Leading articles
Limitations in the use of the NNS data
(See paper by Cobiac et al., pages 152–173)

The information collected in a national dietary survey is a compromise between the ideal and the feasible. The United Kingdom national dietary survey collects seven days of weighed food records from each adult (1) and less from children. The Australian 1995 National Nutrition Survey collected a single 24-hour recall from each indi-vidual and a second 24-hour recall from a 10% subsample (2). Consequently, a more limited range of analyses are possible with our data than with the British data.

What can we do and what can we not do? The decision not to collect repeat dietary data on everyone in the survey is a major restriction; it means that the usual intake of each individual cannot be determined. Luckily, the usual intake of the population can be approximated even when the usual intake of each individual is not known (3). The method used by the NNS—the second day on a subsample— allows the population standard deviation to be largely corrected for the day-to-day variation in intakes. The correction procedure involves multiplying each individuals one-day intake by a constant ratio, (square root of (between person variance/total variance)). This might seem to imply that the intake of each individual is corrected as well but this is not so. Although it moves each individual towards the centre, it does not alter the ranking of individuals relative to each other. As Cobiac et al. note (4), individuals in the lowest tertile for the unadjusted distribution are still in the lowest tertile after adjustment. If this adjustment really did yield the usual intakes of individuals then there would have been substantial reordering of the population with people changing quantiles. Beaton et al. found that a single day of intake for carbohydrate has an error of ±59 % in men and ±72% in women (5,6). Therefore any particular man’s true usual intake could be less than half or more than 150% of his one-day estimate. Although there is general movement towards the population mean, the movement varies among individuals and many people will move away from the population mean. The population adjustment procedure uses the average movement only.

Because the population adjustment method does not alter the ranking of individuals, dividing a population into tertiles using data adjusted this way, as was done in the paper (4), is tantamount to dividing the population into tertiles using the unadjusted data. It does not correct for misclassification at the individual level and the literature warns against assuming that it does (7,8). In fact, if this procedure were appropriate for obtaining the usual intake of individuals, it could be used instead of food frequency questionnaires in epidemiological studies and also for val-idating/ calibrating food frequency questionnaires instead of laboriously collecting large numbers of food records/recalls from each person (9–11).

What is the consequence of dividing up a sample using a characteristic with a large amount of measurement error? It makes the groups more similar as regards all other factors than would be the case if the misclassification had been corrected. Regression slopes and differences, and to a lesser extent correlations, calculated across the groups will be substantially attenuated (3,12,13). Hence the lack of variation in health parameters across tertiles of sugar intake (4) may be due to attenuation masking real associations or it may be due to true non-association but it is not possible to say which. An alternative approach would be to compare the mean of the highly variable item for categories created using a variable with less intra-individual variance such as body mass index, smoking, age or SEIFA index.

The Australian Bureau of Statistics (ABS) has produced some, but not all, of the adjustment factors one might like to use when analysing the data. The authors indicate that they did not calculate their own factors (even though it is feasible with the data) but used the values produced by the ABS. Most of their results are based on percent of energy from various sugars (4) but the ABS has not produced adjustment factors for nutrient density. It is unlikely that the adjustment factor for the percentage of energy from a macronutrient can be calculated by dividing the factor for the nutrient by the factor for energy. Beaton et al. for example, found that the between/total ratio was 0.61 for total carbohydrate, 0.50 for energy and 0.69 for percent of energy from carbohydrate in men (5). Moreover, the assumption in the paper that an adjustment factor for a nutrient can also be used for subcomponents of a nutrient (4) is challenged by Beaton’s finding that the ratios for starch and non-starch carbohydrates are different from the ratio for total carbohydrate (5). There are also some unexpected results in this paper (4). For example, 0% of adult women had two-day adjusted intakes below 70% of the RDI for vitamin A (525 micrograms) whereas Table 76 of the ABS publication shows that this should be more than 10% (2).

It is, however, appropriate to use the adjusted intakes of the population to calculate proportions above or below external criteria. In the last decade there has been a sea change in how external standards, the recommended dietary intake (RDIs) or its equivalent in other countries, are described. Previously, a single figure for each nutrient described the amount thought to meet the needs of most of the population. RDIs were often said to be set at the average requirement plus two standard deviations, but inspection of the background papers for our current set (14) shows that this was not the case. This means that criteria such as 70% of the RDI have different meanings for each nutrient.

In the mid-1980s a committee, which included George Beaton, one of the great minds in nutrition, described how the prevalence of inadequate intakes could be calculated (using an approximating method that allowed for the low-level computer power of the time). This approach is called the probability approach (3) and, among other things, requires knowing the estimated average requirement (EAR). Perhaps this spurred countries such as the United Kingdom (15) and the United States (16) to publish their EARs when revising their dietary references. The US publication also describes a short-cut version of the probability approach—the EAR cut-off method—which can be used if certain assumptions are met (16). To calculate the prevalence of inadequate dietary intakes in the population, one simply calculates the proportion with usual intakes below the EAR. The proportion is correct even though many of the individuals in the population are classified incorrectly (16). The RDI does not come into the calculation. It should be noted that these probability-based calculations are good only for calculating population statistics. They do not allow a determination as to which individuals in the survey were below the cutoff (16). They cannot be used to divide up the survey sample into groups with adequate and inadequate intakes.

The real problem is determining the EAR. There are some intriguing discrepancies between the UK and US documents (15,16) but these do not mean the EAR are less accurate than the RDI/RDA because the RDAs are derived from the EAR. An interesting new feature of the US values is the distinction between adequate intakes and RDAs. Adequate intakes are set when there is not enough information about requirements to determine the EAR (16).

Unfortunately, most of the Australian background papers (14) do not allow us to know what the committee thought the EARs were. If we wish to analyse our data to get good estimates of the prevalence of inadequate intakes, we are forced to choose between the various available overseas EAR values. At the moment, it seems likely that the Australian RDI may be revised. Overt specification of the assumed EARs would allow better use of the data that we have.

Dorothy Mackerras
Senior Research Fellow
Menzies School of Health Research
Darwin, Northern Territory

Ingrid Rutishauser
Honorary Fellow
School of Health Sciences
Deakin University
Geelong, Victoria

Leading articles
Blinded by science - The National Heart Foundation of Australia’s position statement on dietary fat and overweight/obesity
(see pages 174–176)

There can be no doubt that the problem of overweight and obesity has now become one of the most important public health issues for Australia. Figures from the 1999–2000 AusDiab study indicate that the prevalence of obesity throughout Australia has risen further so that 19.1% of men and 20.1% of women are now obese (1). In addition, over half of all adult females and 60% of all adult males are now classified as overweight. More alarming is the fact that in 1995 around one in five schoolchildren in Australia were overweight or obese (2). This situation is even more disturbing within the indigenous population where a 1994 survey found that around 25% of the Aboriginal population and almost half of the people from the Torres Strait Islands were obese (3).

This rapid increase in the levels of obesity has the potential to profoundly affect the current and future health of all Australians. Overweight and obesity have been linked to a number of serious chronic health problems such as diabetes, hypertension, cardiovascular disease and cancers (4). In addition, excessive weight is a major con-tributor to reduced quality of life through its impact on back pain, osteoarthritis, breathing difficulties, sleep apnoea and increased psychosocial problems (4). Cardiovascular disease rates in Australia have been falling consistently since the mid-1970s but it remains one of the major causes of premature death and morbidity in Australia (5). Better control of risk factors such as high blood cholesterol and hypertension, as well as lifestyle change and a reduction in smoking rates, have been iden-tified as key contributors to this reduction in cardiovascular disease rates (6). However, overweight and obesity are also major risk factors for the development of cardiovascular disease and there is a real concern among heart disease experts around the globe that the continued rise in rates of overweight and obesity has the potential to undermine their efforts to control this disease (7). In line with this concern the National Heart Foundation of Australia decided to undertake a review of the scientific literature to assess the relationship between dietary fat intake and overweight and obesity and produced a statement that is published in this issue (8). The statement focusses on the relationship between dietary fat and overweight/ obesity, as the Heart Foundation’s current nutrition guidelines, which recommend a reduction in saturated fat but not total fat, did not include the potential for dietary fat to influence cardiovascular risk through promotion of obesity.

Establishing associations between diet and disease and assessing the true impact of dietary interventions is fraught with a number of complications (9). In particular, assessing the dietary behaviour and patterns of free living individuals and populations and attempting to define associations in relation to weight status is subject to an enormous array of confounding factors. In an attempt to isolate and study the impact of a single factor, researchers are often required to conduct their experiments under totally artificial situations and deliberately manipulate the composition of foods which creates a situation where the physical and food environment bears little resemblance to the real world. For example, it is possible to manipulate the fat content of food products while keeping the energy density and palatability constant (10). Short-term feeding trials that adopt this approach to identify the influence of one factor, independent of all others, are a useful scientific tool but this process rarely reflects real life. The findings of such research, therefore, need to be interpreted within that context and as part of the wider body of scientific research. Science provides the foundations for appropriate nutrition policy development but policy makers must take care that they are not blinded by science to the realities of the prevailing food environment. Unfortunately, after assessing the National Heart Foundation of Australia’s position statement on dietary fat and overweight/obesity it is hard to escape the conclusion that this is precisely what happened.

Research has indicated a number of potential mechanisms by which dietary fat could influence weight gain and the development of obesity. Dietary fat increases the energy density of food (together with water it accounts for almost 99% of the variance in energy density of foods) (11). Reducing the total fat in the diet is a useful method of reducing the energy density and has been asso-ciated with reduced energy intake, modest weight loss and prevention of weight regain (12). In comparison to carbohydrate and protein, fat has a poor ability to suppress hunger (satiety) and bring episodes of eating to an end (satiation) (13). The capacity for fat storage in the body is virtually unlimited but the ability of fat to stimulate its own oxidation in response to increases in fat intake is very poor when compared against other macronutrients (14). Most of these effects of dietary fat were identified in the Heart Foundation's review of the literature but their relevance to the formulation of the position statement was rejected because the objectives of the review stated that the effect of fat needed to be independent of energy intake. The rationale for this objective is perplexing as, although a wide array of potential factors has been implicated in the development of obesity, they all must exert their influence on energy balance through either energy intake or energy expenditure. There is some evidence that dietary fat may exert some small effect on energy expenditure (through efficiencies in storage for example) but its major influence on energy balance is always going to be mediated through energy intake. From a policy perspective it is not important whether the effects of dietary fat on weight gain are mediated through energy density or other mechanisms. The more important question is whether dietary fat intake contributes to weight gain and whether reducing total dietary fat assists weight loss and the prevention of weight gain. Applying the same principles it should not be expected that under controlled situations that reducing dietary fat would result in a greater weight loss than reducing energy intake as this is the mechanism by which a low fat diet impacts on energy balance.

Removing the limitations imposed by the original objective of the review, it is difficult to see how a role for total dietary fat in the prevention of weight gain and treat-ment of overweight and obesity can be rejected on the evidence presented by the Heart Foundation’s own review. It is therefore disappointing to see the final policy statement make a definite recommendation to avoid a focus on reducing total dietary fat based on tenuous logic. The evidence presented to support a return to a level of dietary fat intake equivalent to 35% of total dietary energy is sparse and the assertion that polyunsaturated fatty acid intakes are decreasing and saturated fat intakes are increasing in the Australian diet is not supported by the reference cited (15).

If this statement had been prepared and released by any other organisation then it may soon be forgotten. However, the National Heart Foundation of Australia has tremendous respect and leverage within community and political dimensions. This current position statement on dietary fat and overweight/obesity will be perceived as being at odds with the recommendations of a wide range of national and international organisations including the National Health and Medical Research Council’s guidelines on prevention of diabetes and the treatment of obesity and the Dietary Guidelines for Australian Adults (16–18). The inevitable controversy that this position statement will generate is a further illustration of the often long and torturous passage between science and policy and the problems that may beset those who attempt to shortcut this process.

Tim Gill
NSW Centre for Public Health Nutrition
University of Sydney

Original research
Sugars in the Australian diet: Results from the 1995 National Nutrition Survey
Lynne Cobiac, Sally Record, Phil Leppard, Julie Syrette and Ingrid Flight

Abstract (Nutr Diet 2003;60:152–73)
Objective: To estimate the dietary intakes of total, added and natural sugars (monosaccharides and disaccharides) for both adults and children, and to characterise differences between low, moderate and high consumers of these sugars.
Design: The 1995 National Nutrition Survey is a cross-sectional study using 24-hour dietary recall methodology.
Subjects: 3441 children aged two to 18 years; 10417 adults aged 19 years and over (weighted sample count).
Setting: Australian population sampled nationally.
Main outcome measures: Intakes of total, added and natural sugars; food sources of each type of sugar; tertiles of intake density (%E) of each sugar; characterisation of respondents in each sugars tertile in relation to food and nutrient density (per 1000kJ), percentage of children and adults with intakes below 70% of the RDI; differences in BMI, blood pressure, exercise, and self-reported health status.
Statistical analyses: Linear regression analyses were conducted separately for boys, girls, men and women to test for differences in nutrient and food intake densities across the tertiles. Associations between sugars intakes and BMI were investigated with linear and logistic regressions after adjusting for age, exercise level, total energy intake and whether or not individuals were on a weight-reducing diet or classified as an under-reporter. Similar regressions were performed with systolic and diastolic blood pressures, exercise and self-reported health status.
Results: Total, added and natural sugars contributed 22%, 11.2% and 10.5% respectively to total energy intake for the whole population. Non-alcoholic beverages were the major sources of total and added sugars; and milk and fruit products of natural sugars. The only consistent evidence of nutrient dilution (zinc, iron, magnesium) with higher intakes of sugar was observed for women when intakes were in the upper tertiles (27%E from total and 17%E from added sugars). No significant associations were found between sugars intakes and health variables.
Conclusion: Women consuming very high sugar diets may be increasing their risk of inadequate intakes of some nutrients that are already marginal. Children and men appeared to be at less risk. Intakes of sugars appeared to be poor predictors of health variables.

National Heart Foundation of Australia
Position statement on dietary fat and overweight/obesity

(Nutr Diet 2003;60:174–6)
The National Heart Foundation of Australia’s (NHFA) review of the relationship between dietary fat and cardiovascular disease (CVD) while finding good evidence of a link between the amount of saturated fat in the diet and CVD risk, found little evidence demonstrating that coronary events or death are linked to the amount of total fat in the diet (1). However, it has been suggested that dietary fat intake could increase the risk of CVD indirectly by increasing the risk of overweight and obesity, an independent risk factor for morbidity and mortality related to coronary heart disease (CHD). A review of the relation-ship between dietary fat and overweight/obesity was therefore conducted to determine whether recommendations on total fat intake were required for the prevention of cardiovascular disease.

Objectives

The objectives of the review of the relationship between
dietary fat and overweight/obesity were to:

• Determine whether dietary fat, independent of energy
intake, is a risk factor for the development and pro-gression
of overweight and obesity.

• Assess the effectiveness of fat reduction strategies relative
to other dietary strategies for achieving weight
loss in overweight and obese individuals and weight
maintenance in normal weight, overweight and obese
individuals.

The following summary is based on the evidence contained within the Heart Foundation's background review paper completed in February 2003 (2).

Summary of conclusions

Dietary fat is not an independent risk factor for the development and progression of overweight and obesity (moderate evidence).

Dietary fat may increase the risk of overweight and obesity indirectly by increasing the energy density of the diet, hence facilitating excess energy intake.

The association between dietary fat and obesity reported in cohort studies, conducted in different popula-tion groups, was inconsistent (3–10). Inconsistency in the findings can be partly explained by measurement bias, in particular underreporting of dietary fat intake and confounding from physical activity.

The lack of variation in dietary fat intakes in the study populations and the large within-person variation in dietary intake make it difficult to find an association between dietary fat intake and weight gain in cohort studies (11).

Overall, the effect of dietary fat was small compared to other risk factors for overweight and obesity, such as physical activity level. No study reported a dose response relationship between dietary fat intake and weight gain.

It is unlikely that the metabolic effect of dietary fat on energy expenditure and energy storage increases the risk of weight gain. Several controlled trials in normal weight women and men suggest that high fat diets do not result in excess energy intake when confounding factors, such as energy density and palatability, are held constant (12–14). Several controlled trials in both obese and lean individuals have shown that energy intake is dependent on the energy density of the diet, not the fat content. Energy intake was higher with high energy dense diets (above 6kJ/g) than low energy dense diets (to less than 4kJ/g) (15–18). Since energy density is mainly determined by the water, fat and fibre content of foods, high fat diets may lead to excess energy intake and hence contribute to weight gain. However, other factors such as palatability, the physical form of food, the amount (portion size) and volume of food consumed as well as behavioural and genetic factors, may also influence energy intake.

Energy balance is the major determinant of weight loss. Dietary fat reduction is a simplistic behavioural strategy to facilitate energy restriction. However, without energy restriction, fat reduction alone is not effective for achieving weight loss in overweight and obese individuals (moderate evidence).

Randomised controlled trials in obese individuals reported no significant differences in short-term weight loss between energy-restricted diets (1000 to 1200kcal) varying in fat content from 15% to 75% energy from fat (%E fat) (19–21). An interim systematic review concluded that low-energy, low-fat diets are not more efficacious than low-energy diets, which are not low in fat, in terms of weight loss in overweight or obese individuals (22).

Reducing dietary fat intake to 25%E (30 to 35g/day) appears to be as effective as restricting energy intake to 5040kJ/day for achieving short-term (six months) weight loss in obese women (23). However, a randomised control-led trial suggests that short-term weight loss achieved with both low fat ad libitum dietary interventions (dietary fat reduced to 25%E fat or 30 to 35g fat/day) and energy restricted diets (around 5040kJ/day) in free-living obese individuals is not maintained in the long term, after controlling for confounding factors (24). Other factors, such as physical activity, social support and preference for the dietary regime have been shown to also influence weight loss.

Controlled trials suggest that low-fat ad libitum dietary interventions may reduce energy intake leading to short-term weight loss in overweight individuals (25,26). A meta-analysis of low-fat ad libitum dietary intervention studies, in which weight loss was not the primary aim of the majority of these studies, reported a weighted difference in weight loss between intervention and control groups of 2.55kg (95% CI: 1.5 to 3.5kg; P <0.0001) (23). However, few studies separated the effect of dietary fat from other confounding factors such as physical activity, other dietary factors and behavioural influences.

Randomised controlled trials suggest that there is a wide variation in short-term weight loss achieved in obese individuals on low-fat (30%E fat) ad libitum diets (27,28). Differences in the protein content, and to a lesser extent, the type of carbohydrate in the low-fat ad libitum diet resulted in significant (3.7kg 95% CI: 1.3 to 6.2kg; P =0.0002) and non-significant differences in weight loss, respectively. A randomised crossover study found no significant difference in short-term weight loss in over-weight women on a low-fat diet (22%E fat) compared to those on a high monounsaturated fat diet (36%E fat) (29). It may therefore be possible that dietary fat intakes of up to 35%E fat can be consumed as long as the overall diet is low in energy density (i.e. less than 5kJ/g).

Dietary fat reduction alone may be effective for preventing weight gain in normal weight, overweight and obese individuals (little evidence).

Few controlled trials have measured the independent effect of dietary fat reduction on weight maintenance in normal, overweight and obese individuals. A randomised controlled trial suggests energy intake is significantly decreased in normal weight men (P <0.01) and women (P <0.0001) when dietary fat intake is reduced from 35%E fat to 33%E fat, whereas energy intake is signifi-cantly increased only in women (P <0.01) when dietary fat intake is increased from 35%E fat to 41.2%E fat (30). Body weight was significantly increased on the high fat diet (40%E or 122g/day) in both men (P <0.001) and women (P <0.01) but remained unchanged on the low fat diet (33%E fat or 89g fat/day). Eating behaviour may explain differences in the effect of dietary fat reduction on energy intake and consequent weight change in normal weight individuals (31).

The long-term effect of dietary fat reduction strategies on weight maintenance in normal weight and overweight individuals has not been demonstrated.

It is recommended that public health nutrition strategies for the prevention of cardiovascular disease emphasise reducing saturated fatty acid intake.

Current levels of dietary fat intake in Australia are around 32%E fat (32). Average intakes of saturated and polyunsaturated fatty acids (12.7% and 4.9%E, respectively) are not consistent with the NHFA's policy on dietary fat ( 8%E saturated and 8 to 10% polyunsaturated) (1,33). Intake of total dietary fat and polyunsaturated fatty acids in the Australian diet has decreased since 1983, while intake of total energy and saturated fatty acids has increased (34).

The evidence reviewed suggests that dietary fat intake of 30 to 35%E does not seem to be associated with excess energy intake. Since dietary fat restriction may help to facilitate energy restriction, public health recommendations should ensure that dietary fat intake remains at less than 35% energy from fat.

The emphasis of food-based recommendations for the prevention of CHD, developed by the NHFA, is on reducing intake of foods high in saturated fatty acids, such as full fat dairy products, takeaway meals, pastries, snacks and cakes. Reducing intake of saturated fatty acids is likely to also reduce total dietary fat intake. Furthermore, since many foods high in saturated fatty acids are also high in energy density, restricting their intake also facilitates a reduction in energy intake.

Further research required

Energy density may be a major determinant of energy intake although protein intake may act independently of this. Further research is required to determine the relative effectiveness of manipulating the dietary macronutrient, fibre and water content of foods on reducing the energy
density of the overall diet. Long-term studies are required to determine whether strategies for reducing the energy density of the diet are useful for achieving weight loss in overweight and obese individuals and weight maintenance in normal weight as well as overweight and obese individuals.

Other dietary and non-dietary factors are also implicated in the development and progression of overweight and obesity. Clearly physical activity plays an important role in regulating body weight. Further research is required to determine the combined effect of dietary and physical activity strategies in weight management.

Key words: dietary fat, overweight, obesity, energy balance, energy density

Original Research
Survey of dietetic management of overweight and obesity
and comparison with best practice criteria
Clare Collins

Abstract (Nutr Diet 2003;60:177–84)
Objective: The Dietitians Association of Australia (DAA) is endeavouring to support best practice for dietetic management of overweight and obesity in Australia. The aims of this member survey were to describe current dietetic services and intervention strategies in obesity management and to compare current practice with that reported previously.
Method: All members of DAA were sent a questionnaire, with 287 surveys returned (response rate 14%). The questionnaire was informed by a literature review of evidence-based practice and was pre-tested. Self-reported dietetic management strategies were evaluated against best practice guidelines.
Results: The majority of dietetic treatment consultations were individual (45%). Both dietitian (6%) and multidisciplinary (11%) group programs were offered, with 15% care-planning within general practice. Only 13% of respondents reported the adoption of clinical guidelines for obesity management. Factors that sup-ported or prevented dietetic involvement in obesity management were reported consistently as time, funding, staffing, management support and resources. Gaps in skills and areas identified for continuing professional development included counselling for behaviour change and strategies for implementation of best practice guidelines in a variety of settings. Future research questions included evaluation of models of dietetic care to ensure long-term lifestyle changes and improve clinical outcomes.
Conclusion: This survey indicates that there is a need for the development and dissemination of best practice guidelines for the management of overweight and obesity in Australian adults and children. DAA is well placed to facilitate uptake of current evidence-based treatment recommendations through the ratification and implementation of clinical guidelines that support best practice.

Key words: obesity, weight management, dietetic, clinical guidelines, best practice

Original Research
Evaluation of a tool for rating popular diet books
Louise Williams and Peter Williams

Abstract (Nutr Diet 2003;60:185–97)
Objective: The aim of this study was to develop a tool for use by nutrition professionals to evaluate popular diet books.
Design: A questionnaire was developed incorporating quantified criteria based on current authoritative nutri-tion guidelines. It included 22 questions relating to nutritional adequacy, daily energy allowance, recommended rate of weight loss, flexibility and sustainability, physical activity advice, use of supplements, claims, authors’ credentials, and scientific evidence. The questionnaire was used to rate 35 diets in 20 popular diet books sold in Australia in 2001, and the practicality, validity and sensitivity of the tool were evaluated. A dietary analysis of three days of menus from each book was used to assess the validity of the questions assess-ing nutritional adequacy.
Main outcome measures: Assessment scores for each book and correlation with dietary analyses.
Statistical analysis: Spearman rank correlation was used to compare the nutritional adequacy of the diets assessed by the dietary assessment cores from the questionnaire and the numbers of nutrients likely to be pro-vided at <70% RDI or <100% RDI. One-way ANOVA was used to compare the mean scores of books written by those with nutrition qualifications, medical qualifications, and others.
Results: The scoring of the questionnaire was found to correlate well with the nutrient analysis of the diets. Overall scores for the 20 books tested ranged from 32 to 97 out of a possible 100. Only five of the books were found to have scores over 80, the assessment criteria deemed compatible with current dietary and public health guidelines. Three diets provided less than 4200kJ per day, while five books advertised weight loss results of greater than 1kg per week and promoted or used ‘fast’ weight loss as a selling point. The majority of books relied on testimonials rather than supporting their results with data published in peer reviewed journals. Books by people with nutrition qualifications rated highest.
Conclusion: The questionnaire provides a useful standardised tool for nutrition professionals to rank the nutri-tional adequacy of advice in popular diets books and to evaluate their approach to weight loss.

Key words: weight loss, fad diets, obesity

Original Research
The arrival of Mediterranean recipes and food in Australian magazines
Ann Noah and A. Stewart Truswell

Abstract (Nutr Diet 2003;60:198–204)
Objective: The Mediterranean diet is now a well known concept with Mediterranean dishes and foods widely consumed. It was postulated that one way by which the concept, dishes and foods arrived in Australia was via articles in women’s magazines that emphasise food preparation at the domestic level.
Design: Descriptive study of two of the longest running Australian women’s magazines that emphasise cookery.
Methods: All back issues of Australian Women’s Weekly from 1933 and of New Idea from 1934 to the end of 1995 were searched for Mediterranean recipes. These were then classified by country of origin (there are 18 Mediterranean countries) and by degree of authenticity.
Results: The number of Mediterranean recipes of the different types that appeared each year are presented. These recipes started to appear in the 1960s and increased through the 1990s. Italian recipes were the most fre-quent, followed by Greek, French, Spanish and Lebanese.
Conclusions: Mediterranean recipes were introduced to Australian women from the 1960s by magazines, long before recent conferences on the Mediterranean diet. Nearly all of the recipes in these magazines have been from countries on the northern side of the Mediterranean Sea.

Key words: Australian Women’ s Weekly, New Idea, women’s magazines, Mediterranean diet, Mediterranean recipes

Original research
Competency development in public health nutrition: Reflections of advanced level practitioners in Australia
Roger Hughes

Abstract (Nutr Diet 2003;60:205–11)
Objectives: To investigate the attitudes, experiences and beliefs of advanced level public health nutritionists in Australia in relation to public health nutrition competency development.
Design: Qualitative study using semi-structured interviews.
Subjects: Forty-one advanced level public health nutritionists employed in academic and senior technocratic positions in state health systems.
Setting: Australia.
Main outcome measures: Qualitative data on attitudes, experiences and beliefs of advanced level public health nutritionists.
Analysis: Audiotaped interview transcripts were content analysed by theme using a pre-determined inquiry logic.
Results: Career paths of interviewees were mostly opportunistic rather than planned and reflected individual interests and changing health sector opportunities over the last few decades. Disillusionment with clinical practice was a common motivation for career paths leading to public health nutrition. The most commonly reported landmarks for competency development were exposure to mentors, on-the-job experience and post-graduate training in public health. There was disagreement about the utility of dietetic training in public health nutrition competency development, partly the result of recognition of the post-basic nature of public health nutrition competencies. Most advanced level public health nutritionists, however, identified the existing die-tetic workforce as a priority for public health workforce development because of the strong preparation in nutrition and the privileged access to work opportunities dedicated to nutrition that dietitians have. Under-graduate preparation specific to nutrition, experiential learning and post-graduate specialist training were identified as important features of public health nutrition competency development.
Conclusions: The data collected represent the views of a large proportion of the public health nutrition leader-ship group in Australia. Further research investigating workforce composition, competency needs, practices, continuing professional development needs and strategy effectiveness is required. Dietitians, as a professional group, appear well placed to take a leadership role in the scholarship and development of public health nutri-tion in Australia.

Key words: workforce development, public health nutrition, competency development