| Variable | Comment |
| ADTHFL | If DTHFL is a CRF collected variable, can the variable be added to the ADSL, or is there a need to create ADTHFL?. |
| What is the difference with SDTM.DTHFL? Is this variable required, or would it only be needed when there is a difference between SDTM.DTFL and ADSL.ADTHFL?. | |
| We sometimes collect death information even after the subject is no longer participating in the study (e.g. site enters death from public records on CRF if the site knows about it). Would we assign ADTHFL=N in this case because the death is after the "end of study"? Would we assign DTHA30FL=Y in this case because the death is after 30 days after last treatment?. | |
| AGE | ADaMIG has AAGE as Conditional variable in case the age for analysis is derived differently than SDTM AGE. Consider using AAGE in this spec?. |
| AGEGRy | Where index variables have been specified (e.g. AGEGRy) we would suggest the agency specifies the index (e.g. AGEGR1) for variables where the contents are specified to ensure cross sponsor consistency. |
| The Additional Information column states certain groupings should be presented (e.g., "Age <65 and>=65 should be presented". Is this grouping required for studies with populations such as paediatrics that will not have any patients >=65? Could this be clarified in the Additional Information column or perhaps the general scenario when study populations do not include all groupings can be covered in the guidance on how sponsors should populate these type variables?. | |
| DCTFL DCTADY | Is it possible to clarify the definition? Not clear. |
| Minor typo -Use proper case in label (upcase 'd' in 'day' and 'discontinuation'). | |
| Our understanding is that the date of treatment discontinuation (DCTDT) is the date a subject was last administered study medication and the date that a subject discontinued a study period/epoch in ADaM is the date the decision to discontinue was taken. Is our understanding correct?. | |
| The best we can collect is the investigator's decision day for discontinuation. Agree, we do need to clarify the definition and how to collect them. | |
| Would it be more useful to have an end of treatment date/day (EOTDT/EOTDY), instead of just the date/day of trt discontinuation (DCTDT/DCTADY)?. | |
| DCTDT DCTFL DCTADY DCTREAS DCTREASP | How should these be named if multiple treatments are given within a treatment arm and may have different DC dates? E.g., DCT01FL, DCT02FL, etc…?. |
| DCTFL | ADAMIG has EOTSTT (End of Treatment Status) with controlled terminology of COMPLETED, DISCONTINUED, ONGOING. Would this variable be more useful instead of DCTFL (which only flags discontinuations)?. |
| DCSREAS | Can we have one numeric variable like DCSREASN paired with this variable?. |
| DCTREAS | Where can we find this information? (See OCE/OOD additional information). |
| For the variable DCTREAS that the Agency uses the existing CDISC codelist for Completion/Reason for Non-Completion (C66727) as opposed to just “Adverse Event” and “Other” as this would be consistent with treatment period/study discontinuation data. | |
| Additional info references DCTAESP, but DCTAESP is not listed as an ADSL variable in the table. | |
| DCUTDT | Confirm this should match cut-off date (DCUTDTC) entered into TS dataset in SDTM. |
| DTH30FL DTHA30FL DTHB30FL | We sometimes collect death information even after the subject is no longer participating in the study (e.g. site enters death from public records on CRF if the site knows about it). Would we assign ADTHFL=N in this case because the death is after the "end of study"? Would we assign DTHA30FL=Y in this case because the death is after 30 days after last treatment?. |
| Is it required to use 30 days as the window for these or is this just showing naming convention so if study defined window to 28 days as death window, would we use DTH28 for example? So more like DTHXXFL, DTHAYYFL, DTHBZZFL where XX, YY, and ZZ are study dependent?. | |
| Similar comment as for TRTSDT / TRTEDT / TRTEDY. For combination regimens, would this be based on any component of the regimen (e.g., first/last treatment of any component)?. | |
| Per ADaM IG, corresponding values or codelist should have either 'Y' or 'N' and not null. | |
| Rather than these 3 variables, instead can we have 1 categorical variable. | |
| In the rare case that there are incomplete death dates, is an imputation expected? If yes, would a naming of ADTHxxx more appropriate? This approach is also in sync with ADTHFL. Values should be for subject-level flags "Y", "N". | |
| General comment: Can we put the original variable used for the derivation to facilitate the comprehension of the variable and guarantee no misunderstanding? ie, TRTEDY + 30 days. | |
| DTHCAUS | Variable label should "Primary Cause of Death" instead of "Cause of Death". |
| When working across different TA the proposed codelist for DTHCAUS is not universally applicable. We would propose populating DTHCAUS with the cause of death (e.g. fatal AE coded value) and creating a new variable to categorise these for oncology studies (e.g. DTHCGRy). | |
| DTHCAUSP | Variable label should "Primary Cause of Death" instead of "Cause of Death". |
| DTHCAUSS | Variable label should "Primary Cause of Death" instead of "Cause of Death". |
| Would this come from adjudication?. | |
| DTHDT | Suggest adding DTHDTF as a COND variable, especially with the presence of DTH30TFL, DTHA30FL, and DTHB30FL variables being REQ. |
| Would this include all deaths in the data or do we apply cut-off dates if applicable? Also would the death date from public records be included in this variable even if the subject was no longer participating in the study?. | |
| ECOGBL | FDA has indicated the requirement to submit baseline data as part of ADSL (e.g. baseline ECOG) that potentially leads to circular dependency challenges for Sponsors. Would the Agency consider accepting submissions under the RTOR pilot where baseline data is included in a separate dataset using the ADaM Basic Data Structure (BDS) as opposed to ADSL?. |
| ENRLFL | Can we have as Perm variable - Enrolled Flag, if some patients might be enrolled to a study but not treated due to some reasons. |
| EOSDT EOSTT | Can we add these trial experience variables as well to capture 'End of Study Date' and 'End of Study Status'?. |
| ETHNICN | Minor typo - change AdaM to ADaM under "Source" column. |
| Is the Agency expecting the values to be presented in a particular order?. | |
| COUNTRYN | Is the Agency expecting the values to be presented in a particular order?. |
| Do we really need this numeric variable? If so, do we have the standard to assign the numeric value of COUNTRY. | |
| LSTALVDT | Based on each domain? Any date? |
| Is it possible to clarify? What about the Source variable: RFPENDTC (Date time of participation)?. | |
Is the expectation that this variable will be populated for all subjects, as per ADAMIG, "If this variable is included in ADSL, the best practice is to populate it for everyone."?. | |
| LSTCTDT | LSTCTDT variable must be added to ADSL dataset. |
| General | Minor typo - change STDM to SDTM under "Source" column. |
| Can we add intrinsic and extrinsic factors? Disease-related and drug-related biomarkers? Other subgroups?. | |
| NCTXSDT | Suggest adding the study day variable (Study Day of New Anti-Cancer Therapy: NCTXSDY) for consistency with other variable with study date/day pairs. |
| Would this follow the Sponsor’s definition of anti-cancer therapy per protocol or statistical analysis plan and simply use this naming convention?. | |
| PHASE | Phase, as cond variable? ex- escalation or expansion. |
| RACEGRy RACEGRyN | RACEGRy and RACERyN have been included as permissible variables. In what scenario does the Agency expect these variables to be submitted? Is this only required when RACE in SDTM extends the proscribed list of values?. |
| The Additional Information column states certain groupings should be presented. Could this be clarified in the Additional Information column or perhaps the general scenario when study populations do not include all groupings can be covered in the guidance on how sponsors should populate these type variables?. | |
| RACEN | If RACEGRy and RACERyN are not submitted would RACEN be required?. |
| RANDDT | RANDDT variable must be added to ADSL dataset. |
| REGIONy | Would the Agency consider listing countries considered to be in North America and Western Europe for the REGIONy and REGIONyN variables to ensure cross sponsor consistency. |
| Is the Agency expecting the values to be presented in a particular order?. | |
| SAFFL | We use the definition of the safety population as: patient who received at least one drug (whatever study drug) (we don't use the notion of treated pop). |
| For combination regimens, would this flag be based on receiving any component of the regimen or just the investigational drug of interest? Suggest clarifying this. Should separate flags be provided for each component of a regimen?. | |
| In Codelist, change "Y, Null" to "Y, N". Per ADaMIG, population flags are Y,N and null values are not allowed (ADAMIG section 3.1.4). | |
| SITEID | SITEID required variable in IG. |
| TRT01A TRT01A | There is a row for TRTxxP and TRTxxA. Do we still need an additional entry for TRT01A/TRT01P. |
| TRTSDT TRTEDT TRTEDY | Need to clarify for situations where the investigational drug and other combo drugs are administered on different dates. Do we take the earliest first date regardless of the drug? Same question for the other drug start dates by periods. |
| Need to clarify for situations where it takes longer than 1 day for administering one or more drugs. Do we take the date when we start the administration of the last dose or the date when the administration is complete. | |
| For combination regimens, would this be based on any component of the regimen (e.g., first/last exposure to any component) or just the investigational drug of interest? Suggest clarifying this. | |
| For subjects that are ongoing in treatment, is the expectation that TRTEDT/TRTEDY will be left null or populated with the cut-off date?. | |
| TRTFL | Safety Population defined in protocol as 'patient who received at least one drug' but 'Treated' population not defined. |
| What is the Purpose of this flag?. | |
| Can the FDA clarify if the TRTFL (Treated Population Flag) variable should be based on the planned or actual treatment regimen?. | |
| Perhaps a better label for this would be "Combination Treated Population Flag" and would default to SAFFL for mono-therapy?. | |
| In Codelist, change "Y, Null" to "Y, N". Per ADaMIG, population flags are Y,N and null values are not allowed (ADAMIG section 3.1.4). Other option is to make this just a "Flag" instead of a population flag. | |
| Is this required when there is a combination regimen (i.e. should the OCE/OOD Core column be Cond instead of Perm)?. | |
| WEIGHTBL | Can we have as Perm variable?. |