PDMC Implementation Trial in Kenya

Study design:

A multi-centre, 2-arm, cluster-randomised trial evaluating two health system delivery strategies and a 3-arm nested individually randomised trial evaluating two new intervention strategies to enhance end-user adherence to post discharge malaria chemoprevention (PDMC) with dihydroartemisinin-piperaquine compared to the standard of care without adherence strategies.

Study sites: Health facilities with blood transfusion services in malaria-endemic areas in western Kenya.

Study Population:

Cluster inclusion criteria: health facilities with blood transfusion services offering in-patient care for children with severe anaemia and severe malaria. Exclusion criteria: primary health facilities without subservient lower-level health facilities.

Individual inclusion criteria: convalescent children aged <10 years, weighing ≥5 kg admitted with severe anaemia (haemoglobin<5g/dL) or severe malaria; clinically stable, able to take or switch to oral medication; post-transfusion Hb >5g/dL. Exclusion criteria: blood loss due to trauma, malignancy, known bleeding disorders or sickle cell disease, known hypersensitivity to study drug, known heart conditions, non-resident in the study area, previous participation in the study, known need at enrolment for prohibited medication and scheduled surgery during the 4-month course of the study. HIV infection and cotrimoxazole prophylaxis are not exclusion criteria.

Study Interventions:

Cluster randomised trial: Hospitals will be randomised to one of two delivery platforms for PDMC (A or B). In delivery platform A, the initial course of PDMC will be delivered at discharge at the admitting facility. Subsequent courses will be dispensed during monthly visits to the health facility of their choice.In strategy B, all 3 courses will be dispensed at discharge.

Individually randomised trial: Three different adherence support strategies, including: SMS reminders to caregivers (arm a), Community Health Promoter (CHP) support through home visits (arm b), and a control arm comprising of no additional adherence support (arm c).

All participating children will receive standard in-hospital care for severe anaemia or severe malaria (blood transfusion, often combined parenteral artesunate, followed by a 3-day course of AL (whether they initially had malaria or not), which will be started in-hospital as soon as they are able to take oral medication. Many children are likely to receive parenteral antibiotics as well as part of the standard of care.

Primary endpoints:

Cluster randomised trial: The proportion of eligible participants who were not prescribed the full 3 courses of PDMC.

Individually randomised trial: The proportion of children with incomplete adherence to the 9 doses of PDMC (three courses of 3-day treatments 3x3=9).

Secondary endpoints:

1. Clinical effectiveness (all-cause and malaria-specific sick-child clinic visits, all-cause and cause-specific readmissions and all-cause mortality by the end of week 14 after discharge.
2. Safety (incidence of serious adverse events).
3. Cost-effectiveness.
4. Acceptability and feasibility.

Follow-up procedures:

Children will be followed for 14 weeks (i.e., four weeks after the last PDMC course) through passive surveillance of clinic visits and hospitalisations. Each child will then be seen for an end-of-study visit towards the end of week 14. Children will also be visited at home for unannounced home visits one to three days from the last day of the last dose of each 3-day course of PDMC medication to assess adherence.

Sample size:

Cluster randomised trial: A sample size of 51 participants with complete follow-up recruited in 5 control and 5 intervention clusters (300 per arm) provides over 80% power to detect a risk difference of 20% from 35% in the control arm to 15% in the intervention clusters (RR=0.43), using a two-sided alpha of 0.05, and assuming an intra-cluster correlation coefficient (ICC) of 0.03,a coefficient of variation of 0.39 and a small-sample correction at the analysis stage, thus the t-distribution is assumed implying an adjustment in the degrees of freedom. Similarly, the same size would provide over 90% power to detect a risk difference of 23% from 35.0% to 12% (alpha = 0.05) (RR=0.34). To allow for a 15% loss to follow-up and enough power at the individual-level, we aim to recruit approximately 60 participants per cluster, or approximately 600 participants (300 per arm).

Individually randomised trial: A sample size of 147 participants in the control arm (c) and 147 in each of the two intervention arms (a and b), totalling 441 participants, would provide 90% power to detect a 50% reduction in the primary endpoint (proportion of children with incomplete adherence to the 9 doses of PDMC) from 36.8% in the control arm to 18.4% (RR=0.50) in any of the two intervention arms using a 1:1:1 allocation after accounting for the multiple comparisons (α = 0.025). A total of 519 participants (173 per arm) will be recruited to allow for a 15% loss to follow-up.

Data Analysis:

Risk ratios and corresponding 95% confidence intervals will be computed to compare treatment effects using multi-level mixed models accounting for clustering effects.