This is the old United Nations University website. Visit the new site at http://unu.edu
Abstract
Introduction
Group 1. Nutritional status and biological variables
Group 2. Food intake and food security variables; food behaviour and patterns variables
Group 3. Health status assessment variables
Group 4. Socio-demographic and social behaviour variables
Group 5. Sampling frame and enrolment procedures for sites and individuals within sites
Conclusions
Reference
Rainer Gross and Bridget H.-H. Hsu-Hage
Rainer Gross is affiliated with the Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ) in Jakarta, Indonesia. Bridget Hsu-Hage is affiliated with the Department of Medicine at Monash University in Clayton, Victoria, Australia.
Mention of the names of firms and commercial products does not imply endorsement by the United Nations University.
The process of revising the protocol, based on experience in the pilot phase and logical considerations, was conducted by working groups composed of persons directly involved with the field collection of data and professionals who were experts in specific disciplines although not part of the actual pilot research. Four of the working groups addressed the revision of the specific batteries of questions and measures in the protocol, while the fifth group addressed the sample size and sampling frame. The process led to consolidating and condensing many of the questions into single-line items, to recommendations for some additional questions based on common experience, to requirements for standardization of procedures and measuring instruments, and to explicit sequences for enrolling the population. This effort resulted in the revised protocol
At the 1993 meeting in Wageningen, Netherlands, all parties from all field sites presented their actual experiences and findings from the pilot studies performed with the Reconnaissance protocol and with judgments on the utility and feasibility of the various measures and interview questions. They then began the serious process of editing and revising the protocol for the definitive CRONOS (Cross-Cultural Research on the Nutrition of Older Subjects) phase. Five working groups were assigned the responsibility of editing and revising specific sections of the protocol. Their major conclusions are given below.
It was the general impression that, despite the lack of inter-centre standardization exercises, the biological data collected were consistent, both within centres and across centres, with acceptable values and variance. The 14 anthropometric variables, 2 blood-pressure variables, and 1 haematological (blood) variable contained in the original Reconnaissance manual were deemed appropriate and feasible.
Efforts were made to standardize the instrumentation, the measurement procedures, and the expression of the values for the definitive study. The project manual for the Reconnaissance phase itself provided no instruction for the procedures, and the manual of operation of the EURONUT-SENECA project, edited by Drs. de Groot and van Staveren, that was distributed to all of the centres as the guidebook for anthropometric measurements also provided only limited detail. The draft report of the WHO Subcommittee, “Uses and Interpretation of Anthropometry in the Elderly for the Assessment of Physical Status,” contains much more detailed instructions but basically agrees with the methods used in EURONUT-SENECA and the IUNS study. It was decided to incorporate the more detailed instructions into the manual for the CRONOS phase of the study, with some modifications.
It was concluded that a lightweight dressing gown with slits in the sides would be acceptable in all cultures for making the anthropometric measurements, but in some sites the examiner must be of the same sex as the subject. It was recommended that a weighing scale, knee-height caliper (Ross Laboratories, Columbus, OH, USA), limb circumference tape, and skinfold calipers be available at all sites.
In the original design as specified in the Reconnaissance manual, collection of food intake data was optional for each centre, and if it was done, the methods and instruments were at the discretion of the investigators. As most centres were institutes or departments of nutrition, some attempt to quantify nutrient intake was made at all centres except in Guatemala, although the full disaggregated dietary data were reported by only two centres.
The Reconnaissance protocol contained questions related to food behaviour in a more general sense, but not to the quantitative aspects of exact intakes of specific foods and beverages, micronutrient densities, or the amounts of nutrients and other constituents in the diets of individuals and subgroups. It was recognized that a single 24-hour intake report is not necessarily representative of the usual diet of an individual. At least two random 24-hour recalls or a food-frequency instrument would be needed for greater within-subject stability. However, the practicality of developing and validating such instruments for all sites was low. At least the procedures of a 24-hour recall are universally known. Several other problems remained. There was a question as to which food composition tables should be used. Many food items are native only to the country studied, but the accuracy of the analyses that produced the nutrient-content values from site to site is unknown. In older people, memory of food intake over even the last 24 hours may be unreliable. Because a parallel set of younger adults is included in the CRONOS design, a differential degree of accuracy across generations may confound comparative interpretation by age in the final studies.*
*The UNU/FAO-sponsored International Network of Food Data Systems (INFOODS) is assisting countries and regions to assemble the best available databases for their use, to improve the quality of food composition data, and to make these data freely accessible. Nearly all regions of the world are now covered by the INFOODS network. Information on the current status of food composition databases for any country or region can be obtained from http://www.crop.cri.nz/foodinf/infoods/infoods.htm.Despite these limitations, the working group recommended that a dietary method, specifically a 24-hour recall, be included as part of CRONOS. It would remain optional but would be strongly encouraged.
This group considering the health status assessment variables grouped them into six categories: (1) cognitive function, (2) self-perceived health, (3) illness profile, (4) impairments, (5) functional status, and (6) health risk factors. Minor modifications to variables, with some deletions, were made among the first three categories. In terms of functional status, to make the process even more homologous with the other major survey, questions related to activities of daily living (ADL) from the EURONUT-SENECA manual were added, although it was recognized that the questions were not always relevant to populations in developing countries. As an optional procedure, the physical performance test for the elderly currently used in the United States by the National Health and Nutrition Examination Survey III (NHANES III) was added. In the sixth category, a question on recent weight loss was added.
The group considered socio-demographic and social behaviour and practices variables in addition to some minor categories entitled “soft drug use” and “disabilities.” These constituted a total of 39 of the 85 items in the original Reconnaissance manual, but they were dispersed throughout the manual. It was decided that these represented a logical bloc that should be consolidated into two adjacent sections. In the revised protocol, the demographic section was restructured and extended. A question on the place of birth was added. The subject’s religion, formerly included in the social practices section, became more reasonably part of the initial demographic set of questionnaire items. Finally, more detailed information on the family network of first-, second-, and third-order relatives and in-laws was gathered.
For income and employment issues, 5 rather vague questions in the original protocol were converted into tight structures of 16 pre-coded items in the questionnaire under the heading of “household income”; 8 were information-seeking questions and X were derivative calculations. The one item on educational attainment was also included in this section.
A new section and set of questions were included on “living conditions,” including construction of the home, sources of water and fuel, and facilities for food storage and preparation.
The social practices section was generally left intact, with some disaggregation of formerly combined questions. The emphasis on social activities both within the family, such as babysitting, and outside the home, such as gathering with friends or attending a cultural event, was retained. It is realized that for some of the activities listed, opportunities exist only in urban areas. With the revision, a total of 57 items, 49 questions, and 8 calculations replaced or extended the original 39 items in the demography and social aspects domains of the questionnaire.
It was felt that the procedure for enrolment of subjects must be comparable in all countries and that the appropriate geographic and socio-economic contrasts must be recognized. For instance, in China it was reported that the investigators could not separate urban areas into those of lower or higher income. In Kuala Lumpur, Malaysia, rich and poor people live in the same neighbourhoods but have different-quality housing. In Guatemala the urban upper-income adults were selected from several different neighbourhoods. At the community level, it is now specified that communities of at least 6,000 people will be selected. In a change to facilitate enrolment, the width of the age bands was increased from 5 to 10 years for middle-aged adults (defined as 35-44 years of age) and from 10 to 15 years for the elderly (defined as 60-74 years of age).
According to these fortified and tightened sample selection and enrolment procedures, the persons to be studied in each community will be selected by the following steps: (1) A map of the community will be obtained (or made) showing the blocks, streets, and households. The blocks and then the households will be numbered consecutively, starting from a randomly selected corner. (2) A census of the entire community will be made using family composition questions. (3) All eligible elderly people will be numbered consecutively in ascending order, corresponding to the number of the household; if there is more than one elderly person in a household, the younger will be enumerated first. (4) A systematic sample of elderly people will be taken, using an appropriate interval to select 150 people. The first 50 of these will be approached first, and those who reject participation in the study will be replaced from the remaining 100 until 100 elderly people (50 men and 50 women) have been selected. (5) An identical procedure will be used to enrol the middle-aged adult controls in each community.
The original plan of the multicentre project, Nutritional and Food Security Situation of Rural and Urban Elderly from Selected East Asian and Latin American Developing Countries, was to obtain a full sample of 600 people per site and analyse the results individually, collectively, and comparatively, in order to better understand the influences of ageing and of family migration to urban centres on the nutrition and health of low-income populations in developing countries in selected countries in Latin America and Asia. From the eight countries on two continents, broad insights into senescence and urbanization were expected. The change in plans converted the project into a pilot study, and the format of a pilot study allowed the opportunity for midcourse evaluation. The final phase of the evaluation was the redrafting of a protocol for a revised study, Cross-Cultural Research on Nutrition of Older Persons (CRONOS) [1].
1. Gross R. ed. CRONOS (Cross-Cultural Research on the Nutrition of Older Subjects). 3rd ed. Food Nutr Bull 1997;18:267-303.
