Article Text

Download PDFPDF

OP39 Development of a short food frequency questionnaire to assess diet quality in population studies
  1. SR Crozier1,
  2. SM Robinson2,3,
  3. S Shaw1,4,
  4. HM Inskip1,4,
  5. J Baird1,4,
  6. C Cooper1,4,
  7. C Vogel1,4
  1. 1MRC Lifecourse Epidemiology Unit, University of Southampton, Southampton, UK
  2. 2AGE Research Group, Newcastle University, Newcastle upon Tyne, UK
  3. 3NIHR Newcastle Biomedical Research Centre, Newcastle upon Tyne Hospitals NHS Foundation Trust and Newcastle University, Newcastle upon Tyne, UK
  4. 4NIHR Southampton Biomedical Research Centre, University of Southampton and University Hospital Southampton NHS Trust, Southampton, UK


Background Food frequency questionnaires (FFQs) are a popular tool in nutritional epidemiology, enabling estimates of habitual diet in large populations, but are time-consuming to complete. There is an increasing need for a short, accurate dietary tool that characterises healthy dietary patterns for use in observational and interventional research.

Methods The National Diet and Nutrition Survey (NDNS) is a general population national survey. Randomly-selected participants from across the UK were asked to keep a diary of everything they ate and drank over four days. Diaries were coded and intake frequencies of foods derived using groups of nutritionally similar foods. For the 2% of participants who had three-day diaries, the frequencies were multiplied by 4/3. Principal component analysis (PCA) was performed on frequencies of consumption. Reduced-item prudent diet scores were calculated by multiplying coefficients for the ten most characteristic foods by each individual’s standardized reported frequency of consumption. Prudent diet scores calculated for both the full and 10-item prudent diet scores were standardised. Willing participants provided a blood sample; vitamin C status was measured on the BMG Labtech FLUOstar OPTIMA plate reader and beta carotene status by high performance liquid chromatography.

Results 6090 adults (aged 20 to 96 years) provided dietary data in the first eight years of the NDNS (2008–2016). PCA of intake frequencies of 126 food groups revealed a prudent diet pattern characterised by high intakes of wholemeal bread, vegetables, fruit, water, oily fish, yoghurt and high-fibre breakfast cereals, and low intakes of white bread, sugar, sugar-sweetened soft drinks and chips. Pattern coefficients were very similar when analyses were conducted separately in males and females, and in those aged less than 65 and 65+ years. The 10-item prudent diet score was calculated based on foods with the five highest and five lowest coefficients. Spearman’s correlation between the full and 10-item prudent diet scores was 0.84. The mean difference between the full and 10-item prudent diet scores was 0.00SDs with Bland-Altman limits of agreement -1.89 to 1.89SDs. Notable correlations were seen between the full prudent diet score and vitamin C (rs=0.43) and beta-carotene status (rs=0.45); these were only slightly attenuated for the 10-item prudent diet score (rs=0.40, rs=0.38 respectively).

Conclusion A 10-item prudent diet score based on foods that characterise the prudent dietary pattern describes a healthy diet. A 10-item FFQ would have clear advantages for time and resources, and may provide an appropriate tool to describe UK diets while reducing participant burden.

  • diet
  • food frequency questionnaires
  • principal component analysis

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.