Figure 2.

Deterministic (red line) versus stochastic (blue line) modelling of malaria elimination using a model previously published elsewhere.³ Only one run of the stochastic model is shown for clarity. The phases of the WHO malaria control-to-elimination continuum1 are indicated by the shaded background. The ‘limit of detection by surveillance’ indicates the number of cases below which a malaria surveillance programme is unlikely to detect any malaria, thus suggesting ‘apparent elimination’ (yellow circle). An arbitrary example is shown in the figure. Because of this detection limit, only the upper portion of the figure can be represented by surveillance data (‘data & model’), whereas the lower portion can only be represented by modelling predictions (‘model only’). Improving the sensitivity of surveillance would lower this detection limit. For ‘true elimination’ to occur (green circle), the number of malaria cases must fall below the ‘elimination threshold’ (<1 case). Only a perfect surveillance system detecting every case would have a limit of detection by surveillance equal to the elimination threshold. This is not generally the case in the field where surveillance systems are far from perfect and can miss many cases. Thus, the limit of detection by surveillance is generally above the elimination threshold. If malaria control interventions are stopped inappropriately early when apparent elimination occurs (red dotted line), numbers of cases begin to increase again. Modelling gives an indication of how long control measures would need to be continued to achieve true elimination. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)