Degree Name

Doctor of Philosophy


Smart Foods Centre, School of Health Sciences - Faculty of Health & Behavioural Sciences


Dietary assessment has changed dramatically with time, progressing from face-to-face interviews and hand calculated nutrient intakes to the use of computer technology to automate various parts of the process. The most common application is the use of software packages to calculate nutrient intake data obtained from dietary interviews. The development of technology to automate the interview process will allow for clinicians to spend more time focussing on patient education and counselling. The central hypothesis tested in this thesis was that automated dietary assessment would prove to be a feasible adjunct to the professional consultation in the primary healthcare setting. Development phase A series of studies were conducted examining various aspects of computer-assisted survey technology (CAST) applied to dietary advice in the primary healthcare setting. The research is presented as a case study, using action research methodology. Items in the dietary survey were developed from data reduction of food lists reported in the 1995 Australian National Nutrition Survey (NNS95), in conjunction with professional interpretation and judgement. The opinions and beliefs of patients from focus group interviews shaped development of the user interface and a dynamic website was developed to best allow for a diversity of eating patterns. Testing phase Video-recorded usability testing found the website to be user friendly with the time taken to complete the survey comparable to the time taken for a dietitian to interview and assess a patient’s food intake. The website was then implemented in the primary healthcare setting over a period of twelve months. Computers were set-up in fourteen medical practices in the Illawarra region of NSW, Australia. Doctors recruited patients with metabolic syndrome to use the website. Data was sent to a dietitian in the research team for development of an individualised dietary prescription, which was then sent back to the doctor to discuss with the patient. Implementation phase A cross-section of 200 patients revealed the majority of users were aged between 46 and 65 years, overweight and physically inactive. Computer ownership was identified in 80% of the users, with only 8% of patients having never used a computer previously. The computer located in the medical practice was the least preferred location of use and patients with a higher BMI were 1.9 times (p=0.04) more likely to use the computer in the home or an alternate location than at the medical practice. Reported nutrient data was highly variable. Under-reporting was observed in 46 patients (32.2%), overreporting in 31 (21.7%) of patients and 66 patients (46.2%) reported their intakes on target. No relationships were found for the reporting status of the patients and their age, BMI or gender. A repeatability study with n=38 patients revealed a learning effect which led to increased understanding of the website functions with time. Compared to a 3-day weighed food record, data from the website produced stronger correlations than a face-to-face diet history assessment. Patients using the website achieved an average 25% of their dietary goals within six weeks, despite a preference for face-to-face contact with the dietitian. Stakeholder evaluation established acceptance of the technology by dietitians, doctors and patients and provided insights into their positions within the healthcare system. Evaluation phase The research found that computerised assessment of dietary intake was a feasible addition to daily practice in the primary healthcare setting. Automating the diet history interview via the internet allowed increased patient access to dietitians whilst improving the doctors’ awareness of the nutrition needs of their patients. This is especially important in the growing light of metabolic syndrome worldwide.

02Whole.pdf (8079 kB)