Adherence to the referral advice after introduction of rectal artesunate for pre-referral treatment of severe malaria at the community level: a noninferiority trial in the Democratic Republic of the Congo

Globally, an estimated 435,000 people died of malaria in 2017. Approximately 61% of the victims were children under 5 years (CU5), and 93% of those malaria deaths occurred in the World Health Organization (WHO) African Region. Among the 17 countries that had nearly 80% of the malaria deaths globally in 2017, Nigeria accounted for the highest proportion of deaths (19%), followed by the Democratic Republic of Congo (DRC) (11%) [1]. Approximately six to eight deaths out of 10 occur in the community before that patient reaches a health facility or within 30 min after arrival at the health facility [2].

Severe malaria progresses in children very quickly, and once a patient can no longer be given oral medication, survival depends upon quickly being taken to the hospital, where parenteral treatment can be given. Parenteral anti-malarial treatment is sometimes available at the primary health care level. However, access to parenteral treatment is quite difficult, and the delay can have a negative impact on the progress of the disease, even leading to death. Because the disease progresses rapidly, the risk of death from severe malaria is greatest in the first 24 h of the onset of symptoms [3], yet in most malaria endemic countries, the transit time between referral and arrival at health facilities that can administer intravenous treatment is usually prolonged, thereby delaying the commencement of appropriate anti-malarial treatment.

Therefore, the 2010 WHO guidelines for the treatment of malaria recommend the use of an emergency treatment for patients with suspected severe malaria before their transfer to a health facility. The WHO-recommended pre-referral treatment options include intramuscular artesunate, artemether or quinine, or rectal artesunate (RA) [4].

Evidence from recent studies demonstrates that in situations where IV administration is not possible and intramuscular injection is impractical (and the patient cannot be treated orally), the use of a single dose of RA as pre-referral treatment reduces the risk of death or permanent disability in young children [5]. The use of this intervention in endemic countries remains low, pending evidence about its efficacy, effectiveness and cost-effectiveness.

In line with the WHO recommendation, the DRC adopted the strategy in 2012 but delayed its implementation because its acceptability and feasibility were still unknown. In its effort to reduce the morbidity and mortality of malaria, the DRC in 2005 launched the community-based management of malaria integrated with pneumonia and diarrhoea at community care sites (CCSs) within the framework of the Integrated Management of Childhood Illness (IMCI).

The cornerstone of this strategy is the early detection and prompt treatment of clinical cases, and CHWs are given the opportunity to diagnose and treat malaria. This requires that effective and appropriate treatment with first-line oral artemisinin-based combination therapy (ACT) and guidance on referral criteria are provided at the community level through trained CHWs. Community-based management of malaria allows the coverage of health services for malaria to extend beyond the reach of health facilities. Diagnosis with a rapid diagnostic test (RDT) prior to the treatment of malaria-positive patients with ACT is currently the recommended practice. Severe cases should be referred to an appropriate facility.

CCSs are well-defined geographic areas covering one or more villages that are far from a health facility (more than 5 km) and that receive health care services provided by two trained CHWs. The CHWs are trained in IMCI and are selected and appointed by the communities. They provide the following services under the supervision of the chief nurse of the health facility (primary point of care): community management of childhood diseases (malaria, diarrhea, pneumonia and malnutrition), including the referral of cases with danger signs; community outreach activities; and community-based surveillance of diseases. Some studies have shown that in some countries, it was feasible and acceptable to use rectal artesunate at the community level as a pre-referral treatment for severe malaria [6‐9]. In contrast, in Laos, the rectal route was found unacceptable by the population in general, and health care workers (HCWs) impeded the use of a pre-referral emergency malaria treatment [10].

Although the feasibility and acceptability of the strategy has been demonstrated, its effect on adherence to referral advice has not been thoroughly assessed. Vermeersch et al. noted the need to address the mother’s judgement of the severity of malaria to prevent overuse of the RA as a pre-referral treatment [9]. Moreover, Kaona et al. outlined that gaining insights into the mother’s capacity to distinguish uncomplicated malaria from severe symptoms of the disease is key to successfully implementing pre-referral treatment at the community level [6]. On the other hand, adherence to referral advice is a cornerstone of the success of the strategy. Indeed, concerns have been raised regarding whether the risks of routine use of pre-referral RA at the community level include not only failure to adhere to medical advice, as the child will recover rapidly [8, 9, 11], but also the increasing possibility of artemisinin-based drug resistance [8, 12, 13]. However, there is limited information on adherence to referral advice following the administration of RA for pre-referral treatment of severe malaria [11]. Overall, most studies assessed adherence to referrals for any serious illness conditions from either the community level to health facilities [14] or from primary health facilities to hospitals [15‐19].

The primary objective of this study was to assess the adherence of the mothers/caretakers of CU5 to referral advice after pre-referral treatment of severe malaria with RA at the community level. The secondary objectives were to determine the acceptability of the strategy and the knowledge of CHWs regarding danger signs of malaria to better inform scaling up. The main hypothesis was that the integration of pre-referral RA at the community level will not negatively affect the uptake of referrals of CU5 with symptoms of severe malaria, i.e., adherence to the referral advice after the introduction of the “new intervention” (a pre-referral RA for CU5 with danger signs of malaria) will be noninferior by more than a 15% margin compared to adherence to “reference intervention” (urgently referring CU5 with malaria danger signs without providing any treatment).

The introduction of RA as a pre-referral treatment for severe malaria might pose the risk that the child will not be taken to a health facility because he or she will appear to get better rapidly after the treatment; however, such non-adherence runs the risk of death from resurgence and the development of drug resistance due to incomplete removal of parasites. However, pre-referral RA is more acceptable because it prevents the danger of death and disability, taking into account that every hour without treatment for severe malaria increases the risk of death [3]. This study does not add to the knowledge of the efficacy or cost-effectiveness, but could add some evidence of the effectiveness of RA.

The study findings confirm that the integration of RA for pre-referral treatment of severe malaria into the routine IMCI strategy was feasible and acceptable; acceptability was 79.4% among mothers, 90.3% among CHWs, and 97.8% among nurses. The high acceptability rate of this innovative intervention by the community is very important from a public health point of view, as the National Malaria Control Programme is expecting to deploy the RA for pre-referral treatment of severe malaria on a large scale.

The acceptability rate is similar to those observed in other studies [6‐9]. This may be explained by the fact that although pre-referral RA was new in the selected communities, the use of a rectal route for the treatment of diseases was usual in these settings. This mode of treatment was also reported in other studies [7, 8]. The communities considered treatment via the rectal route to be more effective and easy to administer. These findings corroborate evidence documented elsewhere [8]. In contrast, in Laos, the use of rectal treatments was uncommon and was not considered to be more effective than treatment via other routes, even among health care workers [10].

Another explanation for the high acceptance of RA was that malaria was perceived by the community as a public health emergency and was considered to be responsible for most deaths among children. Moreover, confidence among the community in the health services provided by CHWs was observed.

Results showed that adherence to referral advice after integration of the pre-referral rectal artesunate was not more than 15% inferior compared to the initial adherence. This result strongly advocates recommendations to scale up the strategy throughout the country, as the findings stressed that introduction of the pre-referral rectal artesunate, which is a life-saving strategy, did not negatively affect adherence to referrals. In fact, failure to adhere to referral advice might delay identification of the real cause of the disease, while rapid recovery may lead to caretakers to perceive that no further treatment is necessary [8]. The high adherence to referral advice in targeted communities may be explained by the fact that these communities had longstanding PMI-funded malaria interventions, including community outreach aimed at raising community awareness of the importance of prompt care-seeking behaviours and promoting adherence to referral advice.

At the time the study was conducted, information on adherence to referral advice after receiving RA for the pre-referral treatment of severe malaria at the community level or on factors influencing adherence after pre-referral treatment was limited. Most of studies have assessed adherence to referrals from primary health facilities to hospitals, and it ranged from 28 to 75% [15‐19].

Another study in Afghanistan assessed compliance with referrals for children with serious illness, including acute respiratory infections (ARI), diarrhoea or fever, as provided by any health workers at the community or at the health facility level (primary or secondary) and found that 76% of caregivers complied with the referral advice [14].

One factor that significantly influenced adherence to referral advice was the level of education of the mother/caretakers. Al Fadil found similar results in Sudan [16]. In addition, although the caregiver’s marital status and the child’s symptoms on arrival at CCS seemed to be associated with adherence to referral, the difference was not significant. These findings do not corroborate those of other studies [9, 11, 14‐19]. In addition, for a referral to be completed, the parents of a sick child should seek treatment at the CCS or at health facility. This seems not to be an issue in the targeted health zones, as most respondents (more than 7 in 10) stated that they initially consulted a CHW or a health care provider when they noticed that their child presented a danger sign of malaria. However, as some of them acknowledged that they performed an inappropriate action, it is crucial to reinforce health education to improve care-seeking behaviour for sick children.

Considering that one of the key challenges in delivering health care services in the DRC is geographical accessibility, especially in remote rural areas where some villages are located more than 5 km (almost more than 1-h walking) from the nearest health care facility, the CHWs trained in IMCI offer a good alternative that extends primary health care to communities by maintaining linkages with formal health services.

CHWs are nonmedical volunteers who live in the catchment area they serve and can provide early treatment for diarrhea, pneumonia and malaria at any time of the day. However, this advantage is not always the case in other sub-Saharan African settings, as reported earlier in Malawi, where CHWs are government employees and live outside the catchment area [20]. Overall, in the DRC, the work of CHWs is well appreciated in the selected communities, and interventions such as pre-referral RA that are aimed at reinforcing the effectiveness of their job by decreasing the high mortality rate due to malaria in CU5 are likely to be accepted by both the community and the CHWs themselves.

However, failure for them to recognize danger signs of malaria as per the standardized IMCI case-definition would prevent expected health outcomes of the pre-referral RA strategy. That is, why Newbrander emphasized achieving successful child survival through recognition by health care providers of general danger signs of diseases that need referral [14]. The results showed that less than half CHW had the capacity to correctly identify the five danger signs suggestive of severe malaria as per the standardized IMCI case—definitions appropriate for the community level. With such a low proportion of CHWs who have the capacity to correctly identify all five danger signs of malaria, they may have a tendency to overrefer [14], or in a preventive point of view, they may overuse the pre-referral RA leading to drug resistance.

The risk of misclassification of suspected cases of severe malaria was more pronounced, with the results showing that one reason mothers/caretakers did not adhere to the referral advice is that the child recovered after the RA, which is extremely unlikely if the child had severe malaria and received just one single dose of RA. Finally, although the success of this strategy depends greatly on the adherence of mothers/caretakers to referral advice, the current findings also highlight the need to reinforce CHWs’ capabilities to distinguish symptoms suggestive of severe malaria from all other serious conditions that a CU5 may present to ensure prompt pre-referral treatment and to prevent overuse of the pre-referral RA [9].

Adherence was assessed based on the self-reporting of the mothers/caretakers. Although the effective arrival time of the CU5 at the referral health facility was not checked, the CHWs systematically conducted follow-up interviews with the mothers/caretakers of all children who were given RA and referred. These visits enabled assessment of the child’s status and the collection of self-reported information regarding the referral facilities to which the child was taken. Other studies should more thoroughly explore adherence rates before and after the administration of pre-referral treatment at the community level, for instance, by providing a referral note that will be stamped at the facility on arrival and returned to the mothers/caretakers or by reviewing hospitalization registries at the health facility level.

Integration of the pre-referral strategy for severe malaria into CCSs in the DRC is feasible and acceptable, and it positively affects adherence to referral advice. However, more health education is needed for parents of CU5 to raise awareness of the danger signs of malaria and reinforce prompt care-seeking behaviours. In addition, regular provision of RA and more training is needed for CHWs and nurses to increase their performance in the management of severe malaria at the community level.