Discovery Data Coordinating Center

The Discovery-DCC is directed by Douglas Landsittel, PhD, Professor of Biomedical Informatics, Biostatistics, and Clinical and Translational Science at the University of Pittsburgh. Dr. Landsittel has over 20 years of experience in collaborative, educational, and methods research. He has published over 130 peer-reviewed manuscripts and has served as a principal investigator (PI) on a National Institute of Diabetes and Digestive and Kidney Diseases-funded data coordinating center, an Agency for Healthcare Research and Quality-funded R25 grant on methods in patient-centered outcomes research, and a Patient-Centered Outcomes Research Institute-funded contract on causal inference methods. Dr. Landsittel has been appointed to both National Institutes of Health and Centers for Disease Control and Prevention study sections. He has served as chair on multiple reviews and is a member of several expert panels.

Program Manager: SCCM has a full-time program manager on staff. The program manager’s primary responsibility is to oversee research projects from conception through final reporting, which may include overseeing individual research projects from pre-award through post-award, tracking project timelines and budget compliance, maintaining records and reports on project status, and working as a liaison between the PI, other investigators, and project leads.

Database Administrator: SCCM has a full-time research database administrator who is responsible for the performance, integrity, and security of SCCM’s research-related database. Research database administrator consultation services are offered in the following areas:

• Developing conceptual design for planned research database projects, in conjunction with PIs and research staff
• Developing designs that facilitate back-end data organization and user-facing accessibility and usability
• Developing project-specific procedures to implement and ensure the collection of accurate and usable data
• Assisting PIs and institutions with setup, testing, and implementation of data collection
• Preparing and distributing data reports to PIs

The Discovery-DCC can contribute to the success of grant proposals and the conduct of studies through a number of key functions:

• Translating clinical and scientific questions into appropriate study designs and analysis strategies for funding proposals 
• Calculating sample size or power estimates that are consistent with the associated analysis strategies and specific to the research phase and design type
• Conducting standard and novel statistical analyses and writing methods and results for manuscripts, reports, or presentations Developing data collection forms and web-based data management using Research Electronic Data Capture (REDCap) Cloud
• Maintaining effective data management and producing ongoing recruitment and safety reports for data safety monitoring boards (DSMBs) and other committees
• Providing other logistical support as needed for specific RFAs

The Discovery-DCC shall be responsible for the communication, coordination, storage, maintenance, and integrity of data collected from all sites involved in a study. The Discovery-DCC shall also have primary responsibility for statistical analysis to support the study. The study PI(s) shall work closely with Discovery-DCC staff to provide general direction, oversight of adverse event (AE) reporting, and data presentation, and may carry out additional aspects of administrative, clinical, and technical expertise and leadership in the design and coordination of the study, in collaboration with the data coordinating center and PI(s).

SCCM will maintain a data coordinating center that meets the administrative, physical, and technical safeguards required by the U.S. Department of Health and Human Services’ Health Insurance Portability and Accountability Act (HIPAA) regulations to support programs.

All sites participating in a multisite research program must be HIPAA-compliant and must demonstrate that they have policies on the use of, and access to, any data collected and stored at the site. This includes transferring, removing, disposing of, and reusing electronic media and electronic protected health information (ePHI).

Technical safeguards require access control to allow only authorized people to access ePHI. Access control includes using unique user IDs, an emergency access procedure, automatic logoff, and encryption and decryption.

Audit reports or tracking logs must be implemented to keep records of activity on hardware and software. This is especially useful for pinpointing the source or cause of any security violations.

Technical policies should also cover integrity controls or measures put in place to confirm that ePHI has not been altered or destroyed. Information technology disaster recovery and offsite backup are key to ensuring that any electronic media errors or failures can be quickly remedied and ePHI recovered accurately and intact. Network, or transmission, security is the last technical safeguard required of HIPAA-compliant hosts to protect against unauthorized public access of ePHI. This concerns all methods of transmitting data, whether it be email, Internet, or even private network, such as a private cloud.

The Discovery-DCC can serve in multisite studies as the data coordinating center, although this is not required if the researchers prefer to use a different data coordinating center that meets the criteria of Discovery, as noted below in Data Policies and Health Insurance Portability and Accountability Act Compliance.

The Discovery-DCC uses Research Electronic Data Capture (REDCap) Cloud for research that requires data collection. REDCap is a secure, web-based application designed to support data capture for research studies. REDCap Cloud is compliant with FDA CFR Part 11, HIPAA, and EU Annex 11 regulations.

REDCap Cloud includes the following security features:

• Secure data entry and storage in the cloud accessed through single sign-on
• Protection for data in flight and at rest:
  • Using https encrypts data between the data center and the client browser
  • Checking the PHI (personal health information) box when creating a data entry field ensures that the data are auto-encrypted in the database using a 256-bit RSA key
• Secure transfer of large data files
• Robust audit trail that shows all changes to all records within the system, including a complete audit log of data, time, and user access
• Control of access to REDCap studies through assignment of users, user roles, and permissions

REDCap Cloud has several features that can be incorporated into a case report form (CRF) that ensures data accuracy and quality, as well as tools for medical coding, data monitoring, and documenting CRF completion.

• Detailed response sets for each question, reducing the need for open-ended responses
• Medical coding that allows for the use of established medical terminology
• Data edit checks (minimum and maximum values, integers, whole numbers)
• Ability to query responses to each variable, which can be set up to generate automatic queries to data entry personnel
• Source data validation
• Documenting completed CRFs
• Double data entry at the site
• Rules to limit missing values, field validation errors, outliers, and invalid date entries

Medical coding: Medical dictionaries can also be used to create CRF items on medications prescribed to study subjects. Study users can be assigned the role of medical coder and given permissions to view and modify data entered on medications prescribed.

Data monitoring: source data verification, medical review, and data review: With source data verification, data entered into a CRF has to be separately verified by a someone other than the user who entered the data. This person is typically a site monitor who is given access to the specific data entries that require monitoring. The site monitor confirms that the monitoring duties have been carried out.

Documenting completed CRFs: Data entry users check a box at the end of each CRF to indicate that it has been completed. Data entry users can also be required to provide an e-signature for CRFs completed for each subject.

Role of the REDCap Cloud Systems Administrators
The systems administrators are the gatekeepers to the REDCap Cloud application.

• They provide access to REDCap Cloud by assigning a generic user role to a user. Once in the library, a user can be assigned to a specific study with a defined role, (eg, data entry). Proper roles and permissions for the PIs and other users are assigned to ensure that access is granted appropriately.
• PIs and other study participants designated by the PIs will have access to CRFs and all data entered for the study. However, most users will only be able to see data entered to their own site.
• Users added to the REDCap Cloud library automatically receive login information, online training, and documentation that includes definitions of terms, data queries, data validation, and the data codebook.

Rules for Data Storage and Transfer
All data collected will be maintained and stored in REDCap Cloud. Datasets in CSV format (unless another format is requested) will be transferred using REDCap Cloud’s secure Send-It feature. The Send-It module can be used to send large (up to 500 MB) and/or sensitive files to one or more recipients in a HIPAA-compliant manner. A secure, password-protected link will be emailed to the recipient. The link typically expires within five days. In addition to the CSV data file, a link to a PDF file codebook will also be provided. Before receiving these links, the recipient must sign a data use agreement. Once received, data files must be stored at the recipient’s site in compliance with HIPAA.

REDCap Cloud CRFs will be used to screen, enroll, and randomize subjects. The data coordinating center will track enrollment and other results for transparent reporting, including the number randomized to each study arm and reasons for exclusion or loss to follow-up. The number enrolled by each site and the number randomized into each study arm will be provided in data reports to the study PI. The reports will include:

• CONSORT diagram with actual versus expected enrollment figures
• Data summary tables 
• Data tables that show missing visits and missing CRFs
• Listings of adverse events (AEs), serious adverse events (SAEs), deaths, unanticipated problems (UPs), and protocol deviations and violations

The statistician will perform interim analysis of the study according to an agreed-upon schedule. This analysis may examine whether the study could be considered for early termination and/or other analyses requested by the study PIs.

Study PIs will provide definitions of AEs, SAEs, and UPs and guidance for when they should be reported by participating sites. Site PIs are responsible for reporting AEs, SAEs, and Us that occur at their site. AEs and SAEs will be reported in REDCap Cloud.

A CRF that can be used multiple times (as needed) will be available for reporting AEs and SAEs along with the other CRFs used to collect study data. The data coordinating center will monitor AE and SAE submissions and provide detailed reports to the study PIs. The data coordinating center at the direction of the study PIs, will also provide reports to the central institutional review board (IRB) (if in place), the DSMB, and any other required regulating bodies. Study CRFs may include alerts to participating sites indicating that an ASE or SASE may be required.

The data coordinating center database administrator will monitor AE and ASE form submissions to REDCap Cloud and report all submissions to the study PI within 24 hours. The administrator will also prepare weekly reports on all AE and SAE submissions and will also monitor queries sent to study PIs advising them to submit AE and SAE forms. The study PI is responsible for reporting or directing that reports be made to IRBs, National Institutes of Health (NIH), and U.S. Food and Drug Administration. An email alert system that provides a more immediate means for sites to report SAEs and UPs will be put in place so that PIs are made aware of the most serious issues as they occur.

The Discovery Steering Committee is responsible for developing and implementing procedures to ensure integrity in the following areas:

• Clinical informatics: Oversight and mechanics of collecting, managing, reporting, securing, and protecting Discovery research data and ensuring research data integrity.
• Data and safety monitoring: Ensuring monitoring for each Discovery research activity, if appropriate based on the type of research being conducted.
• IRB, including a central IRB: The purpose of an IRB is to ensure that appropriate steps are taken to protect the rights and welfare of human participants in a research study. A central IRB reviews research studies for multiple sites. A central IRB is a single IRB of record for a given protocol. The Discovery Steering Committee will ensure that an independent central IRB is selected for each project unless an individual IRB approach is taken by the research work group. Before a clinical trial begins, the protocol (study plan) must be reviewed by a group of experts not connected with the research to ensure that it is ethical, provides potential benefits, and does not cause unnecessary harm to participants.
• Data policies and intellectual property: Any and all results and inventions from research and development in connection with Discovery programs will be detailed in a written agreement by all study participants before the project begins, including but not limited to intellectual property rights such as technical information, expertise, patent application rights, patent rights, data ownership, copyright, and trade secrets.
• Creation of public use data: Data from Discovery studies constitute an important scientific resource. To enhance the public health benefit of these studies, public use datasets may be made available to qualified researchers. Discovery public-use datasets will generally be made available after initial study completion, in accordance with Discovery procedures. Discovery datasets will be provided only to those secondary investigators who agree to adhere to the signed research data use agreement (RDUA). Data will also be created for NIH or other repositories as required by specific programs. Execution of an RDUA will require approval by the investigators’ relevant IRB or demonstration of exemption from the need for IRB approval by institutional policy. Dataset creation and distribution will be performed by the Discovery-DCC.

Early collaboration with the Discovery-DCC is critical for optimizing the potential for success, beginning at the earliest phase of study conceptualization.

Depending on the type of study and/or proposal, some of the activities in Table 1 may not be entirely applicable to your study.

Table 1. Outline of phases of collaborative activities for funding submissions

Phase of Activity	Nature of Collaborations	Ideal Timeline
1. Initial concept
1.a. Developing the research question	Discuss the research question, conceptual framework, causal relationships, and the broad implications of the needed design parameters to evaluate that research question.	5-6 months before submission
1.b. Assessing feasibility of the associated design	Consider the feasibility of the associated designs and, if needed, reconsider the question of interest and hold discussions on further iterations of the main concept and potential designs.
1.c. Refining the approach, sample size and power, and drafting a statistical analysis approach	Based on the above discussions, outline the corresponding approach, estimate the sample size and power, and draft a preliminary analysis plan.
2. Refining the approach
2.a. Analysis of preliminary data	If required, provide analysis of preliminary data, which can be time-consuming and can influence the resulting proposal, leading to further iterations of the above steps.	3-4 months before submission
2.b. Continued development of the design and analysis plan	Hold discussions to further develop the approach, including identification of the key variables and temporal relationships between variables, and refinement of optimal statistical methods.	2-3 months before submission
2.c. Finalize the approach and funding proposal	Review and revise approaches as needed. Changes made in later stages of project development often have unintended consequences on the final approach, so this further check is critical.	1 month before submission
3. Study conduct and analysis
3.a. Study initiation	Setup of data collection procedures, including programming in REDCap, outline of statistical programs for subsequent reports, and development of protocols and a manual of operations.	6 months to 1 year
3.b. Data collection	Once data collection is initiated, upkeep and further modification of the REDCap database is often needed as the study proceeds.	Ongoing after first 6 months to 1 year
3.c. Ongoing statistical reports of recruitment and events	Reports are usually needed on an ongoing basis (often monthly) as the study proceeds. Depending on the type of study, reports on safety and adverse events, quality assurance reports, and reports to committees and funding agencies may also be needed throughout.
3.d. Statistical analysis of interim endpoints	Depending on the type of study, formal interim analyses may be needed in addition to the ongoing reports.	Midpoint of the study, or at multiple interim times
4. Final statistical analyses and writing the main paper	Depending on the type of study, the final analysis may follow easily from the previous steps (eg, for an explanatory trial) or may be time-consuming (eg, as for many observational studies).	Near or after completion of the study
5. Revisions to the primary manuscript and follow-up manuscripts	Most studies have follow-up manuscripts on secondary objectives that continue after the end of the formal study period.	After study completion

Having sufficient effort budgeted is critical for several reasons:

1) Proposals that include insufficient effort will likely be penalized in the review process, thus reducing their chances of receiving funding.
2) Insufficient effort, even when funded, will limit flexibility to address changes in the approach, data management, or other key aspects of study operation.
3) Low levels of funding result in personnel time being spread over a large number of projects and limit their ability to think creatively and work efficiently on any one project.

The recommendations in Table 2 represent broad guidance on how to budget for Discovery-DCC personnel.

Table 2. Proposed effort of Discovery-DCC personnel on funding proposals These efforts are approximate.

The final efforts are based on the needs of the individual investigators.

Type of Proposal	Description	Proposed Discovery-DCC Effort
R01 or U-grant large multisite trial	Trial that requires data management and coordination across multiple sites, regular reports, and study monitoring throughout the trial	PhD statistician: 20%-30% MS statistician:50%-100% Program manager:50%-100% Research coordinator: 50% Other effort: XX%
R01 or U-grant large multisite observational study	Observational or quasi-experimental study that requires monitoring and coordination across multiple sites, regular reports, and complex data management	PhD statistician: 15%-25% MS statistician: 25%-50% Program manager:25%-75% Research coordinator: 25% Other effort: XX%
R01 single-site trial	Trial that requires less intensive data management, regular reports, and study monitoring throughout the trial	PhD statistician: 10%-20% MS statistician: 25%-50% Program manager:20%-30% Research coordinator: 50% Other effort: XX%
R01 observational study from a single site or set of related datasets	Observational or quasi-experimental study that requires analysis and merging of multiple datasets, with some reports and potentially complex data management	Ph Dstatistician: 15%-25% MS statistician: 25%-50% Program manager:25%-75% Research coordinator: 25% Other effort: XX%
R21 single-site trial or multisite pilot trial	Trial that requires less intensive data management, regular reports, and study monitoring throughout the trial; outcomes are typically more focused on theprocess or easily collected outcomes	PhD statistician: 5%-10% MS statistician: 20%-30%% Program manager:20%-30% Research coordinator: 20% Other effort: XX%
R21 or R03 observational study	Observationalor quasi-experimental study that requires minimal data management and withone or a small number of easily merged datasets, with most effort focused on a final analysis	PhD statistician: <10% MS statistician: 10%-20% Program manager:<20%

There are several ways to request Discovery-DCC services:

1. Submit a Clinical Investigator Proposal and request specific Discovery resources
2. Submit an SCCM grant proposal (SCCM-Weil Research Grant or Discovery Research Grant) and request data coordinating center services. If a grant is awarded, data coordinating center services can be used to implement the research outlined in the grant proposal.
3. If you already have research funding, you can submit a request to use Discovery-DCC services.
4. If you are submitting an NIH grant for a clinical trials study, Discovery can be subcontracted to serve as the study XXXX

Please contact Discovery for new collaborations with the Discovery-DCC.