Abstract Body

Many patients with tuberculosis (TB) disease are not diagnosed or started on TB treatment. Reliance on TB symptom screening misses patients with TB, especially HIV-infected individuals. We assessed if targeted universal testing for TB (TUTT) in adults at extreme risk of TB would increase the total number of patients diagnosed with TB in primary care clinics.

We conducted a cluster randomised trial, allocating 62 large clinics in three provinces in South Africa to either: augmentation of standard of care (SOC), symptom-based TB testing with the TUTT intervention, or to SOC. In TUTT clinics, we targeted high risk adult (?18 years) clinic attendees, irrespective of the presence of TB symptoms. High risk individuals were: HIV-infected; close contacts of someone with TB in the past year; or prior TB in the past 2 years. Fieldworkers in TUTT clinics collected a sputum sample, which was processed and split for Xpert Ultra and mycobacterial culture. Outcome was the total number of TB patients diagnosed per clinic per month in each arm, measured using counts of laboratory diagnoses of TB at all clinics in the year prior, and during the intervention. We compared numbers of TB cases per month during the intervention between study arms, and also taking into account secular trends.

From March 2019 to March 2020, we consented and sputum-tested 30,500 adults in TUTT intervention clinics. Most (71%) were HIV-infected. Overall 8% of all sputum tests were positive for M. tuberculosis on ?1 assay. There was marked variation in effect of the intervention at individual clinics and by province but cluster- and province-adjusted comparison between TUTT and SOC clinics, restricted to the intervention period, showed a nonsignificant increase of 14% additional patients with TB diagnosed in TUTT clinics per month (95%CI: -6%; +38%). However, difference-in-differences analyses showed TB diagnoses per clinic per month in SoC clinics declined by 8% in the intervention period compared to the year prior, whereas TUTT clinics diagnosed 17% (95%CI: 14%; 19%) more TB patients relative to SoC.

Implementing targeted universal testing in high risk groups increased the number of TB patients by 17% compared to symptom-based TB testing, and strategies such as this may assist in eliminating TB.