Contents - Previous - Next

Use of BMI for monitoring and surveillance, practical issues

Introduction
'Conventional' nutritional surveillance
Proposed nutritional surveillance
Discussion
Appendix 1.
Appendix 2.

F. Sizaret

Nutrition Planning, Assessment and Evaluation Service, Food Policy and Nutrition Division, Food and
Agriculture Organization of the United Nations, Via delle Termi di Caracalla, 00100 Rome, Italy

Introduction

Since 1974, considerable effort has been expended to design, perfect and implement nutritional surveillance. Those responsible for surveillance activities have spontaneously focused mainly on children, particularly children under 5 years old. This is not surprising, as nutritionists have always centred much of their interest on infants, weighing and measuring them for comparison with the reference norms. The advent of computer processing has only accentuated this trend, and thus the nutritional status of a population is today identified with that of its offspring.

Equally important efforts have also been made to try to assess nutritional status through food consumption surveys which included anthropometric data on every household member surveyed. Analysis and interpretation of these consumption surveys a number of problems with calculating and comparing intake and requirements, so much so that surveys of this type tend to be used for market and behavioural studies, or to provide information on food consumption, expenditure, trends, etc.

The anthropometric data section of these surveys, however, had received very little attention, and it was only in the 1990s, at the instigation of the Food Policy and Nutrition Division of FAO and the Rowett Research Institute of Aberdeen, that anthropometric data on adults were given more serious consideration.

'Conventional' nutritional surveillance

For various reasons, anthropometric data on children tend to be collected when mothers bring their children in for a check-up or during school medical check-ups. There is very little to study in these data as very few variables are measured, and a descriptive analysis only becomes explanatory over time. Analysis of child anthropometry data often failed to show an expected association with market or weather failures and the protective role of the parents may be a strong factor (positive or negative).

It is now recognized that nutritional surveillance based on regular measurements of child weight and height, although it may serve as a back-up to other data on food price trends, weather variations and so forth, does not always provide the information needed to orient effective action. These large-scale efforts soon become routine - an increasingly complicated routine in terms of providing timely information. This approach to nutritional surveillance can only offer very partial solutions because changes at the level of a whole population are very slow and hard to pinpoint from one year to the next.

The risk, then, is that nutritional surveillance systems which deliberately ignore adults will offer biased information. The nutritional status of a population should illustrate the nutritional status of each component member. And even though public opinion obviously remains more sensitive to the nutritional problems of the very young, it is the underlying problems which need to be solved, and these will be brought out much more clearly by an epidemiological study of the anthropometric data of each and every household member.

Proposed nutritional surveillance

An individual's nutritional status can change very quickly but (excepting natural or manmade calamities, of course), the nutritional status of a whole population changes very slowly. It would be relevant, therefore, in attempting to monitor the nutritional status of a population, to determine first their nutritional status and then monitor the risk factors, which can in fact change quite rapidly. By analysing these two sets of information together, the necessary action can be charted, particularly for groups at risk, in accordance with the responsibilities and special interests of each intervening sector (health, education, agriculture, social affairs, planning, etc.).

Considerable effort needs to be made to obtain a representative sample of households in the population studied. The size of the sample will depend on the level of aggregation required (village, district, municipality, geographical area, country, etc.). The sample will have to draw on the most recent population census data. To measure and monitor the impact of development projects on the nutritional status of the Sierra Norte region of Mexico, for example, it would be advisable to identify families in at least three types of communities: very remote and thus highly dispersed communities, communities living close together in villages, and categories in between.

The data collected should make it possible to identify and describe individuals at nutritional risk. The data, which can be summarized as follows, should be collected quickly without disturbing the survey population:

For each individual in the household selected, list age, sex (and physiological status for women), weight (in kg), and height (in cm). Depending on the circumstances, other classic variables may be collected: relation to head of household, educational level, whether or not the person is handicapped, ethnic group, the name of the assistance programme in which he/she participates, whether or not he/she has been vaccinated, etc.

The list of suggestions is long and the decision to include such additional data will be dictated by their possible usefulness for future decision-making.

For each household surveyed, the variables may include:

1. The geographical situation: region, department, municipality, district, rural/urban, etc.

2. Housing specifics:

• the roof (straw, tile, sheet metal, etc.)
• walls (board, thatch, etc.)
• flooring (adobe, bare earth)
• total area and number of rooms.

3. Available services: (electricity, drinking water, sewerage, w.c.), for categorization of these dwellings.

4. Durable or semi-durable goods owned by the family, e.g. radio, TV, bicycles, watches, beds, cupboards.

Lastly, it would be advisable to gather additional data for ease of categorization (approximate income in preceding month, head of household's profession). Concerning this subject, in rural areas where individuals do many different jobs in the course of the year, the agricultural characteristics of the household should be more closely identified: main crops, type of husbandry, production for sale and for household consumption, etc. Then these populations can be grouped later according to the professional activity of the head of household.

A questionnaire of this type should be designed to answer all these questions on a single sheet printed on both sides so that the data can be computerized immediately. One or more pilot surveys can be tested for manageability. It should not take longer than 2 days to train people to fill in these questionnaires, and one interviewer can probably interview four families each day, depending on the household's size and the distances to be covered.

The best way to diminish the risk of observer error is to work in teams of one supervisor and five to six enumerators: a team thus composed can collect data from some 500-600 families in 1 month, furnishing a mass of simple and easy-to-analyse data for the region under study. The data can be computerized and processed almost instantly. Once the field work is completed, 2 weeks should be long enough to compile enough preliminary data to present at a workshop bringing together representatives of the various institutions working to combat poverty.

The nutritional status of both adults and children can then be analysed to identify groups at risk. The study of the various factors, using new statistical analytical techniques, will make it possible to identify these groups and their relative importance, and provide vital elements for decision-making. In this manner, BMI is a useful diagnostic tool, and able to look at the different effects on adults and children and thereby discriminate between factors such as food, health and care in its causality.

This concept of nutritional surveillance, which starts with simple anthropometric data on all individuals in a representative household survey of a given geographical area, thus provides valuable elements for evaluating nutritional status from anthropometric data. Adding to the analysis a set of simple socio-economic variables characteristic of these households, answers the basic questions of who is malnourished and where they live. Why they are malnourished will then begin to emerge. The periodic monitoring of anthropometric data will reveal the presence of a positive trend, but this trend will still be difficult to perceive from one year to the next. Monitoring risk factors, however (inflation rates, seasonality, family size, housing conditions, ease of access to specific villages, etc.) offers an extremely reliable indication of whether or not increasingly large numbers of individuals are likely to be at nutritional risk and are much easier to watch on a regular basis.

Sample questionnaires are shown in the Appendix.

Discussion

Allen: What is the current FAO attitude to micronutrient deficiencies; these are more common than energy deficiency?

Sizaret: We don't deny the importance of micronutrient deficiency, but we want to concentrate on housing and hygiene and water and sewage supply first.

Ferro-Luzzi: What micronutrients does Lindsay think are important in affecting BMI?

Allen: I think they cause stunting rather than low BMI, and we could overestimate the energy deficiency.

Scrimshaw: Zinc is one factor affecting growth, but the data are not conclusive: iron supplementation studies stimulate an increase in growth.

Appendix 1.

ANTHROPOMETRY OF HOUSEHOLD MEMBERS AND
RELATED SOCIOECONOMIC AND HOUSING CONDITIONS

QUESTIONNAIRE No.

Department: __________________ District: _______________________
Department: __________________ Village: _______________________

Name and complete address: ____________________________________________

Date of survey: ______________________________________________________

HEAD OF THE FAMILY:

Main activity: __________________________________________________

Ethnic group: ____________________________________________________

Annual income of all the household members (together):_____________

Area of the house: m² _____________________

Materials:

Roof: 1. tiles; 2. straw; 3. zinc; 4. wooden slats; 5. Other _________________
Walls: 1. brick; 2. straw; 3. zinc; 4. planks; 5. Other _____________________

Facilities:

1. tap within; 2. tap outside; 0. none. _________________________________
Is there a septic pit? 1. Yes; 0. No ___________________________________
Are there any drains for sewage water? 1. Yes; 0. No. ____________________
Electricity: 1. Yes; 0. No. __________________________________________
Cooking: 1. Yes; 0. No. ____________________________________________
Is the house along an asphalted road? 1. or along a sandy way? ______________

GOODS OF THE HOUSEHOLD:

Appendix 2.

CHARACTERISTICS OF THE INDIVIDUALS OF THE SELECTED HOUSEHOLD

Contents - Previous - Next