Background

Malaria remains a major global health challenge with 241 million annual cases and 627,000 deaths recorded worldwide in 2020. The bulk of this disease burden falls on Africa with 95% of all malaria cases and all malaria deaths occurring in the World Health Organization (WHO) African Region [1]. Whilst there have been recent successes in eliminating the disease from their territory by some countries, no country in sub-Saharan Africa has achieved elimination to date.

In southern Africa some countries have reduced malaria transmission to levels low enough that elimination has become a realistic prospect [2]. Realizing the need for inter-country coordination in the pursuit of elimination, eight countries in southern Africa established the Elimination 8 (E8) initiative under the Southern African Development Community (SADC) in 2009 [2]. The E8 facilitates the harmonisation of malaria control and elimination strategies between member countries with the ambitious goal of eliminating malaria by 2030. The E8 comprises of two groups of countries: the southernmost so-called frontline four (Botswana, Namibia, South Africa, and Eswatini), where malaria incidence has been reduced to less than 5 cases per 1000 per year, and their northern neighbours, the so-called second-line four (Angola, Mozambique, Zambia, and Zimbabwe) where a high burden of transmission persists with over 25 million cases reported in 2020 [3], see Table 1. In the frontline countries large declines in annual malaria cases were achieved from 2000 to 2015, but progress has plateaued in recent years with a resurgence of cases seen in outbreak years such as 2017 [2]. In second line countries the total number of malaria cases has increased since 2015 [1, 4].

Table 1 Malaria cases and annual incidence per 1000 population in E8 countries in 2018
Full size table

A major challenge facing countries striving for malaria elimination is the cross-border movement of people between countries with a high malaria burden to countries with low malaria transmission [5, 6]. The southern African region is characterised by a high degree of inter-connectedness between countries through economic and cultural ties, population movement as well as shared vector ecologies [7, 8]. The higher level of malaria transmission in second-line countries serves as a perpetual reservoir of infection that is imported into the four frontline countries [9]. To address this impediment towards achieving elimination, the Elimination 8 Strategic Plan of 2015 to 2020 had as one of its primary objectives the need to reduce cross-border malaria transmission [10]. The rationale was that importation of malaria could be mitigated by expanding access to malaria services for residents of remote, often underserved border districts as well as cross-border mobile and migrant populations (MMPs). Thus the E8 in collaboration with National Malaria Control Programmes (NMCPs) and other partners, established 46 border health facilities on five key international borders between high and low transmission districts of E8 countries in 2017 [2].

To better understand the malaria situation in border areas and the role that border posts and other health facilities in border areas have in providing malaria services, a descriptive study was undertaken to determine amongst border populations their access to malaria prevention, diagnosis and treatment; knowledge, perceptions and behaviour related to prevention and treatment-seeking; and patterns of travel.

This paper reports on the findings of the study conducted in the four second-line countries (Angola, Zambia, Zimbabwe and Mozambique).

Methods

The study reported in this paper was part of a larger investigation, the E8 border post impact evaluation study, which was undertaken in seven countries in the E8 region. The methods and objectives differed between the three front-line countries (Namibia, Botswana, South Africa) in which it was conducted, and the four second-line countries. In this paper, only the results of the study conducted in second-line line countries are reported. Related studies in frontline countries were completed later and will be reported separately. Below is a description of the core intervention, and the associated study as carried out in the four second-line countries.

Core interventions

In 2017, the E8 with the help of partner organisations established malaria health posts along international frontiers within the region, supported by The Global Fund to Fight AIDS, Tuberculosis and Malaria (see Fig. 1). The locations chosen for border posts followed a desk review and a consultation with national malaria control programs in each country. Candidate locations were border areas, which were underserved in terms of health facility provision, and where there was evidence of a gradient in transmission from one side of the border to the other. The rationale was that borders with different levels of transmission on opposite sides, as evidenced by district health data for malaria, would be prone to significant parasite importation from the high to the low endemicity sides of the border in the absence of adequate malaria services along the border.

Fig. 1
figure 1

(Source: E8 Secretariat)

E8 Malaria border health posts, and study sites in second line countries

Full size image

The posts offered free testing and treatment to residents of border districts and travellers, and provided outreach services to surrounding communities, as summarized in Table 2.

Table 2 Primary models of service delivery for E8 border health posts (Total 46)
Full size table

Study outcomes

The following outcomes were measured in the study described in this paper: (1) Proportions of respondents receiving timely diagnosis and treatment for febrile illness; (2) Proportions of respondents able to access health facilities for malaria diagnosis and treatment; (3) Proportions of respondents with access to malaria prevention; (4) Respondents’ knowledge of symptoms and malaria prevention, and awareness of border health posts; and (5) Frequency of domestic and international travel.

Study design and implementation

The study in second line countries, was conducted in one site located in the vicinity of a malaria border post in each of the four countries, selected in consultation with national malaria control programmes, at border crossings with high proportions of mobile populations (Fig. 1). The study population included residents of all ages within a 30 km radius of a selected malaria post.

Using census data, each study area was segmented into geographic clusters of roughly 100–125 households each. Three of these segmented clusters were randomly selected, and 25 households were randomly sampled in each segment to achieve a sample of 75 households per site. An additional 20% contingency was included to allow for non-response.

Written informed consent was sought from the head of each selected household, or from a responsible household member acting as their representative. A structured questionnaire was used to obtain information on knowledge of malaria, attitudes toward and access to malaria posts and existing health facilities, the proportion of residents receiving diagnosis and treatment for febrile illness, practices for malaria prevention, and risk factors for malaria associated with local and international travel. Sections of the questionnaire were adapted from the standard Roll Back Malaria Indicator Survey [11]. Questions were answered by the head of household or their representative on behalf of individuals within a selected household; parents or guardians responded on behalf of children under 18 years of age.

It was assumed that 75 households would need to be selected in each site to achieve a sample size of 300 individuals per site, or 1200 individuals for the four countries combined. In sites where average household size differed from 4 individuals per household, the total number of households was adjusted to achieve the target of approximately 300 individuals. The sample size of 300 per site was pragmatically chosen to match the resources that were available to conduct the study. For questions that were only asked of subgroups of participants, for example proportions seeking treatment out of those who had fever within the recall period, the sample sizes were necessarily smaller.

In Mozambique, Centro de Investigação em Saúde de Manhiça (CISM) conducted the study. In Angola, Zambia and Zimbabwe, the study was carried out by their respective NMCPs. The studies were overseen and coordinated by Elimination 8 Secretariat. Fieldwork was conducted within 6 to 18 months of the implementation of the border posts in each country (Table 3).

Table 3 Date of implementation of border health posts and timelines by country
Full size table

Ethics approval was obtained from the University of California San Francisco Committee on Human Research (IRB# 17-23221), and the ethics committees of implementing partner organisations in each country. The E8 Research Subcommittee acted as a study steering committee.

The study was funded by The Bill and Melinda Gates Foundation through a grant to the University of California San Francisco. Additional funding for the study was obtained from The Global Fund to Fight AIDS, tuberculosis and Malaria. The funders had no role in the publication of the study.

Results

Study results are tabulated under the following themes: access to diagnosis and treatment for febrile illness; access to malaria prevention; international and domestic travel; and knowledge of malaria.

Access to diagnosis and treatment for febrile illness

Fewer than 70% of respondents reported seeking care for febrile symptoms within 48 h of onset of fever amongst adults and children in Angola, adults in Mozambique and children in Zambia (Table 4). A high proportion, but not all of those who presented with fever received a blood test. The highest testing rate was reported in Zimbabwe for both adults and children. Overall, due to neither treatment-seeking nor testing of those who do seek treatment being universal, only about half of those who reported fever had a diagnostic blood test, Zimbabwe being the exception with nearly all those who had fever receiving a test. Nearly all those with a positive test result reported receiving medication.

Table 4 Treatment-seeking behaviour and access to diagnosis and treatment: resident respondents seeking treatment for febrile illness (recall < 4 weeks) and respondents who received a blood test for malaria, by country
Full size table

The overwhelming majority of respondents used either a government hospital or other government health facilities as their first choice for seeking treatment for febrile illness (Table 5). Few used private sector facilities including pharmacies, and none reported traditional healers as their first choice for diagnosis of fever. Community health workers were the first choice for seeking diagnosis and treatment for the majority of respondents in Zimbabwe.

Table 5 Access to treatment: among those with fever who sought treatment, first choice for seeking treatment, reasons for first choice for seeking treatment, time taken, distance travelled and treatment given, by country (recall < 4 weeks)
Full size table

Convenience and nearness were the predominant reason given for seeking treatment at a given facility, with quality and cost also mentioned in Angola and Mozambique. Most respondents lived within 45 min travel time of the facility where they sought diagnosis and treatment and, probably due to the design of the study, many lived within 15 min of their chosen facility. At the site in Angola, however, many respondents (44% of adults and 36% of children) lived more than 45 min from a place where they would seek health care (Table 5).

Table 6 Insecticide treated net ownership and usage, by country
Full size table

Access to malaria prevention

Net ownership was high in Mozambique (87%) and Zambia (96%), and moderately high in Angola (65%). The lower ownership in Zimbabwe needs to be seen in the context of this site being in an IRS area, which was not targeted for LLIN ownership. In the three sites where LLINs were the primary means of vector control (Angola, Mozambique, Zambia), the number of nets owned per household (of those who owned nets) was above two. Although this figure was highest in Angola, this needs to be seen in the context that a larger proportion did not own any nets in this site. Unfortunately, data on total number of residents per household were not collected and therefore the proportion of households meeting the universal access provision of at least one net per two individuals could not be calculated. Based on an assumption of 4 individuals per household on average, the overall (average) provision of nets was near to or above the universal access target of at least one net per two individuals. In the three sites where LLINs were the main vector control policy, net use was high amongst those who owned nets (71–91%), and moderately high at an overall population level (59–86%). Nearly all respondents thought that nets were an effective means of preventing malaria (> 96%). In the three LLIN sites, the condition of most nets was good or fair.

Reasons for not using a net were a perceived absence of mosquitoes, nets being in poor condition, insufficient numbers of nets or because it was too hot at night (data not tabulated) (Table 6).

Nearly all household heads in Zimbabwe reported previous IRS (96%), more than half did so in Zambia (57%) and 42% in Mozambique. No previous IRS was reported by respondents in Angola (Table 7). In all three of the sites that reported previous IRS this was conducted within the current malaria season. High proportions of respondents believed that IRS was effective for preventing malaria in locations that also reported high coverage (Zambia and Zimbabwe) but only around 50% thought that IRS was effective in the other two sites.

Table 7 Reporting of indoor residual spraying (IRS) by country
Full size table

International and domestic travel

In the sites in Angola and Mozambique > 40% of respondents said they travelled during the previous 3 months. Travel was both domestic and across international borders. The reasons for domestic travel were mainly personal in Mozambique, i.e., visiting relatives, whereas the main reason for international travel was of an economic nature (for trade or shopping). In Angola 25% of international travel was to seek medical treatment. Sleeping outside whilst travelling was common for adults (ranging from 17 to 43%), particularly for Angolans and Mozambicans. Small numbers reported using any protective measures except in Zimbabwe where 39% reported using protective measures. Bed nets were the most common protective measure used, followed by the use of repellents. Chemoprophylaxis was hardly ever used (Table 8).

Table 8 Local and international travel by country (recall < 3 months)
Full size table

Knowledge of malaria

Mosquito bites were correctly identified as a potential cause for malaria by around 90% of respondents. Fever was named by a majority as a symptom of malaria. In some sites a larger proportion mentioned other symptoms such as chills and headaches as symptoms of malaria. Death was recognized by most as the worst possible outcome of malaria, although a minority (in Angola) said body weakness was the worst outcome. Very high proportions thought that mosquito nets were an effective means of protection. Considerable proportions of respondents mentioned incorrect prevention methods, such as cutting grass and burning of leaves. Not many respondents identified spraying the house with insecticide as a protective measure even in countries where it is the main means of vector control. This may be because respondents were thinking of measures, they themselves could undertake, rather than protective interventions in general (Table 9).

Table 9 Knowledge about malaria and its prevention by country
Full size table

Discussion

Cross-border malaria has been shown to be a major obstacle to malaria elimination [5, 9, 12]. To reduce the level of malaria importation from high to low burden countries in Southern Africa, the E8 made a major investment in setting up malaria border health posts along key borders in the region [2]. Recognizing the importance of border malaria in the quest for malaria elimination, a strategy for district level elimination has been adopted in border districts on both sides of borders separating high and low burden countries in the region [3]. To achieve the goals of reduction in cross-border importation and local elimination in border districts, comprehensive surveillance of malaria in border areas is essential. This study has provided a snapshot of access to malaria prevention, diagnosis and treatment and malaria related knowledge and behaviour in four border districts in so called second line (high burden) countries in the region.

Routine testing for malaria of febrile patients is a cornerstone of case management in settings where malaria is endemic. This study found that amongst border residents care-seeking for febrile illness was generally high. In Mozambique and Zimbabwe all children were reported to have sought care for the most recent febrile episode, but care seeking for children was low for children in Angola and Zambia. Not seeking care was mainly due to respondents judging that the fever episode did not warrant seeking care, but care-seeking may also be dependent on how accessible the nearest health care is. For this study survey sites were chosen to be within 30 km of a border post; hence the majority of respondents said that they lived within 45 min of the facility where they sought diagnosis and treatment. For most respondents the first and most convenient choice for seeking treatment was a government hospital or health facility. In the Zimbabwean site community health workers were most frequently mentioned as the preferred choice for seeking diagnosis and treatment for fever. Given that health care was generally accessible but a minority of respondents did not seek care when they or their children experienced fever, there appears to be a need for more effective messaging to convey the risk of serious disease and death if fever episodes are left undiagnosed and hence untreated. At the Angolan site many residents required more than 45 min to reach health care, which may explain the lower level of care-seeking within 48 h of fever at this site. Of concern is the finding that a proportion of respondents said they did not receive a blood test when they presented with fever. There is, therefore, a need for stronger direction to providers to adhere to testing guidelines in areas where malaria infection cannot be ruled out, even when malaria cases are rare. Supply chain management to prevent RDT stock outs may also require improvement to ensure these do not lead to testing not being done. The clear exception in this regard was the site in Zimbabwe where near universal testing of all those presenting with fever was reported. It is noteworthy that care was mostly provided by community health workers in Zimbabwe. This group of health workers provided high levels of adherence to guidelines when presented with patients experiencing fever; they were clearly the first choice for seeking diagnosis and treatment for the majority of respondents in Zimbabwe, and their presence in communities resulted in high proportions of respondents seeking treatment within 48 h of the onset of fever. This model of providing first line health care for malaria is, therefore, an example that other countries should consider adopting.

The study did not include testing respondents for malaria infection, but respondents were asked the result of the blood test at their last clinic attendance, if they were tested. Many reported a positive result (Table 4), indicating a high malaria burden of malaria, and potential for cross-border transmission given the high incidence of reported travel across borders. This underscores the importance malaria border health posts in reducing the burden and transmission of malaria in border areas, and across borders.

A positive finding from this study is that provision of protection against malaria through either mosquito nets or IRS was generally high. High proportions of households owned nets, particularly in Mozambique and Zambia, and those who owned nets, mostly slept under a net the night before the survey. Overall net provision was close to or exceeded the universal coverage criterion of one net per two persons where nets were the primary vector control intervention. A high proportion (> 76%) of nets were in good or fair condition. There was near universal acceptance of the effectiveness of nets to prevent malaria. For a minority, the obvious barrier to net usage was not owning a net.

In all sites adult respondents reported travel over the previous 3 months both within country, and across borders. Travel frequency was particularly high at the Angolan and Mozambican sites. The Angolan site had the highest proportion of cross border travellers, of whom a significant proportion travelled to seek medical treatment. This was mostly into Namibia (data not tabulated) and underscores the importance of malaria border screening to reduce the importation of malaria into low burden countries nearing elimination. Visiting friends or relatives was a common reason for travel, reflecting close ties between communities across borders in the region. A significant proportion travelled for economic reasons such as trading or shopping.

Sleeping outside whilst travelling was commonly reported, with many doing so without any protection against malaria. This represents an obvious gap in malaria protection for a group that are at high risk of malaria infection. Some reported using nets while sleeping outdoors, but in countries where the provision of IRS rather than nets is standard policy, there is a need to make provision for travellers in some other form, be it through chemoprophylaxis or repellents, or making mosquito nets available for the particular purpose of providing protection whilst travelling.

This investigation had important limitations. (1) The descriptive nature of the design had no comparator as it only provided data from communities in the vicinity of a border posts and there were no data on the period preceding the establishment of border posts. Nevertheless, it provides an insight into the provision of malaria services in these areas. (2) MMPs were not surveyed in the study. (3) The survey sample size was relatively small and confined to one locality in each country. Findings should, therefore, be regarded as illustrative, rather than statistically representative of the entire border. (4) The surveys represent snapshots in time, and not all were conducted at the same time in each country. Consequently, some results may have been confounded by the time of fieldwork in the seasonal malaria cycle.

The overall conclusions were

  • In some settings, a proportion of border residents did not receive a blood test when experiencing fever, either because they did not access health care, or because they were not tested when presenting with fever. Whilst most providers carried out blood tests when individuals presented with fever, there were exceptions that are cause for concern and remedial action.

  • Community health workers can play a key role in ensuring convenient and timely access to malaria diagnosis and treatment. This is a cost-effective option that should be considered in border districts where access is poor, and malaria infections go undetected and lead to cross-border transmission.

  • In general lack of access to health care due to distance or cost or mistrust of the provider was rare.

  • There was a high level of correct knowledge of causes, symptoms and prevention of malaria.

  • Nearly all those who reported a positive blood test result received medication at the place where they sought care.

  • A majority, but not all border residents had access to primary prevention against malaria through either LLINs or IRS. Although overall provision of LLNs was high, a proportion of households did not own any nets, even in sites where this was the main form of vector control.

  • Bed nets were overwhelmingly regarded as an effective method of preventing malaria.

  • Border residents travelled frequently both within their own country and across borders; some cross-border travel was for the purpose of seeking healthcare.

  • Sleeping outside whilst travelling was common and mostly without any protection against malaria; there was a clear gap in the provision of malaria prevention for this group.

Given the high level of travel by border residents and high levels of infection in many border areas, this study demonstrated the importance of border health facilities in providing access to malaria services. Prevention needs to be improved, particularly for people who travel and sleep outdoors. Community health workers can play a key role in providing access to testing and treating individuals with malaria symptoms. Messaging to communities to seek treatment when experiencing fever, and to health care providers to test those presenting with fever, needs to be intensified to ensure that symptomatic infected individuals do not go untreated.