Efficacy of non-pyrethroid indoor residual spraying or intensive behaviour change communication in combination with long-lasting insecticidal nets for malaria control in west Africa: a pragmatic, cluster-randomised, controlled trial.

Background: Since 2015, progress in the control of malaria has stalled owing to multiple factors, including the probable reduced efficacy of long-lasting insecticidal nets (LLINs) caused by insecticide resistance and plateauing LLIN use rates. This study aimed to assess the additional effect of non-pyrethroid indoor residual spraying (IRS) and intensive behaviour change communication (BCC) when combined with LLINs on malaria in rural west Africa.

Methods: This pragmatic, parallel-group, cluster-randomised, controlled trial took place in community settings in Burkina Faso and Côte d'Ivoire. Villages (clusters) with an average population of 300 inhabitants, a minimum distance between villages of 2 km, and year-round accessibility by a four-wheel drive vehicle were eligible for inclusion. A simple randomisation approach with a computer-generated random number sequence was used to assign 39 villages with a total population of 10 750 inhabitants to three groups: LLIN alone (16 villages; the control group), LLIN plus IRS (11 villages), and LLIN plus BCC (12 villages). Field teams that collected epidemiological data and laboratory staff were masked to intervention allocation; it was not possible to mask participants, field teams that implemented the interventions, or study investigators to intervention allocation. The IRS intervention consisted of pirimiphos-methyl treatment of the dwellings and the BCC intervention focused on promoting LLIN use, environmental sanitation, and early health-seeking behaviour through home visits, interpersonal talks, and group talks. The primary outcome was malaria incidence rate in the whole population as measured by passive case detection at health centres. Data were collected for 10 months before (the pre-intervention period) and 10 months after (the post-intervention period) randomisation. All outcome data were analysed by intention to treat and using constrained baseline analyses. The trial is registered with ClinicalTrials.gov, NCT03074435, and is completed.

Findings: Between November, 2016, and August, 2018, 215 000 theoretical person-months of follow-up resulted in 3612 malaria cases recorded by passive case detection during both pre-intervention and post-intervention periods. During the post-intervention period, passive case detection showed a 23% reduction in malaria incidence rate (rate ratio 0·77 [95% CI 0·64-0·93], p=0·0073) in the LLIN plus IRS group and a 22% reduction (0·78 [0·63-0·96], p=0·020) in the LLIN plus BCC group compared with the LLIN-alone (control) group. No IRS-related adverse effects were recorded.

Interpretation: Results show that the addition of non-pyrethroid IRS or intensive BCC to LLINs can effectively reduce the number of malaria cases. The effect of IRS observed in our trial was intermediate compared with that reported in previous trials conducted in Africa. Notably, this study provides the first trial-based evidence supporting the effectiveness of an intensive BCC intervention, which is a promising result but requires confirmation through additional studies.

Funding: Initiative 5%, Expertise France.

Translation: For the French translation of the abstract see Supplementary Materials section.