Integrated Data Acquisition System for Medical Device Testing and
Physiology Research in Compliance with Good Laboratory Practices
Steven C. Koenig1, Ph.D., Cary Woolard1, Guy Drew2, Lauren Unger1, Ph.D., Kevin
Gillars1, M.S., Dan Ewert3, Ph.D., Laman Gray1, M.D., and George Pantalos1, Ph.D.
1Jewish
Hospital Cardiothoracic Surgical Research Institute at the University of
Louisville, Department of Surgery, Louisville, KY 40202
2 US Army Institute of Surgical Research, Fort Sam Houston, TX 78234-6315
3Department of Electrical and Computer Engineering, North Dakota State University,
Fargo, ND 58105
*Funding
for this project was provided by a grant from the Jewish Hospital Heart and
Lung Institute (Louisville, KY).
Running Title
Data Acquisition System
Keywords
Data Acquisition, GLP Compliance, Medical Devices, Physiology
Correspondence
Steven C. Koenig, Ph.D.
Associate Professor
Jewish Hospital Cardiothoracic Surgical Research Institute
at the University of Louisville
500 South Floyd Street, Room 118
Department of Surgery
University of Louisville
Louisville, KY 40202
TEL: (502)-852-7320
FAX: (502)-852-1795
e-mail: sckoen01@athena.louisville.edu
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
ABSTRACT
In seeking approval from the U.S. Food and Drug Administration (FDA) for clinical trial
evaluation of an experimental medical device, a sponsor is required to submit
experimental findings and support documentation to demonstrate device safety and
efficacy that are in compliance with Good Laboratory Practices (GLP). The objective of
this project was to develop an integrated data acquisition (DAQ) system and
documentation strategy for monitoring and recording physiological data when testing
medical devices in accordance with GLP guidelines mandated by the FDA. DAQ
systems were developed as stand-alone instrumentation racks containing transducer
amplifiers and signal processors, analog-to-digital converters for data storage, visual
display and graphical user-interfaces, power conditioners, and test measurement
devices. Engineering standard operating procedures (SOP) were developed to provide
a written step-by-step process for calibrating, validating, and certifying each individual
instrumentation unit and the integrated DAQ system. Engineering staff received GLP
and SOP training and then completed the calibration, validation, and certification
process for the individual instrumentation components and integrated DAQ system.
Eight integrated DAQ systems have been successfully developed that were inspected
by regulatory affairs consultants and were determined to meet GLP guidelines. Two of
these DAQ systems were used in support of 40 of the pre-clinical animal studies to
evaluate the ABIOMED artificial heart. Based, in part on these pre-clinical animal data,
the AbioCor clinical trials began in July 2001. The process of developing integrated
DAQ systems, SOP, and the validation and certification methods used to ensure GLP
compliance are presented in this paper.
Submitted to: Biomed. Instr. & Tech.
1
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
INTRODUCTION
Scientists, engineers, and clinicians record experimental data to evaluate physiologic
responses with medical devices during acute and/or chronic testing. Assessment of
cardiovascular function on a systems level requires the periodic or continuous
measurement and monitoring of pressures, flows, volumes, and/or electrocardiogram.
The output of the transducers used to measure these physiologic waveforms are
typically in the microvolt or millivolt range, and subsequently require signal conditioning
to provide amplification and/or offset to maximize the input range of the recording
device to optimize data integrity. In the 1960’s, many data acquisition methods involved
recording and analyzing data using strip chart recorders (Maloy 1986, Wilkison 1984).
Although acceptable with proper use and analysis, extrapolation of key physiologic
parameters using this approach can be tedious and time consuming. Experimental data
often consisted of handwritten documentation in laboratory notebooks and/or strip chart
recordings. Over the past several decades, there has been a migration from analog
tape and/or standard strip chart recorders toward digital data acquisition and analysis
systems in which data can be streamed directly to a digital storage device. The
primary advantages to the digital approach is the ability to store large volumes of data,
perform waveform analyses, and by taking advantage of processor speed one can
analyze more data in a faster, more efficient manner. A number of turn-key
instrumentation and software packages are now commercially available to provide data
acquisition and analysis, including BioBench (National Instruments, Austin, TX),
PowerLab (ADInstruments, Grand Junction, CO), ARIA-1 (Millar Instruments,
Submitted to: Biomed. Instr. & Tech.
2
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Houston, TX), DADiSP (DSP Development Corp, Newton, MA), PO-NE-MAH (Gould
Instrument Systems, Valley View, OH ), and WinDAQ (Akron, OH).
To protect the American public against fraudulent products that are consumed either in
or on the body, the Congress passed the Food, Drug, and Cosmetic Act in June 1938.
This Act called for the implementation of regulations for the development, testing, and
marketing of many products. The Medical Device Amendment, passed May 28, 1976,
expanded the scope of the Food, Drug, and Cosmetic Act to include the regulation of all
medical devices. These regulations were to be implemented and carried out by the
U.S. Food and Drug Administration. As a part of this implementation process,
guidelines for the conduct of any experiments to generate data to be submitted to the
FDA seeking product approval were drafted and offered for public comment in 1977.
The final version of this guideline was published in the Federal Register on December
22, 1978 as Title 21 of the Code of Federal Regulations (CFR), Part 58, with the title
Good Laboratory Practice for Non-clinical Laboratory Studies. More commonly referred
to as the GLPs or GLP guidelines, this FDA document has been revised twice, most
recently in January 1999.
This article reviews the effort to establish a system for
computer-based data acquisition and analysis that is compliant with GLP guidelines.
According to the Code of Federal Regulations (CFR), all research facilities that conduct
laboratory studies for submission to a regulatory agency such as the US Department of
Health and Human Services and the FDA, are required to establish and maintain a
current management system to assure that GLPs are followed. Title 21 CFR, Part 58,
Submitted to: Biomed. Instr. & Tech.
3
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Good Laboratory Practice for Non-clinical Laboratory Studies describes the standards
for conducting “studies that support or are intended to support applications for research
or marketing permits for products regulated by the FDA, including food and color
additives, animal food additives, human and animal drugs, medical devices for human
use, biological products, and electronic products.” “Compliance with this part is
intended to assure the quality and integrity of the safety and efficacy data filed pursuant
to sections 406,408, 409, 502, 503, 505, 506, 507, 510, 512-516, 518-520, 721, and
801 of the Federal Food, Drug, and Cosmetic Act and sections 351 and 354-360F of the
Public Health Service Act.”
In accordance with 21 CFR, Part 58, specific standard operating procedures (SOP) are
required for each piece of equipment used for data acquisition and monitoring. The
SOP frequently incorporates the specific instructions contained in the equipment
manufacturers’ manual. Details on the methods, materials and schedule for inspecting,
cleaning, maintaining, testing, standardizing and calibrating at the laboratory where the
experiments are being conducted are required, and written records of all of these
procedures must be maintained. Any remedial action that is taken in the event of
equipment failure must also be documented. The designated person(s) responsible for
the performance of each operation described must be qualified and well trained. All
personnel working with the equipment must read, understand and receive in-house
certification to use the equipment.
Submitted to: Biomed. Instr. & Tech.
4
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
In response to advances in technology and paperless record keeping, expanded
guidelines were proposed. 21 CFR Part 11 – Final Rule issued in 1997 sets the criteria
under which the FDA will consider electronic records and electronic signatures to be
equivalent to paper records and the more conventional handwritten signatures,
respectively. The FDA defines electronic records as those records created, modified,
maintained, archived, retrieved, or transmitted electronically. Electronic records that
meet the requirements of 21 CFR Part 11.2 may be used in lieu of paper records.
Computer systems (including hardware and software), control processes, and attendant
documentation need to be well organized and readily available because they are
subject to FDA inspection. The process toward achieving GLP compliance with
electronic record keeping (i.e., digital data acquisition) has presented a significant
challenge due to the complexities in developing SOP, validating, and certifying
computer hardware and software. Further, it is quite common for many investigators to
incorrectly assume that a commercially developed data acquisition and analysis
program is GLP compliant. To the best of our knowledge, we are unaware of any
commercial, digital data acquisition systems that are GLP compliant. The objective of
this project was to develop an integrated data acquisition (DAQ) system and
documentation strategy while maintaining quality assurance for monitoring and
recording physiological data during testing of medical devices that meet GLP guidelines
mandated by the FDA. The process for developing, testing, documenting, and certifying
the integrated DAQ system is presented.
Submitted to: Biomed. Instr. & Tech.
5
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
METHODS
Data Acquisition Instrumentation Rack. Transducers, amplifiers, and signal processors
were purchased from commercial vendors based on extensive performance testing and
evaluation that has been previously reported. The selection criteria for pressure
measurement was 10 V excitation voltage, fixed gain up to 10 V, frequency response
to 5 kHz, ability to perform multiple physiologic calibration procedures, and long-term
stability and reliability (Reister 1998). Flow measurement instrumentation was selected
by comparing electromagnetic, doppler, and transit-time techniques in a large animal
model to evaluate accuracy, reliability, frequency response, and waveform morphology
(Koenig 1996). Instrumentation amplifiers, signal processors, a custom developed
signal conditioner and distribution unit, analog-to-digital converters and data storage,
visual display and graphical user-interfaces, power conditioners, and test measurement
devices were then integrated in a custom designed, DAQ instrumentation rack (Figure
1a). This approach provided a stand-alone system that enabled cabling to be routed in a
structured manner that could be visually inspected, well documented, and minimized
electrical noise in the acquired physiological data (Figure 1b). At the base of the DAQ
instrumentation racks are industrial grade wheels that allow easy transport of these
systems to different locations. The DAQ instrumentation racks are comprised of three
key elements: (1) signal conditioning components, (2) communication ports, and (3)
back-up power supply with LED display features. The layout for the DAQ
instrumentation racks was developed using computer-aided design software (AutoCAD,
San Rafael, CA). The CAD designs (Figure 2) were sent to an outside vendor (Premier
Metal, Bronx, NY) for fabrication of the DAQ rack housing.
Submitted to: Biomed. Instr. & Tech.
6
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
(Figures 1-2)
The signal conditioning components of the DAQ instrumentation rack include a chassis
with six pressure amplifiers (Ectron Model 428, San Diego, CA ), 2-channels of flow
(Triton Technologies, San Diego CA or Transonics, Ithaca, NY), and 2-channels of ECG
(Gould Instruments, Cincinnati, OH. The low-level analog output (V or mV) from these
amplifiers and signal processors is fed into a 16-channel signal conditioning and
distribution unit (to be described later). These analog data are routed through an
analog-to-digital (A/D) accessory (BNC-2090, National Instruments, Austin, TX)
mounted to the DAQ instrumentation rack, and converted to digital format via an A/D
board (AT-MIO-16E-10, or PCI-MIO-16XE-10, National Instruments, Austin, TX) housed
inside a desktop computer (Micron, Boise, ID) and displayed in real-time on the
computer monitor also mounted in the DAQ instrumentation rack. The analog data may
also be displayed on a variety of precision test and measurement instrumentation
devices including a digital multi-meter (DMM, Fluke model 45, Carrollton, TX ) and/or
digital storage oscilloscope (Tektronix model 340A, Beaverton, OR). The amplifiers,
signal processors, and precision test measurement instrumentation devices were fitted
with heavy duty extension slides (Premier Metal, Bronx, NY) mounted to tapped panel
mounting holes on the DAQ rack housing frame set at 17¾” standards.
Two interface panels are located on one side of the DAQ instrumentation rack (Figure
1c). One interface panel provides 16-channels of analog input and 16-channels of
analog output via BNC connectors. The other interface panel includes LPT1 for
printing, COMM1 for RS-232, a category-5 Internet access connector, and an auxiliary
Submitted to: Biomed. Instr. & Tech.
7
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
input for future communication applications. Additional features of the DAQ
instrumentation rack include up to 2 hours of full load battery power back up via an
uninterruptable power supply (Best, Necedah, WI ) and large LED display modules
(Simpson, Elgin, IL). The UPS provides clean power and prevents loss of study data
due to unexpected power surges or power outages. The large LED display modules
provide real-time heart rate, mean arterial pressure, and/or cardiac output that
physicians can easily view at a glance from across the operating table.
Variations of the described DAQ instrumentation rack have been developed in support
of different application requirements. Specifically, DAQ systems were designed with
medical isolation and medically approved amplifiers and signal processors for clinical
intraoperative measurements and monitoring (Figure 3a) that are not required for animal
testing, precision test measurement instrumentation for engineering development
(Figure 3b), and standard instrumentation for universal data collection (Figure 3c). The
clinical monitoring DAQ system includes electrical isolation to minimize risk of
accidental injury to the patient during intraoperative data collection in clinical operating
rooms, catheterization laboratories, or other appropriate hospital settings. The clinically
rated DAQ system (38”h x 22”w x 29”d, 250 lbs) has a chassis with up to eight pressure
and/or ECG amplifiers that are isolated to Association for the Advancement of Medical
Instrumentation (AAMI) standards (Gould Instruments, Cincinnati, OH). In place of a
large desktop computer, a notebook computer (Tecra 8100 Toshiba, New York, NY)
with an A/D card (DAQCard™-AI-16XE-50, National Instruments, Austin, TX) is used.
The notebook computer affords portability of the data beyond the clinical setting. The
Submitted to: Biomed. Instr. & Tech.
8
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
engineering development DAQ system (70”h x 27”w x 29”d, 400 lbs) includes a digital
storage oscilloscope for design and development projects. A universal data collection
DAQ system (78”h x 27”w x 29”d, 500 lbs) provides general data collection capabilities
from a variety of external signal conditioners and/or medical monitoring devices in
support of non-clinical studies.
(Figure 3)
Signal Conditioning and Distribution Unit. A key component of the integrated DAQ
system was the development of a 16-channel signal conditioning and distribution (SCD)
unit to drive multiple peripheral monitoring and recording devices without significant loss
of signal strength and integrity. The SCD unit was also designed with finite fixed gain
and offset to maximize A/D input range of peripheral devices, and low-pass filters to
remove electrical noise and prevent aliasing. Specifically, the design criteria per
channel was defined as follows:

Input Impedance = 10 K
 Negative Offset = up to -4 VDC (1.0 mV)

Output Load Impedance ≥ 2 K
 Positive Offset = up to 4 VDC (1.0 mV)

Zero setting = 0 VDC (1.0 V)
 Pos/Neg gain = 1x, 2x, 3x, 4x, 5x (0.1 %)

Inverter = -1x (1 mV)
 Low pass filter = 60 Hz (24 dB/octave)
A component view of the main amplifier driver for each channel consisting of five
amplifier stages is shown in Figure 4. The analog input signal (J1) is initially fed into a
differential input amplifier (U6-AD620AN) that can be configured to operate in singleended or differential mode (Analog Devices, Norwood MA). The second stage consists
of positive or negative offset and zero adjust networks that are switch selectable (SW11 to SW1-3). These networks also include voltage regulators (U2-MC78L05ACP and
U3-MC79L05ACP) to minimize conducted/radiated external electrical noise (Motorola,
Submitted to: Biomed. Instr. & Tech.
9
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Austin, TX). For example, the negative offset network includes a resistor combination
(R12-R3-R11) that can be configured to provide up to -4 VDC offset. The offset and zero
networks are then fed into a second amplifier (U1-AD620AN, Analog Devices, Norwood
MA). The third stage provides antialiasing with low pass filters (U4-D74L4B60Hz) that
have a cutoff frequency of 60 Hz with a 24 dB/octave roll off (Frequency Devices,
Haverhill MA). Assuming a minimal sampling rate of 400 Hz, which is common for most
medical applications, then by the Nyquist Criteria (sampling frequency greater than
twice the cutoff frequency of the filter) the physiological data will not be distorted in gain
and phase due to aliasing. The user also has the option to invert (SW2) the filtered
output before it is fed into a third amplification stage that provides user-selectable fixed
gain steps (SW3-1 to SW3-4) from 1 up to 5 in steps of 1. This gain stage allows
the user to optimize the resolution of the data by maximizing the input range of the
peripheral monitoring/recording device. The output of stage three is then fed through a
series of buffer drivers containing precision bipolar amplifiers (U7-AD704JN) that can
drive multiple output devices (Analog Devices, Norwood MA).
(Figure 4)
The 16-channel SCD unit is wired into the integrated DAQ instrumentation rack to
provide additional signal conditioning and to distribute physiological analog signals to
multiple output devices. The back panel of the SCD contains a single row of BNC input
connectors and three rows of BNC output connectors aligned in 16 columns,
representing 16 channels of signal conditioning. In other words, each channel of the
SCD has a column of BNC connectors for each input signal and three outputs. In
addition to the 164 matrix, there is a column of 3 BNC connectors that allow any of the
Submitted to: Biomed. Instr. & Tech.
10
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
16 output channels to be distributed to any of three precision test measurement devices
(i.e., digital multimeter, oscilloscope, etc.) that are user-selectable. The front panel of
the SCD contains 3 rows of user-select switches and LEDs aligned in 16 columns
(Figure 5). Subsequently, each channel represents a column of three user-select
switches that enable the conditioned signal output to be distributed to one of three
peripheral monitoring/recording output devices.
(Figure 5)
The front panel switches have been configured such that only one output channel per
row of switches can be routed to a peripheral monitoring/recording output device.
Additionally, the switch enable scheme was configured to allow multiple switches to be
turned ON, however, only the switch corresponding to the highest channel number will
be enabled. This can be confirmed visually by the LED corresponding to each channel
and output monitor device channel. For example, if channels 2 and 7 are switched ON
for monitor output device 1 (i.e., multimeter), then only the signal conditioned output
from channel 7 will be enabled and it’s corresponding LED illuminated. This
configuration allows the user to easily check the conditioned output for a large number
of signals without constantly flipping switches ON and OFF with the possibility of
missing a recording for a particular channel. For example, during calibration recordings
for 6 channels of pressure, the user turns ON the switch for output channel 1 to record
the output voltage on a DMM then repeats the procedure for output channels 2-6
without having to turn their corresponding switches OFF. Subsequently, this provides a
form of error checking by allowing the user to visually confirm that each channel has
been properly selected.
Submitted to: Biomed. Instr. & Tech.
11
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Data Acquisition Software. Since 1994, the authors have been developing a data
acquisition program for analog-to-digital conversion, real-time graphical display, and
storage of cardiovascular hemodynamic data in support of non-GLP experimental
protocols. The Cardiovascular Data Acquisition Software (CDAS) program (Drew 2000,
Drew 2002) was developed in LabVIEW (National Instruments, Austin, TX), an industrystandard programming development package commonly used by many
mathematicians, engineers, and scientists. It is a graphically oriented programming
environment that enables development of indicators and controls for real-time data
acquisition compatible with industry-standard analog-to-digital (A/D) hardware (National
Instruments, Austin, TX). The CDAS data acquisition program is configured to run
using four menu-driven screen modes. In mode 1, Measurement and Automation
(National Instruments, Austin TX) is used to assign channel names (i.e., Aortic
Pressure), identify analog input channels (i.e.,channel 1), convert physical input values
to physiologically equivalent units (i.e., 0-2 V = 0-200 mmHg), identify data acquisition
device, and analog-to-digital conversion format (i.e., non-reference single-ended input)
as shown in Figure 6.
(Figure 6)
The user can document the experimental data by annotating fields in the Program
Profile Menu screen (mode 2, Figure 7). Documentation for experimental data sets
include: (1) Data File Parameters to assign a filename and length of data set; (2) File
Header Information to annotate laboratory, organization, study title, subject number,
date of experiment recording, IACUC/IRB protocol number (medical ID number), DAQ
operator, and filename (extension automatically includes date yy/mm/dd and time data
Submitted to: Biomed. Instr. & Tech.
12
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
collected); and (3) DAQ Channel Setup to annotate abbreviations for individual channel
names assigned in Measurement and Automation (mode 1), which can be configured to
display data in a variety of formats. Other user-selectable options in mode 2 include the
ability to simultaneously record two separate data sets (i.e., two test subjects, Figure 8),
select fixed data collection epochs (i.e., 30 second files) and/or continuous data
recording (i.e., on-off toggle control), and/or load a previously defined program set-up
file. Indicators and pop-up warning menus provide real-time user feedback in mode 2 to
ensure no errors have been inadvertently made during the configuration process. Upon
successful completion of the set-up and configuration processes, the “run profile”
control can be selected to initiate waveform monitoring and data collection (mode 3).
(Figures 7-8)
Following “run profile” a continuous waveform chart display of up to 16 channels (one
subject) or 8 channels (two subjects) can be initiated (Figure 8). Selected patient
parameters can be monitored continuously and/or stored in automatically incremented
data files (i.e., one data file recording of 30 seconds length every hour for 48 hours).
Additional documentation for each data set can be logged in a “Notes” indicator. An
indicator also identifies whether the continuous (i.e., on-off switch) or epoch data (i.e.,
30 second file) function has been selected and illuminates when enabled. A split screen
feature enables simultaneous monitoring and/or data collection from two different
experiments. There is an overlay feature that allows multiple waveforms from the same
experiment to be displayed in the same graph (i.e., overlay Aortic and Left Ventricular
Pressure). A “freeze” indicator allows the user to stop continuous display and study
individual waveform characteristics without interrupting data collection. An exit indicator
Submitted to: Biomed. Instr. & Tech.
13
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
allows the user to return to the Program Profile Menu.
A Data Viewer option (mode 4) allows the user to retrieve and review previously
recorded data sets. The header information and ASCII data for each channel are
automatically displayed. The entire data set or individual epochs can then be replayed
through data display graphs. Indicators located below the data display graphs identify
starting and ending data points and length of data set displayed. ASCII data sets can
easily be imported into a spreadsheet (i.e., Excel) or loaded in a data analysis package.
GLP Compliance Implementation. The development and extensive testing of the DAQ
systems alone are not sufficient to meet GLP standards mandated by the FDA. Further,
it is incorrect to assume automatic compliance by purchasing individual instrumentation,
integrated systems, and/or data acquisition and analysis software from a commercial
vendor. Documentation consisting of standard operating procedures, certification of
testing, calibration, maintenance, and validation of instrumentation, and a quality
assurance unit are required to develop, monitor and audit the conduct of the study. The
documentation must be readily available for FDA inspectors enabling them to determine
whether procedures have been followed to ensure study integrity. The documentation
and quality assurance processes are described next.
Standard Operating Procedures. A standard operating procedure (SOP) is a written set
of instructions for completing a specific task or operating a specific piece of equipment
that has completed a rigorous review and been signed as approved by the author,
Submitted to: Biomed. Instr. & Tech.
14
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
supervisor, documentation coordinator, quality assurance manager, and facility
reviewer. The SOP pertaining to the individual instrumentation units and the integrated
DAQ system contain written instructions with procedural methods, materials, schedules
of inspection, cleaning, maintenance, testing, calibration, and/or standardization.
Instrumentation SOP assign designated personnel, list materials and equipment,
provides instructions for inspection, calibration, maintenance, and certification, and
contingency plans to complete the task. Personnel must document their qualifications
through education, experience and training records and for each individual SOP they
are responsible for using. A sample engineering instrumentation SOP is presented in
the Appendix.
Calibration and Maintenance Procedures. All pieces of equipment and instrumentation
are calibrated and maintained as defined by their corresponding SOP. All in-house and
off-site calibration and maintenance is documented and duplicated in Calibration and
Maintenance notebooks, and entered into a certified GLP compliant database for the
laboratory (The Calibration Manager® Database, Blue Mountain Quality Resources,
State College, PA). The Calibration Manager® Database is validated periodically to
ensure effective functionality of the software. Calibration and maintenance labels are
generated and affixed on the front of all instrumentation for quick visual inspection. In
the event of a failed calibration or maintenance check a Failed Calibration Data Integrity
Report is filled out identifying the studies effected, impact on studies, and recommended
course of action. Any resulting repair, maintenance, and subsequent re-calibration
Submitted to: Biomed. Instr. & Tech.
15
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
and/or repair are documented before the piece of equipment or instrumentation is
released for use.
Validation Testing and Certification. A critical step toward achieving GLP compliance of
the DAQ system is the validation of the individual measurement instrumentation and the
integrated system. This is accomplished using regression, stress and performance
testing methods that are documented and certified. The individual instrumentation
components of the DAQ system were tested and certified according to their individual
standard operating procedures (SOP), as described earlier. The integrated DAQ
system containing each of these validated and certified components was then tested as
a stand-alone unit. First, a GLP compliant voltage standard (DVC 8500, Calibrators,
Inc., Mansfield, MA) was used to generate static analog input voltage steps in finite
increments from –10 V to + 10 V (range of A/D converters). The displayed and
recorded voltages were compared to the voltage of the data recorder as well as
readings taken from a GLP compliant digital multimeter (Fluke 45 DMM, Carrolton, TX).
Second, a GLP compliant function generator (Tektronix CFG 280, Gaithersburg, MD)
was used to generate square, sawtooth, and sinusoidal waveforms of varying amplitude
(10 V) and frequency (1-1000 Hz), and the displayed and recorded waveforms
compared to the settings of the function generator. Third, a GLP compliant patient
simulator (medSim 300B, Dynatek Nevada, Carson City, NV) was programmed to
produce static (0, 40, 80, 100, and 200 mmHg) and dynamic pressure waveforms (atrial,
ventricular, and arterial) of varying amplitudes (0, 40, 80, 100, and 200 mmHg) and
heart rates (30, 60, 80, 120, 160, 200, and 240 bpm). The displayed and recorded
Submitted to: Biomed. Instr. & Tech.
16
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
waveforms were compared to the settings of the patient simulator. Test procedures
were performed for short (24-hours) and long (one-week) evaluation periods. The
experimental results for all test conditions were then documented using test procedure
and validation forms archived in our GLP storage facilities. GLP QAU personnel
performed audit validation testing procedures and support documentation, for
compliance with GLP guidelines. Any discrepancies found during the validation testing
were documented, a course of action recommended, the actions were implemented,
and validation re-testing were conducted and documented.
RESULTS
In 1998, the Jewish Hospital Cardiothoracic Surgical Research Institute at the University
of Louisville (UofL) established a GLP program that includes an in-house Quality
Assurance Unit (QAU). All laboratory personnel received extensive GLP and SOP
training by regulatory affairs consultant (Kathleen Zajd, Prologue Research
International, Westerville, OH) to aid in the Institute’s development of the GLP program.
Engineering staff completed the calibration, validation, and certification process for the
individual instrumentation components and integrated DAQ systems. Eight integrated
DAQ systems have been successfully developed and audited by several quality
assurance specialists (NAMSA, Northwood, OH) and were determined to meet GLP
guidelines. These GLP compliant DAQ systems have successfully supported industrysponsored and federally funded research projects for the past three years. We continue
to maintain an active GLP program, performing periodic reviews of our SOP and
support documentation to ensure compliance.
Submitted to: Biomed. Instr. & Tech.
17
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
For example, our GLP program has been instrumental in contributing to the pre-clinical
animal studies (Figure 9) required for the ABIOMED AbioCor™ totally implantable
replacement heart (Danvers, MA). Those efforts resulted in FDA approval and initiation
of the multi-center clinical trial for the AbioCor™ resulting in the world’s first two clinical
implants at Jewish Hospital (Louisville, KY). In support of this study, discrete
experimental data points (i.e., heart rate, and systolic/diastolic blood pressure) were
displayed using two DAQ systems outfitted with medical monitoring instrumentation
(Hewlett-Packard medical monitor, Andover, MA). The data were transcribed manually
onto data record forms that were entered into a GLP-validated database developed by
Advertek, Inc. (Louisville, KY). The signed and dated hardcopy printouts have been
identified as the "raw data" and are maintained in conventional archives to support the
submitted study reports. The data acquisition and analysis software currently used for
non-GLP studies is approaching full compliance with the FDA’s regulations pertaining to
electronic records and electronic signatures.
(Figure 9)
DISCUSSION
The concept of designing an integrated DAQ system arose from the need to consolidate
measurement instrumentation in already overcrowded surgical suites and postoperative holding facilities and the desire to optimize the data acquisition process.
The
introduction of federal regulations that pertain to electronic records motivated us to
reassess development and documentation procedures for data acquisition to satisfy
compliance requirements used in GLP laboratory studies. Accomplishing the objective
of developing a compliant customized data acquisition system would allow us the
Submitted to: Biomed. Instr. & Tech.
18
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
unique opportunity to provide comprehensive support to research investigators seeking
FDA approval for non-clinical laboratory safety studies.
Investigators at the Jewish Hospital Heart and Lung Institute and other academic
institutions include surgeons, physicians, physiologists, scientists, and engineers. They
actively conduct medical research designed to characterize new surgical techniques,
test innovative medical devices, and evaluate pharmacological agents. Investigators
require accurate data collection of physiological measurements of cardiac, systemic,
and pulmonary function for post-processing analyses. Laboratory studies must comply
with GLP standards for FDA submission of study data and approval for clinical trials.
Approximately 25 academic institutions with membership in the international Society of
Quality Assurance (SQA) are reported to be involved in GLP studies (Hancock, 2002)
that could benefit from an integrated GLP compliant data acquisition system. The
FDA’s release of guidelines for electronic data recording in 1997 provides the
opportunity to implement electronic data recording and documentation strategies, with
many attractive features over current techniques that are limited to reporting discrete
data points and archiving written documentation. The primary features of electronic
data recording and documentation are the ability to record and analyze continuous
waveforms, analyze larger data sets in a more efficient manner (less labor intensive),
document and archive data, and improve accuracy of experimental results by
minimizing measurement error. With advances in computer technology and information
management techniques, we affirm that an integrated data acquisition system with
digital data acquisition and analysis software capabilities for GLP testing of medical
Submitted to: Biomed. Instr. & Tech.
19
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
devices will provide investigators with a valuable research capability that will meet FDA
guidelines for electronic data recording.
CONCLUSION
The design of integrated DAQ systems and the step-by-step process for developing the
support documentation and quality assurance program required to meet FDA guidelines
for GLP compliance is presented in this paper. Advances in computer technology
combined with the FDA’s recently released guidelines for electronic data recording and
record keeping provide the opportunity to improve the efficiency and integrity of digitally
acquired experimental data while reducing resource requirements and expense. Our
group has developed eight integrated DAQ systems for animal, clinical, and engineering
applications that meet GLP guidelines, including two systems used to support the
successful pre-clinical testing of the AbioCor artificial heart. We welcome anyone
interested in developing electronic data collection and storage systems to contact us.
We believe the future for pre-clinical testing of medical devices seeking FDA approval is
with electronic data acquisition and analysis strategies.
Submitted to: Biomed. Instr. & Tech.
20
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
REFERENCES
1. Drew GA and SC Koenig. Biomedical patient monitoring, data acquisition, and
playback with LabVIEW. Chapter 2 (pp 92-98): In LabVIEW for Automotive,
Telecommunications, Semiconductor, Biomedical, and other Applications. Prentice
Hall PTR, Upper Saddle River, NJ, 2000.
2. Drew, G. A. and Koenig, S. C., “Biomedical Patient Monitoring, Data Acquisition, and
Playback with LabVIEW®,” in Virtual Bio-Instrumentation: Biomedical, Clinical, and
Healthcare Applications in LabVIEW®, Olansen, J. B. and Rosow, E., 180-186,
Prentice Hall, 2002.
3. Hancock, S., Meeting the Challenges of Implementing Good Laboratory Practices
Compliance in a University Setting. Qua.l Assur. J. 6, 15-21, 2002.
4. Koenig SC, CA Reister, J Schaub, RD Swope, DL Ewert, and JW Fanton.
Evaluation of transit-time and electromagnetic flow measurements in a chronicallyinstrumented non-human primate model. J. Invest. Surg. 9(6):455-461, 1996.
5. Maloy L and RM Gardner. Monitoring systemic arterial blood pressure: Strip chart
recording versus digital display, Heart Lung 15, 627-635 (1986).
6. Reister C, SC Koenig, J Schaub, DL Ewert, RD Swope, and JW Fanton. Evaluation
of dual-tip pressure catheters during chronic 21-day implantation in goats. Med.
Eng. & Phys., 20:410-417, 1998.
7. Wilkison, DM, KC Preuss, and DC Warltier. A microcomputer-based package for
determination of regional and global cardiac function and coronary Hemodynamics,
J. Pharmacol. Methods 12, 59-67 (1984).
Submitted to: Biomed. Instr. & Tech.
21
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
LIST OF FIGURES
Figure 1.
Photographs of an integrated data acquisition (DAQ) system illustrating
location of (a) signal conditioning components, (b) electronic cabling
layout and power conditioning, and (c) communication ports.
Figure 2.
Computer aided design of an integrated DAQ system.
Figure 3.
Photographs of three multi-functional integrated DAQ systems for (a)
clinical intraoperative measurements and monitoring, (b) engineering
development, and (c) universal data recording applications.
Figure 4.
Schematic for one channel of the signal conditioning and distribution
(SCD) unit. Each channel has five stages: (1) differential mode selection,
(2) positive/negative offset and zero network, (3) antialiasing filter, (4)
invert signal selection, and (5) buffering.
Figure 5.
Illustration of front panel display for signal conditioning and distribution
(SCD) unit containing 16 columns and 3 rows of ON/OFF switches and
LEDs.
Figure 6.
Measurement and Automation (mode 1) to map physical (i.e., volts) to
physiological units (i.e., mHg), and assign analog-to-digital (A/D) device
and input mode (i.e., non-reference single-ended).
Submitted to: Biomed. Instr. & Tech.
22
Koenig, et al. ‘Data Acquisition System’
Figure 7.
April 17, 2003
Program Profile (mode 2) for documenting experimental data, assigning
waveforms, configuring graphical display, and enabling data collection
options.
Figure 8.
Medical monitoring, waveform display, and data recording graphical user
interface (mode 3).
Figure 9.
Application of integrated DAQ system used in support of an animal study
to evaluate an implanted cardiovascular device.
Submitted to: Biomed. Instr. & Tech.
23
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Figure 1.
(a)
Submitted to: Biomed. Instr. & Tech.
(b)
(c)
24
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Figure 2.
Submitted to: Biomed. Instr. & Tech.
25
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Figure 3.
(a)
Submitted to: Biomed. Instr. & Tech.
(b)
26
(c)
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Figure 4.
Submitted to: Biomed. Instr. & Tech.
27
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Figure 5.
Submitted to: Biomed. Instr. & Tech.
28
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Figure 6.
Submitted to: Biomed. Instr. & Tech.
29
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Figure 7.
Submitted to: Biomed. Instr. & Tech.
30
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Figure 8.
Overlay Select
Subject
Name/Number
Next Data Set
Sequence
Number
Channel Name
Mean Value
Exit Graphs and
Return to Program
Menu
Freeze Select
nd
Continuous
Data Save
Data Set Notes
and Comments
Mean Pressure
and Flow
Display Select
Epoch Data Set Save
Submitted to: Biomed. Instr. & Tech.
When selected, 2 subject control
31
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
Figure 9.
Submitted to: Biomed. Instr. & Tech.
32
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
APPENDIX
SOP F11 REV 01 COPY 01
WRITER: S. C. KOENIG
ELECTRICAL
CALIBRATION OF
THE ECTRON MODEL
428 AMPLIFIER
EFFECTIVE DATE: ________
PREVIOUS UPDATE: 06/10/99
1.0
PURPOSE
The purpose of this Standard Operating Procedure (SOP) is to describe the procedure for
calibrating an Ectron model 428 amplifier.
2.0
SCOPE
This SOP pertains to electronics personnel who are authorized to operate an Ectron model
428 amplifier. The Ectron model 428 amplifier is a precision, chopper-stabilized dc
amplifier with a selectable-voltage excitation power supply. The Ectron model 428
amplifier is primarily used for amplification and signal conditioning of pressure transducers
(i.e. Millar micromanometer catheters) for measuring cardiac and circulatory pressures. It
will be calibrated in-house annually.
3.0
DEFINITIONS
3.1
4.0
None
REQUIRED MATERIALS
4.1
Components
4.1.1 Ectron Model 428 Amplifier and Accessories
4.2
References
4.2.1 Certificate of Calibration for Ectron model 428 amplifier
4.2.2 User’s Manual for Ectron model 428 amplifier
4.2.3 SOP F22, Electronic Calibration
4.2.4 SOP F23, Electronic Equipment Maintenance
4.3
Attachments
4.3.1 Not Applicable
Submitted to: Biomed. Instr. & Tech.
33
Koenig, et al. ‘Data Acquisition System’
5.0
April 17, 2003
SOP METHODS
5.1
Inspection
5.1.1 CAUTION: The Ectron model 428 amplifier should only be used in
non-human applications, unless proper electrical isolation
techniques have been established. Use of the Ectron model 428
amplifier in human applications could result in serious injury or
death to the patient and/or operator as a result of microshock
and/or macroshock.
5.1.2 The Ectron model 428 amplifier should be checked for proper
calibration annually (refer to SOP F22, Electronic Calibration).
5.1.3 Prior to use of the Ectron model 428 amplifier, the last recorded
calibration date (labeled on control unit) should be verified.
5.1.4 Prior to use of the Ectron model 428 amplifier, the system (control
unit and accessories) should be inspected for contaminants (i.e.,
dirt, blood, etc.) and any visual damage (i.e., broken wires, broken
display, etc.).
5.2
Cleaning
5.2.1 CAUTION: Do not use chemicals containing benzine, benzene,
toluene, xylene, acetone, or similar solvents.
5.2.2 CAUTION: Do not use abrasive cleaners on any portion of the
Ectron model 428 amplifier.
5.2.3 If the Ectron model 428 amplifier requires cleaning, use a soft cloth
dampened in a solution of mild detergent and water. Do not spray
cleaner directly on the instrument, since it may leak into the cabinet
and cause damage.
5.3
Maintenance
5.3.1 When the Ectron model 428 amplifier is not in use, all components
(control unit and accessories) should be cleaned and properly
stored.
5.3.2 If any physical damage to the Ectron model 428 amplifier (control
unit and/or accessories) is observed, it should be sent to the
engineering section along with written documentation (refer to SOP
F23, Electronic Equipment Maintenance).
Submitted to: Biomed. Instr. & Tech.
34
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
5.3.3 Certified engineering staff should refer to Chapter 5, pages 5-1 to
5-12 of the Ectron model 428 amplifier User’s Manual for identifying
and completing appropriate maintenance procedures (also see the
SOP F23, Electronic Equipment Maintenance).
5.3.4 Repairs called in to the Manufacturer require a RMA number before
shipping equipment out for servicing.
5.4
Testing
5.4.1 A series of performance tests may be applied to the Ectron model
428 amplifier, as part of maintenance procedures or if engineering
staff suspect a device is out of calibration, by referring to Chapter 5,
pages 5-1 to 5-12 of the Ectron model 428 amplifier User’s Manual.
5.4.1.1 If performance tests show device failure, the device may be
sent to Ectron, Corp. for servicing.
Contact:
5.5
Ectron, Corp.
8159 Engineer Road
San Diego, CA 92111-1980
Ph: (800)-732-8159
Fax: (619)-278-0372
e-mail: sales@ectron.com
Web: http://www.ectron.com
Calibration
5.5.1 If an Ectron model 428 amplifier is out of calibration, it should be
sent to the engineering section (refer to SOP F22, Electronic
Calibration).
5.5.2 Certified engineering staff should refer to Chapter 5, pages 5-1 to
5-12 of the Ectron model 428 amplifier User’s Manual for identifying
and completing appropriate calibration procedures (also refer to
SOP F22, Electronic Calibration).
5.6
Certification
5.6.1 Verification of electrical calibration of Ectron model 428 amplifier
should be performed by a certified technician or engineer who has
been properly trained on these procedures as defined by SOP F22,
Electronic Calibration.
5.6.2 Ectron model 428 amplifier calibration certification should be
maintained on file as defined by SOP F22, Electronic Calibration.
Submitted to: Biomed. Instr. & Tech.
35
Koenig, et al. ‘Data Acquisition System’
April 17, 2003
5.6.3 In-house documentation of requests to Ectron for electrical
maintenance and/or calibration of an Ectron model 428 amplifier
should be maintained on file as defined by SOP F22, Electronic
Calibration and/or the SOP F23, Electronic Equipment
Maintenance.
6.0
7.0
PERSONNEL RESPONSIBLE FOR ASSURING COMPLIANCE
6.1
Documentation Coordinator- The Documentation Coordinator is
responsible for maintaining all GLP-related documentation, files, and
archives.
6.2
Engineering Support Staff- Engineering support staff are responsible for
adhering to all guidelines as specified in the SOPs pertaining to
Engineering.
6.3
Institute Director- The Institute Director is responsible for ensuring overall
compliance with GLPs and SOPs.
6.4
QAU Manager- The QAU Manager is responsible for periodically
monitoring all procedures, and reporting all findings.
6.5
Study Director- The Study Director is responsible for ensuring that all
personnel under his/her supervision are properly trained, are familiar with,
and follow all applicable SOPs pertaining to his/her study.
CONTINGENCIES
7.1
8.0
When personnel find circumstances that do not permit compliance with this
SOP, they shall immediately consult their supervisor who determines what,
if any, follow-up is needed.
APPROVAL
_______________________________________________________________________
Documentation Coordinator
Date
______________________________________________________________________
Institute Director
Date
QAU Manager
Date
______________________________________________________________________
Writer
Date
Submitted to: Biomed. Instr. & Tech.
36