1,328)

26% (145 of 567)

CDRH

13% (193 of 1,506)

23% (62 of 271)

7% (47 of 688)

15% (84 of 547)

Source: OIG analysis of CDER and CDRH data, 2006.

CDER and CDRH also revised some of ORA’s NAI recommendations. In
other words, center reviewers sometimes identified violations when a
BiMo investigator found no significant violations at the inspection site.
CDER reclassified 26 percent of the inspections that ORA identified as
40 We could not conduct this analysis for CBER because it does not record ORA’s
recommended classification.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

13

F

I N D I N G 
S

NAI. CDRH reclassified 15 percent of the inspections that ORA
identified as NAI.
These reclassifications suggest that ORA and the centers sometimes
interpret the regulations and guidance for BiMo inspection
classifications differently. Although the centers ultimately classify
inspections, they rely on BiMo investigators to identify violations at the
inspection site and collect evidence to support inspection classifications.
If BiMo investigators interpret the regulations and guidelines
differently than center reviewers do, they may miss evidence to support
an OAI classification when one would be appropriate. Conversely, BiMo
investigators may spend time gathering evidence to support an OAI
classification that the center does not support.
BiMo supervisors and investigators reported that a lack of feedback
from the centers and inadequate training may contribute to the
differences in interpreting inspection classifications. Eighteen of
nineteen BiMo supervisors we spoke with stated that the centers rarely
or never provide feedback on the inspection reports that the districts
submit to the centers. Further, 19 percent of the BiMo investigators we
surveyed (32 of 170) responded that they would like more information
on the centers’ final classifications or feedback on their
inspection-related performance.
FDA relies on voluntary compliance to correct violations of regulatory
significance
The centers send warning letters for 70 percent of inspections classified
as OAI. The centers can respond to OAI inspections by issuing an
untitled letter, issuing a warning letter, disqualifying a clinical
investigator, or disqualifying data gathered in the clinical trial. Our
analysis of center data shows that the centers chose to respond to
70 percent of inspections they classified as OAI during the
FY 2000–2005 period by issuing warning letters. (See Table 3 on the
next page.) The centers’ next most common response to OAIs was
untitled letters, which they sent for 16 percent of OAI inspections. Less
frequently, the centers took more formal actions, including initiating
disqualification of inspection subjects or disqualifying data from clinical
trials.

Warning letters and untitled letters rely on voluntary compliance, so
FDA must reinspect individuals who receive these letters to ensure that
they do not repeat violations in future clinical research. If the centers
do not track the response to warning letters and untitled letters, they
OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

14

F

I N D I N G

S

cannot ensure that voluntary compliance occurs. According to FDA’s
“Regulatory Procedures Manual,” “Warning Letters are issued only for
violations of regulatory significance. Significant violations are those
violations that may lead to enforcement action if not promptly and
adequately corrected. A Warning Letter is the agency’s principal means
of achieving prompt voluntary compliance with the Federal Food, Drug,
and Cosmetic Act.”41

Table 3: Centers' Official Actions for BiMo Inspections Classified as OAI
Percentage of
OAIs That Resulted
in a Warning Letter

Percentage of
OAIs That Resulted
in an Untitled Letter

Percentage of
OAIs That Resulted
in a Disqualification
Letter

Percentage of
OAIs That Resulted
in Disqualified Data

Unknown Action

CBER

82% (55 of 67)

6% (4 of 67)

12% (8 of 67)

0% (0 of 67)

0% (0 of 67)

CDER

57% (38 of 67)

16% (11 of 67)

18% (12 of 67)

2% (1 of 67)

8% (5 of 67)

CDRH

71% (151 of 214)

20% (42 of 214)

3% (6 of 214)

1% (1 of 214)

7% (14 of 214)

70% (244 of 348)

16% (57 of 348)

8% (26 of 348)

1% (2 of 348)

6% (19 of 348)

Center

Total

Source: OIG analysis of CDER, CDRH, and CBER inspection files, 2006.

Although CBER and CDER fail to track inspection followup, CDRH
data show that that center conducts few follow-up inspections. Center
databases fail to track whether the center conducted a followup to an
inspection classified as VAI or OAI. CDRH’s database indicates
whether an inspection was initiated as a followup to a prior violative
inspection. However, the database does not link the followup inspection
to the violative inspection. Nor does it indicate in the record for the
violative inspection that followup occurred. CDRH data show that the
center conducted 3 follow-up inspections for every 100 inspections
classified as OAI or VAI in the FY 2000–2005 period (36 follow-up
inspections for 1,048 inspections classified as OAI and VAI).

41 FDA, “Regulatory Procedures Manual,” Warning Letter Procedures, p. 4-1-1, March

2007. Available online at http://www.fda.gov/ora/compliance_ref/rpm. Last accessed
April 9, 2007.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

15

F

I N D I N G

S

Uncertainty and lack of coordination impede FDA’s ability to conduct BiMo
inspections
In interviews, center and ORA staff reported that inspections triggered
by new product applications affect the conduct and timing of other
inspections. The centers cannot predict how many new product
applications they will receive each year and therefore cannot plan for
the inspections that these applications trigger. At the beginning of each
fiscal year, the centers plan and budget for BiMo inspections. During
the year, the centers also receive new product applications, which
generate BiMo inspection assignments with short due dates because of
the overall performance goals of PDUFA and MDUFMA.42 Assignments
based on new product applications are usually retrospective reviews
that focus on verifying clinical trial data used in the application. These
assignments generally take priority over all other BiMo inspections.
Because each center has a set budget for all inspections, an unexpected
rise in new product applications may mean that the ORA districts have
more inspections overall and more inspections with short due dates
than they originally planned.

Because of the uncertainties of timing and prioritization involved with
BiMo inspections, districts sometimes must delay some of the other
inspections that ORA conducts, particularly those with later due dates.
In fact, 11 of 19 district supervisors we interviewed stated that
assignments for inspections related to marketing applications create
challenges for managing resources. Staff in five district offices reported
asking the centers for additional time to complete BiMo inspections; two
offices estimated that they request extensions for about 20 percent of
BiMo assignments. Staff in four other districts reported that they have
to cancel other inspections when they have too many unexpected BiMo
assignments.
Additionally, the centers generally fail to coordinate assignments with
one another or the districts. When asked if they faced any challenges
conducting BiMo inspections, staff from 7 of the 19 district offices
reported that lack of planning and coordination by the centers created
challenges for the district. For example, staff in a district that oversees
42 PDUFA requires pharmaceutical companies to pay fees to FDA when submitting
Preapproval Marketing Applications and New Drug Applications. These fees are dedicated
to expediting the review of product applications for CDER and CBER. MDUFMA has the
same requirement for device applications. The BiMo inspection serves as one step of the
evaluation process. 21 U.S.C. §§ 379g(4), 379j.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

16

F

I N D I N G 
S

a large geographic area reported that inspection assignments for distant
sites would be more manageable if they could combine multiple
inspections in a single trip. However, this is not feasible without
coordinating across the assigning centers.
Failure to coordinate BiMo assignments across centers may also impede
BiMo Steering Committee goals. In interviews, BiMo Steering
Committee members reported that they want to increase the number of
inspections of ongoing trials and inspections that focus on high-risk
sites. One committee member we interviewed told us that the current
system, in which each center independently issues assignments and
tracks data, does not allow for the timely coordination necessary to
target high-risk and ongoing inspections across centers.
FDA guidance and regulations do not reflect current clinical trial practices
According to interviews with center officials and ORA supervisors,
FDA’s guidance and regulations for clinical trials have fallen behind
industry practices. FDA officials reported that when the agency
developed clinical trial regulations, a single investigator at a single site
ran each clinical trial. Since then, clinical trials have grown
increasingly complex. Trials are larger and involve multiple sites
within and outside the United States.43

FDA officials told us that clinical investigators frequently delegate tasks
that include direct care to human subjects to colleagues or subordinates.
Current FDA regulations do not address colleagues or subordinates.
When FDA conducts an inspection and finds significant deficiencies
related to individuals other than the clinical investigator, the agency
may only take action against the clinical investigator.
Finally, although sponsors increasingly conduct clinical trials outside
the United States, FDA authority over foreign trials remains limited.
For example, one center official estimated that 20 to 25 percent of the
trials for products that FDA oversees occur outside the United States.
According to the official, centers are often unaware that foreign trials
43 Regarding the increasingly complex management structures of clinical trials, see

Department of Health and Human Services, OIG, “The Globalization of Clinical Trials:
A Growing Challenge in Protecting Human Subjects,” OEI-01-00-00190, September 2001.
See also Iain M. Cockburn, “The Changing Structure of the Pharmaceutical Industry,”
Health Affairs, 23(1) pp. 10–22, 2004; and R. A. Rettig, “The Industrialization of Clinical
Research,” Health Affairs, 19(2), pp. 129–146, 2000. Regarding the growing size and
complexity of clinical trials, see Institute of Medicine, “The Future of Drug Safety:
Promoting and Protecting the Health of the Public,” Chapters 1 and 2, 2006.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

17

F

I N D I N G

S

have taken place because FDA’s investigational product regulations
generally do not apply outside the United States. When ORA does
inspect foreign clinical trials and finds violations, the centers can only
disqualify data from consideration in a new product application.

As previously noted, FDA lacks a
complete database that includes
information on all BiMo
inspections. However, FDA also
lacks registries that contain complete clinical trial and IRB information.
Failure to track all clinical trials and IRBs limits the centers’ ability to
oversee trials through BiMo inspections. First, the centers cannot
assign BiMo inspections of trials and IRBs that they cannot identify.
Second, the centers cannot identify the percentage of trials and IRBs
that BiMo inspections reach. To identify the percentage of trials and
IRBs that BiMo inspections reached during the FY 2000–2005 period,
we used estimates of clinical trial and IRB populations from
Government sources outside FDA.

We estimate that FDA inspected
1 percent of clinical trial sites during the
fiscal year 2000-2005 period

The centers reported 2,856 BiMo inspections that required a clinical
trial site visit for the FY 2000–2005 period.44 (See Table 4 on the next
page.) However, the centers do not have data to identify the percentage
of trial sites that these inspections reached. Therefore, we used the
number of INDs and IDEs the centers received as well as estimates
from the clinical trial registry that NIH maintains to estimate the
percentage of trial sites that the BiMo inspections reached.45

44 We excluded IRB and sponsor inspections from this analysis because they generally do

not take place at a clinical trial site.
45 Section 113 of the 1997 FDAMA directs the Secretary of the Department of Health and

Human Services to act through NIH to establish and operate a databank of information on
clinical trials for drugs to treat serious or life-threatening diseases or conditions. Included
within this databank is information on studies to treat serious or life-threatening diseases
and conditions conducted under FDA IND regulations. Although the NIH registry does not
include all the clinical trials FDA oversees, it provides the best available estimate of the
average number of trial sites associated with clinical trials. The registry is available at
www.clinicaltrials.gov.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

18

F

I N D I N G

S




T able 4: C enters' B iM o Inspections W ith Site Visits by Fiscal Y ear
Fiscal Year

CDER

CD R H

C BE R

Total

2000

296

91

143

530

2001

186

113

79

378

2002

194

153

109

456

2003

226

167

103

496

2004

219

189

75

483

2005

221

183

108

512

1,342

896

617

2,855

Total

S ource: O IG Analysis of C D E R , C D R H , and C B ER data, 2006.
W e included only clinical investigator, contract research organization, and sponsor m onitor inspections.

Although FDA cannot identify the total number of clinical trial sites
operating during the FY 2000–2005 period, the centers did track the
15,268 INDs and IDEs that they received during this period. However,
each IND or IDE may involve many trial sites, so the number of INDs
and IDEs cannot substitute for the total number of trial sites.
Therefore, to estimate the total number of trial sites, we calculated the
average number of clinical trial sites associated with a random sample
of clinical trials listed on the NIH clinical trials registry. We found that
clinical trials listed on that registry had an average of 23 sites per
trial.46 Based on these figures, we estimate that the 15,268 INDs and
IDEs the centers received included about 350,000 trial sites.
Using the estimate of 350,000 trial sites, we estimate that FDA’s
2,855 inspections reached just under 1 percent of the trial sites
associated with the INDs and IDEs the centers received.

46 A 95-percent confidence interval for our sample is +/- 8 trial sites, or between 15 and

31 sites per trial. Using this range, we calculate that FDA’s 2,855 inspections reached at
most just over 1 percent of trial sites associated with the 15,268 applications received by the
centers.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

19

F

I N D I N G 
S

Centers assign more surveillance inspections than directed inspections
For the FY 2000–2005 period, 75 percent of BiMo inspections were
surveillance inspections. Surveillance inspections generally focus on
verifying the clinical trial data that sponsors submitted with their new
product applications. Further, most surveillance inspections target
completed trials. Data verification inspections serve an important
purpose in enabling FDA to assess the quality of the data used to
support new product applications. However, because these inspections
take place after trials conclude, they cannot ensure that sponsors,
clinical investigators, and IRBs are taking the necessary actions to
protect human subjects during the trials.

Some of the directed inspections that the centers assigned during the
FY 2000–2005 period likely focused on ongoing trials. For example,
CBER annually targets a high-risk category of clinical trials for
inspection, which may include ongoing clinical trials.47 CDRH
implemented a program in the past year to increase the number of
inspections that focus on higher-risk clinical trials that may be ongoing.
However, the three centers do not track whether inspections target
ongoing trials, so we could not estimate the proportion of BiMo
inspections that reached ongoing trials. Several senior FDA officials
reported to us that they recognize the need to inspect more ongoing
clinical trials to protect human subjects.
Centers inspect few IRBs
Our analysis of center data shows that in the FY 2000–2005 period,
FDA conducted an average of 214 IRB inspections each year. The
number of IRB inspections has decreased annually since 2002. (See
Table 5 on the next page.) IRB inspections offer the centers
considerable oversight of human subject protections. Because IRBs
usually oversee a variety of clinical trials, the centers can review
numerous trials across the centers with each IRB inspection.

At most, we estimate that during the FY 2000–2005 period, the centers
inspected an average of about 6 percent of IRBs each year, or less than
40 percent of all IRBs in the 6-year period we evaluated. Because the
centers do not track the total population of IRBs, we used the IRB
database maintained by OHRP to estimate the population of IRBs.
OHRP oversees all research that the Federal Government sponsors.
47 Past examples include studies focusing on pediatric populations and gene therapy
studies.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

20

F

I N D I N G

S

The OHRP database includes all IRBs that oversee research sponsored
by the Government, so it likely underestimates the population of IRBs
overseeing all clinical trial research. In February 2007, OHRP reported
3,579 IRBs in its IRB database.48

T ab le 5: IR B In sp ectio n s as a P ercen tag e o f T o tal B iM o In sp ectio n s
N u m b er o f IR B
In sp ectio n s

F iscal Year

To tal B iM o P ercen tag e o f Insp ectio n s Th at
In sp ectio n s
Are IR B insp ectio n s

2000

198

795

25%

2001

187

633

30%

2002

278

812

34%

2003

251

854

29%

2004

207

779

27%

2005

163

769

21%

1,284

4,642

28%

To tal

S ource: O IG A nalysis of C D E R , C D R H , and C B E R data, 2006.

IRB inspections assigned to ORA are allowed longer timeframes than
inspections of clinical investigators or sponsors. A center’s inspection
assignment includes a due date. In many cases, the centers allow ORA
up to 1 year to complete IRB inspections. By comparison, ORA
generally has 30 to 90 days to complete inspections of clinical
investigators.49
Center data show that despite these longer timeframes, ORA frequently
failed to meet the requested due dates for IRB inspections. ORA failed
to meet CDRH’s deadlines for 60 percent of IRB inspections (214 out of
359); 24 percent (85 of 359) were more than 60 days late. Similarly,
ORA failed to meet CBER’s deadline for 38 percent of that center’s IRB
inspections (24 of 63).50

48 OHRP oversees research that is conducted or supported by the U.S. Department of

Health and Human Services. See 45 CFR Part 46.
49 In some cases, IRB inspections are directed because of a past OAI or some other

concern. In these instances, the timeframes are shorter than the timeframes presented
here.
50 We could not determine the timeliness of CDER inspections because CDER does not
include the inspection due date in its database.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

21

Δ

R E C O M M E N D A T I O N S

FDA conducts BiMo inspections to ensure the quality and integrity of
clinical trial data and to determine whether sponsors, clinical
investigators, and IRBs are complying with regulations that protect
human subjects enrolled in clinical trials. Our evaluation identified a
lack of comprehensive reporting and tracking systems, a lack of
coordination among the centers and ORA, and regulations that do not
fully reflect current clinical trial practices.
No Federal requirements prescribe the number or type of BiMo
inspections FDA must conduct. However, FDA relies on BiMo
inspections as a principal mechanism for overseeing clinical trials after
it reviews the protocol. We recognize that resource limitations may be a
factor in determining the extent of FDA’s oversight of clinical trials
using BiMo inspections. However, to ensure its effectiveness in meeting
BiMo objectives, the inspection program needs comprehensive
management and reporting systems.
FDA has taken steps to improve the BiMo inspection program. For
example, in 2004 it created the BiMo Steering Committee to identify
and address weaknesses in the BiMo program. We identified additional
steps that FDA can take to improve oversight of clinical trials. These
changes will help FDA plan and conduct BiMo inspections more
effectively. These changes will also help FDA meet the BiMo objective
of “determining that human rights and the welfare of human research
subjects are adequately protected.”
FDA should:
Improve Information Systems and Processes
FDA’s BiMo inspections are a principal mechanism to ensure the
quality and integrity of clinical trial data and to determine that human
subjects are protected. However, the weaknesses we identified in the
BiMo information systems and management processes inhibit FDA’s
ability to effectively oversee and manage BiMo inspections across
centers. To improve the effectiveness of the BiMo program, FDA should
take the following actions:
Develop a clinical trial database that includes all clinical trials. FDA should
develop a comprehensive database to use as an internal management tool
to more effectively identify ongoing clinical trials for inspection. Although
NIH currently maintains a clinical trial registry, that registry was not
intended to be a comprehensive listing of clinical trials. It was created to
OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

22

R

E C O

M M E N D A T

I O N

S

inform the public regarding potentially new life-saving treatments.
Because the registry contains clinical studies to treat serious or
life-threatening diseases and conditions conducted under FDA’s IND
regulations, it excludes most other clinical trials subject to FDA
regulation, such as medical device studies.
Create an Institutional Review Board registry. By identifying all IRBs
overseeing clinical trials, FDA could target IRBs more effectively for
inspection. FDA has acknowledged the importance of creating an IRB
registry and published a proposed rule in July 2004. Recently, the agency
set a timetable of March 2008 for issuance of a final rule.51
OHRP currently maintains an IRB registry for studies under its
auspices.52 FDA could build upon OHRP’s existing IRB registry to
create a larger and more inclusive database while minimizing the
administrative burden for IRBs.
Create a cross-center database that allows complete tracking of Bioresearch
Monitoring inspections. FDA cannot efficiently manage BiMo inspections
across centers without a database that includes timely and complete
information about all BiMo inspections. A single, consistent database
would require the centers to agree about which fields are essential and
how they should be defined. At a minimum, the database should include:
(1) specific information on the reason for an inspection, (2) whether the
inspection targets an ongoing or a completed clinical trial, (3) ORA’s
recommended classification, and (4) follow-up inspection information.
Establish a mechanism to provide feedback to Bioresearch Monitoring
investigators on their inspection reports and findings. Improved feedback
between the centers and BiMo investigators could lead to a common
understanding of the regulations and guidelines. This would minimize the
likelihood that an investigator would miss evidence to support an OAI
classification when one would be appropriate or spend time gathering
evidence to support an OAI classification that the center does not
corroborate.
Seek legal authority to provide oversight that reflects current clinical trial
practices. FDA should consider seeking additional authority that covers
all of the stakeholders in the management and conduct of clinical trials.
51 72 FR 22,520 (April 30, 2007).
52 OHRP oversees research that is conducted or supported by the U.S. Department of

Health and Human Services. See 45 CFR § 46.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

23

R

E C O

M M E N D A T

I O N

S

In particular, FDA should consider seeking authority to include the
colleagues and subordinates of a clinical investigator if they participate in
the conduct of a clinical trial.

AGENCY COMMENTS AND OFFICE OF INSPECTOR GENERAL
RESPONSE
FDA concurred with four of our five recommendations. These
recommendations aimed to improve information systems and processes
related to the BiMo program. FDA did not address our recommendation
to establish a mechanism to provide feedback to BiMo investigators on
their inspection reports and findings.
Specifically, FDA concurred with our recommendations to develop a
clinical trial database that includes all clinical trials, to create an IRB
registry, to create a cross-center database that allows complete tracking
of BiMo inspections, and to seek legal authority to provide oversight
that reflects current clinical trial practices.
FDA pointed out that BiMo inspections make up only one part of its
efforts to ensure human subject protections, noting that it views its
protocol review before a clinical trial commences as the most important
step in protecting human subjects. We recognize the important role
that FDA’s protocol review plays in protecting human subjects and
made several changes to our report to reflect this point. We do note,
however, that this report addresses another important part of the
system for protecting human subjects: oversight of the trials once they
are actually underway.
The agency highlighted the efforts of its Human Subject
Protection/Bioresearch Monitoring Council in identifying key issues that
needed to be addressed, such as coordination, tracking mechanisms,
regulations, and guidance. FDA also emphasized the importance of
risk-based approaches to BiMo inspections rather than committing to
inspecting a specified percentage of clinical trials.
Where appropriate, we made changes to the report based on FDA’s
technical comments.
See Appendix B for complete agency comments.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

24

Δ

A P P E N D I X

A

METHODOLOGY
Scope
Our study focused on Bioresearch Monitoring (BiMo) inspections of
sponsors, clinical investigators, and Institutional Review Boards (IRBs)
in clinical trials from fiscal year (FY) 2000 to FY 2005. We reviewed all
of the inspections that BiMo investigators conducted for the Center for
Drug Evaluation and Research (CDER), the Center for Biologics
Evaluation and Research (CBER), and the Center for Devices and
Radiological Health (CDRH) that are listed in Food and Drug
Administration’s (FDA) inspections databases for this 6-year period.
We excluded BiMo inspections that focused on the bioequivalence and
laboratory compliance programs.53
Data Sources and Analysis
We used seven data sources for this study.

Food and Drug Administration’s Bioresearch Monitoring inspections data.
We analyzed FDA inspections data to evaluate how FDA tracks and
manages BiMo inspections. We also calculated the number and type of
inspections FDA conducts.
We analyzed data from six FDA databases for BiMo inspections. Each
center we evaluated and the Office of Regulatory Affairs (ORA)
maintain separate databases to track BiMo inspections. CDRH and
ORA each maintain a single database. CBER and CDER each maintain
two databases.
From each database, we analyzed the records for BiMo inspections of
clinical investigators, sponsors, and IRBs for CBER, CDER, and CDRH
during the FY 2000 to 2005 period.
We analyzed fields in the databases that indicate the type of inspections
that the centers conducted in the FY 2000 to 2005 period. (See Table 6.)
Each center tracks inspections differently, so we could not gather
information on every variable for all centers.

53 We excluded laboratory BiMo inspections because they were not clinical in nature. We

excluded bioequivalence BiMo inspections because they do not necessarily follow the same
investigational product processes described in the Background on p. 2.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

25

A

P

P E N D

I X

~

A

Using the fields listed in Table 6, we evaluated the number and types of
inspections that FDA conducted from FY 2000 to 2005. We conducted
all database analysis in SAS®.
The following is a description of each of the databases:
‰

Field Accomplishment and Compliance Tracking System (FACTS):
We received FACTS data from ORA in July 2006. The FACTS data
included 5,850 records. We removed 472 records that either
duplicated other records or targeted subjects outside our scope (for
example, manufacturers). We excluded 66 records that ORA coded as
“Cancelled” or “Returned” or that the Center for Veterinary Medicine
requested. Our analysis includes 5,312 FACTS records.

‰	

CBER data: We received data from CBER’s two databases in
July 2006. CBER dedicates one database to clinical investigator and
sponsor inspections. This database originally contained 714 records.
We removed records that did not fall within the scope of this study,
including five manufacturing inspections, four inspections classified
as “no inspection,” three inspections classified as “out of business,”
and eight inspections classified as “washout.” We also removed
11 inspections that did not include inspection start dates. Our
analysis includes 683 inspection records from this database.
CBER uses a second database to track IRB inspections. This
database included 73 records. We used all of these inspection records
in our analysis.

‰	

CDER data: We received data from one CDER database, the Center
Office Management Information System (COMIS), in October 2006.
CDER uses COMIS to track closed inspections of CDER drug
products trials that started before FY 2004. The COMIS data
originally included 4,237 inspection records. However, many of these
records were duplicates. After we removed the duplicate records,
COMIS included 1,196 inspection records.
We received data from a second CDER database, the Division of
Scientific Investigations database (DSI), in December 2006. CDER
uses the DSI database to track all inspections starting in FY 2004 or
later. The DSI database originally included 1,137 inspection records.
We removed records that did not fit within the scope of our study,
including inspections that started after FY 2005 ended. Our analysis
included 1,107 inspection records from the DSI database.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

26

A

P

P E N D

I X

‰	

~

A

CDRH data: We received data from CDRH in November 2006.
CDRH’s database originally included 2,762 inspection records. We
excluded 1,168 inspection records, including records that ORA
classified as “no inspection made” and records that CDRH classified
as “washout,” “out of business,” or “cancelled.” Our analysis included
1,594 CDRH records.

Table 6: Variables Used To Analyze FDA BiM o Inspections Data
Variable

CBER

CDRH

CDER

Yes

Yes

Yes

Reason for Inspection

Not available

Yes

Yes

ORA Recom m ended Classification

Not available

Yes

Yes

Center Final Classification

Yes

Yes

Yes

Official Actions

Yes

Yes

Database 1: Not
available
Database 2: Yes

Assignm ent Date

Yes

Yes

Yes

Inspection Start Date

Yes

Not available

Yes

Inspection End Date

Yes

Yes

Not available

Not available

Not available

Not available

Inspection Subject

Trial Status: Ongoing or Closed

Source: OIG analysis of CDER, CDRH, and CBER data, 2006.

File review. We used FDA’s databases to identify all inspections that
either ORA or a center classified as OAI. We reviewed 248 CDER
inspection files, 99 CBER inspection files, and 321 CDRH inspection
files, for a total of 668 files.
We reviewed the files using a Microsoft Access® protocol. Our protocol
included inspection assignment, classification, official actions, and
correspondence between the center and the inspection subject. We
analyzed data from the file reviews using SAS®.
E-mail survey of Bioresearch Monitoring investigators. ORA headquarters
provided us with a list of the 231 ORA field investigators who had spent
100 or more hours conducting BiMo inspections in the previous 5 years.
We removed eight investigators from the original sample because they

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

27

A P P E N D

I X

~

A

no longer conduct inspections or had not conducted an onsite inspection
in several years or we could not obtain contact information for them.
Our survey solicited respondents’ experiences with FDA guidance,
inspection and classification recommendations, inspection followup with
the assigning centers, and challenges in conducting BiMo inspections.
Before sending the survey, we solicited comments from ORA officials
and a BiMo investigator about the survey’s content and clarity. We
incorporated their feedback into the final survey.
We e-mailed each investigator an introductory letter explaining the
survey and the study in August 2006. Later in the month, we e-mailed
the survey. In September 2006, we e-mailed a second copy of the survey
to those who had not responded. At the end of September, we called the
investigators who had not yet replied. We received responses from
170 BiMo investigators for a 76-percent response rate.
Interviews with Food and Drug Administration officials. We conducted
telephone interviews with all 19 ORA directors of investigation (DIB) who
oversee BiMo investigators. Several DIBs invited supervisors in their
districts to participate in the interviews.
The interviews focused on the DIBs’ experiences with the supervisory
role, interactions with BiMo investigators, interactions with the
assigning centers, classification recommendations, and their challenges
in supervising BiMo inspections. Before the interviews, we solicited
comments from ORA officials and a DIB about the interview guide’s
content and clarity. We incorporated their feedback into the final
interview guide. We conducted the interviews in October and November
2006. At least two OIG staff participated in each interview.
We also interviewed senior FDA officials at FDA headquarters, ORA,
CBER, CDER, and CDRH.
Observation of Bioresearch Monitoring inspections. We observed three BiMo
inspections: two clinical investigator inspections and one IRB inspection.
FDA initiated the two clinical investigator inspections because of new
product applications. The IRB inspection was a directed inspection.
We based our observations on a protocol that focused on the process that
the BiMo investigator followed as well as the interactions between the
investigator(s) and the inspection subject(s). At least two OIG staff
participated in each inspection visit.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

28

A P P E N D

I X

~

A

Clinical trial registry site estimate. To compute an estimate of the average
number of clinical trial sites associated with a clinical trial, we used data
from the clinical trials listed on the National Institutes of Health’s clinical
trials registry as of February 2007. We selected a stratified random
sample of 150 trials. Because the registry is separated by trial phase, the
three strata used for the sample were Phase I, Phase II, and Phase III of
the clinical trial. (See Table 7 for our sample design.)
Table 7: Clinical Trial Registry Site Estimate Sample Design
Stratum

Population Size

Sample

Phase I

1,866

40

Phase II

4,566

50

Phase III

4,863

60

Source: Office of Inspector General analysis of the National Institutes of Health's Clinical Trial Registry.

We first counted the total number of clinical trial sites for each of the
three clinical trial phases. This served as our population size. Using
the OIG Office of Audit Service’s RAT-STATS software, we chose
random numbers to select the clinical trial. Because each trial in the
registry was assigned a number, we were able to use the random
numbers to identify the sampled trial. After sampling the trials, we
counted the number of sites associated with each sampled trial.
We excluded 43 of the 150 clinical trials sampled. There were three
situations in which the sampled trials were unusable and therefore had
to be excluded. First, some of the trials appeared in the population
multiple times because they were named Phase I/Phase II or
Phase II/Phase III. If a trial was listed as Phase I/Phase II, it was
counted as Phase II. Similarly, if a trial was listed as Phase II/Phase
III, it was counted as a Phase III. Eighteen occurrences of a Phase
I/Phase II trial in the Phase I population were excluded, as were seven
occurrences of a Phase II/Phase III trial in the Phase II population. In
addition, no sites were listed for 16 trials. Finally, terminated trials
occurred twice in the sample. We used the remaining sample of 107
trials to compute the average number of sites per trial. The sample was
weighted using standard statistical formulas for a stratified sample.
Based on these formulas, the estimated average number of sites for a
clinical trial is 23. This estimate can vary by +/-8 trials at the
95-percent confidence level.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

29

A

P

P E N D

I X ~

A

Review of Food and Drug Administration policies, procedures, and guidance
documents. We obtained and reviewed all relevant policies, procedures,
and guidance documents issued by FDA for conducting BiMo inspections.
We used these documents to better understand the process for conducting
inspections of clinical trials.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

30

A

P

Δ

P E N D

I X

~

A P P E N D I X

A

B

Agency Comments

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

31

A

P

P E N D

OEI-01-06-00160

I X

THE FOOD

~

AND

B

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

32

A

P

P E N D

OEI-01-06-00160

I X

THE FOOD

~

AND

B

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

33

N D
A
P P
I XE N ~ D

Δ

I BX

~

A
B

A C
P P
K EN NO DW I L X E ~ D B G M E N T S

This report was prepared under the direction of Joyce M. Greenleaf,
Regional Inspector General for Evaluation and Inspections in the
Boston regional office, and Russell W. Hereford, Deputy Regional
Inspector General.
Danielle Fletcher served as the team leader for this study. Other
principal Office of Evaluation and Inspections staff from the Boston
regional office who contributed include Rosemary Borck, Chris Galvin,
and Robyn Sterling, and central office staff include Ayana Everett.

OEI-01-06-00160

THE FOOD

AND

D R U G A D M I N I S T R AT I O N ’ S O V E R S I G H T

OF

CLINICAL TRIALS

34