The failure to use isoniazid (INH) preventative therapy (IPT) in persons living with HIV (PLHIV) in Sub-Saharan Africa represents one of the single biggest implementation gaps between evidence and practice in today's response to the HIV epidemic. In PLHIV, TB is a lead cause of death, and IPT reduces TB incidence by 40%. Yet in Africa, less than 2% of eligible individuals receive IPT. Given the existence of both country guidelines recommending IPT, as well as simple clinical algorithms to identify IPT eligible persons, a remaining critical requirement for scale-up is strengthening the link - mediated by middle management in most health systems - between health ministry policy and clinics. In Uganda, District Health Officers (DHOs) serve as key middle managers working at the nexus between policy and implementation. We propose to test a countrywide multi- component “SPIRIT” (Simplified INH Preventive Therapy) intervention targeting DHOs – whom we view as critical dissemination agents. SPIRIT is based on the PRECEDE model of behavioral change that deploys predisposing (teaching collaborative); enabling (INH/B6/septrin single pill combination and SMS from DHO to provider); and reinforcing (reporting collaborative) components. For this resubmission, we provide data to show feasibility of SPIRT through a pilot study of 5 DHOs and their clinics. The DHOs engaged in the mini- collaborative and implemented key components of SPIRIT including bidirectional text messaging to front line providers. The number of HIV+ adults prescribed IPT increased from zero at baseline to 300 at 8 weeks. Aim 1: Determine if the SPIRIT intervention increases IPT initiation. We will form 20 groups of 5 District Health Officers and randomize 10 to the SPIRIT intervention and 10 to control (country standard) in a cluster- randomized trial. The primary outcome is proportion of IPT-eligible adults initiating IPT. For secondary outcomes, we will measure changes in knowledge, attitudes and practices regarding IPT among DHOs and front line health workers to assess mechanisms through which the intervention achieves outcomes. Aim 2: Evaluate the effect of the SPIRIT intervention on IPT completion and TB incidence. Even if the intervention increases IPT use, quantifying actual use of IPT by patients and effects on population health status (e.g. reduction in TB), provides an important impact measure that can enable policy makers to prioritize this intervention more widely. A two-stage survey sampling approach will be used to identify a probability sample of patients eligible for IPT in which to measure adherence through hair levels of INH (direct measure of pill consumption/adherence) and TB incidence (population health measure) Aim 3: Assess the cost and cost-effectiveness of SPIRIT. Using effectiveness measures obtained in Aims 1 and 2, standard time and motion and costing methods, we will estimate the cost and cost-effectiveness of SPIRIT vs standard of care in our sampled population from Aim 2. Outcomes of interest will include program costs per: a) IPT initiation; b) IPT completion; and c) TB case averted.


Funding Source

Project Period