Interorganizational SEND - Scenarios White Paper
This page holds the evolving content for the Interorganizational SEND group's white paper on data scenarios.
Strategy for Paper
To describe the "end product":
- Scenarios description
- Insights gained from modeling use case scenarios.
The focus of the paper should be the results, i.e., what people can get out of it. So less description on the process of selection and some of the chronological events that led us to the results so much as the takeaways.
(What do we mean by interorganizational)
Scenarios for generating SEND datasets for nonclinical safety studies with components contributed by separate organizations
The FDA is expected to issue draft guidance sometime in 2013 (check date before publishing) addressing the submission of standardized nonclinical data in the SEND (Standard for the Exchange of Nonclinical Data) format for non-human studies conducted as part of pharmaceutical drug development. The representation of nonclinical study data in SEND is described in the CDISC Standard for Exchange of Nonclinical Data Implementation Guide (SENDIGv3.) [link?]. In addition to SEND conformance, the FDA may include additional expectations in its Study Data Specification guide for the submission of nonclinical study data.
The upcoming regulatory e-data submission needs are coinciding with a time period where there is increased preclinical outsourcing to Contract Research Organizations (CROs). Additionally, outsourced preclinical studies are conducted in myriad ways, and more than one CRO may be involved with conducting parts of a study. Compliance with the upcoming FDA mandates will therefore require new workflows, data flows and data stewardship processes to be established at both sponsor companies and their CROs. However, the process by which SEND datasets are generated and managed to assure data integrity and audit traceability is not prescribed by FDA nor indeed should it be prescribed.
The purpose of this paper is to address SEND implementation scenarios for preclinical studies that are conducted in an inter-organizational setting involving more than one organization. The example scenarios presented include those where one organization is responsible for the complete SEND dataset, and others where more than one organization is responsible for contributing components to the ultimate end product – a complete SEND dataset per study (should an example or two be included here to clarify what we mean?). This paper is the outcome of collaboration among sponsors, CROs and vendors participating in the Interorganizational SEND Project, as part of the PhUSE workgroup (refine title and link to wiki).
Challenges and Needs
(Outline the scope of challenges and needs, as basis of driving scenarios – set the stage and why these were developed and tested.)
One of the first tasks for the group was to develop a framework for describing, classifying and prioritizing scenarios for aggregating data across multiple sources. The following attributes were identified as key criteria to differentiate scenarios that were considered:
- Study Director's Organization (CRO, Sponsor)
- Organizations involved in collecting the data (Sponsor, CRO, Subcontractor)
- SEND Capability Levels of the Organizations involved (Sponsor, CRO, Subcontractor)
- Controls for Edits to the Data by Respective Organizations
- Interim data transfers versus only a final data transfer (is this relevant"?)
To develop an understanding of practical workflows, we felt it was necessary to select a few example scenarios. The following three were selected as representative use cases ranging from very simple scenarios to somewhat more complicated ones.
- CRO study with SEND dataset fully assembled by the CRO
- CRO study with SEND dataset fully assembled by Sponsor
- A study with the Sponsor performing toxicokinetic concentration and parameter analysis, with the rest of the SEND dataset assembled by a CRO
Based on the collective experience of the group, the selected scenarios were determined to be representative of likely real-world situations. Each scenario is self-contained and includes a description accompanied by a flow diagram summarizing the likely sequence of activities including study start-up, data flows within an organization and between organizations, reviews, and decision gates. At each of these gates, options are indicated for both success and failure conditions. The final step for each of the scenarios, of course, is the submission of datasets to the FDA. Since this step occurs across all of the scenarios, it is called out as a common regulatory workflow described in depth within the first scenario below.
It is important to point out that the inter-organizational workflows described in the three scenarios below are only exemplars of how specific activities might occur in practice, and not intended to prescriptive. It will be up to each sponsor and CRO to define their specific use cases, and consider other controls and steps that they may wish to implement.
CRO Study with SEND Dataset Fully Assembled by the CRO
This scenario represents the situation where a sponsor outsources the entire study to a single CRO which is responsible for executing the study and transmitting a complete SEND dataset to the sponsor. This scenario assumes that the dataset will be submitted once towards the end of the study, or at periodic intervals based on milestones negotiated between the sponsor and the CRO. At the time of completing the draft study report, the CRO would package the SEND dataset including a define file, protocol data and collected data normalized and harmonized to the requirements for SEND. The dataset would be reviewed by the sponsor; discrepancies, if any, would need to be addressed by the CRO at which point a complete, signed SEND dataset would be transmitted to the sponsor.
The completed dataset would then enter the sponsor company's regulatory submission workflow which we have assumed would be common to all of the scenarios as summarized in the following diagram ("How are we dealing with call-outs to the figures?"). In practice, completed SEND datasets that form part of an eventual IND, NDA and/or BLA package will flow into the sponsor company's regulatory submission workflows, specifically as part of the electronic Common Technical Document (eCTD). In other cases, additional datasets in support of an already submitted regulatory package may be sent directly to the FDA outside of the eCTD process. Sponsor submissions are received by FDA's electronic gateway, at which point the SEND datasets are identified and passed through a SEND Validator to check for conformance with the SEND data model and any additional requirements specified by the FDA in its Study Data Specifications Guide ("e.g., <provide link to an example one here from the CDER website>").
Three outcomes can result from the SEND Validation check: (a) the full dataset passes validation successfully, (b) parts of the dataset pass validation, and (c) the entire dataset fails validation. The latter will presumably result in a notification sent by the FDA to the sponsor detailing reasons for failure. In all of the three scenarios contemplated in this white paper, this will require the sponsor to directly address the validation issues reported by the Agency (i.e., for Scenario 2 where it is responsible for dataset assembly) or transmit the details to the CRO which in turn will be required to address the issues (i.e., Scenario 1).
CRO study with SEND dataset fully assembled by Sponsor
To be completed!
- Basis for selecting these scenarios
- Industry drivers – amt of outsourcing, GLP…. These are common scenarios
- criteria for selecting scenarios
Among our first challenges were to decide on how SEND data should be handled. Our thought processes are already plublished and will be maintained on http://www.phusewiki.org/wiki/index.php?title=SEND_Implementation_Wiki_-_FAQ_-_Handling_of_SEND_in_Study_Documentation (depending on where this is published, we will need to adjust this link). Before developing the flow charts for these scenarios we agreed on the following:
- We recommend SEND be listed in the Protocol.
- SEND datasets should be given similar QA and QC treatment as would the individual listings.
- The Study Director signature on a study report indicates accountability for the validity of the content of the datasets.
- If the decision to create/modify SEND datasets is made after report finalization, a memo to file may be prepared for inclusion with the study data documenting the SEND datasets prepared and a description of the review status (QA, QC, etc.).
- Description – concrete examples, flow chart
- Description – concrete examples, flow chart
- Description – concrete examples, flow chart
Some of these may need to be introduced in the scenario descriptions
- Challenges getting data combining data from disparate source systems
- non-major collection systems may not capture exclusion reasons
- Key insights from early discussions from our FAQ contributions
Impact on Drug Development
Some ideas to consider for this section
- Organizational Readiness to prepare and review SEND datasets
- Use of SEND for interrum project review
- Data connectivity with Clinical Study
- Enpoint predictivity development
(Influential points for future directions of SEND and Implementations)
Last revision by Snath, 2013-02-22