Background and objective:
The AIRE project aims to improve the detection of respiratory distress in children under five years of age by introducing the routine use of pulse oximeters (POs) during consultations in primary health centers (PHCs) in four countries (Burkina Faso, Guinea, Mali, Niger). A wide range of activities (training of health workers, distribution of essential drugs, supervision visits, etc.) were also carried out. Assessing whether the activities planned by the project were actually implemented and for as long and as often as planned is very important in order to link intervention activities to results; and to analyze implementation challenges in the field.

Methods:
Our research used the Carroll et al. (2007) and Perez et al. (2016) framework for analysing fidelity and adaptation to implementation. We identified the core activities planned for this project both through a review of the AIRE project documentation and by interviewing the operational teams. We then interviewed the actors involved in the project in16 PHCs (health workers, heads of centers, etc.) to learn about the actual implementation of the activities.

Results:
In all countries, essential activities such as the training of agents and the distribution of POs have been carried out, but have sometimes been modified compared to what was planned. We note that the implementation of some activities varies between countries. For example, in Guinea and Burkina Faso, visits to supervise the use of the PO were made more often than planned, while in Mali visits were less frequent in some centers due to safety issues. In one country, technical support to the PHCs on drug management could not be carried out due to difficulties in recruiting a logistician.

Conclusion:
Differences in implementation fidelity between countries and between health centers may help explain variations in project outcomes.