Rapid reduction of malaria following introduction of vector control interventions in Tororo District, Uganda: a descriptive study

Despite scale up of interventions geared towards control and eventual elimination of malaria, the disease is still endemic in 91 countries worldwide with 3.3 billion people at risk of developing malaria in any given year. In 2015, 212 million cases of malaria occurred globally resulting into 429,000 deaths of which 92% were in Africa [1]. High malaria burden has been identified as one of the causes of poverty in many developing countries [2]. However, with scale-up of interventions in most endemic countries, global reduction of malaria burden has been reported [1].

Uganda is among the 15 countries most affected by malaria in the world [3, 4]. The national malaria indicator survey of 2009 reported that 45% of Ugandan children aged below 5 years carried malaria parasites. According to the Ministry of Health (MoH), malaria accounts for 25–40% of outpatient visits in public health facilities and is responsible for half of inpatient paediatric deaths. Plasmodium falciparum is the predominant species of malaria in Uganda [5, 6]. The major malaria parasite vectors in the country are Anopheles gambiae complex and Anopheles funestus [7]. Malaria transmission in Uganda is perennial with two peaks in April to June and October to December, corresponding to the rainfall seasons during which the vector biting density is high [8, 9]. The highest global entomological inoculation rate has been reported in Uganda [8]. The country is among others still in the attack stage of malaria control [9, 10]. The goal of Uganda’s malaria control policy is to reduce morbidity and prevent mortality attributable to malaria [11]. The US Present’s Malaria initiative supported strategies including proper case management of non- complicated cases using artemisinin combination therapy (ACT), intermittent preventive therapy in pregnancy (IPTp), Integrated Community Case Management (ICCM), distribution of long-lasting insecticide-treated nets (LLINs) to communities and indoor residual spraying (IRS) in epidemic prone and hyper-endemic districts have been scaled up in Uganda. In 2013, the MoH achieved universal distribution of LLINs to all communities by offering a free bed net to every two people in a household. In 2006, IRS was started in 10 districts of Northern Uganda as a complementary strategy to LLINs and ACTs to achieve faster reduction of malaria incidence in the region [12]. In 2014, IRS was scaled up to more 14 high malaria prevalence districts in Northern and Eastern Uganda [13].

Evaluating effectiveness of such malaria interventions over time often requires well designed studies such as controlled randomized trials [14‐16]. However, such studies are usually not feasible due to inadequate resources. Health Management Information System (HMIS) was introduced to collect routine surveillance data in health facilities useful for monitoring and evaluation of interventions [17]. The HMIS captures all data required to determine all malaria indicator variables such as incidence and test positivity rate. Since the roll out of the District Health Information Software, version 2 (DHIS2) in 2012 [18], routine malaria surveillance has improved. All public and high volume private health facilities report malaria cases weekly to the MoH using HMIS 105 form. With improving completeness of reporting, available routine surveillance data can be analysed and used by districts to monitor trends of malaria incidence and evaluation of malaria interventions [19]. However, due to inadequate malaria epidemiology expertise in districts [20, 21], these data are not fully utilized. The study analysed surveillance data of a high malaria endemic district which implemented universal distribution of LLINs in 2013 and advanced to IRS in combination with LLINS in December 2014. The aim of this study was to use routine malaria surveillance data to assess change in malaria infection using incidence, test positivity rates and outpatient (OPD) attendance due to malaria as key indicators at three time-points (when there were no vector control interventions; when LLINs were used alone and when LLINs were used in combination with IRS). Malaria incidence is a measure of the rate of new malaria cases in the population and an important indicator of the frequency of the disease transmission.

This study analysed three malaria burden indicators; incidence, test positivity rate and OPD attendance due to malaria to assess change in the disease magnitude in Tororo District since the introduction of vector control interventions. The findings show consistent reduction in malaria incidence, test positivity rate, and OPD attendance due to malaria following universal LLIN coverage was achieved, but most importantly, after universal LLIN coverage was combined with IRS. This suggests that malaria control interventions that combine LLINs with IRS can be more effective than those that use LLINs alone [36]. Findings also show that during 2014 universal coverage of communities with LLINs was followed by a reduction in malaria incidence, but the monthly reduction trend was not significant. This could suggest that LLINs alone might not change malaria endemic level of a district [37].

The lack of significant reduction in malaria incidence between 2013 and 2014 following the universal distribution of LLINs in Tororo District in 2013 might mean that when LLINs were used alone, there was inadequate protection of communities from malaria vector bites to cut down transmission [12]. This could be explained by several factors. First, during the implementation of universal LLINs, two persons in a household received one net. It was difficult to ascertain whether everyone slept under the same net every night. Second, adherence to LLIN use may have been low, as suggested by data in the Uganda National Malaria Indicator Survey 2014. Third, increasing resistance of vectors to pyrethroid insecticides used in LLINs [7] may have led to reduced effectiveness of the treated nets. Assessing bed net ownership and usage was beyond the scope of this study because relevant data were not captured by the routine surveillance system. Nevertheless, these findings provide more evidence to earlier susceptibility studies [7] which reported failure of LLINs to reduce indoor biting Anopheles density and randomized trials which found insignificant malaria reduction when LLINs were used in communities without IRS [36].

The study also found that combining LLINs with IRS was followed by rapid decline in malaria incidence and a significant reduction of the disease in 2015. Indoor residual spraying is known to rapidly reduce malaria vectors and protect communities from infective bites [22] and might have protected all people who did not sleep under LLINs and gave additional protection to those who slept under treated nets [22]. The average house coverage with IRS in Tororo was 90% in 2015, according to a report by the District Health Office [unpublished]. A combination of the two interventions, therefore, might have achieved greater reduction of malaria vectors resulting into significant reduction in malaria transmission.

The decline of malaria in Tororo District is similar to the rapid malaria reduction which IRS and LLINs achieved in Northern Uganda [38]. These findings are also consistent with studies which described LLIN and IRS combination as synergistic malaria control strategies [12, 36]. The reduction observed is in line with the reported progress of malaria reduction in African countries by WHO [3]. However unlike the findings by Jagannathan et al. who found increasing trend of malaria in a cohort study of 100 children following initial vector control using LLINs in Tororo [39], this study found insignificant decrease in malaria incidence following universal distribution of LLINs alone. Results obtained by this analysis might be more reflective of realities in the communities at large because the data used were about the whole district.

Variation in monthly reporting by health facilities (70–99%), missing and incomplete data were among the limitations of this study. Data were only available from July 2012 when DHIS2 was rolled out in the district. We used 2012 to demonstrate the monthly trend before LLINs, but did not compare the data to other years. Inadequate data on other key malaria indicator variables such as mortality rate and malaria in pregnancy limited our analysis to three parameters. There was lack of data 3 years before implementation of vector control interventions as baseline for most accurate descriptive comparison. Assessing demographic and environmental factors which could have contributed to decline in malaria was beyond our scope due to lack of event records. The number of malaria cases used in computing incidence comprised of clinically diagnosed and the laboratory confirmed because the roll out of the test and treat policy in the district was gradual between 2012 and 2015. However, because health workers used same clinical guidelines for diagnosis, the incidence we obtained might give a fairly accurate estimation of the malaria reduction.

A rapid reduction in malaria incidence was observed in Tororo District following the introduction of IRS in addition to LLINs. There was no significant reduction in malaria incidence following universal distribution of LLINs to communities before introduction of IRS.