POLICY BRIEF
Key messages:

1. CHWs are capable of assessing respiratory rate and prescribing appropriate treatment.
2. Improved tools for assessing respiratory rate would strengthen rational use of antibiotics.
3. A three day rather than five day course of antibiotics could improve caregiver adherence, reducing risk of drug resistance and cost, however, additional evidence is needed to support use of a short course of antibiotics for pneumonia in Africa.

Introduction:
Community‐based interventions for health, such as integrated community case management (iCCM), increase access to care and have potential to reduce child mortality by reaching marginalised populations. Through training, the use of effective clinical algorithms and active supervision, they offer an opportunity to improve rational use of antibiotics and limit the development of drug resistance in resource‐poor settings. This study provides evidence on rational use of antibiotics for treatment of pneumonia symptoms to inform future implementation of iCCM, safeguarding effectiveness of current treatments whilst continuing to maximise access to care.
Methods:
1,497 CHWs’ consultations were directly observed by non‐clinical researchers in CHWs’ usual place of work, with measurement of respiratory rate by CHWs recorded on video in 538 consultations. Videos were used by experts to conduct a retrospective gold standard assessment of respiratory rate. Caregivers whose children were prescribed antibiotics were followed up to assess adherence, through questionnaires and pill counts. Focus group discussions and key informant interviews were also conducted.
Results:
CHWs:

1. 65% [55,74]* of antibiotics were correctly prescribed for fast breathing (suspected pneumonia)
2. 74% [69,79] of children received appropriate treatment
3. qualitative data indicate that CHWs have a good understanding of pneumonia diagnosis and treatment, and although caregivers expected and sometimes applied pressure to receive drugs, CHWs’ treatment decisions were not influenced
4. CHWs adhered to guidelines for 92% [87,95] of children.

Caregivers:

1. 66% [50,79] of children completed full five day course of amoxicillin (pill count)
2. self‐reported adherence is high for the correct number of doses given per day (93% [82,96]), but only 54% [36,64] gave treatment three times a day for five days
3. adherence for three times a day, for three or more days is far higher at 76% [63,85]

*Numbers in brackets [x,x] indicate 95% confidence intervals
Acknowledgements: Kirstie Graham, Chomba Sinyangwe, Sarala Nicholas, Rebecca King, Karin Kallander, Helen Counihan, James Tibenderana, Mark Montague, Samuel Mukupa, Sylvia Meek, Prudence Hamade