Intermittent Preventive Treatment in Infants - IPTi
Swiss Tropical Institute London School of Hygiene and Tropical Medicine

IPTi Decision-support tool

IPTi Decision-Support Tool

Data sources

Transmission intensity

Allocation of first administrative levels to malaria transmission intensity and seasonality categories was based on data extracted from systematic reviews of the malaria literature in 2006 [6]. To categorise transmission intensity, the most recent and methodologically robust estimates of annual entomological inoculation rate (EIR) were used. Transmission was categorised as:

1. Low: EIR<10 bites per person per year (pppy)

2. Medium: EIR 10-100 bites pppy

3. High: EIR>100 bites pppy

Given large gaps in entomological and parasitological data, we present estimates of the mean and 95% confidence intervals for P. falciparum parasite prevalence by first administrative level, from modelled surfaces for the year 2007 developed by the Malaria Atlas Project [2-3]. Parasite prevalence for children aged 2-5 years old can be related to EIR using a modelled relationship [4], and we have used the following unadjusted cut-offs to convert these data:  0.1-49.9% = EIR<10, 50.0-70.0% = EIR 10-100, >70.0% = EIR>100.

Seasonality

Seasonality of malaria transmission was extracted from the published literature where available and was categorised as being markedly seasonal (75% or more of cases are concentrated in six or less months of the year) or not markedly seasonal [5]. Where this was unavailable, the MARA (Mapping Malaria Risk in Africa) database provided a classification based on season of climate suitability for malaria transmission.

As some of these data may not accurately reflect the current transmission setting that applies to a given area, there is also an opportunity for the user to input data and/or personal opinion, based on professional experience in the field.

Age-patterns

Continuous probability distributions were fitted to six transmission scenarios using data on the age-distribution of malaria morbidity and mortality from a systematic literature review of epidemiological research studies in Sub-Saharan Africa [6]. See the Acknowledgements for sources of data. The proportion of clinical malaria fevers, hospital admissions with malaria and malaria-diagnosed deaths (defined as acute-febrile illness minus other obvious causes) that would be targeted by IPTi, were estimated from the best fitting distribution.

Cases averted

An individual-based stochastic model of IPTi [7] was used to predict the number of cases and deaths likely to be averted with IPTi using sulphadoxine-pyrimethamine (SP) for these same six transmission scenarios. The cases averted are presented as ranges for low (EIR=1-10 bites pppy), medium (EIR=10-100 bites pppy) and high (EIR=100-200 bites pppy) transmission settings. The following assumptions were kept constant across all settings: 4% of fever cases and 48% of severe cases per five day interval were treated effectively and other interventions were considered to be absent.

Simulations were run for 100%, 80%, 50% and 0% wildtype infections, with the remainder of infections being equally distributed (50:50) between intermediate and strongly resistant mutations. The model component for SP action was based on the Hastings & Watkins data for dhfr108 mutations [8]. SP was assumed to clear all wildtype infections and to provide 50 days of prophylaxis against new wildtype infections. SP was assumed to clear all intermediate resistance infections and to provide 10 days of prophylaxis against these infections. SP was assumed to clear 50% of strongly resistant mutations and to have no prophylactic effect against these infections. It is not possible to vary the ratio of intermediate and strongly resistant infections in this tool, or to consider the frequency of specific mutations.

The model was run separately for each of the different national EPI schedules, and these figures are then adjusted for expected IPTi coverage assuming a linear relationship. EPI coverage rates to inform estimates of this parameter were obtained from: WHO/UNICEF Immunization surveillance, assessment and monitoring.

Cost effectiveness

Having calculated the morbidity and mortality averted, the tool goes on to estimate the costs and savings to the health system associated with introducing IPTi. We recognise that up-to-date costs for every setting are not always easily accessible, therefore default costs are available for all cost inputs, and you are encouraged to use your own data when possible. The costing estimates [9] were based on costs of IPTi originally collected in Southern Tanzania [10] and WHO Choice health facility costs. Specific sources of data and assumptions are presented on the relevant pages of this tool.

Next: About the tool