Four theoretical dietary modelling procedures were undertaken utilising the nutrient profiling data. Two models involved the incorporation of seafood into a 14-day meal plan consistent with the number of food group servings in the AGHE available for 60+ year olds (Table 1) (23). The older age group was chosen due to the frequent low intake of seafood (12–15, 18) and their higher prevalence of heart disease (6). Theoretical energy and macronutrient intakes from these diets were compared against current Nutrient Reference Values (NRVs) for 51–70 year olds (7). Two further models were carried out to identify the number of weekly servings of lower fat and higher fat seafood required to meet LC n–3 PUFA recommendations.

The AGHE was used to build the 14-day menu based on food group guidelines for 60+ year olds. The AGHE was developed in 1998 and the meal plan was devised so that the different gender and age groups could meet 70% of the Recommended Daily Intake (RDI) (i.e. RDIs established in 1990). As new NRVs have since been made (7), the current study used the AGHE as the basis for a 14-day meal plan; however, assessment of the theoretical diet was assessed against current NRVs for men and women aged 51–70 years. Specifically, macronutrient recommendations were based on the Acceptable Macronutrient Distribution (AMDR) for protein, carbohydrates, and fat and/or AI for fibre and LC n–3 PUFA (160 mg/d men, 90 mg/d women) (7). The SDT (SDT: 430 mg/d women, 610 mg/d men) (7) and the HF recommendation of 500 mg/d EPA+DHA (4) were also used in determining the percentage of adults who would meet these LC n–3 PUFA targets.

Commonly consumed foods (n=207) were chosen from the AUSNUT database and their energy and nutrient content was extracted using Food Works (FoodWorks 2009, version 6.0.2539, Xyris Software Brisbane). Twenty-six foods were included in the breads and cereals group, in which 11 were white/plain/refined options, 12 were wholegrain/wholemeal options and 3 were classed as ‘other’ (e.g. fruit bread, crackers). The number of individual foods in the breads and cereals food group was based on the percentage that these foods were consumed by adults >19 years from the 1995 NNS. For example, ‘bread and bread rolls’ provided approximately 40% of mean daily intake for breads and cereals in that survey (17). Therefore, 8 choices of bread and bread rolls were included in the breads and cereals category. For the 40 vegetable options, 11 were starchy vegetables (potato and sweet corn), 7 were orange/red vegetables (carrot, pumpkin and sweet potato), 5 were darker green and brassica vegetables (broccoli, spinach, green bean, cauliflower and cabbage), and 17 were other types of vegetables (celery, cucumber, lettuce, mushroom, onion, beetroot, capsicum, squash, tomato, broad bean, snow pea, parsnip and zucchini). For the 22 fruit options, 16 were fresh (4 were apples and pears), 3 options were fruit juices, and 3 were canned varieties. Within the 26 dairy options, 11 were various milk types (e.g. skim, low fat, full fat), 8 were cheese, 4 were yoghurt and 3 were other (custard, ice cream, condensed milk). For the meat and alternative options, 14 were red meat (including veal and pork), 6 white meat (chicken and turkey), 7 beans/legumes, 3 options for eggs, and 7 options for nuts and seeds. Sausages, bacon, salami-type deli meats were included in the extras category due to their higher energy, fat, and sodium content, even though they were reported in the NNS under the meat and meat dishes group. An option of ice cream was included in the dairy and extras category as it contains calcium, but also can be high in fat/sugar. There were 58 options for the extras category in which the majority contained approximately 600 kJ/serving (e.g. 400 ml regular beer, 596 kJ; 40 g iced apple cake, 607 kJ; 40 g pure cream, 664 kJ; 60 g party sausage roll, 799 kJ), as well as 6 options of unsaturated fats and oil foods (3 margarine, three oils).

Twenty-one different seafood species sampled between July 2010 and July 2011, from the 2012 Australian Seafood Compositional Profiles study (22) were included in the models (including both wild and farmed seafood for 3 of the species, totalling 24 different groups of seafood). In the profiles study, most seafood products were analysed raw (except burrowing black fish and the farmed prawn species) and all values were reported in 100 g portions (Table 2). Due to changes in food weight upon cooking, a weight change factor for the raw products was used (24). Weight change factors (reported as a percentage) were listed for ‘finfish, fresh, frozen’ that had either been baked (−24% weight change), grilled (−26%), microwaved (−19%), poached (−16%) and steamed (−17%). A weight change factor of −25% (i.e. 100 g fish–25% loss=75 for calculation) was used as an intermediate between baked and grilled weight loss as this was the most frequently reported method of cooking in a previous study that we conducted (18). In addition, to account for any changes in fatty acid content following weight loss during cooking, a retention factor was used to convert fatty acid content of raw fish to fatty acid content in cooked fish (0.8 for fish with >5% fat content, 0.95 for fish with <5% fat content and 1.0 for crustaceans and molluscs). It was assumed that the change in any particular fatty acid content would be similar to the changes in total fat content upon cooking. As there are no current data on retention factors for fatty acids, values were adopted from retention factors reported in a German publication (25). The content of LC n–3 PUFA ranged from 11.3 mg/100 g (cooked burrowing blackfish) to 3,755 mg/100 g (cooked ocean trout). Three of the species were considered to be high fat (‘higher fat’, ≥10% total fat: Atlantic salmon, ocean trout, yellowtail kingfish), according to the HF in Australia (26), one was moderate in total fat (barramundi, 7.4%) and the remaining were considered to be low fat (<5% total fat). The moderate-fat and low-fat seafood species were grouped (‘lower fat’).

The Risk Solver programme is an add-on feature to the Microsoft Excel computer programme (Risk Solver Premium V9.0.4.0 program; Frontline Systems, Inc.). The programme allows the random selection of foods from a list of pre-determined foods in which energy and nutrient profiles can be assessed. Diet simulation models were based on AGHE food group recommendations, and the food groups (iterations) that were included were usual food items consumed with percentage consumption frequencies based on the 1995 NNS (17). Seafood consumption frequencies were based on a recent Australian survey (18) that involved a similar age group to what the current theoretical models were performed against. Running the simulation procedure randomly selects each single food item according to the imputed number of daily serves in each food group (e.g. 5 random bread/cereal items from the list of 26, 2 random fruit items from the list of 22 fruits). This results in a mean intake of energy and nutrients for the whole day and for each food group. Based on a 14-day menu (varying in the number of serves for each food group/d and to reach overall AGHE recommendations), the simulation process runs the varying combinations 10,000 times (e.g. 10,000 combinations of 5 items of breads and cereals from the 26-item category) and mean energy and nutrient intakes are produced for each day and for the 14-day period. The likelihood (%) in which the theoretical diet would meet nutrient recommendations for a particular gender and/or age group can be calculated using estimated means calculated from the 14-day model.

The aim of Model 1 was to determine whether a simulation model which included a random selection of seafood types will meet the AI, SDT, and HF recommendation for LC n–3 PUFA in men and women 51–70 years of age, as well as the AMDR for protein, fat and carbohydrate, and AI for fibre. Consumption of seafood was entered at an average of 0.3 servings/d for women and 0.4 servings/d for men; over 1 week this reaches the recommendation of 2–3 servings (4). Each type of seafood was assigned a selection probability factor such that the more frequently consumed seafood (e.g. Atlantic salmon) was more likely to be selected than those less frequently consumed (e.g. abalone). Selection probability factors were estimated based on our recent survey (18) and are provided in Supplementary Tables 1–3. Due to the relatively frequent consumption of tuna (18), an option for canned tuna and for grilled/barbequed tuna was also included. The nutrient composition for tuna was taken from the NUTTAB 2010 database; therefore, a total of 26 seafood groups were available for selection in the modelling process. The number of servings for all other food categories was also entered over the 14 days according to AGHE recommendations (e.g. Day 1: 5 servings of breads and cereals, two servings of fruit; Day 2: 4 servings of breads and cereals, 2 servings of extras, etc.).

The aim of Model 2 was to determine whether including a random selection of the lower fat seafood (<10% total fat) will achieve LC n–3 PUFA recommendations for men and women aged 51–70 years in accordance with the AGHE. The seafood in Model 2 included Australian sardine, barramundi, blue sprat, burrowing blackfish, farmed prawns, wild prawns, gummy shark, oysters, southern rock lobster, and wild abalone, plus the two tuna options. A total of 23 lower fat seafood groups were included in the modelling process.

Two simulation models were undertaken to determine the number of servings of lower fat (Model 3) or higher fat seafood (Model 4) to meet LC n–3 PUFA recommendations from only seafood sources. Each lower or higher fat seafood was randomly included in the model in a continuous fashion until LC n–3 PUFA recommendations were met by 100% of men and women.

Appropriate numbers of food group servings from the AGHE were entered into a theoretical 14-day menu plan (Table 1), in which the numbers and types of foods from these food groups were randomly selected. The mean intake of energy, macronutrients and LC n–3 PUFA, and the percentage of men and women who would meet recommendations when including 0.3–0.4 servings/d (2–3 servings/week) of higher and lower fat seafood in an AGHE diet are described in Table 3. For women, including 2 servings/week of any type of seafood would result in 89% meeting the LC n–3 PUFA AI, while 66% of men would meet the AI with 3 servings/week. For both men and women, 37% would meet their respective SDT. The percentage of men and women that would meet the HF recommendation of 500 mg/d of combined EPA+DHA was 36%.

Model 2 describes the mean intake of energy, macronutrients and LC n–3 PUFA following random selection of the lower fat seafood, at a frequency of 0.3 servings/d for women (2 servings/week) and 0.4 servings/d for men (3 servings/week) (Table 4). Eighty-three percent of women and 47% of men would achieve the AI for LC n–3 PUFA. Only 0.3 and 0.4% of men and women, respectively, would achieve the SDT and none would achieve the 500 mg/d HF recommendation.

Random selection of 100 g servings/week of lower fat seafood would provide approximately 54 mg/d of LC n–3 PUFA (Table 5). For women, 2.5 and 10 servings/week of lower fat seafood would be required for about 90% to meet the respective LC n–3 PUFA AI and SDT. For men, 4 and 14 servings/week of lower fat seafood would be required for about 90% to achieve the AI and SDT. In contrast, random selection of 100 g servings/week of higher fat seafood would provide 399 mg/d of LC n–3 PUFA, such that only 0.5 serving/week of higher fat seafood would be needed for men and women to meet the AI and only 1.5–2 servings/week to meet the SDT. In both men and women, consuming 2 servings/week of higher fat seafood would enable the HF recommendation of 500 mg/d to be met.