Changes in health workers' malaria diagnosis and treatment practices in Kenya

Background: Change of Kenyan treatment policy for uncomplicated malaria from sulphadoxine-pyrimethamine to artemether-lumefantrine (AL) was accompanied by revised recommendations promoting presumptive malaria diagnosis in young children and, wherever possible, parasitological diagnosis and adherence to test results in older children and adults. Three years after the policy implementation, health workers' adherence to malaria diagnosis and treatment recommendations was evaluated. Methods: A national cross-sectional, cluster sample survey was undertaken at public health facilities. Data were collected using quality-of-care assessment methods. Analysis was restricted to facilities with AL in stock. Main outcomes were diagnosis and treatment practices for febrile outpatients stratified by age, availability of diagnostics, use of malaria diagnostic tests, and test result. Results: The analysis included 1,096 febrile patients (567 aged < 5 years and 529 aged >= 5 years) at 88 facilities with malaria diagnostics, and 880 febrile patients (407 aged < 5 years and 473 aged >= 5 years) at 71 facilities without malaria diagnostic capacity. At all facilities, 19.8% of young children and 28.7% of patients aged >= 5 years were tested, while at facilities with diagnostics, 33.5% and 53.7% were respectively tested in each age group. Overall, AL was prescribed for 63.6% of children aged < 5 years and for 65.0% of patients aged >= 5 years, while amodiaquine or sulphadoxine-pyrimethamine monotherapies were prescribed for only 2.0% of children and 3.9% of older children and adults. In children aged < 5 years, AL was prescribed for 74.7% of test positive, 40.4% of test negative and 60.7% of patients without test performed. In patients aged >= 5 years, AL was prescribed for 86.7% of test positive, 32.8% of test negative and 58.0% of patients without test performed. At least one anti-malarial treatment was prescribed for 56.6% of children and 50.4% of patients aged >= 5 years with a negative test result. Conclusions: Overall, malaria testing rates were low and, despite different age-specific recommendations, only moderate differences in testing rates between the two age groups were observed at facilities with available diagnostics. In both age groups, AL use prevailed, and prior ineffective anti-malarial treatments were nearly non-existent. The large majority of test positive patients were treated with recommended AL; however, anti-malarial treatments for test negative patients were widespread, with AL being the dominant choice. Recent change of diagnostic policy to universal testing in Kenya is an opportunity to improve upon the quality of malaria case management. This will be, however, dependent upon the delivery of a comprehensive case management package including large scale deployment of diagnostics, good quality of training, post-training follow-up, structured supervisory visits, and more intense monitoring.