Review of nutrition and mortality indicators for Integrated Phase Classification

Summary of technical review¹

The Integrated Phase Classification (IPC) Technical Working Group and the Standing Committee on Nutrition (SCN) Task Force on Assessment, Monitoring and Evaluation recently commissioned a technical review of morality and nutrition indicators. The purpose of the review was to ensure that these indicators, in combination with others, help in making a single statement on the food security situation. Malnutrition and mortality indicators have been included in the IPC, along with those of food security, as 'Key Reference Outcome Indicators' since its inception in 2004. Reference levels for each indicator (thresholds) are attributed to each of the five phases of the classification.

Indicators reviewed that are already in the IPC were crude death rate, 0-5 death rate (deaths/10,000/day), low weight-for-height and low height-for-age. Indicators reviewed but not yet included in the IPC were under five mortality rates (deaths/1000 live births), mid-upper arm circumference (MUAC), underweight and chronic under-nutrition (low body mass index (BMI)). Preliminary findings and recommendations from the review were discussed at a workshop in Rome on 14-15th July 2009.

The main findings and recommendations of the review were as follows:

Acute Malnutrition: Weight-for-height

It is appropriate to continue to use weight-forheight (WH) in all phases of the IPC and to recognise its particular relevance in phases 3-5. Severe acute malnutrition is not recommended as a key reference outcome because of the small numbers and therefore wide confidence intervals (the same argument applies to oedema).

Weight for height reference levels

There is a need for the standardisation of reference levels between different stakeholders and systems based on agreement and consensus beyond the IPC institutions. These reference levels need to reflect more clearly the exponential relationship between malnutrition and mortality, i.e. the increase in prevalence of low WH in each of the IPC phases should reflect the exponential increase in malnutrition that is seen with increases in background mortality and vice versa.

The possibility of using overlapping WH reference levels between phases should also be considered. The actual phase would be decided by considering prevalence of wasting as well as other indicators. Until new reference levels are established, the continued use of the current WH reference levels in combination with an analysis of trends is recommended. The use of relative reference levels or an analysis of trends is recommended for all phases. IPC should include mean WH z score (WHZ) where this information is available and encourage monitoring of means in food security information systems. The IPC should consider using the WHO reference levels for mean WH.

MUAC

The prevalence of low MUAC (<115mm) is appropriate in the emergency phases of the IPC as supporting evidence to indicate mortality risk in the population and to identify need for feeding programmes. Low MUAC is unlikely to be a good indicator of food security. Prevalence of low MUAC are similar to prevalence of severe acute malnutrition (WHZ?-3) based on the 2006 WHO growth standards. One percent low MUAC is the reference level recommended as supporting evidence to indicate mortality risk in phases 4 and 5.

Guidance for interpretation of acute malnutrition

It is crucial that public health factors, disease incidence and expected seasonal patterns of disease and food insecurity are also taken into account in the interpretation of reference levels (or outcomes). Seasonal calendars describing the usual or expected seasonal changes for different regions within a country should be developed and used to help interpret mortality and malnutrition data.

Monitoring trends in malnutrition by age group is crucially important for an understanding of changes in the food security situations. It is recommended that the IPC encourages the reporting of ratios of older (length 85 cm and above) and younger (length <85cm) children mean and prevalence.

Guidance is needed that explains how as food insecurity evolves and deepens, the underlying causes of malnutrition change and interact with each other. The importance of food insecurity as a cause of malnutrition increases and drives the other two groups of underlying causes.

The IPC should also consider differentiating between different levels of diseases and public health crisis. Population density, crowding and shelter might also be relevant reference outcomes needed to help differentiate public health risks that might exacerbate the contribution of malnutrition to mortality.

Height for age (HA)

The inclusion of HA in the early phases of the IPC is appropriate, particularly in phases 1 and 2, but also phase 3 as recommended by WFP. It is recommended as an indicator of underlying vulnerability, rather than as a key reference outcome since there is no consensus on the latter. Stunting is likely to be of most relevance to the IPC when considered among children between 0 and 24 months.

HA reference levels

Given the recent evidence of its association with food security, there is potential of HA as a reference outcome in phases 1-3. The authors recommend continuing to use the HA reference levels already applied in phases 1 and 2 and introducing the WFP recommended reference level in phase 3. An analysis of the distribution of low HA globally and where available in emergency contexts is urgently needed.

Guidance for interpretation

It is possible that prevalence of low HA may increase while prevalence of low WH remains fairly stable. In this situation, increasing prevalence of HA may be taken as supporting evidence of deteriorating food security.

Weight for age (WA)

There was general consensus at the workshop to use WA in phases 1-3 but not 4 and 5 as an indicator of underlying vulnerability, not a reference outcome. Given recent evidence of its association with different categories of food security among children < 24 months and its availability where there is often a lack of data on wasting, it is worthy of further research.

Reference levels and guidance for interpretation

As with HA, the only available reference levels are those published by WHO (1995) which clearly need reviewing. It is likely to be more helpful to consider deviations from regional, national and local baselines. A deterioration in HA and WA can be used as confirmation of deteriorating food security. This can also be used for monitoring long term trends, measuring impact over time and for advocacy purposes.

Body Mass Index (BMI)

The IPC workshop participants broadly agreed that BMI should be included in all five phases of the IPC with an emphasis on low BMI in the upper phases. For the lower phases, it has the potential to capture the double burden of overand under nutrition. Therefore it was recommended that low BMI (<18.5 among non pregnant women aged 15-49) are included in all phases using the widely adopted cut-off of <18.5 to estimate prevalence.

Reference levels and guidance for interpretation

The BMI working group at the workshop recommended that the IPC should consider relative reference levels, based on national or more local figures, with a multiplier of 1.5 of baseline a guide to shift to a higher phase. The authors recommend that this should only apply to those baselines that are less than 20%.

Several factors other than nutritional status influence BMI. One of these most important of these is body shape, and these therefore need to be taken into account when interpreting prevalence of low BMI.

Mortality

Crude death rate (CDR) and 0-5 death rates (0-5 DR) should remain indicators within the IPC and are of particular relevance in the emergency phases to detect rapid changes in severity of crisis. Excess mortality, especially if reported by age, should be included if accurate baseline mortality rates are available as they are a good indicator of the impact of an emergency. A threshold is not needed as it involves a direct count. IMR and U5MR are not appropriate for use in the emergency phases as the estimates cover the past 5 years and are centred about 2.5 years in the past. Trends can how ever be monitored if prospective surveillance systems are present and this should be encouraged by the IPC.

Reference levels and guidance for interpretation

Reference levels for CDR and 0-5 DR used by the IPC are widely accepted and broadly standardised among the humanitarian community. The authors recommend using >5 per 10,000/day for CDR for phase 5. Alternatively, in situations where baseline mortality is relatively low, using doubling of the baseline CDR to identify an emergency may be applied.

Duration as an indicator

There is an urgent need for the IPC to include duration, or time spent within a particular phase, as part of its analysis.

Further research needs

Reference levels need to be revised based on an analysis of the distribution of estimates globally and more specifically in emergencies for the following indicators:

• Prevalence of global acute malnutrition (GAM), and wasting and mean WH, for children <5 years, and for children above and below 2 years
• Stunting, including disaggregating results for children < 2 years
• Low MUAC (<115 mm)
• BMI of non pregnant women aged 15-49 years (<18.5 BMI)
• Obesity among non pregnant women aged 15-49 years (>30 BMI)

Available databases that could be used for this review include the Nutrition Information in Crisis Situations (NICS) database, the WHO Global Nutrition Database and possibly the Centre for Research on the Epidemiology of Disasters (CRED) database. Previously, there were three approaches for determining reference levels:

1. arbitrary categories based on existing reference levels adapted to fit the number of scales required
2. a grouping of the available prevalence data for stunting and wasting in developing countries pre 1993 from the WHO database
3. the NICS categories based on reviews of the malnutrition prevalence data combined with mortality data in humanitarian emergencies, including internally displaced people (IDP) and refugee camps, and acute food insecurity and famine crises between 1992-4

A new review of globally available data in emergencies and more stable contexts is needed. The earlier reviews were undertaken more than 15 years ago, were based on low WH (not GAM) and the data used by the RNIS /NICS covered extreme crises. Since that time, the availability of data has increased, the quality improved, while data are now available for GAM and not only for WHZ <- 2.

The application of the new reference levels should be monitored in a number of pilot countries and compared with food security indicators. The proposed changes to the use of stunting and underweight as indirect evidence in the IPC should also be evaluated as part of the above pilots.

Research is needed to establish the association between severity of food insecurity and nutritional indicators, and with mortality. Research is also needed to investigate how these associations differ between older and younger children, seasonally and according to body shape.

All recommendations for changes from the current version of the IPC Reference Table (Technical Manual, IPC Global Partners, 2008) are highlighted in bold in Table 1. Table 2 contains recommendations for supporting or indirect evidence.

WHZ: weight-for-height z score; HAZ: height-for-age z score; BMI: body mass index
All recommendations for change from the current version of the IPC Reference Table are highlighted in bold.

WAZ: weight-for-age z score; MUAC: mid upper arm circumference

¹Young.H and Jaspars. S (2009). Review of nutrition and mortality indicators for Integrated Phase Classification. Reference levels and decision-making. Study commissioned by the SCN Task Force on Assessment, Monitoring and Evaluation and the Integrated Phase Classification Global Partners. September 2009.

²Refugee Nutrition Information System, that became NICS in 2004.

³Maternal under-nutrition reference levels are based on the 1995 WHO Expert Committee, which gives no reference levels for extreme food insecurity, i.e. famine/ humanitarian catastrophe phase

⁴The use of severe acute malnutrition (SAM), if available, could be considered as supporting evidence for phases 3-5