An economic evaluation of the LINKEDin study: An intervention to reduce initial loss to follow-up among tuberculosis patients in South Africa.

Tuberculosis (TB) patients who are diagnosed but not registered and initiated on TB treatment are categorised as initial loss to follow-up (ILTFU). ILTFU is a key driver of morbidity and mortality associated with TB and is a contributing factor to high TB transmission rates. LINKEDin was a quasi-experimental study which evaluated two specific interventions for reducing ILTFU in three high-TB burden provinces in South Africa, conducted from October 2018 to December 2020. As part of LINKEDin, we undertook a micro-costing from the healthcare provider perspective using an activity-based costing approach. Cost estimates included the cost of the operation of an integrated provincial health data centre in the Western Cape, apportioned to the TB activities it supported in the province. Cost estimates were linked to intervention outcomes to understand the incremental cost of the intervention per additional patient linked to care compared to rates of ILTFU in the absence of the interventions. Sensitivity analyses were conducted to account for uncertainty in the intervention outcomes, and for periods where the implementation of the intervention was interrupted due to COVID-19 related disruptions. Costing data were collected between August 2020 and March 2021. The total cost of implementing the LINKEDin intervention in the WC and KZN was $7 534.42 per month. The cost of implementing LINKEDin in the Western Cape accounted for 56% the total cost of the intervention - 8% from the operations of the PHDC and 48% from the cost of running the intervention - while only 44% of the total cost was accounted for by the intervention run in KwaZulu-Natal. The primary cost driver of the interventions were staff salaries, with the cost of data extraction and in-hospital activities low relative to primary healthcare (PHC)-based follow-up activities. In terms of cost effectiveness, the LINKEDin interventions in KZN was cost $377.28 per additional person linked to care, and $243.62 in the WC, per additional person linked to care. In the Western Cape, systematically tracking persons with TB using an automated system proved highly cost efficient compared to the more labour intense approach adopted in KwaZulu-Natal. Optimising the curation and management of data and increasing the effectiveness of tracing systems and processes can result in cost-savings.