case study  Investigational drug service operations

case  study

Improving investigational drug service

operations through development of an innovative
computer system
Burgunda V. Sweet, Helen R. Tamer, Rivka Siden, Scott R. McCreadie,
Michael E. McGregory, Todd Benner, and Roberta M. Tankanow

Purpose. The development of a comput-
erized system for protocol management,
dispensing, inventory accountability,
and billing by the investigational drug
service (IDS) of a university health system
is described.
Summary. After an unsuccessful search
for a commercial system that would
accommodate the variation among in-
vestigational protocols and meet regula-
tory requirements, the IDS worked with
the health-system pharmacy’s information
technology staff and informatics phar-
macists to develop its own system. The
informatics pharmacists observed work-
flow and information capture in the IDS
and identified opportunities for improved
efficiency with an automated system. An
iterative build–test–design process was
used to provide the flexibility needed for
individual protocols. The intent was to de-
sign a system that would support most IDS
processes, using components that would
allow automated backup and redundan-
cies. A browser-based system was chosen
to allow remote access. Servers, bar-code
scanners, and printers were integrated into
the final system design. Initial implementa-
tion involved 10 investigational protocols
chosen on the basis of dispensing volume
and complexity of study design. Other pro-
tocols were added over a two-year period;
all studies whose drugs were dispensed
from the IDS were added, followed by
those for which the drugs were dispensed
from decentralized pharmacy areas. The
IDS briefly used temporary staff to free
pharmacist and technician time for system
implementation. Decentralized pharmacy
areas that rarely dispense investigational
drugs continue to use manual processes,
with subsequent data transcription into
the system. Through the university’s tech-
nology transfer division, the system was
licensed by an external company for sale
to other IDSs.
Conclusion. The WebIDS system has im-
proved daily operations, enhanced safety
and efficiency, and helped meet regulatory
requirements for investigational drugs.
Index terms: Clinical studies; Computers;
Drugs, investigational; Hospitals; Pharma-
ceutical services; Pharmacy, institutional,
hospital; Protocols; Regulations
Am J Health-Syst Pharm. 2008; 65:969-
73

Problem gan Hospitals and Health Centers cal research studies, many of them
The investigational drug service (UMHHC) manages the drug com- involving multiple drugs. In recent
(IDS) at the University of Michi- ponent of approximately 280 clini- years, both the level of complexity of

Burgunda V. Sweet, Pharm.D., FASHP, is Director, Drug Informa-
tion Service and Medication-Use Policy, University of Michigan
Hospitals and Health Centers (UMHHC), Ann Arbor, and Clinical
Associate Professor of Pharmacy, University of Michigan College of
Pharmacy (UMCP), Ann Arbor; at the time of writing she was also
Director, Investigational Drug Service, UMHHC. Helen R. Tamer,
Pharm.D., is Clinical Pharmacist, Investigational Drug Service,
UMHHC, and Clinical Assistant Professor of Pharmacy, UMCP.
Rivka Siden, Pharm.D., M.S., is Clinical Pharmacist, Investigational
Drug Service, UMHHC, and Adjunct Clinical Assistant Professor
of Pharmacy, UMCP. Scott R. McCreadie, Pharm.D., M.B.A., is
Strategic Projects Coordinator, UMHHC; Adjunct Clinical Assistant
Professor, UMCP; and President, The McCreadie Group, Ann Arbor,
MI. Michael E. McGregory, Pharm.D., BCPS, is Strategic Projects
Coordinator, UMHHC, and Adjunct Clinical Instructor, UMCP; at
the time of this project, he was Specialty Resident in Pharmacy In-
formatics and Technology, UMHHC. Todd Benner is Applications
Programmer, Department of Pharmacy Services, UMHHC. Roberta
M. Tankanow, M.S., is Clinical Associate Professor of Pharmacy,
Emeritus, UMCP.
Address correspondence to Dr. Sweet at the Department of
Pharmacy Services, University of Michigan Health System, UH
B2D301/5008, 1500 East Medical Center Drive, Ann Arbor, MI
48109-5008 (gsweet@med.umich.edu).
The authors acknowledge James G. Stevenson, Pharm.D., FASHP,
for his support, which made this project possible.
Copyright © 2008, American Society of Health-System Pharma-
cists, Inc. All rights reserved. 1079-2082/08/0502-0969$06.00.
DOI 10.2146/ajhp070212

Am J Health-Syst Pharm—Vol 65 May 15, 2008 969

 case study  Investigational drug service operations
the research protocols and the pro-
tocol turnover rate have increased.
Approximately 8000 prescriptions
are filled annually by IDS staff, with
175 transfers of inventory monthly
to satellite pharmacies. All of this
translated into a workload that could
no longer be effectively or efficiently
supported with a manual bookkeep-
ing system. Lacking an automated
system, the IDS staff faced daily chal-
lenges such as generating prescrip-
tion labels with a programmable
typewriter and recording all inven-
tory transactions, whether receipt of
drug, transfer to satellite pharmacies,
or dispensing to patients, manually
on paper drug accountability logs.
The result was transcription and
accountability errors that required
extra time and effort to resolve.
Background
UMHHC comprises an 850-bed,
tertiary care university teaching
hospital with several affiliated health
centers and outpatient clinics. In
1983, the dean of the medical school,
in conjunction with the hospital
administration at the University of
Michigan, mandated that all clinical
trials involving the use of investiga-
tional drugs be coordinated by the
Department of Pharmacy Services.1
This mandate led to the development
of the IDS.
The IDS plays a major role in the
conduct of clinical trials in human
subjects that involve the use of in-
vestigational drugs.a The IDS reviews
study protocols before institutional
review board (IRB) submission to
assess the safety and feasibility of the
treatment plan, and it develops de-
tailed procedures for the preparation
and dispensing of investigational
drugs. The IDS is also responsible
for the receipt and appropriate stor-
age of study drugs, management of
drug inventory, and maintenance of
accurate drug accountability records.
Most (approximately 75%) medica-
tion orders for investigational drugs
are dispensed by the IDS pharmacy,
970 Am J Health-Syst Pharm—Vol 65 May 15, 2008
but some are dispensed from small
supplies of investigational drugs
maintained in the health system’s
decentralized pharmacy areas. IDS
staff members prepare and submit
the billing for pharmacy-related
activity (e.g., protocol setup, dispens-
ing, monthly maintenance fee) for
all studies coordinated by the IDS,
whether the drugs are dispensed by
IDS staff or through a decentralized
pharmacy. These activities are billed
to study accounts rather than to
third-party payers or patients.
In its first two years of existence,
the IDS managed 122 protocols and
dispensed over 4000 prescriptions
annually.1 Over the next five years,
the number of investigators conduct-
ing clinical trials increased substan-
tially, as did the number of protocols.
To manage the large number of drugs
dispensed to study patients and
maintain detailed inventory records,
a computer system using dBase III
software (Ashton-Tate, Culver City,
CA) was developed internally for
inventory control and billing.1 By the
late 1980s, this system was outdated
and was no longer supported. The
increasingly complex drug studies
did not work well with the dBase sys-
tem because it allowed only uncom-
plicated data entry. Lacking a system
that could perform the necessary
functions, the IDS decided to return
to a manual inventory accountability
and billing system.
Billing for services in the manual
system required that a technician
collate dispensing activity from all
decentralized pharmacy areas. Along
with data on the dispensing activity
by the IDS staff, these data were then
manually transcribed into an Excel
spreadsheet (Microsoft Corporation,
Redmond, WA) and forwarded to
UMHHC’s financial operations de-
partment for billing for the services
provided. The billing process was
time consuming, typically taking
approximately 40 hours per month,
and it was prone to transcription
errors. A decision to bill quarterly
improved efficiency somewhat, but
billing still required almost 100 hours
of technician time per quarter. Fur-
thermore, delaying billing for up to
three months increased the potential
for lost revenue because of closed
accounts and for missed activity
because of closed studies. Because
workload statistics were derived from
the same data set as were the bill-
ing data, the manual system did not
provide timely information on work-
load. Therefore, workload analyses
were conducted infrequently, which
made it difficult to assess trends in
productivity and to adjust staffing.
Beginning in the 1990s, several at-
tempts were made to either develop a
new system internally or purchase a
program from a commercial vendor,
but none of the attempts succeeded.
Internal resources for developing a
program were inadequate, and no
commercial product that met the re-
quirements was available. Therefore,
the IDS functioned for more than
15 years with a manual inventory ac-
countability and billing system.
Analysis and resolution
In 2003, the IDS revisited the use
of an automated system. Although
some of the dispensing functions in
the IDS parallel those for commer-
cially available drugs, several unique
aspects needed to be addressed.
Within the many protocols the IDS
manages, there is great variability in
protocol design and in how drugs are
provided for studies (e.g., patient-
specific kits, bulk bottles, blinded
bottles). In addition, services are
billed to grants and not to patients or
third-party payers. For all protocol
activity, there are regulatory require-
ments, including drug accountability
and data tracking. These regulatory
standards are set forth in a Food and
Drug Administration (FDA) docu-
ment known as the Good Clinical
Practice Consolidated Guidelines
(GCP)2 and in the Code of Federal
Regulations.3 Standards have also
been set by the Joint Commission4

and the American Society of Health-
System Pharmacists5 to ensure that
investigational medications are con-
trolled and safely administered.
The IDS team explored commer-
cially available automated systems
that might meet its unique needs.
All of the systems evaluated were
designed for other applications (e.g.,
narcotic control, blood product
tracking, outpatient dispensing)
but could have been adapted for the
IDS. However, commercially avail-
able systems designed for dispensing
prescriptions lacked (1) the flexibility
needed for the range of protocol de-
signs, (2) a mechanism for tracking
inventory, expiration dating, and lot
numbers, and (3) a means of billing
grant accounts. Commercially avail-
able systems for inventory manage-
ment did not work for dispensing
functions and label production.
Although some of the desired
capabilities were available, no system
offered all of the key functions that
the IDS believed were critical for its
operations. The IDS performed an
analysis to determine whether to
compromise functions and move
forward with a commercial program
or to dedicate internal resources to
building a system. After thorough
consideration of the advantages
and disadvantages of each option,
including the effect that dedicating
resources would have on other de-
partmental priorities, a decision was
made to design and build a system
internally. For the next two years, the
IDS staff worked closely with depart-
mental information technology staff
and informatics pharmacists to de-
velop the system that is now known
as WebIDS.
Building the WebIDS applica-
tion. The informatics pharmacists
had considerable expertise in inpa-
tient and ambulatory care pharmacy
systems and software development,
but they had limited experience with
the unique IDS environment. One of
the design goals of the project was
to look for opportunities to improve
case study  Investigational drug service operations
efficiency and safety rather than just
create an electronic version of the
paper-based processes. To meet that
goal, the project developers had to
learn the regulatory and procedural
details of the IDS environment. The
informatics pharmacists met regular-
ly with IDS staff and observed their
workflow, paying special attention to
the flow of information and the type
of information that needed to be
captured. Several opportunities were
identified for improving the opera-
tional efficiency of the IDS through
process reengineering in conjunction
with the new application.
After evaluation of the IDS work-
flow, functional requirements, and
existing data sources, the informat-
ics pharmacists presented initial
recommendations to the IDS team
for feedback and refinement. Process
flow charts, functional specifications,
and sample screen images were used
to drive these discussions. This exer-
cise revealed additional nuances of
IDS decision-making and introduced
unique situations that had to be
supported by the new system. Given
these intricacies and the likelihood
that other unanticipated scenarios
would arise as the project progressed,
the decision was made to use an
iterative build–test–refine method
based on an accepted framework and
required-functions list, rather than
develop a comprehensive scope and
design at the start. This saved con-
siderable planning time and gave the
project flexibility to address newly
discovered needs.
The application development
framework followed several design
principles. First, the system had to
support most IDS processes; this was
to prevent the emergence of parallel
paper and electronic processes and
preclude the need to use multiple
systems to complete regular tasks.
The software had to perform most
of the functions of other pharmacy
systems, including preparing high-
quality and consistent labels, allow-
ing for role-specific security with the
Am J Health-Syst Pharm—Vol 65 May 15, 2008
concept of pharmacist verification
of technician activity, and providing
for a simple and easy-to-use inter-
face, Health Insurance Portability
and Accountability Act compliance,
reporting, patient management, drug
management, and connections to
other electronic systems. The system
also had to support IDS-specific re-
quirements: protocol and IRB man-
agement, randomization, blinding
and enrollment of patients, billing
to study accounts instead of patients
or third-party payers, virtual inven-
tories, patient-specific inventories,
lot number-specific dispensing, and
an automatic audit trail as required
by FDA.
The second design principle was
to use enterprise-level components
such as server technology to allow
for automated backups to protect
against data loss. It was decided that
the system would be browser based
and would not require installation of
software on workstations; this was to
maximize availability of the software
and minimize the challenges of work-
ing with client software on managed
workstations. This decision would
allow authorized staff to access the
software from areas outside the IDS,
including decentralized pharmacies
and locations outside the institution.
The final design principle was that
the system had to be easy to use and
had to introduce overall efficiencies
into the IDS processes. All the func-
tions required for daily operations
were integrated into the application to
minimize the need for work-around
processes or for using additional sys-
tems to accomplish daily business.
Integrating all the required func-
tions into an easy-to-use system was
difficult and expensive. Building
the application required substantial
technical effort. These realities were
accepted because of the lack of viable
alternatives. The final system was de-
veloped using ASP.NET technologies
and an SQL Server database (Micro-
soft Corporation). Additional equip-
ment integrated with the system
971
 case study  Investigational drug service operations
included servers, bar-code scanners,
and thermal printers for labels.
Implementing the WebIDS sys-
tem. The application was first released
in February 2004. Ten protocols were
included in the initial implementa-
tion phase. The protocols chosen for
the initial phase had a high volume
of dispensing activity and a complex
study design. After initial use, bugs
and functional issues were resolved
with timely upgrades.
Once the initial phase was com-
pleted, the IDS staff began adding
more protocols to the system. Enter-
ing all new and existing studies at the
same time might have maximized
the benefit of the application, but
doing so would have been very labor
intensive. Instead, protocols were
added slowly to allow the application
to stabilize. A decision was made to
first enter studies, both existing and
newly IRB approved, in which drugs
were dispensed only by the IDS phar-
macy (phase 1, 162 studies). Once
these studies were in the system,
studies that were dispensed from the
IDS or decentralized pharmacy areas
were entered (phase 2, 108 studies).
A temporary pharmacist and techni-
cian were hired to assist with daily
operations for four months at the
start of each phase. These temporary
personnel freed up 1 full-time equiv-
alent (FTE) IDS technician and 0.4
FTE IDS pharmacist to work on the
implementation of WebIDS. When
temporary help was not available,
it was challenging for the IDS team
to sustain timely implementation of
the system while maintaining daily
operations.
Protocol entry into the system was
initiated by the IDS technician; the
IDS pharmacist checked the entry
and confirmed that the study activity
was ready to be managed by WebIDS.
Full implementation was spread over
two years, with parallel manual and
electronic systems being used for
protocol management and dispens-
ing. Throughout implementation of
WebIDS, the team encountered stud-
972 Am J Health-Syst Pharm—Vol 65 May 15, 2008
ies with complex designs and unique
product packaging; close work with
the informatics pharmacists to im-
prove the flexibility, efficiency, and
ease of use of WebIDS was needed to
accommodate these studies.
Experience with the system
For the past two years, protocol
management (for the drug compo-
nent of the protocol) for all studies
handled by IDS has occurred entirely
within the WebIDS system; some of
these studies have been managed by
the system for more than four years.
The WebIDS system has increased the
accuracy and efficiency of IDS opera-
tions. The system alerts the dispenser
if IRB approval for a study has expired.
It records all drug accountability data
and dispensing activity in real time.
Labels are designed at the time a study
is entered into the system, so subse-
quent prescriptions are filled with
consistent labels, thereby minimizing
potential errors. The risk for account-
ability errors has been reduced, since
inventory can only be logged out of
the study to which it is assigned, and
prescriptions can only be filled for pa-
tients who are enrolled in the study. If
a technician or pharmacist tries to en-
roll a patient in two studies simultane-
ously, the system alerts the dispenser
to contact the investigator. Because
there is a list of authorized prescribers
for each study, the system ensures that
the investigational drug is ordered by
an individual with prescribing privi-
leges on the protocol.
Billing is automated and occurs
monthly. This has allowed significant
improvement in billing efficiency and
an overall increase in revenue. Work
by IDS pharmacy staff is recorded
in the system at the time the work is
done, and work done in the decen-
tralized pharmacy satellites is cap-
tured within a month of the activity
(i.e., when data transcription from
the manual to the automated system
occurs). This means that workload
reports can be generated at any time
to reflect current data. The system
also can generate other reports, such
as drug supplies nearing or past the
expiration date or drugs at or below
the minimum reorder level, on a
regular basis to help the IDS manage
drug inventory.
All of these improvements have
increased the efficiency of IDS opera-
tions. The IDS has attained a level of
technology commensurate with other
areas of pharmacy operations. Since
workload data were rarely captured
in the manual system, differences in
workload since system implementa-
tion cannot be measured. WebIDS
has added a workload component in
the initial study setup to ensure that
the protocol, investigators, and study
supplies are established in the system
for proper dispensing and billing.
As is true with most automation, a
shift has occurred in where effort is
expended. IDS business continues to
grow both in volume and complex-
ity, which adds to the difficulty of
assessing the effects of the system on
workload and staffing.
Given the benefits at our institu-
tion and the similar challenges faced
by other IDSs, the application was
disclosed to the university’s technol-
ogy transfer division. The applica-
tion was subsequently licensed by an
external company (The McCreadie
Group, Ann Arbor, MI) and is being
sold to and installed at other IDSs
around the country.
Ongoing challenges. Although
the overall impact of the system on
IDS operations has been positive,
there are ongoing challenges. At
UMHHC, only the IDS staff and staff
of the outpatient pharmacy in the
cancer center are currently trained
on the WebIDS system. The volume
of investigational drugs dispensed
in the cancer center outpatient clinic
is high, so training the staff in that
area made sense and allowed IDS
dispensing by manual methods to
be reduced. Decentralized pharmacy
areas that dispense investigational
drugs less frequently continue to
use manual accountability logs, and

IDS staff retrospectively transcribe
orders into WebIDS for billing and
record-keeping. Thus, a dual system
(manual and WebIDS) continues to
exist for some protocol activity. It
might be ideal to have satellite phar-
macies use the WebIDS system for
IDS dispensing, but it may not be
practical to train all pharmacists on
a system they would use only occa-
sionally. Furthermore, it is not prac-
tical in daily operations for satellite
pharmacists to dispense using two
different computer systems, one for
commercial drugs and the other for
study drugs.
Another chal lenge for the
UMHHC IDS is working toward an
entirely paperless system in which
all drug accountability records are
available only through the WebIDS
system. A paperless system has
been acceptable to many study
sponsors, but other sponsors in-
sist that company-specific forms
be completed. This should change
in the future, since the clinical trials
industry is gradually moving toward
electronic records through the devel-
opment and adoption of standards.6
As more IDSs across the country
move to paperless systems, this prac-
tice should become better accepted
by study sponsors.
case study  Investigational drug service operations
Expansion in system capability.
WebIDS has allowed improvement
in the efficiency and effectiveness of
daily IDS operations. The system ca-
pabilities continue to expand as new
ideas are brought forward and new
challenges are faced. For example,
when the UMHHC IDS recently
moved its drug inventory to a vertical
carousel system (Omnicell Pharm-
Central, Mountain View, CA), an
interface was developed to allow use
of WebIDS technologies for bar-code
verification of picking and stocking
transactions. Such technologies are
not foolproof, and humans must
still place items into the storage unit
correctly, but they are another step
in improving the medication-use
process and minimizing the potential
for medication errors. Through the
external vendor, the application has
continued to mature, with increased
functions and refinement of existing
ones.
Conclusion
The WebIDS system, codeveloped
by experts in the area of the IDS and
pharmacy informatics, has improved
IDS daily operations, enhanced safe-
ty and efficiency, and helped meet the
stringent regulatory requirements for
this environment.
Am J Health-Syst Pharm—Vol 65 May 15, 2008
a
At UMHHC, an investigational drug is
defined as a chemical or drug that is not FDA
approved for use in humans, a commercial
product or an FDA-approved drug provided
at no charge by the study sponsor for use in a
comparative trial with other drugs or placebo,
or an FDA-approved drug supplied at no
charge to the subject to be used for a non-
FDA-approved indication for the purpose of
a study.
References
1. Ryan ML, Colvin CL, Tankanow RM.
Development and funding of a pharmacy-
based investigational drug service. Am J
Hosp Pharm. 1987; 44:1069-74.
2. Food and Drug Administration, Depart-
ment of Health and Human Services.
International Conference on Harmo-
nization: Good Clinical Practice Con-
solidated Guidelines. Fed Regist. 1997;
62:25692-709.
3. Food and Drug Administration. Code
of Federal Regulations. Title 21, foods
and drugs. Subchapter D, drugs for hu-
man use; part 312, investigational new
drug application. www.accessdata.fda.
gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.
cfm?CFRPart=312 (accessed 2007 Apr 4).
4. Medication management. Section 1. In:
Comprehensive accreditation manual
for hospitals. Oakbrook Terrace, IL: Joint
Commission on Accreditation of Health-
care Organizations; 2004:MM1-20.
5. Guidelines on clinical drug research. In:
Deffenbaugh JH, ed. Best practices for
health-system pharmacy: position and
guidance documents of ASHP 2006–2007.
Bethesda, MD: American Society of
Health-System Pharmacists; 2006:318-24.
6. Clinical Data Interchange Standards
Consortium. Draft domains posted for
review. www.cdisc.org/models/sdtm/v1.1/
index_no_registration.html (accessed
2007 Apr 24).
973