Can individuals’ beliefs help us understand nonadherence to malaria test results? Evidence from rural Kenya