How much more evidence on community case management of pneumonia in sub-Saharan Africa?

By Joan Kalyango

Pneumonia is the leading cause of death among children under five years old worldwide, despite being preventable and treatable, mostly due to poor access to care. Sub-Saharan Africa (SSA) contributes more than half of the total deaths due to pneumonia. Community case management of pneumonia (CCMp) has been recommended to improve access to effective treatment. CCMp uses informally trained persons referred to as ‘community health workers (CHWs)’ to treat simple pneumonia cases. In settings where malaria and diarrhea are common killers of children, integration of CCMp with community case management of these illnesses, is recommended.

When well implemented, CCMp reduces child deaths in settings with insufficient health care. However, most of the evidence regarding CCMp has come from studies done in Asia. Could these findings be replicated in SSA? Although both settings are resource constrained, they differ in characteristics such as: HIV prevalence, which may affect the distribution and characteristics of pneumonia; the burden of malaria, a disease whose symptoms closely resemble those of pneumonia and would therefore complicate the diagnosis and treatment. It is important to know the effects of CCMp on illness and deaths in SSA, and whether its implementation is sufficient.

A recently published paper by Druetz et al. reviewed the current state of knowledge regarding the effects and implementation of a CCMp intervention in SSA. The areas addressed by the literature included: performance of CHWs, effects of the CCMp intervention on treatment failures, health seeking practices, as well as child deaths.

Performance of CHWs

CHWs experience difficulty in assessment of fast breathing and chest-in-drawing and yet these are key symptoms in the classification of pneumonia. The CHWs also have difficulty indicating the severity of pneumonia. Wrong classification of children is mostly due to errors in assessment of symptoms. In addition, many children receive the correct treatment from CHWs, except when there is need to distinguish between severe and non-severe pneumonia. Some dosing errors have also been noted.

Appropriate referral of severely sick children to health facilities is a key aspect of CCMp that should minimize deaths from severe disease. The findings from the paper by Druetz et al. show that when CHWs can classify disease as severe they appropriately refer children but errors in classifying severity of the pneumonia hinder appropriate referral.

What can we discern from these results on the performance of CHWs?

The CHWs experience challenges in the assessment of the child’s illness and these challenges negatively affect their ability to classify the illness, give appropriate treatment, and refer children when necessary.

Should CCMp be abandoned because of these challenges?

Effects of CCMp on child deaths and treatment failure

Although a reduction in child deaths has been attributed to CCMp in SSA, this reduction has not been significant. However, children with pneumonia experienced fewer treatment failures in the group with CCMp compared to the group of children referred for care for pneumonia. Integrated care also increased the promptness of appropriate treatment.

If pneumonia is among the leading causes of child death in these settings, and the main reason for child deaths is poor access to care, why does an intervention that improves access to care show no impact? It is estimated that CCMp could lead to a 70% reduction in pneumonia-related deaths.

Is the lack of impact on child deaths due to the level of implementation or the way these interventions have been implemented?

Is it due to low incidence of pneumonia or pneumonia related deaths in the areas where the intervention has been implemented?

Is the quality of care provided by the CHWs adequate to give improved treatment outcomes and impact on mortality?

Is the method of diagnosis of pneumonia used by CHWs accurate enough so as not to misclassify children into whether they have pneumonia or not?

In order to conclude on whether CCMp does not really impact on mortality, we need to answer several questions regarding the intervention, the settings where it has been implemented, and how it has been implemented. Some light has already been shed on the question of the quality of care provided by CHWs.

Are the errors made by CHWs in assessment and treatment of children of significant magnitude so as to negatively impact on outcomes?

As the authors have highlighted, more evidence on CCMp that was not included in the review is becoming available. But no significant impact on mortality has been shown yet. And as we evaluate the impact of CCMp on outcomes, we should do so in light of the trends in vaccination coverage of measles vaccine, pertussis vaccine, haemophilus influenza type b vaccine, and the more newly introduced pneumococcal vaccine, which could change the disease burden and characteristics. We should also consider the resistance statistics for the drugs used in CCMp in various settings.

Effects of CCMp on health-care seeking

CCMp increases care-seeking for pneumonia symptoms. A considerable proportion of children utilize CHWs with high levels of satisfaction with the care received. Some barriers to care-seeking have been noted, including drug stock-outs, unavailability of CHWs, and caregiver perception that the illness is very severe.

Is the level of utilization of these CCMp services of sufficient magnitude to impact on outcomes at the population level?

If not, how then can the utilization be increased?

Some of the barriers can be addressed. Improving the drug supply management systems can alleviate some of the supply-related impediments to care-seeking. Most of the data is arising from settings of controlled trials where the situation is likely to be better than in the real settings where these CHWs will eventually perform their roles. Drug stock-outs are likely to be a major concern with CCMp programs, just as with formal health systems. In addition, unavailability of CHWs should be tackled as strategies to improve CHW-motivation are devised.

Factors influencing performance of CHWs

There is a need for more data on which factors influence performance. Different CCM programs have used CHWs with varied characteristics, for example, some have used highly educated CHWs versus low level of education, females versus both females and males, among others. It is not yet clear what level of CHWs should be used. Most studies have dealt with small numbers of CHWs, which makes precise assessment of the influencing factors difficult. Surprisingly, supervision and refresher training have not been consistently seen to improve performance of CHWs. In one study, refresher training was found to improve care for severe disease but worsen care for non-severe disease. These findings are paradoxical and the question is,

Why are supervision and refresher training not consistently effective? Specifically, why would training have a negative impact on care?

What characteristics of CHWs are associated with better performance? What factors influence the performance of CHWs?

Antibiotic misuse in CCMp programs

Three sources of antibiotic misuse in CCMp programs have been identified including: over-treatment or under-treatment because of prescribing errors, over- or under-treatment because of illness classification errors, and misuse as a result of misclassification arising from inadequate accuracy of diagnosis method. We still need evidence on whether these errors will result in poor treatment outcomes, or whether they increase resistance.


There is still so much information required to facilitate the effective implementation of CCMp. However, more and more evidence is continually becoming available in SSA and elsewhere. Should we wait to get information on various aspects of CCM specific to SSA while children continue to die? Or can we generalize information from elsewhere about some aspects of CCM so as to get the implementation rolling?

The question I ask you is: what findings can we generalize to SSA and what information specific to SSA is urgently required?’

Joan Kalyango is Associate Professor of Clinical Epidemiology and Pharmacy at Makerere University College of Health Sciences.


No comments yet.

Leave a comment