TB cases (1974-2018) and controls (1998-2018) analytical dataset (Karonga)

The interview data for TB cases from the 3 eras are appended together and merged to the 'tbcpu' file. This latter file is the results of an algorithm that goes through all the TB data and creates TB episodes and assigns them to a category of how likely the episode was actually TB (see documentation on tbcpu algorithm). Only episodes on the tbcpu file are kept so other questionnaire data are dropped. Questionnaire data on controls from the 3 eras are appended to the cases dataset. Existing variables are modified and new ones created using the data from the questionnaires and data on lab results, contacts, GPS coordinates and follow-up and outcomes which are merged in. In the variable labels (S) indicates that the variable is unchanged from the original source and (D) indicates that it is a 'derived' variable, often using more than one source variable. For both types information on where the data come from and how they are processed is found in the 'Recoding and Derivation' section