Making the best use of new technologies in the National Diet and Nutrition Survey: a review

.Background Dietary assessment is of paramount importance for public health monitoring. Currently in the UK, the population’s diets are examined by the National Diet and Nutrition Survey Rolling Programme (NDNS RP). In the survey, diet is assessed by a four-day paper-based dietary diary, with accompanying interviews, anthropometric measurements and blood and urine sampling. However, there is growing interest worldwide in the potential for new technologies to assist in data collection for assessment of dietary intake. Published literature reviews have identified the potential of new technologies to improve accuracy, reduce costs, and reduce respondent and researcher burden by automating data capture and the nutritional coding process. However, this is a fast-moving field of research, with technologies developing at a rapid pace, and an updated review of the potential application of new technologies in dietary assessment is warranted. This review was commissioned to identify the new technologies employed in dietary assessment and critically appraise their strengths and limitations in order to recommend which technologies, if any, might be suitable to develop for use in the NDNS RP and other UK population surveys. Objectives The overall aim of the project was to inform the Department of Health of the range of new technologies currently available and in development internationally that have potential to improve, complement or replace the methods used in the NDNS RP. The specific aims were: to generate an itinerary of new and emerging technologies that may be suitable; to systematically review the literature and critically appraise new technologies; and to recommend which of these new technologies, if any, would be appropriate for future use in the NDNS RP. To meet these aims, the project comprised two main facets, a literature review and qualitative research. Literature review data sources The literature review incorporated an extensive search of peer-reviewed and grey literature. The following sources were searched: Cochrane Database of Systematic Reviews (CDSR), Database of Abstracts of Reviews of Effectiveness (DARE), Web of Science Core Collection, Ovid MEDLINE, Ovid MEDLINE In-Process, Embase, NHS EED (Economic Evaluation Database), National Cancer Institute (NCI) Dietary Assessment Calibration/Validation Register, OpenGrey, EPPI Centre (TRoPHI), conference proceedings (ICDAM 2012, ISBNPA 2013, IEEE Xplore, Nutrition Society Irish Section and Summer Meetings 2014), recent issues of journals (Journal of Medical Internet Research, International Journal of Medical Informatics), grants registries (ClinicalTrials.gov, BBSRC, report), national surveys, and mobile phone application stores. In addition, hand-searching of relevant citations was performed. The search also included solicitation of key authors in the field to enquire about Making the best use of new technologies in the NDNS: a review 4 as-yet unpublished articles or reports, and a Bristol Online Survey publicised via social media, society newsletters and meetings. Literature review eligibility criteria Records were screened for eligibility using a three-stage process. Firstly, keyword searches identified obviously irrelevant titles. Secondly, titles and abstracts were screened against the eligibility criteria, following which full-text copies of papers were obtained and, in the third stage of screening, examined against the criteria. Two independent reviewers screened each record at each stage, with discrepancies referred to a third reviewer. Eligibility criteria were pre-specified and agreed by the project Steering Group (Section 1.6). Eligible records included: studies involving technologies, new to the NDNS RP, which can be used to automate or assist the collection of food consumption data and the coding of foods and portion sizes, currently available or beta versions, public domain or commercial; studies that address the development, features, or evaluation of new technology; technologies appropriate for the requirements of the NDNS RP in terms of nutritional analysis, with capacity to collect quantifiable consumption data at the food level; primary sources of information on a particular technology; and journal articles published since the year 2000 or grey literature available from 2011 onwards. The literature search was not limited to Englishlanguage publications, which are included in the itinerary, although data were not extracted from non-English studies. Literature synthesis and appraisal New technologies were categorised into eleven types of technology, and an itinerary was generated of tools falling under each category type. Due to the volume of eligible studies identified by the literature searches, data extraction was limited to the literature focussing on selected exemplar tools of five technology categories (web-based diet diary, web-based 24- hour recall, handheld devices (personal digital assistants and mobile phones), nonautomated cameras to complement traditional methods, and non-automated cameras to replace traditional methods). For each category, at least two exemplars were chosen, and all studies involving the exemplar were included in data extraction and synthesis. Exemplars were selected on the basis of breadth of evidence available, using pre-specified criteria agreed by the Steering Group. Data were extracted by a single reviewer and an evidence summary collated for each exemplar. A quality appraisal checklist was developed to assess the quality of validation studies. The checklist was piloted and applied by two independent reviewers. Studies were not excluded on the basis of quality, but study quality was taken into account when judging the strength of evidence. Due to the heterogeneity of the literature, meta-analyses were not performed. References were managed and screened using the EPPI Reviewer 4 systematic review software. EPPI Reviewer was also used to extract data.