|
|---|
| Variable | Comment |
| Code list should be 'Y',null. | |
| This flag is subject and parameter level. Shall we consider adding another baseline flag for lab toxicity grade, say ABLTOXFL? This is because multiple labs can be used for the same lab toxicity summary. For example, CK, CKISO, CKMB, MKMBCK all are used for increased CPK summary, and baseline records can be available for each of the lab parameter. While for the lab toxicity summary, the baseline flag will be set to one of them (the one with highest toxicity grade). | |
| ANRIND BNRIND | Abnormal' should also be added in 'OCE/OOD Additional Information' column to match it with CT. |
| Can "abnormal" be removed from the codelist since "low" and "high" are already included? "Abnormal" isn't mentioned in the additional information column either. | |
| APERIOD | APERIOD is required per OOD. Could sponsor include APHASE instead of APERIOD in case aphase concept is used in analysis instead of aperiod?. |
| For single period study(without cross over etc.), whether it is required variable? Consider adding as conditional as it creates overhead for single period studies. | |
| Suggest changing OCE/OOD Core=Cond and adding 'Required if there are multiple treatment periods' (similar to ADAE.PERIOD). | |
| ASEQ PARCAT1 ANL0FL CRIT MCRIT | These are the commonly used variables considered important for analyses. |
| ATOXGR ATOXGRN | This variable has been dropped in latest OOD standards, but was there before and I thought came from ADaM IG so would be interesting to get some input on how to derive it. How should this variable be used to distinguish low vs high grades? e.g., `Should low show as negative -4, -3, -2, -1 and high as 1, 2, 3, 4? Or should we have a description variable to go with this variable to state which direction the grade applies to? Also confirm it will be the maximum of ATOXGRL and ATOXGRH for each record. |
| ATOXGR- BTOXGR- | In CTCAE v5.0, for laboratory tests ALP, ALT, AST, BILI, GGT and Fibrinogen, only post baseline grade derivation is defined based on a Normal or Abnormal value at Baseline (see screenshot to the right). Do you have any recommendations for how to assign grading for baseline values in the ATOXGR-/BTOXGR- variables? For example, can 0 be assigned for Normal baseline, and 1 for Abnormal baseline (in the OOD spec the codelist is 0-4, and null). |
| ATOXGRL ATOXGRH ATOXGRLN AGOXGRHN | As these variables are not specified in ADaM IG, should be indicated as 'FDA'. Same comment applicable for baseline grade variables. |
| These are ADaM 1.2, not ADaM 1.1. What does FDA expect from sponsor who are using ADaM 1.1, and have a different approach for handling bi-directional grading, e.g. ADLBTOX tier 2 dataset. Should those sponsors still add the ADaM 1.2 variables?. | |
| Suggest adding to "additional information" or in general ADLB details that both the HIGH and LOW versions are expected to be populated for lab tests that are analysed bi-directionally. For example, a LOW grade 2 for a bi-directional test would have ATOXGRL/BTOXGRL="Grade 2" and ATOXGRH/BTOXGRH="Grade 0". | |
| To assign grades for parameters with bi-directional criteria, our current approach is to create different parameter codes, eg, low/high hemoglobin are "HGBL" and "HGBH" with the lab data provided on both records. Would this be an acceptable approach as well?. | |
| We received an request from the FDA regarding the sign of the grading and directionality "Include laboratory toxicity grading that indicates the directionality of the abnormality at baseline and post-treatment, for example for potassium, “-3” for grade 3 hypokalemia, “3” or “+3” for grade 3 hyperkalemia." Is this also expected and to be expanded to the baseline grades? | |
| ATOXGRL ATOXGRH BTOXGRL BTOXGRH | Shall code list be 'Grade 0' etc.?. |
| AVALU | We regularly include both SI and Conventional unit observations in our analysis datasets (as well as linear transformations where necessary/appropriate) all with different values of PARAM/PARAMCD. Does this contradict the last bullet of the introduction that only one standard unit should be used?. |
| What is the rationale behind this variable and making it a required when units are captured in PARAM and LBSTRESU is present?. | |
| What if the unit in analysis is not the same as the unit in SDTM? E.g. in our company, LB always contains SI units. If needed for analysis, the conventional units are added in SUPPLB. In ADLB, the unit in PARAM can be different from LBSTRESU (but will be equal to SUPPLB.CONVU). Is that acceptable?. | |
| ADLB contains SI unit measurements. BMS creates dataset ADZL to contain US conventional units. What's the FDA expectation on unit of Lab data? | |
| AVISIT | Text in 'OCE/OOD Additional Information' column does not seem to be relevant for AVISIT. Please consider removing it, if so. |
| BASE | Does the Agency have any guidance regarding how to determine the baseline value in particular when time isn't collected? For example, we typically use the lab value closest to the start of study drug. We assign the lab value on Day 1 as pre-dose and use that as the baseline value. Does that align with the Agency's thinking? Secondly, does the Agency have a preferred way to assign baseline when multiple lab tests are collected on Day 1? For example, using the average value or taking the worst case. |
| BNRIND | As this variable is specified in ADaM IG, should be indicated as 'ADaM' instead of FDA. |
| DTHFL | DTHFL is a SDTM variable in DM, which includes deaths after data cutoff date. Should ADSL.ADTHFL be here to replace DTHFL?. |
| DTYPE | DTYPE variable is missing. This variable is needed as as lab undergo many row imputation. |
| EVLLBFL | We would suggest renaming this EVLLPFL to align with the ADaM IG as from our understanding this is a parameter-level population flag. |
| Do we need to follow the same definition of EVLLBFL? BMS doesn't compare whether LB test is within TEAE defined period. | |
Worsening from baseline may be analysed for different treatment emergent periods (e.g within 30 days of last dose, within 100 days of last dose) for the same study. In this case we would include two treatment emergent flags in ADLB (as per IG conventions). Consider defining EVLLBFL to include all on-study and baseline records for the subject where baseline and at least one on-study value is available for the parameter. Then analysis on worsening can be performed for different treatment emergent periods using EVLLBFL and corresponding flags ONTRTFL/ONTRT2FL. | |
| Should ADLB also include a flag to define the treatment emergent period?. | |
| Codelist values should be Y & Null; Is this variable an equivalent of ANL01FL? What is the purpose of this flag. | |
| General | Would there be interest in adding worst toxicity flag(multiple for bi-directional tests) recommendations to the ADLB specifications?. |
| We have ANL01FL, ANL02FL flags to identify the post-treatment records with highest ATOXGRL/ATOXGRH, respectively, that will be used for the analysis. | |
| Including a sample dataset in the final standards would be helpful to see how these expectations would be implemented, in particular, for the algorithm outlined in the 4th bullet. | |
| About the algorithm outlined in the 4th bullet and sentence 'Patients who move from a baseline low grade to a high grade, or from a baseline high grade to a low grade', does it mean we should display all post baseline grades from the baseline grades (not only the grades that are worsening from the baseline grade)?. | |
| Suggest adding the Agency's expectations regarding CTCAE version to the additional information. Meredith mentioned some of these during our initial walkthrough. | |
| LBSPEC LBMETHOD LBFAST | Consider adding these variable as conditional as lab grading may be dependent on them (e.g glucose fasting vs non-fasting). |
| While this is a minimal set of variables we feel it would be helpful to include these additional variables from SDTM as part of the minimal set as these make up the remaining unique identifiers for a particular lab test (along with LBTESTCD/LBTEST). | |
| LBSTRESC | Values are not controlled by codelist (LBSTRESC). |
| ONTRTFL | Consider adding this variable as required. |
| SHIFTy | Shift variables are needed for analyses to capture the shift from BNRIND to ANRIND and for bidirectional lab tox grading, add 2 SHIFTy variables for the shift b/w btoxgrl and atoxgrl; BTOXGRH and ATOXGRH?. |
| Should a change or shift variable for grades be added? An example (FDA request): Provide a shift column that describes the change from baseline grade to post treatment grade. For example, for change from grade 1 hyperkalemia to grade 1 hypokalemia post-treatment, the shift column for potassium would specify “+1 to -1”, indicating treatment-emergent hypokalemia. | |
| TRT01A TR01SDT TR02EDT TRTxxA TRxxSDT TRxxEDT | Instead consider adding TRTP, APERSDT, APEREDT to identify records for analysis for multiple periods. |