Addressing malnutrition among children in routine care: how is the Integrated Management of Childhood Illnesses strategy implemented at health centre level in Burundi?

Under-five mortality remains a huge problem for sub-Saharan African countries, with an average under-five mortality rate of 79 deaths per 1000 live births in 2016 [1]. Preventable infectious diseases such as pneumonia, diarrhea and malaria remain leading causes of death among children under five years and account for one third of all under-five deaths [1]. Many of these deaths occur in children whose immune systems are already weakened by malnutrition—globally, nearly half of all deaths among children under five are attributable to malnutrition [1–3].

Individual health interventions shown to be effective in reducing child mortality include exclusive breastfeeding, improved vaccination coverage, oral rehydration therapy, treatment of pneumonia and early treatment of malaria in endemic areas [4–7]. However, most children at first-level health facilities have multiple and overlapping morbidities, which can make diagnosis and treatment difficult for the health workers (HWs). It gradually became clear to health care planners in the mid-1990s, that an approach integrating management of the most important health problems in childhood was needed to achieve better health outcomes. The World Health Organization (WHO), in collaboration with the United Nations Children’s Emergency Fund (UNICEF) and other agencies, developed a strategy known as the Integrated Management of Childhood Illnesses (IMCI).

The IMCI strategy aims at substantially reducing deaths due to diarrhea, pneumonia, malaria, measles, and malnutrition among children under five years old [8, 9]. It includes three main components [6, 1) improvement in the case management skills of HWs through provision of locally adapted guidelines on IMCI and activities to promote their use; (2) improvement in the overall health care system required for effective management of childhood illnesses and (3) improvement in family and community health care practices. The main strength of the strategy is that it guides HWs through a systematic evaluation of the conditions which lead to the greatest morbidity and mortality in children, with economically affordable diagnostic tools and medicines and with the active involvement of parents in the care and follow-up of their children.

The IMCI guidelines define steps for assessing, classifying, treating, and referring sick children; counseling the child’s caregiver and vaccinating children. A central tenet of IMCI is to organize the diagnosis/classifications of sick children on signs and symptoms; this allows implementation even by staff with limited medical skills.

Since its introduction, studies have demonstrated that IMCI could effectively improve quality of care for children and reduce child mortality when the three components were introduced and implemented concurrently and adequately [10]. A recent Cochrane systematic review on IMCI [11] concluded that the use of the IMCI strategy may lead to fewer deaths among children under five, but may have little or no effect on nutritional status when the community component is inadequately implemented.

Whether or not the malnutrition component of IMCI guidelines is implemented by HWs in the curative consultations of children under five is the focus of this study.

To our knowledge, there have been few studies describing in detail HWs’ assessments of children with regards to their nutritional status in the context of a weakened health care system and high prevalence of chronic malnutrition. This article presents HW performance during routine clinical practice in rural health centres in Burundi .

The structure of this article is as follows. First, a contextualization of the study with background information on IMCI and the health system in Burundi. Next, an overview of the analytical tool and data. The findings of the study are presented and discussed in the next two sections. The article concludes with recommendations for policymakers and researchers.

Burundi is one of the poorest countries in the World, with a GDP per capita estimated in 2016 at PPP $ 777 [12]. The civil war during the 1990s significantly affected the health system, especially with regards to human resources. Medical doctors are a rare resource in Burundi (in 2010, the ratio was 1 doctor per 19,231 inhabitants) [13] – under such constraints, IMCI is a particularly relevant strategy.

Burundi adopted the two first components of the IMCI strategy in 2003. At that time, Burundi had an infant mortality rate of 129 per thousand live births, one of the highest in sub-Saharan Africa. Severe acute malnutrition among children under five years varied by region and ranged between 1.1 and 4.1% [14]. The third component (improvement in family and community health care practices) was adopted in 2006.

Under IMCI, it was expected that all children who presented to health facilities would be routinely assessed for nutritional status and treated when identified as malnourished.

So far, practices of HWs in routine care with regards to malnutrition in Burundi has not been studied in a systematic way. It should be a priority as Burundi is facing continuing high rates of infant mortality and high caseloads of acute malnutrition. In 2017, the infant mortality rate was still at 47 per thousand live births and the prevalence of severe and moderate acute malnutrition was respectively at 1 and 4% [15].

This study builds on an impact evaluation of the Performance Based Financing (PBF) scheme applied to malnutrition in Burundi’s health centres. A full presentation of the research has been published elsewhere [16]. The research protocol was approved by ethical committees in Burundi and at the University of Antwerp, as well as the Institutional Review Board of the Institute of Tropical Medicine, Belgium. For this paper, we use data coming from the health facility baseline survey, which was conducted in September 2014 in 90 health centres selected randomly among 193 health centres that offer services to treat both moderate acute malnutrition (MAM) and severe acute malnutrition (SAM).

A total of 514 consultations were observed.² These consultations were carried out by 145 health workers. As stated above, our results focus mainly on the observed history and examination.

The IMCI guidelines were developed with the knowledge that many primary health centres in low-income countries lack doctors. It is hoped that thanks to IMCI, nurses, clinical officers, community health workers, etc., who assess and manage sick children perform a sufficiently good level of care to generate real health benefits for the children. Our results cast doubt on such an outcome in Burundi: overall, our study found a consistently poor level of adherence on IMCI guidelines by HWs, which is likely to affect the quality of care for their patients. This low compliance is consistent with evaluations of IMCI case management practices in several other settings [20–23] and other studies in the field of quality of care for sick children [24]. This finding indicates a missed opportunity to benefit Burundi’s children given that some studies have reported that IMCI is associated with improvements in the quality of health care [25–27].

The IMCI guidelines include a malnutrition component. The importance of this being that it is within these curative consultations that screening can be carried out for malnutrition to aid early detection and treatment and indeed prevention if possible.

A nutritional history is essential in all paediatric consultations but is notoriously neglected by HWs in many primary health care settings in low-income countries [28, 29]. IMCI guidelines require that all children under two years get a feeding assessment; this especially as breastfeeding has already been shown to reduce mortality in infants and young children [4, 30–32]. Our study found that only a small proportion of children were asked about their diet and only few HWs consistently asked about (maximum) four out of the seven questions relating to malnutrition. Anthropometric measurements are also essential. However most of the children in our study missed the opportunity of accurate nutritional assessment and detection of malnutrition. Weight was systematically recorded by about 38% of HWs, but anthropometric measurements such as weight alone or height/length alone are themselves not useful unless they are interpreted according to resp. height or age.

Our exploratory analyses also established a link between lack of weight measurement and supervision. This confirms again how IMCI implementation rests on supervision [33]. This strongly suggests the need to train the district teams on IMCI, focusing on improving IMCI guidelines knowledge and supervision skills to implement plans and monitor progress [34].

There were worryingly few children who had oedema assessed (13%), a factor which can completely change the diagnosis of a child from a normal z-score or MUAC or MAM to SAM, which is a diagnosis requiring prompt and comprehensive treatment, especially in light of the fact that SAM children with oedema have higher risks of mortality [35, 36]. The poorest performance was observed with regards to the number of HWs who discussed the growth curve.

Child health interventions including breastfeeding, nutrition counseling about complementary feeding and growth monitoring have been shown to prevent malnutrition and substantially reduce child mortality [4, 37]. Although exclusive and continued breastfeeding rates are high in Burundi, nutrition advice was frequently not given and growth monitoring not performed.

Overall, the results suggest that female health workers perform better than males on every task. We believe that this can be glued to the fact that in our culture, it is the women who care most about the children needs at home (breastfeeding, feeding them, getting them to vaccination, or in curative consultation in case of illness). It is quite possible that these mothers who are used to care children needs better investigate the problems of a child who is brought to curative consultation.

We also found that HWs who have a (permanent) contract with the government perform better, suggesting that the stability these contracts confer does not entail their motivation, on the contrary.

Overall, required tasks are not systematically performed (if ever), and most of the time consultations are limited to the presenting complaint; this could lead to incorrect diagnoses of the children’s condition and to potentially miss a serious and treatable disease. Such a failure in a context of high prevalence of malnutrition raises a lot of questions. One can imagine that the non-systematization of questions and physical examinations is linked to the fact that the tasks of the IMCI guidelines are numerous. Even these minimal tasks are not properly completed.

Our bivariate analysis suggests that being examined by the HW in charge of the activity of management of malnutrition increases the chance for the child to be weighed, while lack of training on monitoring growth and screening and management of malnutrition plays negatively on the weight-for-height/length z-score calculation.

These results suggest that a poorly implemented split of tasks is not conducive to malnutrition screening. Some rotation of staff within the different departments of the HC would probably increase staff versatility and hopefully, lead to a higher attention to malnutrition in routine consultation. While the low performance of HWs is frequently attributed to a lack of knowledge and skills, it appears that the diploma of HWs do not play a role in malnutrition screening. This raises questions about the initial training program of HWs with regards to malnutrition, but also about the strategies undertaken to improve HW performance [38]. Our findings also raise questions on the commitment of the HW to implement the malnutrition IMCI guidelines component. A possible cause could be that IMCI consultations take longer compared to routine consultations. Yet other results (not reported here) [39], obtained from administering vignettes to the same HWs at the end of the day’s activities (ideal conditions without pressure from the patient’s queue), suggest that the main problem is at the level of IMCI knowledge held by staff.

We believe that there is a need for complementary qualitative studies to investigate the possible factors associated to this low performance. One question of interest should be how HWs perceive the IMCI guidelines including the various tasks.

The recent Cochrane systematic review evaluated the effects IMCI strategy implementation in terms of death, nutritional status and quality of care [11]. However, the certainty of the evidence has been assessed as too low to allow an affirmative conclusion on whether or not IMCI has an effect on the way of HWs treat common illnesses. Our results are consistent with those of two studies focused on febrile children in Ghana [40] and Zambia [22]. In these two middle-income countries, the studies by Baiden et al. and Lunze et al. respectively also highlighted that adherence to IMCI case management guidelines were rather poor and below expectations. Our study complements their works by showing that the problem also arises for malnutrition, which is very rarely the presenting complaint in curative consultation.

Our study shows that the systematic integration of malnutrition screening into the ambulatory care package is problematic. In Burundi, management of acute malnutrition has for a long time been considered as the ‘turf’ of international NGOs. It could be interesting to explore whether other countries with a weak health system and a high prevalence of malnutrition present a similar pattern of limited integration.

Our study has limitations. First, the mere fact of being research subjects and the attention from the surveyors may have temporarily altered the behavior of the observed HWs. This could have changed their usual practices in curative consultation by trying to prove that they are doing their job properly.

If this effect has played any role, our results actually underestimate the depth of the problem.

Second, we were also unable to determine the adherence to IMCI guidelines between trained and untrained HWs on IMCI guidelines given that the period since the last training varies from one HC to another and according to staff changes. That said, some studies have reported a positive impact of IMCI training on quality of care [21, 41], however others found no effect [42, 43].

This study has found poor compliance by HWs to IMCI guidelines in health centers of Burundi. This leads to a substantial proportion of children not receiving early and appropriate care, especially malnutrition care. The incomplete implementation of IMCI prevents Burundi from seizing the maximum benefits from the strategy in terms of child survival.

Our findings call for vigorous action, by the Ministry of Health of Burundi and all its partners committed to child health and involved in the IMCI strategy. Obviously, what has been done until now is not satisfactory. We believe that the country needs a multi-faceted approach to ‘save’ the IMCI strategy. Training has to be improved, but more consistent and more meaningful supervision is the key element needed to address this non-compliance.

Our study also rings an alarm for the international and national actors involved in the ‘free health care + performance-based financing‘strategy in Burundi [44]. Through the IMCI metric, our study captures a central component of quality of care at health centre level. The observed low compliance to IMCI shows that a serious general problem with quality of care persists, despite attention put upon quality through PBF quality checklists. Possible actions include a revision of how quality of care is measured under the national PBF program (see for example, [45]) and the embedment of PBF into a more comprehensive approach to the determinants of quality of care.

But the implications of our study go also beyond Burundi. As shown in our discussion, our findings are consistent with observations in other countries. There seems to be a serious problem with the implementation of the IMCI strategy in many low-income countries. This is very worrying as the strategy, designed for countries with high child mortality and low availability of medical staff, is the cornerstone of pediatric care for hundreds of millions of children.