A mix-up during submission led to the wrong version of table 3 being included in the full version of this paper (see bmj.com) by Chantal M Morel and colleagues (BMJ 2005;331:1299-302, 3 Dec). The R₀ value for case management with chloroquine should be 0.35 (rather than 0.3). The adherence for artemisinin based combination treatment should be 35% (not 40%), and neither that nor the adherence for non-artemisinin based treatment needs a footnote. Values for probability of success when patients were not fully compliant should be 35% for non-artemisinin based treatment and 0% for intermittent presumptive treatment during pregnancy (rather than 35% and 10% respectively, as given). These revised values also apply to table B on bmj.com.