In 2009, the Gombe state Ministry of Health (MoH) with technical assistance from Valid International and UNICEF initiated Community-based Management of Acute Malnutrition (CMAM) programming in three local government areas (LGA): Gombe, Dukku and Nafada.

Although admission numbers have been impressive, from the outset health facilities have also consistently reported high levels of defaulting, sometimes as much as 50% of the total programme exits. The serious, early and widespread nature of defaulting across the CMAM programmes in the state was confirmed by a Semi-Quantitative Evaluation of Access and Coverage (SQUEAC) investigation in March 20101. The defaulter problem was further elucidated by a programme review conducted by Mandalazi2,3 and a review of health facility data in July 20124.

As a standard indicator of CMAM programme performance, default rate ranks alongside mortality and cure rates as one of the principal measures of programme effectiveness. A child is classed as a defaulter after three (3) consecutive weekly absences from treatment follow-up. The Sphere5 standard for default rate is less than 15% of all exits. High default rates also compromise the ability of a programme to respond to need and to achieve good coverage and hence a significant public health impact.

In view of the fact that defaulting has continued to be problematic and that the issues which give rise to it continue to persist, an in-depth investigation to better understand the reasons for and consequences of defaulting was undertaken using a semi-quantitative approach based on the investigative tools used in SQUEAC6. This report details the findings of this investigation carried out in Gombe LGA and Dukku LGA7.

Objectives

The overall objective was to investigate defaulting from treatment by children admitted in the Gombe CMAM programme.

More specifically, the investigation aimed to:

  • Understand the full range of factors (demand side, supply side; behavioural, institutional; cultural, material) that prohibit sustained compliance with treatment

  • Identify key reasons for defaulting, their relative importance and the relationships between different factors

  • Establish factors that promote compliance and those that lead to defaulting

  • Identify measures or procedures that would allow defaulting children to be accurately traced back to their homes

  • Provide recommendations based on findings to reduce defaulting

Methodology

The investigator applied SQUEAC as the main investigative framework. As such, the approach was semi-quantitative in nature and utilised existing and available datasets and information on defaulting from which to build upon further investigation. Specifically, the investigator used the following available datasets / information:

  • SQUEAC investigation 20101

  • Community mobilisation review 20122

  • Defaulter tracing 20123

  • Health facility data review 20124

  • Routine programme data

With these as starting points, the investigator used a variety of SQUEAC tools6 supplemented by ethnographic techniques for an in-depth investigation of the defaulter phenomenon.

The investigation consisted of three interlinked and dynamic elements:

  1. Collection and analysis of quantitative data
  2. Defaulter tracing and iterative qualitative data collection
  3. Visits to OTP catchment areas with particularly high and low defaulting

Routine data on programme admissions and exits in 2012 was requested from LGA nutrition officers and the state nutritionist to explore defaulting trends over the year. Using data extracted from outpatient therapeutic programme (OTP) beneficiary cards, a database of all children admitted to the CMAM programme who had defaulted since August 2012 up until the first week of November was made. This timeframe was chosen because of the new discharge criteria, which OTPs started using in August to address the absence of a supplementary feeding programme (SFP) in Gombe state. The new discharge criteria of MUAC > 125mm for two visits and steady weight gain replaced the standard criteria of a minimum 8 weeks stay and MUAC > 115 and steady weight gain that had been in use since the start of the CMAM programme. This data was mapped and analysed to identify spatial and temporal patterns in defaulting. Specific analyses focused on time to default, time to travel and the mid-upper arm circumference (MUAC) on admission and on the last visit. In addition, a database of all children discharged cured and meeting discharge criteria during the same period was created and a comparative analysis carried out on the MUAC on admission.

Convenience sampling was used for the collection of qualitative data and the selection of key informants such as carers of defaulters, community leaders, herbalists, OTP staff and community members. Villages were selected based on their size, accessibility and the number of defaulters. At the beginning, the investigation focused on areas with the highest numbers of defaulters because it was anticipated that, given difficulties in tracing children, there would be a greater chance of finding at least a few defaulters. Later into the investigation villages with few or even one defaulter were also visited, as long as they were accessible by car. Due to incomplete address data it proved to be difficult to trace children in towns, so unless the investigator had an address or knew that volunteers in that area were active, smaller villages were preferred for defaulter tracing in order to make best use of time. In the end a mix of large and small villages was visited, with some on the main road and others with more difficult access. The number of defaulters in visited villages varied from low (one or two defaulters) to high (ten and more defaulters) and all areas with high number of defaulters were visited.

A combination of case histories, interviews and focus group discussions were undertaken. The information was triangulated by source and method as much as possible before it was considered reliable. Case histories of defaulters were taken to better understand the reasons for defaulting. Group discussions were primarily carried out with groups of men and women separately from the community and carers of beneficiaries. These focused primarily on understanding possible reasons for defaulting, exploring perceptions of malnutrition and treatment seeking behaviour, and on examining understanding and opinions of the CMAM programme. Semi-structured interviews were conducted with carers of defaulters and beneficiaries, service providers, community volunteers, community leaders and traditional healers.

Findings and Discussion

Quantitative data analysis

During analysis of the OTP cards from both Gombe and Dukku LGAs it was noted that many defaulters had been wrongly classified as discharged cured; after missing three visits at the OTP some of those with higher MUACs were marked as discharged cured rather than defaulted. This means that the defaulter rate (proportion of defaulters compared to total exits from the programme) is actually higher than that reported. For the purpose of this analysis only data from OTP cards of children officially classified as defaulters were included and only these children were traced.

It should be emphasised that information on the OTP cards was incomplete which made defaulter tracing particularly difficult (see Conclusions). Gombe Abba OTP (Dukku LGA) stood out for particularly poorly filled in OTP cards – and as no outcome data had been recorded, no children from this site could be entered in the defaulter database. Similarly no data was included for Nassarawo OTP (Gombe LGA) as the programme did not run between August and November due to construction work.

Although all OTPs had adopted the new discharge criteria (MUAC > 125 mm for two visits and steady weight gain) in August, it appeared that very few discharged cured children actually met the discharge criteria. It was also noted that in many OTPs children were being discharged with very low MUAC, on a few occasions with MUAC as low as 100 mm.

Defaulting over time

Routine data on admissions at Gombe state level was incomplete and of questionable reliability as two very different versions were provided for 2012. Data was obtained at LGA level for Gombe only for the period of January to September 2012; admission data was unavailable for Dukku. Defaulting rates in both LGAs are high; however, defaulting patterns in Dukku may be different to those in Gombe. Due to lack of support from the LGA, there are delays in transport of ready-to-use therapeutic food (RUTF) from Gombe to Dukku LGA and distribution from Dukku to OTPs. OTP staff are required to spend their own money for transportation of RUTF and thus stock-outs of RUTF in Dukku are much more frequent. The pattern of defaulting may be linked to the periods when RUTF is out of stock and hence there are no admissions. If the carers of those already in the programme are sent home without any or with insufficient RUTF, they become discouraged and often decide never to come back.

Figure 1: Defaulting rates in Gombe LGA between March and September 2012 compared to the seasonal calendar

In Gombe LGA, the OTPs were out of stock during January and February 2012 and the first admissions were not registered until the 8th and 9th week of the year. The overall average default rate for the period between calendar weeks 10 and 39 was 46%, which is consistent with what was observed during the 2010 SQUEAC in Gombe1 and never dropped below 24% (see Figure 1). The defaulting rate after the first month of resumed admissions in March was very high, reaching 100%. This was because the OTPs just started operating again and there were no discharged cured children at that point.

Comparing defaulting rates to the seasonal calendar it can be seen that the defaulting rates again increase during the rainy season. This can partially be attributed to more limited accessibility for women who have to travel from far away, but also to Ramadan which took place somewhere between weeks 27 and 33. During Ramadan people are less active and less willing to travel and for mothers who had to walk long distances Ramadan increased the likelihood of defaulting. This was also confirmed by the qualitative investigation. The rainy season coincides with the hunger period during which, because people have used their food stocks and won’t harvest the crops in the field for two more months, one would expect to see more malnutrition. September, (which coincides with week 36) when the harvesting starts, women are required to help in the field and are again more likely to miss visits to the OTP.

In general, OTPs with a larger number of admissions also have higher defaulting rates. In Gombe LGA, Pantami and Town Maternity OTP have the highest number of admissions, on average 15 new admissions per week, and also the highest defaulting rates of 41% and 39% respectively. Kumbiya Kumbiya OTP has very few new admissions, less than three per week, and also the lowest defaulting rate at 18%. However, Tudu Wada OTP with an average of 8 new admissions per week stands out with high defaulting rate of 35%. Although admission data was not available for Dukku LGA, Zange OTP stood out for having a very low defaulting rate (one defaulter between August and November) despite a high number of new admissions.

The volume of beneficiaries leads to longer waiting times and has been previously been seen to impact on the quality of service provided at the OTP and leads to purposeful rejection of existing and potential beneficiaries1. This was investigated and discussed with carers of beneficiaries and OTP staff and it still appears to be a big problem in both LGAs (see issues in service provision). On several occasions mothers were refused treatment even during our visit; the persons taking the MUAC or collecting ratios cards would leave early and told them to come back next week. Some mothers reported that they were told their child is not eligible to be in the programme without having their MUAC taken. When a mother who came late was rejected treatment in Town Maternity OTP (Gombe LGA), the health worker explained to us that he referred her to the OTP nearer to her home. He argued that is their standard procedure despite the fact that earlier in the day they several children from very close to other OTPs were admitted.

MUAC on admission: comparison of cured and defaulted children

Figure 2, Figure 3, Figure 4 and Figure 5 provide a comparison between MUAC on admission among defaulters and children discharged cured between August and November 2012. In general, children who were discharged cured were admitted to the OTP with a higher MUAC than the children who defaulted (median 111 mm compared to 103 mm in Gombe and 110 mm compared to 106 mm among defaulters in Dukku).

Low MUAC on admission for defaulters indicates that the children are brought to the OTP when the condition is very severe. Children may already have additional medical complications and are at a higher risk of negative outcomes, potentially resulting in defaulting. These differences in how soon children are brought to the OTP may be related to perceptions of malnutrition and communities’ treatment-seeking behaviour which in turn also impacts on defaulting (see section 4.2).

It should be noted however that the number of children who fulfilled discharge criteria is very small, so the sample size of cured children on admission is much smaller compared to defaulters’ and may not be representative.

Figure 2: MUAC at admission of defaulters from August to November in Gombe LGA

Figure 3: MUAC at admission of cured cases from August to November in Gombe LGA

Figure 4: MUAC at admission of defaulters from August to November in Dukku LGA

Figure 5: MUAC at admission of cured cases from August to November in Dukku LGA

Time-to-default

An analysis of the number of visits to the programme indicated that most children defaulted by their second visit (Figure 6 and Figure 7), with as many as 41% defaulting after the first visit. This tendency remains unchanged since the previous SQUEAC in 2010. For comparison, the children discharged cured stayed in the programme from between four and ten weeks, with most staying no longer than five weeks.

Town maternity OTP has particularly striking number of children who only come once before defaulting and as can be seen in the section on time-to-travel the majority live close to the OTP. Also, all but three OTPs; Kumbiya Kumbiya (Gombe LGA), Zange and Dokoro OTP (Dukku LGA), have more defaulters in the first week.

Defaulting in early stages, combined with the fact that many defaulters present to the OTP with low MUAC, could indicate hidden deaths. Indeed the qualitative investigation confirmed this assumption during defaulter tracing qualitative investigation.

Figure 6: Time-to-default from August to November 2012 in Gombe LGA

Figure 7: Time-to-default from August to November 2012 in Dukku LGA

Time-to-travel

It was commonly assumed by OTP staff in Dukku and Gombe that distance was the major problem and reason for defaulting. Many beneficiaries come from outside their catchment area and were thought to be those who default. A time-to-travel analysis (Figure 8 and Figure 9), however, revealed that the majority of defaulters in Gombe came from the areas closest to the OTP with carers needing less than 30 minutes to reach the OTP. In Dukku many defaulters also came from nearby, particularly in Malala OTP, which is urban. In other OTPs defaulters came from villages further away and the time to travel was increased because of the challenging terrain. This suggests that distance is a bigger factor in Dukku than in Gombe but that it is not the only reason for defaulting (see qualitative investigation).

Figure 8: Time-to-travel for defaulters from August to November 2012 in Gombe LGA

Figure 8: Time-to-travel for defaulters from August to November 2012 in Dukku LGA

MUAC on default

An analysis of the MUAC on the last visit before defaulting revealed that, compared to Dukku, children in Gombe defaulted with a lower MUAC. Given that the median length of stay before default is the same for Dukku and Gombe the difference in the MUAC may be explained by the fact that in Gombe the median MUAC on admission was also lower for defaulters. However, in both LGAs the majority of children who default are still severe acute malnutrition (SAM) cases (median of 106mm in Gombe compared to 110mm in Dukku). Again this underlines the potential mortality risk associated with early defaulting which was confirmed by the defaulter tracing (see qualitative investigation).

Figure 10: MUAC at default from August to November 2012 in Gombe LGA

Figure 11: MUAC at default from August to November 2012 in Dukku LGA

The home locations of defaulters from August to November 2012 were mapped (Figure 12 and Figure 13) and the analysis of the spatial distribution supports the time to travel evidence that the majority of defaulters in Dukku and Gombe actually live close to the urban OTP sites, but further away from OTPs in rural areas (Zange and Dakkoro OTP).

Figure 12: Defaulter mapping in Gombe LGA

The map for Gombe (Figure 12) was created using Google Earth8 due to unavailability of good quality map of Gombe. Mapped are 124 children who defaulted in Gombe LGA between August and November 2012; the size of the circles is proportional to the number of defaulted children (the bigger the circle, the higher the number of defaulters it represents), with different colours representing different OTPs. From around 90 children that couldn’t be mapped about 50 came from outside of Gombe LGA. For about 40 children from Gombe town there is no accurate information on which quarter within Gombe town they live in, which could explain why there is so much empty space in the centre of the map. The map shows however, that the defaulters come primarily from peripheral quarters in Gombe town and that they are not always brought to the nearest OTP. The reason for this could be either because the carers are not aware that there is an OTP closer to them that offers this service or because of (dis)satisfaction with the service provision in some OTPs. Beneficiaries mentioned both reasons during our investigation. Kagarawal, Anguwa Uku and Malay Inna settlements in Gombe (all three lie next to each other and there is no visible border between them) stood out for having a particularly high number of defaulters, representing around 20% of all defaulters in Gombe LGA. These areas are under Tudu Wada OTP catchment. This area in north-eastern part of Gombe metroplis is also characterized by security issues – two army raids were conducted here during our stay in Gombe.

Figure 13: Defaulter mapping in Dukku LGA

On the map of Dukku above, each point represents one defaulted child. The map shows that children who default after first visit don’t necessary come from far away villages – many live relatively close to the OTP. In Dukku, Zange OTP stood out for having a very low number of defaulters (one).

The presence of community volunteers in an area does not seem to be related to low defaulting; volunteers come from areas nearest to the OTPs. What appears to make the difference is whether or not the volunteers are active and carry out follow up. With the exception of Zange OTP (Dukku LGA), however, no other OTPs had volunteers who carry out activities in the community.

The difference between Dukku and Gombe though is in accessibility. In Gombe living close to an OTP meant having easy access because it is an urban area with good road connections. In non-urban areas of Dukku however, the access was often difficult because of the sandy terrain. This explains why the spatial analysis for Dukku differs slightly from the time to travel analysis in that on the map aerial distances between the villages and the OTPs seem relatively small, but in reality it takes a longer time to reach that village compared to reaching an equally distant village by a good road.

Those areas identified with a particularly high or low number of defaulters were visited during the investigation to better understand what factors contributed to the different rates of defaulting (see reasons for defaulting)

Qualitative investigation

It was possible to trace and talk to carers of 14 defaulters in Gombe LGA and 21 defaulters in Dukku LGA.

Defaulter tracing in both LGAs was very slow and challenging for several reasons:

  • Incomplete address information (only name of the village recorded)

  • Lack of community cooperation due to security issues in high defaulter area in Gombe

  • Semi-nomadic Fulani population

  • Communities absent due to harvest period

  • Distance and poor accessibility

Providing the community leaders and active volunteers with lists of defaulters to identify in advance worked to some extent. However, due to distances to travel and protocols that needed to be carried out, it turned out to be very time consuming and given the time constraints and limited manpower (one consultant) it was not always feasible.

Key reasons for defaulting

The most common reason given by carers of defaulters who were interviewed about their experience of the programme and the reasons for defaulting was in fact the beneficiary’s death; 50% of defaulters traced in Gombe and 32% of defaulters in Dukku had died. Although it was not always possible to ascertain the exact date of the demise of the child, given that the median MUAC on admission was low (see MUAC at admission) death may indeed have been the direct cause of default. However, as the median MUAC on default was also found to be low (see MUAC at default) it is also likely that in many cases death was actually the consequence of the default. In either case this clearly underlines the mortality risk identified by the MUAC measurement and associated here with late presentation and early default.

Figure 14: Reasons for defaulting among 35 defaulters traced in Gombe and Dukku LGA

Another major direct reason cited by mothers to stop coming was that the child didn’t eat the RUTF – either because they weren’t able to swallow, didn’t like it or vomited when fed. If the child vomited, the mothers felt that the RUTF gave the child another sickness and decided it wasn’t good for them. This is the result of service failure and OTP staff not communicating key messages to the carers. Children are breastfed for a long time and besides breastfeeding, their main source of nutrition is pulp, which mothers normally make by mixing millet and water. Mothers sometimes feed their children RUTF straight after feeding them pulp, when the child is already full and thus vomits or refuses to eat. Rather than frequent feeds of smaller quantities, mothers try to feed children large quantities in one go, rushing and forcing the child to eat. One of the key messages that staff are supposed to give is that RUTF contains everything a child needs to recover and should be given before any other food – except breast milk. While some OTPs give health education talks on clinic days and tell mothers how to feed their children, others do not. On exit interviews with new beneficiaries it was also found that sometimes mothers are not only not told about the RUTF and feeding practices, but are also not given information on how much RUTF they should give to the child every day. A few times mothers were told to come back after the child finished all the RUTF and some mothers reported not coming to the clinic on the weeks they still had some RUTF left at home. Clear messages on the importance and reasons for a child coming back each week are not being given. In addition, the appetite tests are not being conducted in some OTPs, while in others they are conducted inappropriately – in large groups without proper explanation of what to do in case the child doesn’t swallow. This could also mean that children who should be admitted to the stabilisation centre for in-patient care are not being identified.

Distance, as an aggregate of many reasons, has a big impact on defaulting and is a bigger problem in Dukku than in Gombe, as suggested by quantitative analysis. In Figure 10 distance as a contributory factor is represented by many separate reasons:

  • Enrolled when visiting relatives/migrated: Sometimes mothers would go and visit their parents in other villages. There they would learn of the CMAM programme and enrol the child. They would keep coming for as long as they staying with their relatives, but after the return to their own village they would stop because it was too far to travel.

  • No money for transport and no access to transportation: For some families it was too expensive to pay for the fuel for transportation or, in instances when they had money, there was simply no transportation available. This is especially a big problem during the harvest season where owners of motorcycles go to the farm in the morning. This means that the mothers arrive at the OTPs late or are not able to come at all. If they come late, they will be shouted at by the OTP staff and sometimes even sent back home without any treatment.

  • Mother was ill/gave birth: If the mother was unwell or pregnant, she was less able to walk long distances to bring the child to the OTP.

The attitude and behaviour of certain OTP staff has clearly had a negative effect and has led directly to defaulting in some cases. In many OTPs, if the mothers don’t come by 9 am or in some instances by 10am, they are shouted at and sometimes told to come back next time. In Zange OTP (Dukku LGA) the staff are aware of the challenges and difficulties carers may experience when coming to the OTP and they waited for mothers to come until 2 pm. Mothers also report being shouted at if the child doesn’t achieve a steady weigh gain, if they forget the ration card at home or if they miss one visit. This shouting by the OTP staff creates fear in some mothers and even if they experience other problems (e.g. the child doesn’t eat the RUTF), they are too afraid to return to the OTP.

Several carers of defaulters felt that the child had recovered; had more energy, didn’t cry much and had an appetite, and therefore didn’t need the “food” anymore, so they defaulted. For others, it was slower than expected recovery or vomiting and/or diarrhoea that the child experienced when eating the RUTF that made them not return. Illness perceptions underpin treatment-seeking behaviours and contribute to defaulting. There is a general lack of awareness of malnutrition in the communities. Hunger (chiwoyoma), is not perceived as life-threatening since the child only needs to receive food to recover. The hunger is caused by lack of breast milk, barwa nono or endamre, and/or spoiled milk (nono bakeu, abinchi mekeyeu) due to the mother breastfeeding during pregnancy. Illnesses with aetiologies most closely corresponding to a clinical definition of malnutrition, like dotti, kurga, bayama and tamoa, are believed to be untreatable with western medication and if a child is brought to hospital they will die. In some instances the carers mentioned that while RUTF cannot treat the condition, it can return the child’s strength and energy. These perceptions of malnutrition and common treatment-seeking behaviour could be the reason why mothers default as soon as the child’s condition improves.

Certain differences were apparent between the 2 LGAs in terms of factors cited for defaulting. In Dukku distance and the child’s inability to eat the RUTF were the most common reasons for defaulting. This could be the result of poorly organized appetite tests and no explanation given. Also, health education sessions were held in OTPs in Dukku far less frequently than in OTPs in Gombe.

In Gombe, the main reasons for defaulting were death and perceptions of the child’s illness. A very common reason for defaulting from Tudu Wada OTP (Gombe LGA) was the requirement for carers to bring food (groundnut, soya beans, millet) on their fourth visit to the OTP and threats that if they don’t bring it, they will not be seen. Due to no SFP in Gombe state, carers are asked to bring food for cooking demonstrations where they are taught how to prepare the food and what to feed the child after discharge. As the child’s condition improves and stabilizes, mothers are advised to prepare this food and give it to the child along with RUTF and breast milk, so that by the time the child stops taking the RUTF they are used to the new food. Different OTPs have a different approach to this food – some only talk about it during health education sessions, while others do cooking demonstrations. However, out of all the OTPs only Tudu Wada was reported to have made the practice to bring food mandatory. Many families cannot afford to buy food and making it compulsory results in mothers defaulting from the CMAM program.

What they do at OTPs is after few visits start telling the mothers they should also prepare supplementary food at home and give it to the child so that by the time the child stops taking the RUTF he will be used to the new food.

Wider issues contributing to defaulting

The mind map (Figure 15) shows the full results of the qualitative investigation and details all the factors contributing to defaulting mentioned both by primary (carers of defaulters) and secondary sources (carers of children discharged cured, OTP staff and the community)

On the whole the main issues that cause defaulting can be grouped as follows:

  • Issues with service provision;

  • Lack of community mobilisation; and,

  • Access and associated costs.

Issues in service provision cover a number of inadequacies associated primarily with the quality of care and follow up and the inability to ensure the permanent availability of RUTF at OTP level. The limited explanation of the child’s condition and its treatment by health staff, results in a failure to communicate essential information which would encourage mothers to comply with weekly attendance until discharge. There was no evidence of systematic follow up; once a child had defaulted no efforts were made by OTP staff to trace these children or to enlist volunteers to carry out a home visit to ascertain the reason and to try and encourage a return to the health facility. Once a child has defaulted they are already considered lost to the programme. It also appears that there is no active or regular supervision of OTP practice by MoH staff; this prevents errors from being identified and rectified and staff from being motivated and interested in following the correct procedures and protocols. Frequent RUTF stock outs have also been experienced, particularly in Dukku, which have caused a loss of confidence in the programme by the community.

Despite its central importance to the success of CMAM community mobilisation has not been prioritised. There was little evidence of effective sensitisation or volunteer activity at community level (indeed their role is not clearly defined and communications with OTP staff are poor), or of real engagement with beneficiaries at the OTP. This lack of community mobilisation has resulted in poor understanding of the programme in the community (despite high awareness). CMAM is seen as a feeding rather than a therapeutic programme, even staff refer to the RUTF as a food. A few nurses in charge mentioned that mothers might even register at several clinics at the same time to obtain ‘the food’, but may not be able to continue to attend every OTP which contributes to high defaulting rates, although this claim could be confirmed. Moreover without community mobilisation traditional perceptions of malnutrition and common treatment seeking behaviour which hinder compliance (as already seen in key reasons for defaulting) also continue to persist.

In terms of access, aside from distance, lack of father’s permission was often quoted as a reason for defaulting (particularly by OTP staff and members of the community), but in practice this did not seem to be a direct barrier for many women. Most key informants said that fathers wouldn’t refuse treatment for the child and this was particularly true if the mother had access to her own money and was able to pay for transportation. If the father doesn’t have money to pay for the transport, he may say no to the mother without explaining the reason to keep his pride. Key informants believe that the reason why the husband may not allow the mother to come back to the OTP if she was sent home without the RUTF (e.g. after reporting to the OTP too late) maybe because he suspects that she is going somewhere else. However, none of the mothers of beneficiaries and defaulters reported any kind of problems with obtaining husband’s permission to bring the child to the OTP.

Figure 15: Mind map - factors affecting defaulting in Gombe

Other issues limiting access concerned the status of the mother and the opportunity costs linked to the prioritisation of other activities. A pregnant mother it was suggested by some carers would be too lazy to walk to the OTP. A few carers believe that mothers may stop coming because they were told to by someone else (carers of beneficiaries, relatives or community) to bring the child to the OTP and are annoyed by it. As we talked to the mothers during the harvest season, there were also complaints about the work and preparations that mothers need to complete in the morning before coming to the clinic. This may cause them to come late and if they are shouted at by staff, they are less inclined to come back. In a predominantly Muslim community, fasting during Ramadan also contributed to defaulting, especially if the carer had to walk a long distance to reach the OTP. One volunteer at Kumbiya Kumbiya OTP mentioned that she heard carers of beneficiaries talking and associating CMAM with family planning and this could discourage mothers from continuing because of fear that the treatment would negatively impact their own or their children’s reproductive health.

All these factors identified by the investigation are interlinked and the concept map (Figure 16) below shows the relationships between the different factors affecting defaulting.

In general the absence of community mobilisation, compounded by problems with the quality of care at the OTP level and often associated with issues of distance and the payment of transport costs (particularly in Dukku) combine to produce high defaulting.

As a result of no outreach and sensitisation activities there is low awareness of the causes, signs and symptoms of malnutrition in communities. People’s perceptions of childhood illnesses and treatment-seeking behaviours lead to late admissions and reinforce misconceptions about the severity of the disease. Lack of information from the OTP about the causes and complications of malnutrition reinforces the belief that this is not a serious disease and can be treated easily as long as one gives food to the child. In addition, the lack of inappropriate information given by the health staff to carers also shapes the community’s perception of the CMAM programme. The reason why children were brought to the OTP was most often connected with hunger (lack of appetite, child not wanting to eat anything other than breast milk, not enough breast milk) and loss of weight. Mothers heard of the “magical” food that gives children their appetite back and makes them plump from other beneficiaries in the community. These beliefs also contribute to high defaulting rates because as soon as the child regains their appetite and starts eating other food, mothers feel that their condition has improved and the child no longer needs RUTF especially as they see it more as a food distribution programme not the treatment of an illness. Some children then relapse again, but because of perceptions of malnutrition and misconceptions about the programme they are not brought to the OTP until their MUAC is very low and they are visibly very severely malnourished.

Sometimes mothers don’t return to the OTP due to the distance or often out of fear of being shouted at for negligence. The information given to carers is very limited or sometimes even misleading, which may be the reason why some never return. Mothers are often not given key messages and are not aware what to do if there are any problems with the child taking RUTF or with recovery. As they are afraid they will be shouted at if they go back to the OTP and ask, and particularly if it is a long way, they stop going and seek treatment elsewhere. Due to a lack of communication between the OTP and volunteers there is also no defaulter tracing and children are therefore exposed to a higher mortality risk particularly given the low median MUAC on default (see MUAC at default).

The quality of service is poor; clinics are understaffed and very few health staff currently working in the clinics actually received CMAM training. Due to poor communication and organisation between the LGA and state nutrition offices, there has not been proper supervision conducted for the past year. This means that the staff don’t receive feedback on what they’re doing wrong and the service quality cannot improve. If the quality of service doesn’t improve, children don’t recover, they die or stay in the programme for a long time, morale of staff worsens, they become abusive and cause mothers to default.

Figure 16: Concept map - Causal relationships between factors affecting defaulting in Gombe

What keeps beneficiaries in

While the main focus of the investigation was to identify reasons for defaulting it was also important to try and differentiate factors which contributed to compliance with the treatment. Carers of new and current beneficiaries were interviewed, OTP staff talked to and observations conducted on clinic days therefore to better understand what might explain why some beneficiaries default, while others stay in the programme.

In both LGAs mothers of cured children reported that their main motivator to keep coming back was the child’s improvement they observed since enrolment in the programme or improvement they observed in mothers of beneficiaries they knew in their community. Some mothers also said that they were told at the OTP that if they start the treatment and then stop, it is worse than not starting at all and this created a fear in some of them that if they stop coming before the child is discharged the sickness will come back. This contradicts our finding where some carers default because of the improvement in child’s health. While cured and defaulted children can be found in the same village, the way the carers referred to recovery was different. For carers of cured discharged children, recovery meant that the child gained weight and grew plump. On the contrary, when carers of defaulted children talked about recovery they talked about return of the child’s appetite and the ability to walk/play.

In Dukku, Zange OTP stood out for having an extremely low default rate – a very busy clinic had only one defaulter in the past three months. Malala OTP stood out for a high defaulting rate with defaulters predominantly located in nearby villages. Longer visits were therefore made to the Zange and Dokoro OTPs. Unfortunately less time was spent at Malala OTP (because of Child health week) and in Dukku OTP (because of time constraints) which both had a high number of defaulters coming from urban areas close to the OTP.

Compared to other OTPs visited, the relationships between health staff and beneficiaries at Zange OTP demonstrated mutual respect and understanding. At the time of our visit to the OTP, the clinic was run by three community volunteers and one trained personnel. As in other OTPs visited, mistakes were observed in service provision (summarised in issues in service provision), but the beneficiaries reported higher satisfaction with the service provision than in other OTPs. Many carers come from far away, so it is surprising that the defaulting rate is so low. The staff waited until early afternoon for mothers to come, understanding that during the harvest season it may be difficult for mothers to find transport or that some of them have to walk for hours. This attitude was very different to other OTPs, where mothers were requested to come early in the morning and where the whole clinic finished and closed by 1pm. If the mothers came late they were shouted at and sometimes refused treatment.

The volunteers at Zange OTP also conducted outreach visits to follow up on children who did not come to the clinic, which indicates good communication between them and the nurse in charge. While volunteers at other OTPs complain of not receiving any payment and a lack of motivation, volunteers at Zange believe that without them there would be no CMAM programme and they owe it to their communities to come and do the work. The volunteers themselves are very motivated by the improvement they see in children who exit the programme and organise themselves to encourage mothers to share their stories with the community members.

While in Zange OTP volunteers’ activity and motivation for running the clinic helped keeping defaulter rates low, we did not observe elsewhere that better service provision or better staff attitudes result in low defaulting rates. This may be partially due to limited time spent at OTPs and because every OTP visited (apart from Zange) had one or another problem when it came to staff attitude.

In Gombe one area stood out for particularly high defaulting – about 20% of all defaulted children in Gombe came from Kagarawal/Anguwa Uku/Malay Inna area. We tried to talk to the community about reasons for defaulting, but were not successful in our attempts because of security issues and a lack of willingness to collaborate. Those members of the community we did manage to talk complained of the way mothers are treated at the OTP and of the requirements to bring food to the OTP.

It was not possible to clearly pinpoint the differences between children who default and those who comply. A large majority of the traced defaulters came from Fulani families. However, because of the small sample size and the lack of information about the ethnic group of children who didn’t default, a causal relationship between being a Fulani and defaulting cannot be claimed. It is possible that since a large proportion of the Gombe population is Fulani more children admitted to the programme are also Fulani.

Some differences between children who default and those who don’t were also observed in carers’ perception of the condition. Some mothers of defaulters mentioned dotti (sickness caused by animal spirits) as a condition that the child was suffering from and explained that it is not treatable with western medicine. Mothers of cured children never talked about their child having such an illness, but again the sample size is too small to be able to confidently say that there are real differences between these two groups.

No obvious differences were observed when we compared the mother’s age, the number of children or the distance, but a very low defaulting rate was observed in the OTP where community volunteers act as a link between the community and OTPs which demonstrates the importance of mobilisation activities.

Issues in service provision

Although the aim of this project was to investigate the reasons for defaulting, during visits to OTPs on clinic days several inconsistencies and issues in service provision were observed. Some of them are connected to poor clinical practice and by promoting poor health outcomes in beneficiaries may directly or indirectly impact defaulting. It is important to highlight all these issues in order to enable appropriate supervision and training to be carried out, if needed.

Main issues observed:

  • Inappropriate weight taking in some OTPs (weighing children with clothes on)

  • No double RUTF rations given to twins

  • MUAC measurement taken too tight

  • No key messages given to the carers at the time they are given RUTF

  • Insufficient number of RUTF sachets given to children, particularly to those on borderline between two weight categories

  • Filling in clinical signs on the OTP card without them being taken (e.g. recording temperature and respiratory rate without actually taking it)

  • No checking for oedema unless the child is visibly swollen

  • Mothers requested to recite seven rules (“I am illiterate, ignorant, stubborn…”) in Tudu Wada OTP in order to receive RUTF

  • No supervision from state nor from LGA nutrition officers (visits do not equate to supervision)

  • Appetite tests not conducted appropriately so children who should be in the stabilisation centre are potentially missed

  • Lack of communication between the OTP and volunteers, which results in no follow up visits

Conclusions

Overall the defaulter investigation in the Gombe CMAM programme identified several key and interlinked issues contributing negatively to defaulting, but also others which have a positive effect and can minimise it.

On a positive note, the investigation observed:

  • Very low defaulting rate in areas with active community volunteers
  • Programme appreciation and beneficiaries’ motivation to continue coming after observing rapid recovery in children

Unfortunately, issues observed during the SQUEAC two years ago are still there and have not been addressed. The defaulting rate in the programme remains unacceptably high and immediate action to reduce defaulting and improve the success of the programme is required. The man issues identified as negatively impacting defaulting rates are:

  1. Poor service provision:

    • Inadequate messages given to mothers, especially on what is wrong with the child, why it happened and why it is important for carers to come back every week

    • Referral to RUTF as food

    • Early discharges

    • Threats and refusal to treat if carers come too late, forget the ration card, don’t fully immunise the child or if they fail to bring in the food

    • Key messages are not communicated to mothers

  2. Poor record keeping and database maintenance, including incomplete admissions records and misclassification of defaulters as cured, which may be underestimating the number of defaulters.

  3. CMAM programme is still seen as a special programme from UNICEF. The staff believes they should be paid extra and this impacts on their attitude and motivation.

  4. Lack of active case finding and social mobilisation activities and as a result low community awareness on causes, symptoms and appropriate treatment of malnutrition, which makes mothers seek treatment late, often when the child presents with other complications. This puts the child at higher risk of death so he/she recovers more slowly which is against the mothers’ expectations and they may stop coming.

  5. Poor relationship and lack of communication between community volunteers and the OTP sites, including:

    • Poorly defined responsibilities of volunteers (volunteers believed their role is to refer children if they see any and not active case finding, follow up of defaulters or community sensitisation)

    • Lack of motivation among volunteers

    • No defaulter tracing and inadequate name and address information, which makes it difficult to trace children if they are absent

    • Recruitment of volunteers primarily from areas close to OTPs.

  6. Poor access to the OTP, either as a result of distance or money or both.

Recommendations

Defaulting rates can be most efficiently reduced by addressing issues with service provision at the OTPs and by reviewing and implementing social mobilisation strategy.

To address the main issues, it is thus recommended to Gombe state and the LGAs that:

  1. Roles and responsibilities of the CMAM team at state, LGA and OTP level are clarified and agreed. These roles and responsibilities should include but are not limited to: RUTF transportation/delivery, provision of routine medication, reporting system, supervision, coordination of volunteers and training

  2. Primary health care educators in each LGA hold a meeting with the OTP in charges and those in charge of CMAM to address issues in service provision and provide a platform for problem sharing and help facilitate the formation of an action plan

  3. State MoH and LGA nutritionists provide support through initiation of regular supervision of OTP staff to improve and maintain good quality of work, as well as identify and address problems as they arise

  4. Record keeping is improved and OTPs take information on beneficiaries’ address at the time of enrolment in the programme. The information should include village name, parents’ names, landmarks to reach the house and if possible also a contact number

  5. Key messages are revised in order to improve accuracy and quality of information given to carers of beneficiaries. Messages should be communicated to carers and need to be clear, simple and include information about:

    • What is wrong with the child

    • Why it happened

    • The importance of the child receiving RUTF and returning to the clinic every week until discharge

    • What kind of complications may arise and what to do

  6. Education on Infant and Young Child Feeding (IYCF) – how and what to feed children – is conducted at every OTP during clinic days

  7. It is clarified with all OTP staff that CMAM is not a special programme, independent of other services at the OTP.

  8. Staff in CMAM clinics rotate and are provided with one-to-one supervision by LGA nutrition officers and/or state nutrition team in order to make sure every person is proficient in every part of CMAM

  9. Health staff and community volunteers avoid referring to RUTF as food

  10. Roles of community volunteers are re-defined in order to include defaulter tracing and the provision of feedback to the community leaders

  11. Absent children are followed up immediately and encouraged to return to the OTP the following week

  12. Community mobilisation strategy is implemented and sensitisation campaigns on malnutrition and IYCF conducted

  13. A referral system is set up and staff made aware of other OPTs which provide CMAM. This will allow nomadic communities or people moving back from their relatives to their own homes to take up their treatment at another OTP. This is especially important with the upcoming scaling up of the CMAM programme in eight remaining LGAs in Gombe

Refresher training for OTP staff is not recommended because the staff know what they are supposed to be doing (e.g. are aware of current discharge criteria and treatment protocols), but they don’t implement this knowledge in practice. Rather than organising refresher training where it is difficult to keep an overview of the staff and identify those who need an additional explanation, one-on-one on site supervision is necessary to improve and ensure the quality of the service. Also, there is a need for encouragement and provision feedback to the OTP staff as well as to community volunteers.

Endnotes



  1. Guevarra, E et al. 2010. Semi-Quantitative Evaluation of Access and Coverage (SQUEAC) of the Community-based Management of Acute Malnutrition (CMAM) Programme: Gombe State, Nigeria. Oxford: Valid International.  2 3 4

  2. Mandalazi, E. 2012. Community Mobilisation in the Community-based Management of Acute Malnutrition Programme, Gombe State, Nigeria: Follow-up visit report. Oxford: Valid International.  2

  3. Mandalazi, E. 2012. Defaulting in the Gombe Community-based Management of Acute Malnutrition (CMAM) Programme. Oxford: Valid International  2

  4. Valid International. 2012. Review of facility data relating to defaulters, unpublished data.  2

  5. Humanitarian Charter and Minimum Standards in Humanitarian Response; recognised set of common principles and universal minimum standards for humanitarian response. 

  6. Myatt, Mark et al. 2012. Semi-Quantitative Evaluation of Access and Coverage (SQUEAC)/Simplified Lot Quality Assurance Sampling Evaluation of Access and Coverage (SLEAC) Technical Reference. Washington, DC: FHI 360/FANTA.  2

  7. Nafada LGA was not included due to time and security constraints 

  8. Google Earth 6.2.2.6613. 2011. Gombe, Nigeria 10°16’57.18”N, 11°10’28.39”E, Eye alt 9.36 km [Accessed 14 November 2012].