Malaria continues to occur in many parts of Oromia and SNNPR regions of Ethiopia, and data from our 2007 survey and from the routine surveillance show that malaria is more prevalent in SNNPR than in Oromia. The incidence of reported clinical malaria cases in the year prior to the survey was four times higher in SNNPR than in Oromia, and the parasite prevalence in our survey was six times higher in SNNPR. In both regions, however, the survey results may be an underestimate of the peak prevalence because the survey was done in January, after the high transmission season that usually goes from August through December. The routine surveillance data also indicate however that, in both regions, malaria incidence in the lowest transmission season is about half of that in peak transmission season. There were some differences in parasite species ratio between the survey and the routine data: less P.falciparum was detected in Oromia than expected from surveillance data. This may represent seasonal variation since the survey was done at one point in time rather than over a full year as in the surveillance data. Apart from seasonal variation, these differences may also relate to differences between falciparum and vivax episodes in treatment seeking (e.g. if a larger proportion of falciparum cases report to health facilities than vivax cases, because of more severe symptoms), or in the duration of infection – which will affect parasite prevalence in the survey but not the incidence in the routine system.
The results of our survey are consistent with findings of other studies conducted within the country, although none are representative surveys of malaria parasite prevalence in the Oromia and SNNPR regions. A parasite prevalence survey in three zones near Lake Ziwai in Oromia from July to September 1994 found an average of 6.8% of slides positive over the period, with 4.8% in December only (P.falciparum 4.5% and P.vivax 2.1%) . The malaria prevalence in pregnant women in three sites of unstable transmission, one of which was Jimma town in Oromia, in surveys in December 2000/January 2001 was 1.8% (70% P.falciparum) .
Coverage of key vector control interventions has shown remarkable progress compared to estimates generated by the Ethiopia Demographic and Health Survey (DHS) conducted in 2005 . The Ethiopia DHS did not specifically report on LLINs, which were not commonly available at the time. Therefore we focus the comparison on ITN coverage and use between our survey and the DHS on the use of conventional insecticide treated nets (ITNs), using the DHS definition of either a LLIN or a conventional ITN which had been treated within the last year. In 2005, the proportion of households with at least one net (any type) in Oromia and SNNPR was only 2.8% and 8.2%, respectively, and only 1.9% and 6.6% of households owned at least one ITN. In contrast, in our survey nearly half of the households in each region owned at least one mosquito net and more than one-third of the households owned at least one LLIN. The mean number of ITN/LLIN per household increased from nearly zero in 2005 to 0.5 at the beginning of 2007.
The proportion of people who slept under an ITN/LLIN exhibited a ten-fold increase between 2005 and 2007 (below 3.5% vs. 35.4%). The improvement in net coverage over the period is mainly due to large numbers of nets supplied by the Global Fund for AIDS, TB and Malaria in rounds 2 and 5, and additional net donations plus assistance in procurement and delivery from UNICEF and other donors. For example, 2 million nets were distributed in 2005 alone, and the Global Fund singled out Ethiopia in its 2006 annual report as a country making major strides in scaling up net coverage . Other studies in the Horn of Africa have documented that bed nets confer high protection against malaria infection under field conditions . Assessment of the protective effect of LLIN against malaria in this study relative to other protective and risk factors (including rainfall) are being analyzed separately and will be reported elsewhere.
Similar good news was found for household insecticide spraying. DHS 2005 estimated that 8.5% and 9.1% of households were ever sprayed with insecticide in Oromia and SNNPR, respectively, and only 2.1% of households were sprayed with insecticide in both regions 0–6 months preceding the 2005 DHS . The survey reported here shows that almost one in five households had indoor residual house spraying in the 12 months preceding January 2007.
The current survey was conducted prior to additional planned further mass distribution of LLINs. The baseline results will help to inform the existing gap and gauge progress in LLIN coverage and utilization towards the stated goal of providing a mean of two LLIN per household. The baseline LLIN coverage in these two regions might serve as a reflection of the overall net coverage in the country though regional variations cannot be ruled out. A similar study conducted in Amhara region in December 2006 reported a slightly lower ITN/LLIN coverage .
However, LLIN distribution does not necessary translate into LLIN use, and we also designed this survey with the intent of examining if the onchocerciasis program can provide assistance to the malaria program through promotion of LLIN use during annual Mectizan® distribution activities. Currently the onchocerciasis CDDs are being trained to integrate education about LLIN into their 2008 activities. Our hypothesis is that the CDDs will serve as an 'adjuvant' to the Health Extension Workers (HEWs) who will play a pivotal role throughout Ethiopia in the grass roots implementation of the malaria control program. Such a finding would be further evidence that integration of malaria programs with neglected tropical disease programs (NTDs) is good policy [20, 21].
The results of this survey compare favourably with that of the routine surveillance and indicate that SNNPR (particularly the eastern and north-eastern areas of the region) bears a higher burden of malaria than Oromia. Future malaria control interventions need to take this into consideration.