In order to meet the acute need for data related to COVID-19, the CRDW team has constructed three data marts (de-identified, limited, and identified) to provide the most commonly requested data elements for this patient population.
With nightly refreshes from Epic’s Clarity Data Warehouse, the CRDW team is able to identify patients tested for COVID19 by filtering on the LABMCOVID19 lab component procedure code. Distinction between “COVID Positive” and “COVID Negative” patients is based on the reported lab results from this procedure.
The initial instance of the COVID-19 data mart includes de-identified structured data on patient demographics, encounters, diagnoses, labs, medications, flow sheets, and procedures. Additional data will be added based on resource availability and urgency.
Images from the Human Imaging Research Office are now available in the limited data mart. Chest radiographs are currently available, and chest CT scans will be added soon. Images are in the DICOM format (.dcm file extension). Access to these images requires IRB approval for use of the limited data mart; studies which already have IRB approval will automatically be added to the image data mart. Please remember to add real dates and employee data to your IRB submission to access the limited data mart. For more information, see our Medical Imaging Tip Sheet.
Both the de-identified and limited data marts are now available for use. The identified data mart will be made available on a per protocol basis.
For further details:
The purpose of SEE Cohorts is to provide a secure web-based tool for the initial exploration of de-identified data. It allows researchers to search available data, build a cohort of patients, and view actual de-identified data within the interface.
Important notes about using SEE Cohorts for COVID-19 research:
- All dates within SEE Cohorts have been offset on a per-patient basis. For example, all dates for patient A are offset by a random number (i.e., 65 days), while all dates for patient B are offset by a different random number (i.e., 300 days). Therefore, date comparisons between patients within a cohort are not possible.
- SEE Cohorts does NOT allow researchers to export available data for analysis. Rather, SEE Cohorts provides researchers with the ability to view de-identified data for the purposes of initial discovery. Export and delivery of data sets remain under the control of an official data request process.
- When defining a cohort, users may filter only on recorded numerical values. For example, one can search on patients within hemoglobin >= 12, but one cannot typically search for patients where a recorded test equals ‘positive’ because the recorded test values are non-numeric. A special exception has been made for the results of COVID19 testing. Specifically, users may define patients who tested positive by selecting “COVID19-Positive” under “Labs.” Similarly, users may define patients who tested negative by selecting “COVID19-Negative” under “Labs.”
- The data in SEE Cohorts is refreshed weekly.
We want to encourage and support your use of REDCap for COVID-19 related tracking and monitoring. We have a host of preparedness and tracking templates available to groups in the process of planning COVID-19 monitoring activities.
When creating a new COVID-19 related project, please indicate your project’s purpose as either Research, Quality Improvement, or Operational Support. Additionally, projects will not be moved to Production Status until all required information below is received:
COVID-19 Research Projects: All COVID-19 research projects in REDCap will require IRB approval or a non-human subjects designation. Please provide your IRB number and PI name. If you do not have an IRB#, please send a message to COVID19-IRB@bsd.uchicago.edu to begin your IRB submission process. Additionally, if you are seeking to use Electronic Informed Consent (e-Consent), it is strongly advised to speak with them BEFORE creating your REDCap project.
COVID-19 QI Projects: All COVID-19 QI projects in REDCap will require QI Determination. Please provide your QI Determination number. If you do not have a QI Determination number, please proceed to https://hdsi.uchicago.edu/qi-determination/ and follow the instructions for obtaining QI Determination.
Information on Use of Electronic Informed Consent: Use of Electronic signatures and Electronic Consent has been proposed as one solution to the challenges researchers have encountered during the COVID 19 epidemic to potentially reduce exposure risk to patients, subject and staff. Electronic Informed Consent (eIC) should follow the same regulatory guidelines and provide the same information as paper ICF. It is not meant to replace the informed consent process between research participant and study staff. Please go to the REDCap login page for more information on how to get started with eConsent. Please contact the CRI REDCap Administrator Julissa Acevedo for further assistance.
For emergency REDCap help related to your current COVID-19 projects, please call 773-702-4785. Otherwise, monitoring for regular help request tickets is ongoing daily. Zoom video help sessions are also available. Please contact the REDCap Administrator to request help for all your REDCap COVID-19 related needs.
The CRI’s Associate Director of Clinical Research Informatics, Julie Johnson, PhD, MPH, RN, is available for office hours consultations via Zoom. Julie’s office hours:
*Please note that there will not be office hours on Monday, July 13.
The clinical data science (CDS) arm of the CRI offers data science services for supporting research regarding COVID-19. Researchers will work with informaticians, statisticians, and computer scientists to analyze and visualize their clinical datasets. The CDS group offers the following services:
- Data normalization: Rather than acquiring several linked tables that remain to be explored, BSD researchers will be given normalized, “tidy” datasets that are ready for analysis for their individual research projects.
- Statistical data analysis for patient outcomes research: The CDS team offers expertise in applying a wide range of data analysis, visualization, and statistical techniques for COVID-19 related data.
- Derivation and validation of COVID-19 related clinical prediction models: With experience in longitudinal data analysis techniques, machine learning, and integrating structured and unstructured clinical data, the CDS team offers capabilities for building and validating models aimed at predicting outcomes in COVID-19 patients.
The CRI Scientific Computing team will provide priority access for COVID-19 related computing jobs on the CRI HPC cluster Gardner. Please email email@example.com to request priority access.
Please email firstname.lastname@example.org to set up an appointment for consultation.
The CRI Scientific Computing team will be providing increased office hours (using Zoom) for anyone that needs help to get started on the CRI HPC cluster Gardner. Email email@example.com to set up an appointment.
The CRI Scientific Computing team will also provide consultation on utilizing resources from The COVID-19 High Performance Computing Consortium (https://covid19-hpc-consortium.org/). These donated resources can offer GPU capabilities at a much higher capacity at what the CRI offers at the moment. The consultation may include help with selecting best suitable resources, moving data, building the appropriate software, etc.