The data presented here demonstrate that reported malaria cases remain high in Togo. Indeed this study found an overall increase of malaria cases in Est Mono district of 159% in 2010 from the 2005 level. The proportion of confirmed malaria cases also increased from 17.7% in 2005 to 85.9% in 2010. The government targeted a reduction in morbidity and mortality due to malaria by 50% in 2012 . The national health information reports show an increase of malaria incidence from 80 per 1,000 in 2005 to 160 per 1,000 population in 2010 . Several explanations exist for these observations. With the implementation of the new scale-up policy, there has been increased access to affordable interventions for malaria care even to populations that live more than five kilometres from a health centre. Indeed, the cost of ACT in Togo is relatively low (USD2) for a treatment course compared to the average cost of a case treatment (USD10) before 2007 in the public health centres of the district . This is supported by the increase of the total number of health care visit in Est Mono district during the study period. In Kenya, it was shown that the rate of health care use increased when free high-quality care was offered .
The implementation of the new policy of malaria control in Togo took into account the enhancement of health care . These findings suggest that several factors had impacted on the number of malaria cases reported throughout the six years – including the training of health personnel in malaria case management; the introduction of a new treatment at lower cost; the involvement of the community through the CHWs; improvement in the monitoring and reporting system and the use of new RDTs that are simple and quick to perform and show high sensitivity. The findings are consistent with those of Nyarango et al. from an Eritrean study, where it was reported that the incidence of malaria was influenced by many factors. These factors included the access to diagnosis and treatment, risk perceptions of the net users, the training of health workers, as well as the involvement of community health agents, who increased the proportion of malaria cases treated in the community from 50% to 78% .
A study conducted in Uganda in five selected hospitals over a longer time period (1999 to 2009) reported a significant increase in paediatric malaria admission . The reasons given were the abolition of user fees in the hospitals, low intervention coverage and the high malaria transmission. In Malawi, there was no decline in annual malaria prevalence among children from 2001 to 2010 because of the disparity in treated bed coverage and the low coverage of IRS .
In contrast, a decline in malaria incidence has been reported in many countries [6, 9, 14, 17–21] but in most of these countries the interventions implemented included IRS and the involvement of CHWs. Similarly, in a much larger study in South Africa, where IRS was implemented and use of ACT was introduced sequentially, a reduction in the incidence of malaria was observed after implementing the use of ACT . In Togo, IRS is an intervention included in the strategic plan  but it has not been implemented yet. ITNs remain widely used for malaria control in Togo. Nyarango et al.  and Sharp et al.  state that ITNs remain the intervention of choice for prevention of morbidity and mortality from malaria in areas of high transmission. But the preventive effect of ITNs depends on their insecticidal properties and physical integrity . In a recent study conducted in Ethiopia the combination of ITNs and IRS was found to be more effective as a control measure .
The results of the Togo study suggest that strengthening of the malaria reporting system by the control programme could have increased the number of reported malaria cases in Est Mono. Malaria cases could have been under-reported before the introduction of the new malaria control policy, which included the enhancement of case reporting.
In this study area, there was a decrease in mortality rate from 7.2 per 10,000 in 2005 to 3.6 per 10,000 in 2010 and an increase of laboratory-confirmed cases. These findings are consistent with national data . The steady decline in mortality due to malaria observed in this study has met the national target of reducing mortality by 50% by 2012. The early clinical diagnosis with laboratory confirmation and appropriate treatment using ACT as recommended by the national guidelines are the main factors that could have contributed to the decline in mortality due to malaria in Est Mono district. It is well known that the early clinical diagnosis and laboratory confirmation are essential aspects of effective care and effective case management of malaria .
A study conducted in the province of KwaZulu-Natal in South Africa showed that the use of ACT decreased by 97% mortality due to malaria . Similarly, in Zanzibar the decline by 52% in mortality due to malaria among children under five years old was observed after deployment of ACT . The availability of ACT in Togo and personnel training in malaria case management should be reinforced.
The rainfall was correlated with the prevalence of malaria in the Togo study. The findings are consistent with findings elsewhere [6, 9, 10, 26] and confirm the role played by climatic and seasonal factors in the malaria burden. However, the pattern of rainfall did not change markedly over the six years of the study. The time series analysis shows also that the increase in reported malaria cases could be attributed to the role played by the CHWs trained in malaria case management.
Secondary data were analysed in this study. This could have affected the data quality related to precision and the completeness during the data reporting. The study period of six years was short compared to the period used in other studies and could have affected the power of the ARIMA model. However, previous similar studies using five-year data have provided valid results . Some other factors such as migration, socioeconomic conditions, level of knowledge, community attitudes to and practices of malaria control may have influenced the changes in the burden of malaria that were not taken into account. However, the results of this study are likely to be generalizable to other districts of Togo, since health service delivery and the period of malaria control interventions are similar in all districts. The trends of malaria prevalence and mortality are likely to be representative of those for the entire country.