WFP experiences of vulnerability assessment of Syrian refugees in Lebanon

By Susana Moreno Romero 

Susana Moreno Romero is the Food Security Specialist and responsible of the VAM (Vulnerability Analysis and Mapping) team in WFP Lebanon since 2013, from where she has coordinated the 2013 and 2014 VASyR amongst other assessments in country. She has worked as a food security and nutrition analyst in Rome WFP HQ, South Sudan, Bolivia, Sierra Leone, Niger and Argentina with WFP, INGOs and governmental institutions. She holds a PhD in nutritional anthropology. 

The author would like to highlight that other contributors to the article, providing extensive analysis support, include Catherine Said and Mazen Makarem from the VAM team in WFP Lebanon.

The author extends thanks to all Syrian refugees, stakeholders and field monitors that have participated in and made possible the 2013 and 2014 VASyR. Partners involved in 2013 and/or 2014 VASyR included ACF, ACTED, Care, HI, Intersos, Mercy Corps, MPLP, NRC, REACH, Shield, SI, UNHCR, UNICEF and World Vision.   

Background

Since the outbreak of hostilities in Syria in early 2012, there has been a massive influx of refugees into Lebanon. By the end of July 2014, the official UNHCR figure for registered Syrian refugees had risen to 1,110,863 individuals, not including thousands of Lebanese returnees and Palestinians refugees from Syria (PRS)¹. Lebanon shares the biggest burden in terms of the influx of refugees, hosting 38% of Syrian refugees in the region. In Lebanon, one in five people is now a Syrian refugee. (For comparison, the 2010 pre-crisis population in Lebanon was estimated to be approximately 4.2 million.) The sudden increase in the assistance required, together with increasingly limited resources, required the humanitarian community to focus efforts on optimising the cost-effectiveness of assistance.

To improve knowledge of the living conditions of Syrian refugees, and to inform decision-making and the redesign of programmes, UNHCR, UNICEF and WFP agreed to conduct a joint household survey of the registered and pre-registered Syrian refugee population in Lebanon. The assessment was designed so that accurate, multi-sectorial vulnerability criteria could be derived for the implementation of humanitarian assistance. A concept note for the Vulnerability Assessment of Syrian Refugees (VASyR), complete with the methodology and a multi-sectoral questionnaire, was agreed upon by the United Nations (UN) and Government of Lebanon (GoL) partners, and was shared and discussed with stakeholders through regular multi-agency and multi-sectoral meetings and workshops. The first VASyR was conducted in 2013 and the second one in 2014.

The article considers two aspects of the VASyR:

1. A description of the approach and methodology, how this has evolved in response to the Syria crisis situation in Lebanon, and lessons learned from implementation.
2. Findings r elevant to food security and nutrition from the 2013 and 2014 VASYR

VASyR methods

VASyR 2013

More than 1,400 Syrian refugee households were interviewed in May and June 2013, following: 1) a two-stage cluster random sampling proportional to population size, and 2) a stratified sample according to registration date: awaiting registration, registered between zero and three months, registered from three to six months, and registered for more than six months. A total of 350 households in each stratum were interviewed.

Sector-specific criteria were discussed and agreed upon at the sector working group level (water, sanitation and hygiene (WASH), education, food security, protection, and economic), or through internal discussions (shelter, health, non-food items (NFIs)). According to the criteria agreed by the eight sectors, households were classified under four categories of vulnerability: severe, high, mild and low. The classification of households according to their food security situation is based on a composite indicator that considers food consumption, food expenditure share and coping strategies (see Box 1). In addition, extensive data were collected on the health and nutritional status of 1,690 children between six and 59 months (52% males; 48% females) including mid upper arm circumference (MUAC) measurement. Infant and young child feeding (IYCF) practices were assessed for 618 children under two years of age (6 - 23 months).

VASyR 2014

The main objective of the 2014 VASyR was to provide a multi-sectorial overview of the vulnerability situation of Syrian refugees in Lebanon one year after the original 2013 VASyR. The study analysed the main changes in the Syrian refugees’ living conditions compared to 2013, taking into consideration the major factors affecting any change and recommends steps forward.  The target population was Syrian refugees in Lebanon registered and awaiting registration by UNHCR, considering those included and excluded for assistance. It took place exactly one year later (May/June 2014), to ensure comparability.

For the VASyR, there is a variation in the population stratification. The sample was stratified geographically, using five regions and taking into consideration governorate administrative boundaries, operational areas and numbers of Syrian refugees registered in each region.  This approach allowed for information to be collected at administrative and operational levels so that it may be used for decision making and to maintain consistency with the UNHCR-led sixth Regional Refugee Response Plan (RRP6) for Lebanon. The sample of 1,750 households (350 per strata) is representative of each of these stratums and followed a two-stage cluster random sampling methodology.

VASyR 2013: Key findings on nutrition and food security 

Food security and coping strategies

Nearly 70% of the households had some degree of food insecurity, with the majority falling under the mild food insecurity classification. Some 12% households were classified as moderately or severely food insecure. Food insecurity seemed to decrease with the length of stay in Lebanon. Most households showed acceptable food consumption and dietary diversity however there was a risk of a micronutrient deficiency. Nearly half of the sampled refugees had applied coping strategies in the previous month; around 90% applied coping strategies related to their food consumption. The most common food-related coping strategies were:

1. Relying on less preferred or inexpensive food (89% of households)
2. Reducing the number of meals and portions sizes per day (69% of households)
3. Reducing portion size of meals (65% of households) 
4. Restricting women or adult’s food consumption so that children may eat (8% and 49% respectively)

Most of the refugees surveyed relied on the assistance of friends, family or humanitarian organisations to meet their basic needs. Adult food consumption patterns implied a risk of micronutrient deficiencies.

Health and nutrition of children

Almost half of the surveyed children under the age of five years (45%) were reported as having been sick during the two weeks prior to the survey. The most common symptoms were fever (63%), coughing (51%) and diarrhoea (35%), while 19% of the sick children showed other symptoms like allergies, infections, asthma and measles. Children under two were significantly more likely to be sick, including a much higher incidence of diarrhoea.

The prevalence of acute malnutrition amongst survey children was very low; out of 1,690 children between six and 59 months, 22 (1.0%) were found to be moderately acute malnourished (MUAC 124-115 mm) and 0.4% severely acute malnourished (MUAC <115 mm). There had been no increase since 2012 (SMART survey).  

Out of the 618 children between six and 23 months old that were included in the survey, only 6% had a minimum acceptable diet according to WHO IYCF indicators³. About 50% of children between six and 23 months were breastfed the day prior to the survey. Breastfeeding practice decreased significantly with child age; three-quarters (75%) of infants under the age of one year were breastfed, dropping to about half of children between one and one and a half years old, and decreasing to 25% in children between one and a half and two years old.  

Infant and young child feeding practices were found to be poor among Syrian refugees in Lebanon representing a risk factor for malnutrition due to some of the following issues:

• Delayed introduction of complementary foods (after the recommended 6 months of age) was common.  Over 40% of children under the age of one, and 25% of children between one and one and a half years old had not received complementary foods (based on 24 hour recall). Of the children between one and a half and two years, 10% had not received complementary foods.  
• About three quarters of children surveyed did not meet recommended minimum meal frequency and 85% of the children surveyed did not meet the minimum dietary diversity requirements the day prior to the survey.
• Only 5% of children under the age of two consumed vitamin A rich fruits and vegetables and meat or fish. The food groups most consumed by children were dairy products (54%), grains, roots and tubers (46%), followed by fruits and vegetables not rich in Vitamin A (26%) and eggs (24%). This child food consumption pattern inferred a risk of micronutrient deficiencies. 

How VASyR 2013 informed programming

The 2013 VASyR was used as a basis to determine the level of vulnerability in the population and informed targeted assistance interventions. WFP along with UNHCR started targeting assistance in Lebanon in during September and October 2013 refocusing assistance on vulnerable families. As a result, 70% of registered Syrian refugees continued to be assisted monthly with food assistance from WFP, as well as baby and hygiene kit assistance from UNHCR.  

VASyR 2014: Key findings on nutrition and food security 

Food security and coping strategies

According to the 2014 VASyR, 13% of Syrian refugees are moderately or severely food insecure, 62% are mildly food insecure and some 25% are food secure. These results show a decline in food secure households by 7% compared to 2013, mainly due to the fact there is a higher percentage of households that need to cope because of lack of food or money to buy food. The food security situation is worse in Akkar (North Lebanon) and the Bekaa Valley, where 22% and 16% of households respectively were found to be moderately and severely food insecure. The situation is best in Beirut and Mount Lebanon where 6% of households were found to be moderately and severely food insecure. 

In 2014, 28% of Syrian refugee households had to apply crisis or emergency coping strategies, which is 6% more than last year. The percentage of households spending savings as part of their coping strategies has decreased significantly compared to 2013; it moved from the most important assets-depletion coping strategy to the third most important, after borrowing money or reducing essential non-food expenditures like education or health. 

The majority (82%) of Syrian refugee households borrowed money in the last 3 months, which is 11% more than last year. Half of the households have debts amounting to US$400 or more. Thirteen per cent of households have poor and borderline consumption in 2014, which represents a 6% increase as compared to 2013. 

These results highlight a trend towards a worsening of the general food security situation of Syrian refugees, without dramatic changes.   

Health and nutrition of children

Nearly 70% of surveyed children under the age of 5 years were reported as being sick during the 2 weeks prior to the survey. The most common symptoms were fever (51%), cough (45%) and diarrhoea (35%); 14% of the children who were sick had other symptoms including allergies, infections, asthma or measles. Approximately 48% of children were reported to be sick with more than two symptoms. Children under 2 years old were significantly more likely to be sick, mainly due to diarrhoea and fever.  

IYCF practices continued to be poor, much like 2013, with the meal frequency and diet diversity being the main limiting factors. The minimum acceptable diet was met by 4% of children aged between 6 and 23 months. Half of the children in this age range were breastfed, 63% received complementary feeding, 18% had the minimum acceptable meal frequency and 18% had the minimum diet diversity of four food groups. Similar to 2013, the most consumed food groups for children were cereals and tubers (56%), dairy products (54%) and eggs (26%). The risk of micronutrient deficiencies continues to be an issue due to the low consumption of Vitamin A rich vegetables and fruits and meat and fish that were consumed by 9% and 6% of children, respectively.

How VASyR 2014 will inform programming

The 2014 VASyR is being used as a basis to refine the level of vulnerability in the population and further inform targeted assistance interventions. VASyR results have also been the key source of information on refugees’ household living conditions, for the Regional Refugee Resilience Plan 2015-16, which is currently under discussion. At the same time, the regional multi-sectorial vulnerability profile provided by the VASyR allows activities and objectives within sectors to be prioritised.  

Evolution of the VASyR

Context of the VASyR assessments 

Since the 2013 VASyR took place in May/June 2013, the context in Lebanon and the situation of Syrian refugees in-country may well have been affected by the following factors:

• The number of Syrian registered refugees in Lebanon has surpassed 1 million. The Syrians currently in-country could account for one quarter of the population living in Lebanon, which may clearly have further implications on the increasing tension with the host community, the strain on the infrastructure in Lebanon and access to shelter, employment and essential basic services (health, education, water, sanitation, electricity). 
• As part of responsible programming, various types of assistance (food, hygiene and baby kits) shifted from blanket to targeted assistance during September and October 2013. Targeting of assistance was aimed at households most in need, with some 70% of the Syrian refugee population thus continuing to receive the above assistance. Although 30% of the registered population was deemed as able to cover their basic needs without engaging in irreversible coping strategies (and thus no longer qualifying for assistance), it is also part of responsible programming to monitor how the targeting of assistance affects the Syrian refugee population as a whole. 
• The time spent by Syrian refugees in Lebanon may have positive or negative effects. Refugees may have increasingly adapted to the new context, may have a better awareness and network facilitating access to some services, and may have a better knowledge of assistance benefits. On the other hand, time implies a higher risk of exhaustion of resources (e.g. savings and/or assets) and difficulties to continue coping through loans.

VASYR 2014 provided the follow-up to the 2013 study to explore the impact of these issues. 

Stratification by region

Since the 2013 VASyR, there was evidence of regional disparities within Lebanon for different indicators, but a lack of comprehensive and representative information at regional level based on sound assessments or standard methodology. There was mounting interest coming from the humanitarian community to better understand these regional differences in the refugees’ situation and fill this critical information gap. This geographical stratification was used in the 2014 VASyR.

Stratification by registration date 

Stratification by registration date was included in VASyR in 2013 but not in 2014. One of the main reasons behind the stratification by registration date in the 2013 VASyR was to explore whether this variable affected household vulnerability and could therefore help better define the need of assistance. The 2013 VASyR showed that refugees awaiting registration or recently registered did tend to show poorer living conditions for some indicators compared to those registered for a longer period of time Yet overall, vulnerability was not significantly different among these strata. Information about living conditions by registration date is available from the 2013 VASyR, and if repeated in 2014, strata would have changed given the disproportionate number of refugees in each strata in 2014, most of them registered over 6 months ago. Thus the analysis by registration date was carried out with the 2014 VASyR data, but with no representativeness by registration group. 

Nutrition indicators

MUAC and oedema results in the 2013 VASyR indicated a 1% prevalence of moderate acute malnutrition (MAM) and 0.4% severe acute malnutrition (SAM) (1.4% global acute malnutrition GAM)). These results were lower than malnutrition prevalence determined by weight for height in the SMART nutritional survey of 2012 (4.4% GAM), as well as the results that were later released in the 2013 SMART nutritional survey⁴. The decision to remove MUAC from the 2014 survey was based on the following reasons:

• In this population, MUAC underestimates acute malnutrition compared to weight for height. 
• Given the low acute malnutrition prevalence in the population based on MUAC, the precision needed to track potential changes would have required a larger sample size than needed for the VASyR purposes. 
• Due to the lack of significant changes in acute malnutrition rates found in the 2013 SMART nutritional survey compared to 2012, it was not deemed worthy to introduce, to the 2014 VASyR, the added complexity of training and implementing the MUAC exercise (including the standardisation test for enumerators). 

Weight for height and micronutrient status data were not collected in 2013 or 2014, as this would have added undue complexity to the VASyR which is meant to be an emergency multi-sectoral assessment and given the availability of results from the SMART nutritional surveys conducted in 2012 and 2013.  The nutrition component of non-communicable disease (NCD) was not assessed as this was not selected at the sector working group level, although chronic diseases (self-reported) are included in the ‘specific needs’ module of the VASyR. 

Conclusions

The VASyR provides a very valuable comprehensive picture on living conditions for Syrian refugees to better inform decision-making. The assessment is statistically sound with representative data at different levels (registration date, regional). At the same time, it is operationally feasible to undertake in an emergency context when information is needed in a short period of time so as to re-design programmes according to evolving needs. It strongly contributes to identifying main needs as well as areas where more detailed information would be required to better address any sector-specific concerns.    

The VASyR has a set of implementation challenges to overcome and one broad limitation. The main limitation is that the VASyR does not provide all the detailed information needed for each sector; it does not replace in-depth sector-specific surveys. Only the most critical indicators are selected per sector so that the overall questionnaire can be feasibly rolled-out. 

The approach was to conduct a wide-ranging multi-sectoral, higher-level survey that can be carried out without requiring an overly long assessment of interviewees. Challenges and means to address these are as follows;  

1. Improve on information collected, through identifying key sector-specific questions that provides the essential information needed for decision-making and help better define the thresholds that more accurately identify vulnerability. This process requires intra and inter-sectoral discussions with each sector attempting to attain the most information possible for their own purposes⁵. Although the questionnaire should be contextualised and revised in line with lessons learnt from previous assessment exercises, efforts carried out at the international level to standardise vulnerability questions, categories and thresholds would facilitate this process significantly. Such work should take account of specific contexts like urban or semi-urban areas, refugees not residing in camps, and situations in middle income countries.
2. Further enhance data quality. The number of enumerators needed for an assessment of this scale where field data collection takes about 2 weeks, ranges between 64 and 82. These enumerators need to be trained in different sector-specific questions, as well as in the VASyR methodology. Training of trainers has been identified as the best approach but this requires extensive efforts in standardising training modules, providing clear guidelines on the methodology, process and questionnaire along with close supervision at different levels. These three factors are key to minimising regional differences in interpreting questions, methodology and in standardising how to manage unpredictable situations. One of the main objectives of 2014 VASyR has been to improve data quality by introducing these elements but it is a continuous process. In addition, in VASyR 2014, quality monitors from UN agencies accompanied the enumerators during the field data collection. There were two monitors per region, to strengthen and support the supervision role, and it especially revolved around quality of information collected during the interviews. 
3. Clarity around the definition of households used. For the VASyR, a household is considered to consist of family members that live together or in different living structures, eat out of the same pot, and share the same budget that is managed by the head of the household. The definition of households registered with UNHCR is more stringent and considers protection factors so that registration cases are considered as separate households regardless of the common expenditure shared. Since the household definitions are not the same, this implies that some ‘VASyR households’ have more than one UNHCR registration case number. Establishing the limits of the household remains a challenge due to the high number of combinations that are found in the field. 

The food security situation of Syrian refugees in Lebanon has deteriorated in the previous year. As savings and assets are being exhausted or becoming more limited, households engaged in more severe strategies to cope with the lack of food or money to buy food. These coping strategies included reducing expenses on health or education. The average household size is 6.6 members and generally, only one individual is able to work, mainly in temporary employment. This is insufficient to cover the US$762 on average that a given household reportedly spends on a monthly basis. Also, about one fourth of households do not have any member working. Almost half of refugee households live below the poverty line of US$3.84 per person day. Compared to last year, refugees depend more on external sources of cash like WFP’s food vouchers or loans, and less on skilled work or their own savings. Borrowing money is occurring more frequently and debt amounts are higher than last year. Female-headed households and single-headed households with dependents have also increased compared to 2013, exacerbating the difficulties to access work. Despite the fact that households are employing coping strategies, food consumption of most food groups as well as diet diversity has also reduced; this year, households are less likely to have acceptable food consumption. Expenditures on health, water and hygiene items have increased. This has occurred possibly in response to the reduction in hygiene and baby kits in-kind assistance and also to the water scarcity situation in Lebanon. In 2014, there are proportionally more refugee households without access to bathrooms, sufficient access to water, soap or hygiene items. The security situation is also deteriorating for Syrian refugees who experience an increasing harassment and extortion. 

As the conflict in Syria continues and there is no expectation of an early resolution, the number of refugees in Lebanon continues to increase. It is estimated to reach 1.5 million registered by the end of 2014. It is expected that the Syrian refugees’ living conditions will continue to deteriorate and the impact of the crisis will also worsen the situation for the most vulnerable Lebanese population. This will be compounded by the security situation, which is tenser in the last months due to the increasing number of refugees but also to the recent events in the northeast part of the country (Aarsal) as well as in Iraq. The combination of these ingredients constitutes a risky context for Lebanon’s stability, especially if overall assistance is reduced by any given funding constraints. 

For more information, contact: Susana Moreno Romero, email: Susana.moreno@wfp.org

¹  UNHCR Lebanon Portal: http://data.unhcr.org/syrianrefugees/country.php?id=122

²  http://54.225.218.247/wfp/documents/Lebanon/VASyR.pdf

³ WHO 2008. Indicators for assessing infant and young child feeding practices. Part 1: Definitions and Part 2: Measurement. http://www.who.int/nutrition/publications/infantfeeding/9789241599290/en/? 

⁴  The SMART survey results were subsequently reviewed and corrected. This found a lower prevalence of acute malnutrition than initially estimated. For more details, see article by UNICEF in this issue.

⁵ Each sector proposed their key questions based on international agreements and tools but also considering the nature of the assessment and context (assessment at household level in urban and semi urban refugee population in a middle-income country) as well as the indicators used by the sectors for the regional response plan.