In response to the huge burden of malaria in sub-Saharan Africa
 and the call by the WHO for scaled-up control efforts
, coupled with unprecedented availability of resources, targets for malaria control and elimination have been established
[25–27]. Attaining these goals require continuous surveillance, monitoring and evaluation of malaria control programmes for adaptation of intervention policy, procedures and methods to optimize the impact of interventions and rationalize resources.
In this study, the average number of malaria cases, proportional malaria mortality and case fatality rates due to malaria in Zambia declined by 31%, 63% and 62% respectively in children less than five years of age. During this period, IRS using pyrethroids and DDT was associated with a significant overall reduction in both proportional malaria mortality and CFR (P < 0.05) but the impact of LLINs was not statistically significant (P > 0.05). These findings are consistent with those of other studies
While Zambia has made appreciable progress in malaria vector control (Table
1), the observed difference in intervention effect could reflect the challenge of inconsistent bed net utilization
 and justifies the need for enhanced Information Education and Communication/ Behavioural Change Communication (IEC/BCC) and timely replenishment of worn out LLINs. Despite the difference in efficacy, both IRS and LLINs have had a significant impact on malaria cases, proportional malaria mortality and case fatality rates in Zambia.
The overall reduction in mortality and morbidity observed here cannot exclusively be ascribed to vector control, as ACT was simultaneously being implemented evenly across the country
[19, 29] and these would have contributed particularly to malaria outcomes but also potentially to transmission
. The ACT introduction appears to have resulted in improved treatment seeking behaviour by people and fewer stock outs of anti-malarials in health facilities
. There has been improved definitive diagnosis of cases with the roll-out of RDTs
. It is plausible that the CFR has reduced (P < 0.05) as a result of improvement in case management of severe malaria
, even if the vector control interventions were of no benefit.
Although other studies have reported impact of combined interventions on morbidity and mortality of all age-groups
, routine surveillance data have often been considered inadequate for monitoring control programmes
, and parasite prevalence surveys are most commonly used for assessing impact
. Importantly, the reliability of malaria prevalence surveys diminishes with declining prevalence, as the sample size becomes very large
[33, 34]. While routine surveillance systems have limitations
, the use of data from both malaria parasite prevalence survey and routine surveillance is important, particularly in areas where parasite rates are below 5%
In Zambia, ongoing monitoring of programme delivery and malaria incidence is becoming even more important as the reduced malaria infection rates create zones that are potentially prone to malaria outbreaks
[4, 35, 36]. Such data assists in planning effective response measures. There was marked heterogeneity in the average deaths and case fatality rates recorded in the IRS and ITN areas (P < 0.05) and this probably results from inter district heterogeneity in intervention coverage. The malaria control policy striving towards a malaria-free Zambia has facilitated homogenous coverage of integrated malaria control interventions including vector control. This precludes the availability of localities devoid of interventions that could act as control areas since people cannot be denied access to them
In Zambia, routine surveillance data are available across the country and country-wide scaling up of definitive diagnosis using microscopy and RDTs, promotion of IEC/BCC
, monitoring of the number of laboratory tests undertaken and trends in the malaria (slides or RDT) positivity rate, have assisted in providing more comprehensive data on malaria trends in the country, based on complete HMIS records supported by information from nationally representative household surveys
[5, 38]. Thus routine surveillance data are a useful resource for monitoring progress and impact of malaria control interventions.
Most malaria control programmes are being monitored and evaluated using clinical and entomological surveys that include parasite prevalence
[13, 39, 40]. This is the first evaluation of the impact of large scale IRS and ITNs on morbidity and mortality in children below the age of five using routine surveillance data at operational population level. The results indicate a marked impact with some variation between the two interventions, although there may well be other important confounders between predominantly rural and urban settings. The decrease in malaria cases, proportional malaria mortality and case fatality rates provide compelling evidence of the reduction of malaria in Zambia following the scaling-up of interventions.
Although there was an overall reduction in deaths and cases in children <5 years of age, there were a number of districts where these indicators remained persistently high. Pin-pointing precisely the factors responsible for persistence of high deaths and cases in these districts could be difficult, as the low impact of LLINs in operational settings could in large part be attributed to waning ownership, use and the physical and insecticide net durability. While high coverage was attained during the scaling-up programme, some nets were distributed as early as 2005. This situation underscores the need for a net replenishment programme and IEC/BCC programme on net use to maintain a high effective coverage
The comparatively high impact observed in IRS districts could be as a result of a combination of both IRS and LLINs, as rural parts of these districts may also have received LLINs through the country-wide mass distribution programme. IRS implementation has encroached into rural areas in some districts. In urban and peri-urban areas where IRS is confined, the uptake and utilization of anti-natal and child clinic, and commercially distributed LLINs have also improved markedly in the wake of enhanced IEC/BCC campaigns. This view is further supported by the fact that LLINs coverage in Zambia was similar for the poorest (63%) and richest quintiles (65%) and in urban (59%) and in rural areas (64%)
By April 2009, overall proportional malaria mortality reported from health facilities had declined by 66% in Zambia following scaling up of LLINs and IRS between 2006 and 2008, when proportional malaria mortality declined by 47% and nation-wide surveys showed that parasite prevalence declined by 53% (Table
1). The universal coverage with ITN, IRS and ACT is likely to achieve an even greater decline in malaria burden. In moderate to low transmission setting countries like Zambia, the Roll Back Malaria (RBM) target of reducing global malaria cases by 75% (from 2000 levels) may be attained even several years before 2015
 as long as high coverage, as well as effective service delivery, is maintained
This impact assessment was conducted for the period between 2007 and 2008 which may be too short a period to generalize on the observed temporal effects of LLINS and IRS on malaria control. However, the observed reductions in malaria cases, deaths and case fatality rate, in children under-five years of age following scale-up of these interventions are noteworthy findings. As such the use of routine surveillance data in determining the temporal effects of malaria control is an important methodological way forward for malaria monitoring and evaluation.