Consumption of yoghurt and dairy desserts in Australia is rising steadily. In 2007, sales were $AU 982 million(1). Yoghurts for adult consumption held the largest market share (69 %), a segment growing by 7·6 % per annum(1). Just under 40 % of Australian adults now consume yoghurt at least once weekly(Reference Chubb, Boorman and Baines2). Yoghurt, moreover, has been promoted in Australia as one of the healthiest convenience foods(Reference Riley3). Certainly, traditional natural yoghurt manufactured with cultures of Lactobacilus bulgaricus and Streptococcus thermophilis (Reference Adolfsson, Meydani and Russell4) has many nutritional advantages. It is a source of high-quality, readily digested protein, contains many micronutrients, notably Ca, and may be better tolerated than milk by those with lactose maldigestion(Reference Adolfsson, Meydani and Russell4). Yoghurt consumption may improve absorption of micronutrients such as Zn(Reference Rosado, Diaz and González5), reduce the duration of diarrhoeal disease in children(Reference de Mattos, Ribeiro and Mendes6) and improve cellular immune function(Reference Mayer, Micksche and Herbacek7). Mediterranean diets, well known for their health benefit, commonly include plain natural yoghurt daily(Reference Kush, Lenart and Willett8).
However, in Western societies yoghurt has changed markedly from the traditional semi-firm curd product made from milk with perhaps salt (leben) or water (lassi) added(Reference Robinson9). Manufacturers now alter texture through the addition of milk solids, starch or gelatine(Reference Soukoulis, Pangiotidis and Koureli10), alter taste with fruit and/or sweetening agents(Reference Robinson9) and remove fat(Reference Robinson9). Health benefits are sought through fortification with vitamins and minerals(Reference Pirkul, Temiz and Erdem11) or addition of probiotic bacteria(Reference Shah12) or prebiotics such as inulin(Reference Kip, Meyer and Jellema13). Manufacturers are also making more desserts and yoghurts targeted at children(1).
We have previously reported on the nutrient content of Australian snack foods(Reference Walker, Woods and Rickard14), defining ‘snacks’ as portable foods readily consumed outside main meals (breakfast, lunch and dinner)(Reference de Graaf15). We thus had excluded yoghurt as it required utensils to eat. Yet in Australia yoghurt is often eaten between main meals(Reference Chubb, Boorman and Baines2). We therefore now report on Australian yoghurts, fromage frais and dairy desserts, having examined packet and serving size, nutrient content, additives and nutrient claims. As earlier(Reference Walker, Woods and Rickard14), we assess the proportion of these dairy foods that can be deemed ‘healthy’ using UK ‘traffic light’ criteria(16), advocated for use in Australia(Reference Kelly, Hughes and Chapman17) although it does not include a specific category for dairy foods. An alternative system (the ‘Ofcom model’)(Reference Lobstein and Davies18) was therefore also used with both positive and negative criteria and a score for protein content (as a proxy for essential nutrients).
Experimental methods
Data collection
The Australian food supply is dominated by two supermarket chains controlling around 80 % of all food sales. In 2005 and 2008, surveys were undertaken to record information given on the nutrition information panel (NIP), and elsewhere on the packaging, of all yoghurts and dairy desserts presented for sale in a single large supermarket in metropolitan Melbourne. Data were collected and recorded on standardised entry sheets as described elsewhere(Reference Woods and Walker19). In 2008, information in the ingredient list was also noted with codes translated according to the Australian Food Additives Code(20). Yoghurt is defined by Standard 2.5.3 of the Australia New Zealand Food Standards Code (ANZFSC) as a milk product fermented by lactic acid-producing micro-organisms to which other food (such as fruit) can be added(21). Yoghurt drinks were not surveyed. Fromage frais comprised products marketed in Australia as Frûche™, while dairy desserts included custards and crème caramel, mousse, dairy rice puddings and cheesecakes often consumed in Australia as snacks. In accordance with the Code of Practice on Nutrient Claims in Food Labels and in Advertising (CoPoNC), yoghurt described as ‘full fat’, ‘reduced fat’ or ‘no fat’ had >3 g fat/100 g, <3 g fat/100 g or <0·01 g fat/100 g, respectively(22).
Data analysis
Data were analysed using the SPSS for Windows statistical software package version 15·0·1 (SPSS Inc., Chicago, IL, USA). Items in each category were assessed as the sum of products and product varieties (flavour alternatives). Due to non-symmetrical distribution of data, aggregates are presented as the median and interquartile range (IQR). Dairy snacks were assessed using the green/amber (low/medium content) or amber/red (medium/high) boundaries of the UK ‘traffic light’ system(Reference Kelly, Hughes and Chapman17). As ‘added sugar’ was not listed on the NIP, the amber/red boundary for sugar was determined as ‘total sugars’ minus 4·8 g/100 g, an average sugar content for natural yoghurt as reported by Australian Food Composition Tables. The Ofcom nutrient profiling system(Reference Lobstein and Davies18) was also used with foods assessed according to both positive criteria (% fruit, AOAC fibre, protein g/100 g) and negative criteria (kJ, saturated fat, total sugar and Na per 100 g).
Results
In 2005, 248 dairy snacks were surveyed while 140 were surveyed in 2008. Yoghurts constituted about two-thirds of these foods (Table 1). Between 2005 and 2008, packet size rose from (median (IQR)) 200 (300) g to 350 (250) g (P < 0·001). This reflected an increase in the size of individual tubs rather than an increase in the proportion of family-sized tubs of similar composition (sold as 1 kg or 6 × 200 g; data not shown).
Table 1 Nutrient content of dairy snacks surveyed in a large supermarket in metropolitan Melbourne, Australia
IQR, interquartile range.
Median values were significantly different from those in 2005: *P ≤ 0·05, **P < 0·01, ***P < 0·001.
Median value was significantly different from that of yoghurt surveyed in the same year: †††P < 0·001.
‡Large interquartile ranges for some nutrients reflect heterogeneity within the group with respect to fat content or fortification. Here, nine high-fat products (>6 g/100 g) are admixed with nineteen low-fat products.
Table 1 indicates the nutrient content of surveyed items. For yoghurt, median energy and total fat content increased significantly between 2005 and 2008 (both P < 0·05) while median protein content decreased (P < 0·01). For fromage frais, the median content of carbohydrate (P < 0·001), sugars (P < 0·01) and Na (P < 0·001) rose over the same period. For dairy desserts the median levels of total and saturated fat also rose significantly (both P < 0·05).
In 2005, 36 % of yoghurts and dairy desserts were ‘full fat’ (>3 g fat/100 g). By 2008 this proportion had risen to 46 % (P < 0·05). The contribution of sugars to total energy content in ‘reduced-fat’ yoghurts was notable (Table 2). Indeed, of all ‘reduced-fat’ yoghurts surveyed in 2008, 7 % had energy content above the median for ‘full-fat’ products (Fig. 1). The lowest energy content found among ‘full-fat’ products was in a European-style natural yoghurt (3·7 g fat/100 g). The majority (90 %) of ‘reduced-fat’ yoghurts contained more energy than this ‘full-fat’ product. As seen in Table 2, the median carbohydrate content of ‘full-fat’ yoghurt increased by 2008 (P < 0·001) while median Na level decreased (P < 0·05). In ‘reduced-fat’ yoghurt over the same time, total fat and saturated fat content increased (P < 0·001 and P < 0·05, respectively) while protein content decreased (P < 0·001). Figure 2 examines the energy density of dairy snacks in relation to fat and sugar content. There was not only a strong relationship between fat content and energy density (r = 0·820, P < 0·001) but also a strong relationship between sugar content and energy density (r = 0·690, P < 0·001).
Table 2 Nutrient content of yoghurts surveyed in a large supermarket in metropolitan Melbourne, Australia
IQR, interquartile range.
Median values were significantly different from those in 2005: *P ≤ 0·05, **P < 0·01, ***P < 0·001.
Fig. 1 Energy content of ‘full-fat’ compared with ‘reduced-fat’ yoghurts offered for sale in a large supermarket in metropolitan Melbourne, Australia (combined data from 2005 and 2008; horizontal lines indicate the median)
Fig. 2 Energy density of dairy snacks offered for sale in a large supermarket in metropolitan Melbourne, Australia in 2008 as related to (a) fat content per 100 g (r = 0·690) and (b) sugar content per 100 g (r = 0·820)
Table 3 assesses yoghurts and dairy desserts against UK ‘traffic light’ criteria for the ‘green’ labels that indicate a low content of fat, saturated fat, total sugars and salt(16). Although all yoghurts and dairy desserts met the relevant criterion for salt content, very few met either the criterion set for sugar or all four criteria. Moreover, the proportion meeting four ‘green’ criteria declined between 2005 and 2008, while the proportion that would receive one or more ‘red light’ labels (indicating a high fat, saturated fat, sugar or salt content) increased significantly from 12 % to 23 % (P < 0·01). When yoghurts and dairy desserts were categorised according to a nutrient profiling model that included a score for protein content(Reference Lobstein and Davies18), only 21 % (in 2005) or 28 % (in 2008) were deemed ‘healthy’.
Table 3 Percentage of dairy snacks surveyed in a large supermarket in metropolitan Melbourne, Australia complying with the boundary criteria of UK signpost labelling(16) or the Ofcom nutrient profiling system(Reference Lobstein and Davies18)
Percentages were significantly different from those in 2005: *P ≤ 0·05, **P < 0·01, ***P < 0·001.
‡Using both positive and negative criteriaReference Lobstein and Davies(18).
In 2008 ingredients for yoghurts and dairy desserts were also surveyed (Table 4). Products had up to fifteen separate additives, 50 % had six or more added ingredients. In yoghurts, modified starch, gelatine, pectin, agar and locust or carob bean gum were common thickeners. Fromage frais were more often thickened with modified starch while dairy desserts were more often thickened with carrageenan. About half the yoghurts contained acidity regulators (sodium citrate and/or sorbic and citric acids). Few yoghurts were fortified with micronutrients, n-3 fatty acids or inulin, in contrast to dairy desserts where 17 % contained added vitamins and/or minerals and 10 % contained inulin. Fromage frais remained unfortified. Main colourings used were cochineal and annatto.
Table 4 Proportion of dairy snacks surveyed in a large supermarket in metropolitan Melbourne, Australia in 2008 containing additives and/or sweeteners
Sugar (sucrose) was the most common sweetener, found in levels up to 29 g/100 g. In yoghurt, the most common alternative/additional nutritive sweeteners were honey, fruit juice concentrate and glucose. In fromage frais, sweetening with apple/fruit juice was common while glucose was often added to dairy desserts. Nearly 20 % of yoghurts and 10 % of dairy desserts contained non-nutritive sweeteners (acesulphane K or aspartame). A few dairy desserts contained sorbitol. Sucralose was used rarely.
Although few dairy snacks could be described as ‘healthy’, a great many made nutrition-related claims (Table 5), many of these directed towards parents with children. The most common related to Ca, protein or fat content. Others related to the absence of artificial flavours or colours or to the presence of ingredients with potential to cause allergy. Few referred to sugar content.
Table 5 Claims made on packets for dairy snacks surveyed in a large supermarket in metropolitan Melbourne, Australia in 2008
Discussion
The current survey of yoghurts and dairy desserts from an Australian supermarket has shown that between 2005 and 2008, change has largely been detrimental. Packet size as well as energy, total fat, saturated fat or sugar content increased. Many ‘reduced-fat’ products supplied more energy than ‘full-fat’ products due to the high level of added sugars. The energy density of yoghurts and dairy desserts was thus related strongly to both fat and sugar content. There is no evidence for compositional change or legislative change over this time to explain these results. Overall, observed changes in yoghurts and dairy desserts contribute to deterioration in the food environment facing the consumer, particularly given the known importance of high availability of energy-dense snacks in large portion sizes(Reference Kerr, Rennie and McAffrey23) in promoting increased body weight.
As with our previous survey of non-dairy snack foods(Reference Walker, Woods and Rickard14), it is evident that an Australian consumer entering a supermarket to select a yoghurt or dairy dessert is faced with a bewildering choice of available products, of which only a minority are ‘healthy’ despite numerous claims suggesting health or nutritional benefits. Numbers of products declined slightly from 2005 to 2008, most likely due to the introduction of supermarket ‘home-brand’ products (currently 15 % of the market and increasing by 3–5 % per annum). This strategy has been employed to increase margins, build customer loyalty and improve retailer buying power(24). It also gives retailers more power to influence availability and nutritional composition(Reference Burch and Lawrence25).
We previously reported that 91 % of snack foods would carry at least one ‘red’ label indicating a high content of fat, saturated fat, salt or sugar as judged by UK traffic light criteria(Reference Walker, Woods and Rickard14). Yoghurts and dairy desserts compare favourably with non-dairy snacks: only 23 % in 2008 passed one or more criteria for a ‘red traffic light’ label. Nevertheless, the proportion of ‘less healthy’ yoghurts and dairy desserts overall is substantial and appears to be growing. We have also applied the Ofcom model(Reference Lobstein and Davies18) that scores a wider range of nutrients although its primary focus is children’s foods. Using this model, the proportion of ‘less healthy’ yoghurts and dairy desserts becomes even higher.
A clear interpretive front-of-pack nutrition label on dairy snacks can greatly help identify optimal choices and many Australians therefore support the introduction of a ‘traffic light’ system(Reference Kelly, Hughes and Chapman17). Yet while better selection of snacks has potential to improve diets, in making their choice, people often place taste and price above nutrition-related criteria(Reference Bush and Williams26) or are more concerned about additives than nutrients(Reference Williams, Stirling and Keynes27). Regrettably, those with less healthy intakes are least likely to consult an NIP(Reference Lin, Lee and Yen28). There is therefore a need for reformulation of many dairy foods so that healthy choices constitute a larger proportion of the snacks available and the healthy choice may be more probable(Reference Bush and Williams26). Such change may occur when consumers become educated to demand more nutritious choices and when governments support and encourage food reformulation(Reference Bush and Williams26). While some manufacturers may be willing to voluntarily self-regulate or change product formulation, market pressures and adverse retailer influence may make government legislation necessary to ensure that beneficial change is uniformly implemented. Governments can also legislate to restrict imprecise claims, to ensure optimal food labelling, to control food advertising to children, and to define school canteen policies. Taxation can also be used to improve consumer choice(Reference Lobstein and Davies18). Taxes on high-fat and/or high-sugar foods have been proposed but not adopted in Australia(Reference Zimmet and James29). Yet since reducing the price of healthier food options can increase their purchase(Reference Faith, Fontaine and Baskin30), taxes gained from ‘less healthy’ foods might also usefully subsidise healthy food choices and support health promotion.
Acknowledgements
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. No conflicts of interest are known to the authors. Each author has seen and approved the contents of this manuscript. K.Z.W. developed the study concept and contributed to data collection, data analysis, preparation of the first draft document and editing of the final text. J.W. developed the study concept and made a detailed contribution to manuscript drafts and initiation of further data analyses. She also gave detailed comments on the final manuscript. J.R. contributed to data collection and made comments on the final manuscript. R.H. made a large contribution to data collection.
