UNHCR feeding programme performance in Kenya and Tanzania

Summary of research¹

Settlements of new arrivals in the outskirts of Dadaab

Routine monitoring data are available from the many nutrition programmes operating in camps supported by UNHCR, typically growth monitoring programmes, under the Expanded Programme on Immunisation, MUAC² based household surveillance, Supplementary Feeding Programmes (SFPs) and community based therapeutic feeding programmes (TFPs). However, operational challenges and varied reporting structures across implementing agencies have limited the utility of this information for the assessment of changes in population health status or comparisons of trends across regions. In 2005, UNHCR launched a Health Information System (HIS) to enhance the quality and consistency of routine health information available in protracted refugee situations (see Box 1), becoming the first source of routine feeding programme data collected with standardized case definitions and reporting formats across refugee settings.

A recently published paper reviews UNHCR’s HIS data to assess the coverage and effectiveness of selective feeding programmes implemented by partners in Kenya and Tanzania from January 2006 to May 2009. It examines the extent to which routinely collected HIS data can be used to inform nutrition programme strategies in post-emergency situations. Kenya and Tanzania were chosen as case studies because of their geographic proximity, similar restrictions on livelihood opportunities, distinct population trends, and the fact that together they account for more than half of the global refugee population living in camps reported in the UNHCR HIS.

Since the early 1990s, Kenya and Tanzania have hosted hundreds of thousands of refugees fleeing war and insecurity in neighbouring countries. In Kenya, the Dadaab and Kakuma camps currently host approximately 340,000 refugees from Somalia, Sudan, Ethiopia, and other countries. A mid-2006 survey in Kenya showed that despite widespread distribution of WFP rations (~2,100 kcal/day) since 2003, global acute malnutrition rates remained well above internationally recognised emergency thresholds at 22.2% in the Dadaab Camps and 15.9% in Kakuma Camp. The nutritional status of refugees in these camps has improved significantly in the past few years. Recent surveys show that global acute malnutrition rates have fallen from their peak in 2006 to 5.6%-10.7% in Dadaab Camps and 9.2% in Kakuma Camp in 2010.

In Tanzania, refugees from Burundi, Rwanda and the Democratic Republic of Congo are housed in clusters of camps in far western Kigoma, Kasulu and Kibondo districts in the Kigoma Region and the Ngara District in the Kagera Region. As in Kenya, the Government of Tanzania prohibits refugees from seeking formal employment, travelling more than 4 km outside their camp, cultivating land outside the camp boundaries, trading in markets or grazing livestock. A nutrition survey in 2006 showed that despite these restrictions and limited food rations, the rate of global acute malnutrition was low, with a weighted prevalence rate of 3.1% across the three camps. A joint UNICEF, UNHCR, and WFP survey in late 2008 showed further improvements in the nutrition situation, with the prevalence of global acute malnutrition decreasing to 1.9% and with no severe acute malnutrition (SAM). The survey conducted in September 2010 showed that these improvements have been sustained, with 2% prevalence of global acute malnutrition.

Data on camp population, growth monitoring, and nutrition programmes were exported from the UNHCR HIS database. The data included camp-specific information about the month of reporting, total camp population, and population size by age group (less than 5 years of age, 5 years of age or older). Nutrition programme data included the number of children admitted to feeding programmes, the number remaining enrolled in the programme at the end of each month, mean length of stay, average weight gain in therapeutic programmes, and reason for discharge (recovery, death, default, or referral). Information on nutrition programme admissions and enrolment was combined with population data to calculate admission rates and enrolment proportions where possible. Analysis was limited to children under five years who were admitted to supplementary or therapeutic feeding programmes for malnutrition.

Results

SFPs

There were a total of 39,899 new SFP admissions in Kenya and Tanzania between January 2006 and May 2009, accounting for approximately 63% of SFP admissions reported in the UNHCR HIS. Of these, 95% of admissions (n = 37,741) were in Kenya camps and only 5% (n = 2,158) were in Tanzania. Globally, readmissions, defined as moderately malnourished children enrolling in SFP within 1 month of successful discharge, account for 8.6% of all SFP admissions. In Kenya and Tanzania, the proportions of readmissions were lower at 5% and 2%, respectively.

Although monthly growth monitoring attendance was much higher in Tanzania (78%) than Kenya (42%), children in Kenya were 11.60 (95% CI, 10.61-12.68) times more likely to be admitted to SFPs and 9.77 (95% CI, 8.49-11.25) times more likely to be enrolled in SFPs than children in Tanzania. In Kenya, SFP performance consistently exceeded all UNHCR performance standards, with the exception of the average length of enrolment in SFPs, which ranged from 11 to 20 weeks (standard < 8 weeks). All Kenyan camps had recovery rates above 90% (standard >75%), mortality rates at 0.1% or below (standard <3%), and default rates below 8% (standard < 15%). All camps in Tanzania also met overall exit rate standards, with the exception of Kigoma Camps, where an average of 20% of cases were referred for further treatment, and Ngara Camps, where an average of 19% of exits were due to default and 1% to death.

Table 1 summarises SFP performance by country and camp. With the exception of length of enrolment, the HIS data showed improvements over time in all SFP performance indicators in Kenya camps. Between 2006 and 2009 in Kenya, the average recovery rate increased from 89% to 96% and the average default rate decreased from 7% to 2%. In Tanzania, the average recovery rate was lowest at 69% in 2007 (and 92% in 2009), and the proportion of exits due to default increased over time, with a default rate of 13% in 2008.

SFP, supplementary feeding programme

TFPs

Between January 2006 and May 2008, there were a total of 4,347 new admissions of children under five years to TFPs in Kenya and Tanzania camps. Admissions in these two countries accounted for approximately 30% of TFP admissions reported in all countries covered by the UNHCR HIS during this time period. Of these, 79% (n = 3,417) were for acute wasting and 21% (n = 930) were for oedema (OR, 3.83; 95% CI, 3.55-4.12). In general TFP admissions and enrolment rates in both Kenya and Tanzania decreased over time, and higher admission and enrolment rates were observed in Kenya than in Tanzania. Acute wasting admissions were substantially higher in Kenya (n = 3,014, 70% of all admissions across the two countries) than in Tanzania (n = 403, 9% of all admissions). The same was true for oedema, with 690 admissions (16% of all admissions) and 240 admissions (5% of all admissions) reported in Kenya and Tanzania, respectively.

TFP performance met UNHCR standards for both acute wasting and oedema cases at the country level, with the exception of mean length of stay for acute wasting cases in Tanzania, which was 34 days (standard < 30 days). As Table 2 shows, the main factor limiting the recovery rate in TFPs was referral. In Kenya, Hagadera was the only camp with average recovery rates below the standard of 75%; referrals accounted for a high proportion of both acute wasting and oedema exits during 2007 in Kenya. In Tanzania, belowaverage recovery rates in Kigoma Camps were associated with an increasing proportion of acute wasting referrals over time and a high proportion of oedema deaths in 2007 (4 of 23 cases).

TFP: Therapeutic Feeding Programme, SAM: severe acute malnutrition

In Kenya camps, the average TFP enrolment period ranged from 22 to 27 days for acute wasting and from 23 to 27 days for oedema (standard < 30 days). In Tanzania, the average enrolment period was longer, ranging from 25 to 43 days for acute wasting and from 21 to 34 days for oedema. Most camps met the > 8 g/kg/day standard or average weight gain, and average daily weight gain was substantially greater in Kenya than in Tanzania for children with both acute wasting and oedema. Camps that did not meet UNHCR standards included the Ngara Camps in Tanzania (average weight gain of 5 g/ kg/day for children with acute wasting), Dagahaley Camp in Kenya, and the Kibondo and Kigoma Camps in Tanzania, all which reported average weight gains of 7g/kg/day for children with oedema.

Discussion

During the 3 and a half year period that was analysed using HIS data, close to 45,000 malnourished refugee children under five years were treated in UNHCR-supported selective feeding programs in Kenya and Tanzania. With average recovery rates of 77.1% and 84.6% in the therapeutic and supplementary feeding programmes, respectively, mortality rates of less than 1%, and an average readmission rate below 5%, the HIS data suggest that selective feeding programmes had a beneficial effect on enrolled children. Even the camps with higher SFP enrolment rates had very low TFP enrolment rates (= 0.5%), which suggests that growth screening and SFP objectives are being met and that SAM is relatively well controlled in both the Kenya and the Tanzania camps. Any comparison and analysis of trends, however, must take into account the myriad of contextual factors that influence nutritional status, feeding programme size, and coverage and outcomes of targeted feeding programmes.

Limitations

Conclusions are constrained by several factors. One challenge is that HIS data on the SFP programme includes local populations that are enrolled in feeding programmes but are not represented in denominators; thus SFP coverage among refugee populations may be overestimated. Average weight gain and duration of enrolment for discharged cases are the strongest indicators of TFP performance. However, these measures are also the most complicated to accurately measure and are prone to omissions and reporting errors, which decrease their reliability as performance measures. In addition, defaulters limited recovery rates in both Kenya and Tanzania. However, there is no mechanism in the HIS for tracking the reason for default or outcomes of children who default. Consequently, it is difficult to determine whether programme strategies should prioritize efforts to prevent defaulting by increasing accessibility, addressing opportunity costs, or improving reporting.

Conclusions

Inferences from the comparison of HIS admissions and enrolment data suggest that increasing admission and enrolment in SFPs was successful in preventing cases of severe malnutrition in the Dadaab Camps. Modifications to routine data collection and reporting, such as inclusion of the proportion of moderate and severe wasting cases detected at growth monitoring, follow-up on defaulter and referral cases, and additional information on Community-based Management of Acute Malnutrition (CMAM), could also shed more light on areas for improved programme quality in both the Tanzania and the Kenya camps. In addition, linking nutritional surveillance data from the HIS with other relevant information sources (e.g., nutrition surveys, household food economy studies, local market surveys, and joint UNHCR/WFP assessment reports) would allow users to triangulate the results and generate a more comprehensive understanding of the predictors of increased admission and enrolment needs within selective feeding programmes. UNHCR has made the first step toward achieving this aim through the recent launch of an online version of the HIS (called ‘webHIS’ http://his.unhcr.org). This integrated platform hosts public health data from multiple sources.

¹Tappis. H et al (2012). United Nations High Commissioner for Refugees feeding programme performance in Kenya and Tanzania: A retrospective analysis of routine health information systems. Food and Nutrition Bulletin, vol. 33, no. 2 ©2012, The United Nations University

²Mid Upper Arm Circumference