Ambulatory treatment of severe malnutrition

Severely malnourished child with father

Commentary by Dr. Steve Collins

Dr. Steve Collins is a medical doctor with a doctorate in nutrition during emergency operations. He is a director of Valid International, a company which aims to improve the quality and accountability of humanitarian assistance. Since 2000 he has been directing the CTC programme, a multicountry, multi-agency programme to research and develop Communitybased Therapeutic Care.

This article describes the experiences of an MSF-B nutritional project in Faryab that was forced by cultural and geographical constraints to treat children with severe acute malnutrition as outpatients. Given the constraints and difficulties inherent in the context, the results are encouraging and many of the lessons that MSF-B have learnt are useful.

The most important of these lessons is that community-based mobilisation and profound outreach are essential in order to maximise programme impact. This is also one of the main lessons that we have learnt during the first three years of the Community-based Therapeutic Care (CTC) programme. Implicit in this lesson is the necessity to move away from 'Ambulatory/Home Treatment', towards a model of 'Community-based Care'. Ambulatory/Home Treatment is a hang-over from the therapeutic feeding centre (TFC) approach to severe malnutrition, emphasising the medicalised treatment of a severely malnourished individual, rather than the management of malnutrition as a complex socio-economic condition. Malnutrition is not a disease you catch that requires mere medical/nutritional treatment, rather it is the end-point of a complex interaction between social, cultural, economic and physiological processes. Unless project design acknowledges this multi-factorial aetiology, impact will continue to be limited by high default, poor coverage and a lack of compliance.

From the data provided, it is not possible to draw firm conclusions about the MSF-B Faryab programme. Coverage data is essential in order to examine impact. In addition, more data on those who died and defaulted (age, concurrent medical diagnoses, timing of death and attendance record) would be useful. However, the results presented appear to show that the quality of medical and nutritional treatment of those that attended was not a major factor limiting impact. Although no 'phase one' inpatient care was provided, the death rate amongst those who attended the programme was 6%¹. This is well within the Sphere standards and similar to MSF standards for a well run TFC². I would guess that this mortality rate is better than those in many of the TFCs currently operating in Afghanistan³. The rate of weight gain of 6.1 g/kg/day and mean length of stay of 57 days are also reasonable given the limited opportunity costs and risks entailed by being in the programme.

From the report, it appears that the main factor limiting impact is the high default rate. Overall, 48.2% of those investigated defaulted because the opportunity costs of programme attendance were perceived as greater than the benefits derived from that attendance. Another 34% defaulted because they or their carer were ill; either unable to access the programme or under the impression that the programme would not treat them adequately. Thus programme design and implementation impacted on at least 82% of defaulters. In particular, the physical access to distribution points and waiting times at distribution points, understanding of the programme, and the quality of service during consultations would have been important factors⁴.

It is a shame that MSF-B do not report programme coverage. Given the reported "difficulties for mothers in reaching the centres" and the perception of high opportunity costs amongst those attending the programme, I would guess that the programme coverage was low. Low programme coverage is usually the most important factor limiting selective feeding programme impact.

I question the validity of some of the lesson that MSF-B have learnt. The lesson that all children with severe malnutrition should be admitted to residential care appears at odds with the results presented. The programme's mortality rate is already close to the MSF norms for residential TFCs, and I find it unlikely that the costs and effort of constructing residential "phase 1" care for all patients would be repaid in impact. MSFB's first lesson includes an aspiration to have a doctor for all the weekly follow-up consultations. Given that the mortality rate amongst those attending is already low and the rates of recovery are reasonable, it is unlikely that the large human resource and cost implications of having a doctor at each distribution would achieve a commensurate increase in impact. Indeed, by diverting resources away from achieving access, coverage and mobilisation, these two "lessons" are likely to decrease programme impact.

The main limitations to impact appear to be the high default rate and the problems of access to the distribution centres (presumably coexistent with low programme coverage). Efforts to resolve these problems, rather than aspiring towards a more medicalised 'TFC approach' would be likely to be more effective. Increasing the number of distribution sites would improve access and reduce the opportunity costs of attendance. These sites could be temporary, where possible utilising existing health / social structures and mobile distribution teams. This would improve coverage, reduce default and facilitate transition towards longer-term programming. Over the past three years, experiences with Community-based Therapeutic Care in Ethiopia, Sudan and Malawi have demonstrated that with improved access malnourished cases present earlier. The clinical course of malnutrition is one of a gradual increase in severity and concurrent medical complications, therefore those who present earlier are easier to treat and the results are better.

MSF-B's other lessons (more outreach, community mobilisation, improved education, more systematic protocols and greater integration with other structures) are important steps forward. To improve sustainability and cost-effectiveness, community mobilisation can use positive role models within communities by harnessing existing capacity, whether this be traditional practitioners, traditional birth attendants (TBAs), 'wise women' etc, in addition to the more usual paid outreach workers. These are now being incorporated into CTC project design.

MSF-B's experiences in Faryab highlight important issues surrounding the management of severely malnourished children in the community. Maximising impact in these programmes requires the design of a Community-based Therapeutic Care programme specifically to address issues relevant to the care of the malnourished in their homes and villages, not ad hoc adjustments to the basic TFC model. Sending people home with RUTF, medical treatment and didactic education will not maximise the impact of these programmes. Maximising the impact of CTC requires radical changes to the prioritisation, human resources, protocols, monitoring and data collection and logistics of selective feeding protocols.

¹Even when you add an estimate for the number of deaths amongst those who defaulted to this (114 * 17.3% = 20 deaths), the death rate is still 10%. Note that the SPHERE standard does not include the number of defaulters from TFCs that die.

²It is similar to the rate in Concern's pilot outpatient treatment programmes (OTPs) in Ethiopia in 2000 (see research summary, Community based therapeutic care, in this issue of Field Exchange) and Oxfam's experience in Bolosso Sorie, Ethiopia where they conducted one of the first OTPs.

³As there are currently 156 children in the programme and mortality rates tend to be highest shortly after admission (I would be very interested in the timing of deaths), this rate is likely to decrease further by the end of the programme.

⁴Although information on food security is not included, it is likely that this programme operated during both the planting and the harvest periods in Faryab and this would be an important factor influencing default.