Data collection

Demographic and clinical data were collected from the TB laboratory registers and TB treatment registers at all three facilities and entered into two separate databases using EpiData v 3.1 (EpiData Association, Odense, Denmark). Entries in the laboratory register marked as ‘New’ and ‘Diagnostic’ were considered to be ‘presumptive TB cases’. All data were double-entered and validated. The travel time to the BMUs was estimated by a transport official and defined as the estimated time taken for a round trip from the ward to the BMU. Each address was assigned to one of three categories of travel time to BMU: <3 hours, ⩾3 hours and unknown.

We used the standard definitions of treatment outcomes of cured, completed treatment, LTFU, died, treatment failure and not evaluated as defined in the PNG TB guidelines.7 Unfavourable outcomes included LTFU, treatment failure, not evaluated and died. Favourable outcomes included cured and treatment completed. Paediatric cases were defined as any patients aged ⩽14 years.