REFERENCES: 1. Jimenez-Maggiora, G., Thomas, R., Hong, P., & Aisen, P. (2012). ADCS EDC. Alzheimer's & Dementia, 8(4), P584. 2. Thomas, R., Jimenez-Maggiora, G., & Aisen, P. (in preparation). The Alzheimer's Disease Cooperative Study (ADCS) Informatics System: An Open-Source, Web-based, Comprehensive Data Management Environment for Multi-Center Clinical Trials. ADCS Electronic Data Capture (EDC) - Smart Electronic Case Report Form (eCRF) Framework CONCLUSIONS: The ADCS EDC system supports the development of eCRFs that provide a sophisticated set of usability features that facilitate both data collection and analysis activities at every stage of a study’s lifecycle. This approach has been used to design and manage databases for more than 20 large, multi-site AD clinical trials as well as the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and the Dominantly Inherited Alzheimer’s Network (DIAN). RESULTS: Well-designed CRFs improve data quality. Clear instructions and question prompts, layout and formatting (e.g. fonts, colors, etc.) are among the static elements used to create robust data collection instruments. Electronic versions of these instruments, created via the ADCS EDC system, retain many of these elements while allowing study designers to develop interactive eCRFs that react to user inputs and provide instant feedback. The combination of these features yields an intuitive user interface (UI) which is conducive to efficient and accurate data capture. Importantly, we achieve this without significantly increasing the complexity of the eCRF design process by using the study’s data dictionary to automatically configure the vast majority of these features. Gustavo A. Jimenez-Maggiora, Ronald G. Thomas, Jia-Shing So, Stefania Bruschi, Hongmei Qiu, Phuoc Hong, Paul S. Aisen Neurosciences department, University of California at San Diego, La Jolla, CA, USA 2. Advanced Checks If the automatic checks and built-in modules prove insufficient for defining a logic check for a field, the EDC system can be extended to reference routines written in Perl, R, or C. This solution can be employed to model highly-complex logic checks which can be executed from within an eCRF during data capture. Query & Audit Trail Show & Hide Question Logic Basic eCRF Example: a. UI Features a. Date field with “date-strict” tags requires the use of the system’s calendar widget for collecting date data. b. Flag icon notifies users if there are pending queries with a field, and allows users to resolve or issue queries directly from the eCRF view. c. Time icon displays historical transaction data for the field. d. Showing/Hiding appropriate questions for entry guides users through the eCRF when entering data. b. Logic Checks a. By default, all non-text fields are checked for valid entry when required (not hidden). b. Dates fields are automatically checked for date validity. c. Logic modules can be used to augment basic checks (e.g., date_range_compare will verify if an entered date is the in the valid range for the field). METHODS: Data quality in clinical studies begins at the source. Well designed case report forms (CRFs), frequent investigator training and evaluation, and rigorous adherence to sensible and well documented processes are among the many factors that support the implementation of high data quality standards. Faster, higher- quality data flows support real-time tracking of study scientific and operational metrics and facilitate data- driven decision making while simultaneously raising data quality standards and accelerating study event timelines. Data-intensive analytical activities, such as adaptive randomization, interim analyses, and, ultimately, database lock and publication of study results can be realized in a more rapid fashion. Having identified these benefits, the ADCS has used several web-based technologies to improve data quality at the source. A specific example is the ADCS EDC’s smart electronic case report form (eCRF) framework. This framework allows study designers to develop interactive eCRFs by defining a concise data dictionary which is used as source metadata for both data collection and analytical activities. The resulting eCRFs allow investigators and study staff to capture, validate and query study data in real-time over the web while also supporting the export of metadata enriched datasets for analysis and reporting. Though originally designed to replicate paper CRFs, these eCRFs have become increasingly sophisticated in terms of the level of usability and quality enhancing features supported. Features such as real-time range checks, field dependency mapping, skip patterns, dynamic word lists as well as real-time scoring of neuropsychological and functional scales such as the Alzheimer’s Disease Assessment Scale- Cognitive Subscale (ADAS-Cog) and the ADCS-Activities of Daily Living (ADL) have reduced the number of data queries required to attain complete and reliable data. Additional enhancements, such as links to manuals and supporting documentation, also provide valuable information to investigators both collecting and analyzing data. 4. Question Presentation Logic The eCRFs have some UI enhancements to help users capture data efficiently and minimize data entry errors. All eCRFs support show/hide logic for questions so that users can quickly determine whether or not a field requires entry. Also as part of the default setting, a flag has been added to each field so that users can quickly determine if additional follow-up on a field is required. INTRODUCTION: The Alzheimer's Disease Cooperative Study was founded in 1991 in response to an NIA RFA. The primary aim of the ADCS is to ‘advance research in the development of interventions that might be useful for treating, delaying, or preventing AD'. Toward that goal the ADCS has designed, managed, and analyzed over 20 large, multi-site AD clinical trials. The psychometric, clinical, and biological data from these trials have been collected using a locally-developed, web-based electronic data capture (EDC) system (Jimenez- Maggiora, Thomas, Hong, & Aisen, 2012; Thomas, Jimenez-Maggiora, & Aisen, in preparation). This system serves as a central hub used to coordinate the data management, quality assurance, and monitoring activities of thousands of affiliates across several continents. 3. Real-Time Scoring For some of the more advanced instruments, (e.g, the ADCS ADAS-Cog) the entire eCRF form has been customized to provide users with the most straightforward way of capturing the data. The ADAS utilizes multiple wordlists for different versions of the form, and by utilizing dynamic wordlist selections, users are able to choose the wordlist that they are entering data for and have all of the fields text updated to match those on their paper CRF. Currently 8 (4 English, 4 Spanish) wordlists are supported. In addition, calculations of checkbox totals, and score components have been built-in so the form to provide real-time feedback when entering data. Users are able to verify their total counts after each section of the form. 1. Automatic Checks When creating eCRFs, a standard set of logic checks are automatically generated based on the input type of a field. Some examples are: a. Make all non-text fields required b. Enforce numeric ranges if they are defined for the field c. Ensure that dates are equal or prior to the current entry date These checks can be overridden and customized by study designers to ensure that appropriate validation checks are in place for each field. Furthermore, for many commonly requested UI customizations and checks, modules have been developed to allow creators to quickly utilize semi-complex to complex logic in their field definitions. For example, date fields have several different modules that can defined to alter the logic and UI behavior of a given date field on an eCRF. P9 English Spanish Real-Time Scoring Advanced eCRF Example (ADAS-Cog): a. UI Features a. Dynamic wordlist selection updates all field labels for the selected wordlist. Supports both English and Spanish wordlists. b. Total scores and score components are calculated in real-time as users enter data. b. Logic Checks a. Score calculation can be performed by using several different methods: i) by using a widget on the eCRF that communicates with a web service to receive the calculated score; ii) by translating calculations into a custom Javascript file for eCRFs like the ADAS-Cog; iii) server-side after eCRF form submit which can store the calculated scores directly to the database. b. All scores are run through a server-side check to ensure that erroneous data from the browser is not saved directly. *Google and the Google logo are registered trademarks of Google Inc., used with permission. www.iadcs.org Download PDF View publication stats View publication stats