View Report Details

Calculation and collation of typical food portion sizes for adults aged 19-64 and older people aged 65 and over

Project Code: N08026

30/09/2005

School of Pharmacy and Life Sciences, Robert Gordon University, St Andrew Street, Aberdeen, AB25 1HG
Wrieden, W ;
Centre for Public Health Nutrition Research, Division of Clinical & Population Sciences & Education, University of Dundee, Ninewells Hospital and Medical School, Dundee DD1 9SY
Barton, K

1.1 Aims and Objectives

The overall aim of the study was to produce a set of typical food portion weights for younger and older adults.

The individual objectives were:
To extract food portion information from recent dietary surveys of adults (National Diet and Nutrition Survey (NDNS) of adults aged 19-64 years and people aged 65 years and over).
To collate current information on portion sizes of packaged foods and fast foods commonly eaten.
To produce a list of typical food portion sizes for each age range.
To trial the use of typical food portion weights in dietary surveys.
To produce list of portion sizes for foods for adults and older people.

1.2 Approach

1.2.1 Calculation of Typical Portion Weights

NDNS databases were obtained and food portion data was extracted. Similar foods (e.g. Cornflake type cereals) were grouped and re-coded in order to facilitate processing.
A comprehensive analysis of the factors affecting portion size, including age, gender, illness, time of day and (in the elderly) free-living or institutional was then carried out. Age was grouped by widely used categories (19-34, 35-49, 50-64, 65-74 and 75+). The results of the analysis showed that it was necessary to restrict the number of foods for which portion size is recorded for the gender and age groups as some numbers were insufficient to determine a reliable estimate of portion size. The data were highly variable, with a significant number of extreme values. These could not be investigated, nor could they be deemed as clearly erroneous. For this reason median portion sizes were calculated and non-parametric methods were used to look at the factors that caused a significant difference in portion weights. Where these differences occurred and numbers were sufficient individual medians were calculated for age group and / or gender. In addition where there were significant differences in portion weight between free-living and institutionalised subjects over 65 year olds, a separate portion weight was estimated.

1.2.2 Testing of Typical Portion Weights

To test the use of the calculated portion sizes, weighed food diaries of Scottish mothers aged 25-46 years (n 35) collected 1996-1997 and men and women aged 40-74 (n 64) collected 1999-2002, were reanalysed for energy and nutrients using the actual and calculated weights (medians) for each food. The methods of Bland and Altman were used to investigate the agreement between nutrient intakes calculated from actual and calculated portion weights.

1.2.2 Manufactured and Catered Portion Weights

A range of companies including food manufacturers, supermarkets and fast food/ restaurant chains were contacted and information on portion sizes obtained. Supermarkets were visited and an extensive search of websites carried out to obtain further information.

1.3 Key Findings

In total, information on 3411 individuals, 2999 free-living and 412 institutionalised, was available from the two surveys (age and gender data not available for 94 individuals in the 19-64 age group). Similar foods were grouped into 751 food groupings. Some of the food groupings with larger numbers were broken down according to mode of consumption upon further examination e.g. milk in tea/coffee, in cereal, as a drink etc. No cut-off points for inclusion were used and all similar foods regardless of the number of times they had been consumed were included in the food groupings. Those food groupings with 10 or more consumers (in either of the 19-64 or 65+ NDNS databases) were included in the list of foods for which portion sizes were calculated (food groupings with less than 10 consumers in either NDNS database were coded as insufficient numbers).

Mean daily energy and nutrient intakes were calculated from the food diaries using calculated median portion weights (using an age/gender specific portion weight where available). Differences between the energy and nutrient intakes from the actual and calculated portion weights were small for females but considerably larger for males but most individual values lay within 2 standard deviations of the mean difference.

Portion weights collected for manufactured foods, included bread, biscuits, ready meals, burgers, sausages, meat pies, yogurts, savoury snacks, confectionery and soft drinks.

1.4 Technical Evaluation and Interpretation

Other work hasfound that standard portion sizes gave lower mean estimates of energy and nutrient intakes than reported portion sizes, with greater differences in men than in women. This was despite a much larger sample size than in the current project. A suggested method was given to adjust portion sizes which could be further explored.

It is recommended that further work is carried out to address any anomalies due to small number of diaries being tested and to cover more of the foods for which weights have been calculated. The calculated portion sizes, together with the extensive list of manufactured and catered portion sizes will be useful in assessing the diets of groups of adults. In addition they can be used as a guide for researchers devising dietary assessment tools for adults.

Some of the files on this site may be in a format that your computer can't read. However, you can download Readers and Viewers for the following document types below: