Reduced paediatric hospitalizations for malaria and febrile illness patterns following implementation of community-based malaria control programme in rural Rwanda

Background. Malaria control is currently receiving significant international commitment. As part of this commitment, Rwanda has undertaken a two-pronged approach to combating malaria via mass distribution of long-lasting insecticidal-treated nets and distribution of antimalarial medications by community health workers. This study attempted to measure the impact of these interventions on paediatric hospitalizations for malaria and on laboratory markers of disease severity. Methods. A retrospective analysis of hospital records pre- and post-community-based malaria control interventions at a district hospital in rural Rwanda was performed. The interventions took place in August 2006 in the region served by the hospital and consisted of mass insecticide treated net distribution and community health workers antimalarial medication disbursement. The study periods consisted of the December-February high transmission seasons pre- and post-rollout. The record review examined a total of 551 paediatric admissions to identify 1) laboratory-confirmed malaria, defined by thick smear examination, 2) suspected malaria, defined as fever and symptoms consistent with malaria in the absence of an alternate cause, and 3) all-cause admissions. To define the impact of the intervention on clinical markers of malaria disease, trends in admission peripheral parasitaemia and haemoglobin were analyzed. To define accuracy of clinical diagnoses, trends in proportions of malaria admissions which were microscopy-confirmed before and after the intervention were examined. Finally, to assess overall management of febrile illnesses antibiotic use was described. Results. Of the 551 total admissions, 268 (48.6%) and 437 (79.3%) were attributable to laboratory-confirmed and suspected malaria, respectively. The absolute number of admissions due to suspected malaria was smaller during the post-intervention period (N = 150) relative to the pre-intervention period (N = 287), in spite of an increase in the absolute number of hospitalizations due to other causes during the post-intervention period. The percentage of suspected malaria admissions that were laboratory-confirmed was greater during the pre-intervention period (80.4%) relative to the post-intervention period (48.1%, prevalence ratio [PR]: 1.67; 95% CI: 1.39 - 2.02; chi-squared p-value < 0.0001). Among children admitted with laboratory-confirmed malaria, the risk of high parasitaemia was higher during the pre-intervention period relative to the post-intervention period (age-adjusted PR: 1.62; 95% CI: 1.11 - 2.38; chi-squared p-value = 0.004), and the risk of severe anaemia was more than twofold greater during the pre-intervention period (age-adjusted PR: 2.47; 95% CI: 0.84 - 7.24; chi-squared p-value = 0.08). Antibiotic use was common, with 70.7% of all children with clinical malaria and 86.4% of children with slide-negative malaria receiving antibacterial therapy. Conclusion. This study suggests that both admissions for malaria and laboratory markers of clinical disease among children may be rapidly reduced following community-based malaria control efforts. Additionally, this study highlights the problem of over-diagnosis and over-treatment of malaria in malaria-endemic regions, especially as malaria prevalence falls. More accurate diagnosis and management of febrile illnesses is critically needed both now and as fever aetiologies change with further reductions in malaria.