Objective and scope ------------------- The overall objectives of this project are to improve the specification and, to the extent possible, the standardization of endpoint and estimand definitions across protocols, analysis plans and reports from late phase pulmonary TB therapeutic trials. To this end, we are using the estimand framework, as outlined by the ICH E9(R1) addendum, since it contains recommendations that are binding for trials conducted under ICH GCP. A TB trial may require the development of more than one estimand, given differing objectives of regulators, sponsors and physicians and their patients, with all five components in alignment with the particular needs of each stakeholder. This proposal is intended for use for all late-phase clinical trials of new and repurposed regimens for the treatment of active pulmonary TB in which participants are followed up after the end of treatment for long-term outcomes; this includes both phase III and IIC. It is not expected that every sponsor, trialist, or investigator designs a trial in a wholly standardized way, nor is it realistic to think that all “consumers” of the trial results have the same value judgement in what they consider important. Estimands must necessarily be constructed depending on, and specific to, the objective of each trial. However, because the estimand framework does not limit a trial to one estimand, several estimands addressing the specific needs of different consumers are entirely appropriate and will be presented. The suggested strategies for dealing with intercurrent events will be tailored to the estimand being defined. The overarching goal of this proposal is to provide both flexibility (in the definition of multiple estimands) and consistency (in the way that outcomes and ICEs are specified and dealt with) that will simplify the comparison of data across trials and repair the disconnect between clinical trials and practice guideline development. Inasmuch as specification and standardization are separate objectives, the proposal first describse definitions of intercurrent events and we put forward some proposals for estimand specification, limiting our recommendations for standardization until the final section. Ongoing analyses ------------------------ Reanalyses of previous trials and a simulation study are ongoing to explore the effect of estimand proposal(s) on likely TB trial outcomes. This estimand proposal will be updated to include results from these anlayses (and relevant references to peer-reviewed publications) and the estimands themselves may also be updated accordingly. Request for comments and suggestions ------------------------------------