HIV care · Lancet Reg Health Eur · 2024
The summary shrinks uncontrolled HIV’s death risk and inflates TB’s by quoting the wrong table lines
Death rate as a multiple of the reference group’s; bars show likely ranges
The paper says
The paper’s summary says that among 20,364 people with HIV, uncontrolled HIV (few immune cells, virus not kept down) came with 4.2 times the death rate of well-controlled HIV, and tuberculosis (TB) before or at joining the study 3.33 times that of no TB.
Its own data show
Its table gives 7.61 times for uncontrolled HIV (true figure likely between 6.68 and 8.68) and 1.45 (likely 1.26 to 1.67) for TB before or at joining. The summary’s numbers, ranges included, belong to two other lines, as its results section confirms: people missing an immune-cell or virus test, and TB during the study.
Each table line compares one group’s death rate with a reference group: well-controlled HIV for the HIV lines, no TB for the TB lines. The summary, the part most readers use, puts the uncontrolled-HIV figure almost half too low and the TB figure more than twice too high. Both still go with more deaths; only the sizes are wrong.