Chapter 5. Survey methodology

METHODS

The assessment of nutritional status is based on simple anthropometric data and limited to children of preschool age, who serve to represent the general population. The data to be collected are weight, height, sex, oedema and age (if available).

The assessment is limited to protein-energy malnutrition without attempting to assess other nutritional deficiencies. No further variables should be added without considering the additional workload and delay involved.

Weight-for-height is recommended as the main or only indicator of malnutrition by most manuals and guidelines issued by UN agencies, governments, and nongovernmental organizations. It is robust, is independent of age for children, has an internationally accepted reference population, and its interpretation is based on wide experience in many parts of the world.

The indicator is formed from weight and height measurements by comparing the weight of each child to the distribution of weights of reference children of the same height. Boys and girls are treated separately, although in the field a quick analysis can be done using the table for combined sexes in Annex 2.

For each height, the weights of the reference children are distributed as an approximately normal bell-shaped curve (Figure 1), with most weights arranged around the middle of the curve, which is the mean or median of the reference weights. In order to determine the position of a measured weight in relation to the distribution of the reference weights, the distance in kg from the median of the reference curve is determined and expressed as the number of standard deviations of that distribution. This is called a standard deviation or Z-score. The standard deviation of a distribution is a measure of the width of the distribution around the mean. Standard deviation scores of children of different heights and sex are biologically equivalent and can be compared, pooled or treated statistically.

By convention, children with a weight-for-height of less than -2 s.d. or 2 Z-scores below the median of the reference are called seriously or acutely malnourished. This limit is called the cut-off point. In the normal distribution of the reference population, 2.5% of the children are below -2 s.d. by definition. There is a 1 in 43 chance that a child with weight-for-height below this point is not malnourished but is thin for other reasons. The percentage of 2.5 is considered a baseline indicating that there is no malnutrition in a population. From now, we will refer to Z-score rather than standard deviation.

A second cut-off point, -3 Z-scores (or -3 s.d.) below the median, is often used when screening malnourished children for therapeutic feeding and treatment of infections. At this cut-off there is no baseline, since at this level all children are critically ill and severely malnourished.

A similar approach in the past used a cut-off point of 80% of the median weight-for-height. Below this, children were considered malnourished. This method is now outdated because the 80% line does not follow the reference distribution and has different meanings at different height values. It is now replaced by the use of Z-scores. Results obtained by the two methods are not equivalent at different ages and cannot be compared or converted into each other.

It should also be noted that oedema is additional weight. Children with oedema are malnourished even though their weight may not fall below -2 Z-scores. Therefore oedema must be checked for and noted on the data sheet.

MEASURING TECHNIQUES AND RECORDING

Weight: A suitable instrument for weighing a child is a 25 kg hanging spring scale marked out in steps of 0.1 kg. After weighing pants are attached to the lower hook of the scale, the instrument is adjusted to zero. The weighing pants are then taken off and handed to the weigher. The child is freed from all heavy clothing and the weighing pants are put on. The child is then suspended from the weighing scale by the handles of the pants. It should hang freely. The weight is read to the nearest 0.1 kg with the scale at eye level. The measurer reads the value out loud, the assistant repeats it and writes it down on the recording form.

Every morning the scale should be checked against a known weight of 10 kg or less and adjusted, if necessary. If the reading is incorrect, and the scale cannot be adjusted, the springs of the scale must be changed or the scale replaced. Portable electrical scales marked in 100 g steps are also becoming available but need further testing for sturdiness under field conditions. Such a scale can be set to zero while an adult stands on it. The adult then holds the child while both are weighed, which reduces the child’s distress.

Height: Children up to 2 years (23 months or 85 cm) of age are measured on a horizontal measuring board. Shoes should be removed. The child is placed gently onto the board, the soles of the feet flat against the fixed vertical part, the head near the cursor or moving part. The child should lie straight in the middle of the board, looking directly up. The assistant holds the feet firmly against the footboard and places one hand on the knees of the child, while the measurer gently holds the child’s head, places the cursor against the crown of the head and reads out the length to the nearest 0.1 cm.

Children over 2 years of age (or over 85 cm) are usually measured standing on a horizontal surface against a vertical measuring device. The assistant makes sure that the child stands straight, with the heels, knees, and shoulders against the wall, while the cursor is lowered onto the crown of the head, compressing the hair. The height is read out as before, to the nearest 0.1 cm.

An easier way to measure height consists in taking the “lying down” or recumbent length of all children from 6 months to 100 cm (59 months). This method is preferred by many field workers as it avoids scaring children and making them struggle. The recumbent length is on average 0.5 cm greater than standing height. Although the difference is of no importance for the individual child, the effect on the prevalence in a population is significant, increasing the prevalence of malnutrition by 2 to 7% for prevalences between 5 and 50%. This may have to be taken into account when comparing prevalences. A correction may be made by deducting 0.5 cm from all lengths above 84.9 cm or, if this is not possible, by correcting the calculated prevalence by using the table in Annex 4.

Age: An assessment of the ages of the children is important for two reasons:

1) malnutrition is often most marked between 6 and 18 months, which is why the age groups below and above 2 years of age should be considered separately for relief action; 2) if the height of older children is measured when they are standing, the dividing line is 2 years (see “Height”). When birth records or other documents are available, the birth date should be entered on the recording form for later computation of the exact age, or when the age is known by the mother it should be recorded in months in the appropriate space. However, in emergencies, it is often very difficult to obtain ages. If the age is uncertain, no effort should be made to estimate it (for example by a local calendar). Instead, lengths and heights should be used to group children by approximate age: 60 to 84.9 cm is equivalent to 6 months to 2 years, 85 to 100 cm is equivalent to 24 to 59 months (85 to 110 cm if population is not stunted).

Oedema: Oedema is the presence of abnormally large amounts of fluid in the intercellular tissue. It is the key clinical sign of a severe form of protein-energy malnutrition carrying a very high mortality rate in young children. To diagnose oedema, moderate thumb pressure is applied to the back of the foot or the ankle for a few seconds. If there is oedema, an impression remains for some time where the oedema fluid has been pressed out of the tissue. Only if both feet show oedema is this recorded. Cases with oedema are separated from the rest during the analysis and are counted as severe malnutrition. A prevalence of oedema of 1 or 2% is a sign of widespread malnutrition. Children with oedema are severely ill and need immediate treatment.

Dehydration: In some circumstances recording of dehydration may be indicated. This may be important where diarrhoeal disease plays a major role and may especially affect children with evidence of wasting and weight-for-height below -2 Z-scores. The physical signs include loose skin, easy “tenting” of skin and very dry mucous membranes. These children will need immediate attention. Similarly, it may be desirable to record current diarrhoea in certain surveys.