<Compound Name>Legacy Data Conversion Plan

<Compound Name>Legacy Data Conversion Plan

Legacy Data Conversion Plan

<Sponsor Name>

Compound Name

Version 2014-06-30xx-xx

Legacy Data Conversion Plan

Contents

1.1Purpose

1.2Acronyms

1.Introduction

1.1Purpose

This document provides context for the conversion of raw data to standardized data for <submission, pre-NDA>. In addition, this document summarizes the conversion results along with any issues encountered and the resolution of the issues. Any outstanding issues will also be summarized.

1.2Acronyms

2.Legacy Data Summary

Identify studies included and phase[M1]

Assumptions on the data – such as, could only convert some of the data

Was data converted from actual datasets or did it have to be entered manually from the CRF into the data entry system? Did a vendor manually enter the data?

Any translations from non-English to English? (Protocol, CSR, data)

Type of validation run on the legacy data, particularly on data that had to be entered manually (such as double data entry, controls, etc.). Examples: (Mario to provide text).

Were all SDTM domains created or a subset such as safety or efficacy? Was all data converted[M2]?

Phase I sometimes only requires that a subset of the data is submitted. Was this done? For example, for the following studies only the following domains were created: AE, EX, etc.

Point to where more information can be found (such as SDSP, SDRG[M3])

3.Converted Data Summary

Include text on the process for how the data was converted, including version of SDTM.

Data converted that could affect CSR, such as race.

Could include diagram as well as text.

Mention if upcode data, such as MedDRA version or WhoDRUG version plus other dictionaries if applicable. Refer to SDRG for more information.

Dictionaries used, such as converting test codes to text, unit codes to text, etc.

Mention if saved raw data in supplemental domains.

How was the SDTM data validated aside from Open CDISC? Were reports run to ensure numbers match the CSR from the analysis datasets that were created by the sponsor? Reviewers spend a lot of time matching number of subjects and data related to the subject.

Was raw data submitted or bridging document created? Depends on what is agreed to in the Type C Briefing Document and what the reviewer wants to see. Bring it up as soon as possible.

Link to FDA guidance on submitting datasets:

Is ADaM created from the SDTM or were these ‘dead end’ conversions (meaning, nothing else is done with the data at the study level)? Was analysis rederived based on the SDTM? There is a FDA slidedeck about what not to do on the PhUSE Wiki. May have to submit legacy data.

Was controlled terminology followed from legacy to SDTM? Was it updated to a particular version of controlled terminology?

What type of data standards were used in addition to controlled terminology? For example, if there are 1’s and 0’s used for Yes and No. Body systems, too. Didn’t match current SOC but had to rename in SDTM.

4.Conversion Results

If common theme, provide it here. For example, if certain trial domains were not created or if SDTM domains could not be fully populated. Or if certain SDTM required variables, such as EPOCH, could not be populated. Refer to SDRG for more details.

Table with old variable name (from legacy data), SDTM name, comments (include data collected but not converted[JAL4]). Sounds like a bridging document from raw data to SDTM. If don’t send raw data, is this meaningful[M5]? Raw data should be submitted.

If not creating study level ADaM from SDTM, be prepared to submit 2 tables: 1 from legacy data to SDTM, and 1 from legacy data to analysis datasets.

Refer to SDRG for open CDISC issue summary.

Look at .2 of Mario’s example.

Issues Encountered and Resolved[M6]

Outstanding Issues

PDF created 2014-06-30Page 1 of 7

<Compound Name>Study Data Reviewer’s Guide

Appendix I: Conversion Matrix

Appendix II: Issues Details

(Text here)

PDF created 2014-06-30Page 1 of 7

[M1]Claudia: additionally study title and number of subjects should be added for each study

[M2]Claudia: could this be a tabular overview, which SDTM domain is used in which trial

[M3]Claudia: to avoid duplicate/redundant information it needs to be stated clearly that LDCP and SDRG are complementary

[JAL4]Ask Jingyee for more information.

[M5]Claudia: does the FDA wants to have one specification document how legacy data was converted/mapped to SDTM in each trial of this compound?

[M6]Claudia: please be aware that the issues for each single study are documented in the SDRG