Journal of the American Association for Laboratory Animal Science Copyright 2011 by the American Association for Laboratory Animal Science
Vol 50, No 5 September 2011 Pages 660–664
A Computerized Data-Capture System for Animal Biosafety Level 4 Laboratories Dennis A Bente,1, 5,* Jeremy Friesen,1,3 Kyle White,2,4 Jordan Koll,2 and Gary P Kobinger1 The restrictive nature of an Animal Biosafety Level 4 (ABSL4) laboratory complicates even simple clinical evaluation including data capture. Typically, clinical data are recorded on paper during procedures, faxed out of the ABSL4, and subsequently manually entered into a computer. This system has many disadvantages including transcriptional errors. Here, we describe the development of a highly customizable, tablet-PC-based computerized data-capture system, allowing reliable collection of observational and clinical data from experimental animals in a restrictive biocontainment setting. A multidisciplinary team with skills in containment laboratory animal science, database design, and software engineering collaborated on the development of this system. The goals were to design an easy-to-use and flexible user interface on a touch-screen tablet PC with user-supportable processes for recovery, full auditing capabilities, and cost effectiveness. The system simplifies data capture, reduces the necessary time in an ABSL4 environment, offers timely reporting and review of data, facilitates statistical analysis, reduces potential of erroneous data entry, improves quality assurance of animal care, and advances the use and refinement of humane endpoints. Abbreviations: CADT, computerized animal data-tracking system, ABSL4, Animal Biosafety Level 4.
The recent expansion of Biosafety Level (BSL) 3 and BSL4 facilities worldwide has led to an increase in research using highly hazardous infectious agents within biocontainment. Animal models are a vital part of infectious disease research and are used extensively to study the pathogenesis of highly hazardous agents and the usefulness of therapeutic countermeasures and vaccine candidates. In vivo replication of many infectious agents provokes a spectrum of clinical manifestations. Therefore, establishing reliable humane endpoints to protect the animals from unnecessary pain,2 and increasing endpoint precision is vital and can be achieved by standardizing observer scoring through auditing.5 In addition, new agents and animal models are being discovered constantly, and systematic documentation of clinical data is crucial. Nonetheless, the restrictive nature of the high-containment laboratory environment makes handling, clinical evaluation, and data capture from animals at the cage-side challenging.1,7 Traditionally, computerized data capture systems have not been well adopted in the Animal Biosafety Level 4 (ABSL4) setting, especially because of their nonconformance with biosafety requirements and obtuse user interfaces. Biosafety regulations prohibit items leaving the laboratory without receiving appropriate decontamination such as autoclaving, which severely limits bringing paper records out of the ABSL4. Therefore, paper-based records typically are faxed out of the ABSL and then manually transcribed into an outside computer; only at this point do the data captured become available for meaningful review and analysis. Although the process of data capture can be quick, easy to establish, and inexpensive, it has many disadvantages including data loss due to technical issues with
Received: 02 Feb 2011. Revision requested: 28 Feb 2011. Accepted: 05 Apr 2011. 1Special Pathogens Program and 2Business Operations, Public Health Agency of Canada, 3College of Computer Science, University of Manitoba, and 4College of Applied Computer Science, University of Winnipeg, Winnipeg, Manitoba, Canada; 5Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas. *Corresponding author. Email:
[email protected]
fax machines (inside or outside the ABSL4), handwriting recognition problems, time-consuming transfer of data, no audit of captured values, and risk of erroneous data entry. Furthermore, transfer of a complete set of clinical data to the attending veterinarians is demanding. An electronic data-capture system is a computerized system designed for the collection of clinical data in electronic format.8 These systems have found their way into biomedical research and have been deployed especially successfully in clinical trials.6 Electronic data-capture systems allow direct data entry at site, with the benefits of increased accuracy, reduced queries, decreased paper record storage, and timely population of the study database.3,6,8 Previously, attempts have been made to capture animal data digitally at the cage-side by using a personal data assistant;4 however, the system functionality is limited and is not useful in an ABSL4 setting due to limitations in screen size and resolution as well as responsiveness to user interaction and flexibility of function. Here, we describe a new computerized animal data-tracking system (CADT) with a small footprint designed to comply with biocontainment requirements. The system has redundant backup features to prevent data loss and an intuitive interface for simple and rapid data entry. The computerized clinical scoring sheets are highly customizable, allowing accommodation to any type of study design and animal species. The system provides a simple interface by which data capture can be done easily and intuitively; it also stores all data captured safely by providing many backup features and ensures data integrity by auditing all data passing through the system. As a result, CADT guarantees refined endpoint scoring and reliable clinical data capture and considerably reduces time spent on data entry.
Materials and Methods
The objective of the project was to establish a CADT that combines the following characteristics. The CADT must comply with biosafety standards, such as the ability to disinfect all surfaces
660
jaalas11000014.indd 660
9/28/2011 9:50:44 AM
Capture of data in ABSL4 settings
with obligatory ABSL4 disinfectants and the lack of sharp edges or parts that might break and create sharps. The system must have reliable processes and backups to prevent data loss. A small footprint is desirable to assure mobility and the ability to work in diverse ABSL4 settings, where space is often very limited. An intuitive interface is vital for rapid and easy data entry and ensures that the user can focus on animals and procedures. Data capture sheets have to be highly customizable by users to easily accommodate user’s desired experimental designs and animal species requirements. Furthermore, the captured data have to be available to users after exiting the containment laboratory so that data reports and statistical analyses can be generated. Lastly, steps were taken to ensure that other technologies including barcoding, radiofrequency identification tags (RFID), scales, and other telemetry equipment can be easily integrated into the data-capture process or the underlying database of the CADT. After an initial needs assessment of the ABSL4 suite located in the Special Pathogens Program, we decided that tasks such as planning and access to the database for reporting purposes should be provided through a Web-based portal. In addition, the actual recording of observations and clinical data entry within the ABSL4 suite would be performed on a tablet PC to ensure a stream-lined workflow. To comply with Public Health Agency of Canada corporate computing infrastructure requirements, a Web-based technology platform (Windows 2008 Server, Internet Information Server, and SQL Server 2008; Microsoft, Redmond, WA) was selected as the network operating system, associated software, and services to be installed and configured on computers located in the server room. The Microsoft .NET platform was used for both the Web application and tablet interface, allowing both to connect seamlessly with the central database. This platform was chosen because of its proven infrastructure, licensing model, and large standing knowledge base. Microsoft SQL Server was chosen as the database platform because it is robust, inexpensive, and its pervasive use in the business community. Finally, by using the Microsoft .NET 3.5 platform, we were able to connect easily to an existing authentication system (ASPMembership, Microsoft), which allows seamless performance of user management tasks of the CADT system. The CADT system encompasses 3 main components (Figure 1): 1) a Webbased interface for planning and reporting animal data; 2) an SQL server database engine with a central data repository for the system; and 3) a tablet-PC-based capture interface to gather data within the ABSL4. Workflow within CADT was divided into 3 phases. The PLAN phase encompasses score sheet set-up, cage locations, and their identifiers; in addition, users and their passwords are managed. Because this interface had to be as accessible as possible, it was built as a Web application. The CAPTURE phase covers all parts of the workflow in which experimental data are acquired. This interface was chosen to reside on a tablet PC as a client-side application to allow for portability. Due to the nature of the ABSL4 labs and the need for portability of our data-capture device, the application residing on the tablet was set to capture data in a disconnected state. The tablet stores data safely until the information is ready to be sent out of the laboratory after synchronization of all new data with existing data within the CADT repository. The REPORT phase was developed to include tasks such as generating and exporting data reports that describe the data recorded in a usable format (charts, tabular views, and so on). Lastly, to address data security, access to all 3 processes requires the user to login with a unique username
Figure 1. System infrastructure. The CADT has 3 components: (A) a secure, intranet-accessible Web interface is used during the study planning and reporting steps. The experimental design and animal information is uploaded into the database at a workstation computer, which allows web-browser-based access to the database; (B) a SQL server database engine acts as the central data repository for the application. The SQL server is the hub between the database, workstation, and tablet PC; and (C) a touchscreen-enabled application is used during the capture of animal data within the ABSL4. Animal observations are recorded inside the ABSL4 on a tablet PC. Once clinical data are gathered, the tablet PC is returned to its docking station and connects immediately to the server. Data from ongoing and completed studies can then be accessed through the web portal.
and password. In this way, users can be given different access and administrative rights to the system. PLAN process. By using a Web browser (for example, Microsoft Internet Explorer), the CADT PLAN interface can be accessed from any computer connected to the corporate network as well as remotely through a Virtual Private Network connection. The software was designed so that multiple users can log on from the location of their choice (office, laboratories, home). Study design details such as animal information, experiment, and score sheet components are all set up and maintained within the PLAN interface of the CADT system (Figure 2). Users have the ability to create and define the limits of their experiment, including animal types and strains, endpoint definitions, and data-capture frequency. The planning workflow includes the following functions: 1) entering animal information (ID, species or strain, date of birth, cage assignment, and so on); 2) entering 661
jaalas11000014.indd 661
9/28/2011 9:50:45 AM
Vol 50, No 5 Journal of the American Association for Laboratory Animal Science September 2011
Figure 2. Study setup using the PLAN interface. Scoresheets can be (A) created and (B) previewed before the experiment starts.
Figure 3. Screen shots of an endpoint scoring sheet run in CAPTURE application. (A) Overview of clinical scoring sheet of animal 336 with clinical parameters on the left and predefined values in the center of the screen. (B) Weight, temperature, and other data can be entered as dynamic values in a pop-up window.
experimental design and scoring capture frequency; and 3) viewing and customizing score sheets. CAPTURE process. Once an experiment and associated score sheet has been entered in the PLAN application, it is ready to be used in an experiment through the CAPTURE interface (Figure 3), which is designed to run on a tablet PC within the ABSL4. The key consideration when designing the CAPTURE interface arose from environmental restrictions in the ABSL4, such as limitations in field of vision, dexterity, and decreased sensitivity to tactile pressure. The CAPTURE interface’s screen layout was designed to present to the user only the information required at this stage of the clinical observation. A recognizable ‘stoplight’ visual paradigm was implemented across the application to aid in application navigation as well as continuity. An item marked red indicates that the item has been completed (no longer requires additional actions), yellow indicates that an item still requires additional actions, and green indicates that
an item has not yet been examined. An example of how this paradigm is looking at a collection of animals within a cage: any animals not yet examined are indicated by the color green; any animals with some observations are marked yellow; and any animals that have been completely examined and had all data points recorded are marked in red. In addition, information is presented in ways to capitalize on the amount of screen space available to it (for example, text and buttons sized to fit the space provided) as well as to show only the information that the user needs to complete at any step in the process at one time. Furthermore, the application can be used in landscape or portrait mode, thereby providing additional flexibility when capturing data in a wide variety of environments. We modeled the design on existing paper-based score sheets to allow users to work in a familiar paradigm. For the tablet PC, we use a Toughbook CF-H1 mobile clinical assistant from Panasonic (Secaucus, NJ; Figure 4). This tablet PC
662
jaalas11000014.indd 662
9/28/2011 9:50:48 AM
Capture of data in ABSL4 settings
Figure 4. A tablet PC as an input device within the ABSL4. (A) The graphic depicts the tablet PC being tested for factors such as touch sensitivity, control size and placement, and color schemes. (B) A pole cart supports data input without the need for the operator to hold the device. Here the tablet PC and cart are shown while factors like portability and maneuverability are available. We chose to use the Anthro Pole cart (Anthro, Tualatin, OR) because of its customizability. A port replicator (Gamber-Johnson, Stevens Point, WI) was designed to fit this particular model of tablet PC, which was easily mounted to the cart.
(10.4 × 10.6 × 1.3 to 2.3 cm) is lightweight (3.4 lbs) and features a bright 10.4-in. dual-touch display. The larger size of a tablet PC compared with previously used personal digital assistants (for example, Palm Pilot) allows for easier data entry, which is crucial due to the optical aberrations and visual field restrictions produced by the faceplate of a positive-pressure suit (Figure 4 A). The tablet PC allows stylus-entry mode as well as touch-screen entry mode, which is pressure-based and therefore functional even when the user wears multiple layers of gloves. The tablet PC has a sealed case with an International Protection Rating 65-certified design, which accommodates wiping down with alcohol or quaternary detergents. We are currently conducting studies to demonstrate that the tablet PC remains functional after BSL4 decontamination procedures such as formaldehyde fumigation or vaporized hydrogen peroxide treatment. This benefit allows for repair of the tablet PC outside of the ABSL4 and reuse of tablet PCs. In addition to the International Protection Rating 65 design, a shock-mounted hard drive prevents data loss after dropping of the tablet PC from as high as 3 ft. The battery life is approximately 6 h; however, the user has the ability to replace the batteries in the tablet PC while it is being used. Additional features are a radiofrequency ID reader (ISO 15693 compliant), a 2.0 megapixel camera, and 2D barcode reader. The tablet PC in the ABSL4 cannot be connected continuously to the database because the Public Health Agency of Canada considers wireless technology within the ABSL4 to be a data security risk, and a cord connection presents a tripping hazard. Therefore, for capturing data, a synchronization system allows the data to be stored locally on the tablet PC and consequently synchronized with the database upon docking the tablet PC. The tablet PC has an integrated handle and ergonomic rubber hand-strap. To offer a hands-free solution, we mounted a docking station on a pole cart (Figure 4 B) that was modeled after
the ‘crash carts’ often used in emergency rooms. The tablet PC can be charged and communicate with the database while in the docking station. Because troubleshooting and debugging within the BSL4 setting is challenging, the CAPTURE application was designed and tested to be stable and fault tolerant. Furthermore, BSL4 personnel are able to perform application update installs and recover after an application or software failure by using ‘thumb drives.’ Because scoring data are recorded continuously and written to disk throughout the scoring session, the information is unaffected if the application or operating system software should fail. In case of a hardware failure, fully functional and tested standby units are on-hand and can be brought into the BSL4. REPORT process. Data captured within the ABSL4 during or after an animal study can be reviewed online within the REPORT interface of the Web portal. Data are fully searchable by study, experiment, date, user, or animal ID. The REPORT application displays a dashboard of current experiments and a chart or tabular report of score data from current or past experiments (Figure 5). The data can be exported into more comprehensive reporting software (such as Cognos) or statistical analysis software (such as SAS, SPSS, Microsoft Excel).
Results
One of the challenging aspects of the BSL4 environment is that hands-on troubleshooting of software or hardware related problems will not be performed by Information Technology staff when the BSL4 is in use. To ensure that a fully functional and proven system is available in the ABSL4 and that no valuable animal data are lost, the CADT was tested intensively in a mockup ABSL2 setting by using the same workflows as used ABSL4. We demonstrated that the application is stable and reliable. The 663
jaalas11000014.indd 663
9/28/2011 9:50:52 AM
Vol 50, No 5 Journal of the American Association for Laboratory Animal Science September 2011
The CADT will improve scientific productivity by speeding the availability and increasing the accuracy of data for analysis and ensuring data integrity. In future versions of this system, these benefits will be enhanced by integrating the tablet PC’s camera, barcode, and radiofrequency ID reader and connectivity to scales. The full-audit capability of the CADT makes it an interesting technology to support Good Laboratory Practices-like studies. Due to the organizational requirements and environment, data are saved temporarily on a tablet PC’s hard drive and then synchronized with the database. Nevertheless, the CADT can also be used in a wireless or permanently hardwired environment. The flexibility of the CADT’s user interface to accommodate different experiment designs makes it a valuable application for noncontainment or field animal studies. It is also useful for extreme environments such as epidemic investigations in which Information Technology support is limited at best.
Acknowledgments
Figure 5. Data reporting of a completed study in the REPORT application. Data from completed or ongoing studies can be accessed through the Web and displayed (A) graphically or (B) in tabular form.
CADT then was installed in the ABSL4. During the first 2 experiments in the ABSL4, we conducted a side-by-side comparison of the paper-based scoring, and the CADT was used to evaluate the time spent on data entry. Additional tests were performed to assess the simplicity of application updates as well as data recovery in the event of a software failure. The data recovery test was completed with no difficulties, and the data were retrieved and later captured by the application successfully. The update process was also successful. A survey was performed with the ABSL4 users to assess user acceptance of the system. During use of the CAPTURE interface in the ABSL4, users requested new features to simplify the CAPTURE workflow. These requests included visual refinements (such as identifying the animal ID on a cage label, when only one animal is in the cage) to allow smoother navigation through the application. Other than from refinements and advice, user feedback was positive from all of those who participated in the survey.
Discussion
Here, we describe the design and implementation of a CADT for BSL4 laboratories. The system ensures reliable clinical data capture by using redundant backup features that minimize the likelihood of data loss. Computerized animal data capture at the cageside avoids human error during data transfer and provides an auditable log of scoring data. The CADT has been accepted readily by the ABSL4 users, and its intuitive interface considerably reduces time spent on data entry compared with the paper-based system. The data recording interface is highly customizable to different species and study designs. The system helps with user training by providing scoring baselines and greater endpoint precision, thereby promoting humane animal use.
We thank our executive sponsor, Hank Krueger (Director of Business Operations), for his ongoing support and funding the CADTS project. We also thank the BSL4 users in the Special Pathogens Program, especially Mike Gray, for their time spent testing the application and Drs T Ksiazek, C Klages, and A Freiberg (all UTMB) for their comments on the manuscript. The Public Health Agency of Canada has not licensed the software for use at this time; however, inquiries can be directed to Darren Fast (Senior Policy Advisor, Department of Business Outreach and Promotion, Public Health Agency of Canada; darren.fast@ phac-aspc.gc.ca).
References
1. Abraham G, Muschialli J, Middleton D. 2002. Animal experimentation in level 4 facilities, p 343. In: Richmond JY, editor. Anthology of biosafety V: BSL4 laboratories. Mundelein (IL): American Biological Safety Association. 2. Copps J. 2005. Issues related to the use of animals in biocontainment research facilities. ILAR J 46:34–43. 3. Dawson J, Kushniruk A. 2009. Improving the efficiency and accuracy of a tablet PC interface for computerized provider order entry through usability evaluation and provision of data entry strategies. Stud Health Technol Inform 143:447–452. 4. Hampshire V. 2001. Handheld digital equipment for weight composite distress paradigms: new considerations and for rapid documentation and intervention of rodent populations. Contemp Top Lab Anim Sci 40:11–17. 5. Hau J. 1999. Humane endpoints and the importance of training, p 71–74. In: CFM Hendriksen, DB Morton, editors. Humane endpoints in animal experiments for biomedical research— Proceedings of the International Conference, 22-25 November 1998, Zeist, The Netherlands. London (UK): The Royal Society of Medicine Press. 6. Pavlovic I, Kern T, Miklavcic D. 2009. Comparison of paper-based and electronic data collection processes in clinical trials: costs simulation study. Contemp Clin Trials 30:300–316. 7. Powell N. 2006. Meeting the challenges of high-containment animal care, p 187. In: Richmond JY, editor. Anthology of biosafety X: animal biosafety. Mundelein (IL): American Biological Safety Association. 8. Welker JA. 2007. Implementation of electronic data-capture systems: barriers and solutions. Contemp Clin Trials 28:329–336.
664
jaalas11000014.indd 664
9/28/2011 9:50:53 AM
