Older people represent an increasing proportion of the population(Reference Kinsella and Velkoff1). Health or its absence in this rapidly increasing population not only affects the individuals themselves, but also has serious implications for demands on health care and other social resources. Cross-cultural studies(Reference Wahlqvist, Kouris-Blazos and Hsa-Hage2–Reference Wahlqvist, Darmadi-Blackberry, Kouris-Blazos, Jolley, Steen, Lukito and Horie4) examining lifestyle variables and health suggest that diet is an important predictor of survival. For elderly individuals, inadequate nutrition can increase the incidence and severity of disease, hastening loss of independence. It is thus important to understand what influences older people’s food choices to help ensure healthier diets(Reference Kennedy, Ohls, Carlson and Fleming5).
Studies suggest that Mediterranean-style diets as well as limited intake of red meat and high cereal-fibre consumption reduce CVD and cancer in older people, suggesting that eating a diet that is high in fruits, vegetables, nuts, cereals and low in red meat is healthy(Reference Mozaffarian, Kumanyika, Lemaitre, Olson, Burke and Siscovick6–Reference Trichopoulou and Critselis9). Thus a varied diet is seen as a healthy diet, as food variety enables food component needs to be met adequately and comprehensively(Reference Wahlqvist, Kouris-Blazos and Hsa-Hage2, Reference Hollis and Henry10–Reference Horwath, Kouris-Blazos, Savige and Wahlqvist12).
The ability to choose a diet with sufficient quality and variety to meet daily nutrient needs may be adversely affected by the pathological, physiological, economic and societal factors that accompany ageing(Reference Sharpe, Huston and Finke13). For example, gradual loss of health due to the effects of chronic diseases such as arthritis or diabetes can impair the ability to obtain, prepare and enjoy nutritious foods(Reference Hendy, Nelson and Greco14). The decline in the ability to taste and smell may result in changes to food selection and preference, leading to changes in energy and/or nutrient intake(Reference Rolls and McDermott15, Reference Westenhoefer16). This is confirmed by nutritional surveys showing a reduced variety in older people’s diets(Reference Fanelli and Stevenhagen17). In addition, Duffy et al.(Reference Duffy, Backstrand and Ferris18) found that women with impaired ability to smell not only had lower preferences for some nutritious foods and higher intakes of sweets and fats, but also had less interest in food-related activities such as cooking. However, while the decline in sensory-specific satiety may influence variety of food choice, social and environmental resources and older people’s food-related goals may override the importance of the physiological factor in ensuring a varied and balanced diet.
Past research suggests that high level of income, good physical and mental health, mobility in terms of having access to a car and the number of social contacts are resources that may affect older people’s eating habits(Reference Sharpe, Huston and Finke13, Reference Duffy, Backstrand and Ferris18–Reference Sahyoun and Zhang20). Lower levels of economic resources are associated with greater risk of experiencing hunger and food insufficiency(Reference Brown21, Reference Sahyoun and Basiotis22). Some research(Reference Blaylock, Smallwood, Kassel, Variyam and Aldrich23) suggests that it costs more to eat healthily; thus low income restricts not just quantity but also the nutritional quality of the food purchased. Research also shows that health problems related to inadequate nutrition are more prevalent in rural areas, where transportation to and from shops is mentioned as a structural barrier to obtaining adequate food(Reference Hendy, Nelson and Greco14, Reference Lee, Templeton, Marlette, Walker and Fahm24, Reference Wallace, Pascarella and Campanella-Voica25). Thus higher economic status and access to a car may be resources that may contribute to older people having a more varied diet.
Loneliness due to loss of spouse or friends can diminish the social reasons for and pleasure associated with eating(Reference Shifflett and McIntosh26, Reference Walker and Beauchene27). Eating regular meals and having an adequate diet have in part been found to depend on eating with others(Reference Shifflett and McIntosh26–Reference Paquet, St-Arnaud-McKenzie, Ma, Kergoat, Ferland and Dube28). However, Walker and Beauchene(Reference Walker and Beauchene27) showed that people’s number of social contacts bore no relationship to food choice. Revenson and Johnson(Reference Revenson and Johnson29) claimed it is the dissatisfaction with available relationships rather than the number of social contacts that is a powerful indicator of loneliness and cause of reduced intake, suggesting that for social resources the perceived quality of a resource is more important than the actual level.
Studies show that knowledge-based resources such as food knowledge and cooking skills also impact on food choice(Reference Baker and Wardle30, Reference McDonald and Webster31). That is, those who think they have a good understanding of foods and feel highly skilled eat a more varied a diet than those who feel that their knowledge and skills are limited. This implies that older people’s perceptions of food-related resources influence their food intake over and above actual resources.
Research has shown that the goals people have in life affect their food choice and the satisfaction they feel with their life(Reference Oaks and Slotterback32–Reference Dean, Grunert, Raats, Nielsen and Lumbers35). People’s judgement of the healthiness of food was influenced by whether or not they had dieting as a food-related goal(Reference Oaks and Slotterback32). Diener and Fujita(Reference Diener and Fujita34) showed that people tend to have goals relevant to their strongest resources and people who have the resources relevant to their goals/strivings exhibit the highest subjective well-being. Similar findings were observed in relation to food and older people by Dean et al.(Reference Dean, Grunert, Raats, Nielsen and Lumbers35).
Through testing the following hypotheses, the present study aimed to investigate the influence of resources and food-related goals on the variety of older people’s food choice:
1. Income, health status, access to a car and living arrangement will affect the level of dietary variety.
2. Perceived level of different food-related resources will impact on the consumption of a varied diet over and above actual resource levels.
3. Food-related goals will contribute to variety of food intake that is not accounted for by the amount of material resources possessed or the social and other resources perceived to be possessed.
Method
A survey was conducted in eight European countries: Poland, Portugal, United Kingdom, Germany, Sweden, Denmark, Italy and Spain. Questionnaires were administered face-to-face by trained interviewers. The questionnaire was developed in English, translated and back-translated before it was piloted in each country.
Sample
Data collection took place in the autumn of 2005. In total 3200 participants (400 from each country), who were above 65 years of age and living in their own homes, were sampled using telephone recruitment from senior day centres and some snowballing by marketing companies (Table 1). The questionnaire was administered face-to-face by trained interviewers using computer-aided personal interviewing. All respondents participated at least to some degree in either procurement or preparation of food in the household. The samples were quota samples, eight groups of fifty in each country, based on gender, age (<75 years v. ≥75 years) and living circumstances (living alone v. living with a partner), reflecting the diversity of each of the national populations based on education, income and urbanization of living environment (urban/suburban/rural). The sample was taken from at least three geographical locations in each country. Non-response was not logged.
Characteristics of participants in the eight countries
Measures
The selection process of resources and food-related goals potentially relevant for seniors’ satisfaction with their food-related life was based on an analysis of eighty in-depth interviews with seniors (ten participants from each of the above eight quota groups) in each of the eight countries. The most frequently named food-related resources and goals were collected from these qualitative data, resulting in the extraction of eleven goals (rated for importance on a 5-point scale from not important to extremely important) and twenty-two resources (assessed through agreement with the statement ‘Do you agree or disagree that you have a good (resource name)?’ on a 5-point scale from strongly disagree to strongly agree) which were included in the questionnaire (Table 2).
Multiple hierarchical regression of summed eating habits v. perceived resources and goals among older people (n 1484) from eight European countries
R 2 = 0·07 for step 1; ΔR 2 = 0·13 for step 2; ΔR 2 = 0·03 for step 3.
Coefficient was statistically significant: *P < 0·05, **P < 0·01, ***P < 0·001.
Nutritional adequacy was measured using a weekly food variety score(Reference Savige, Hsu-Hage and Wahlqvist36) that measures food variety according to the biological/botanical origins of the food, e.g. all citrus fruits are grouped together. Foods can be added up, with each biologically distinct food group scoring only once, no matter how often the foods within this group are eaten, but a minimum quantity of about 2 tablespoons must be consumed before it can be scored. Foods from these various sources can be scored and the total used as an indication of adequate nutrient intake. A food variety score of at least 15 over one week is generally nutritionally adequate. A score of 30 over a week or 12 in a day is considered excellent. The weekly food score has been used in different countries such as Australia, Japan and Vietnam(Reference Savige, Hsu-Hage and Wahlqvist36–Reference Ogle, Hung and Tuyet40). As the score measures food groups rather than particular foods, this measure was chosen as it was thought to be an ideal tool to compare nutritional intake across different countries.
Further, participants’ levels of actual resources were measured with respect to car availability, living arrangement, monthly income, and physical (PCS8) and mental health (MCS8) as measured by the SF-8 Health Survey(Reference Ware, Kosinski, Dewey and Gandek41). Demographics (e.g. gender, age, education, weight and height) were also obtained.
Results
Correlations between the predictor variables (only r > |0·3| are reported here) showed high positive correlations of monthly income with car availability (r = 0·40), good dental health (r = 0·29) and perception of a good income (r = 0·53) and negative correlation with the goal of keeping expenditure low (r = −0·40). Further, measured physical health (PCS8) was positively correlated with measured mental health (MCS8; r = 0·37), perception of good health (r = 0·58) and perceived mobility level (r = 0·53). In addition, measured mental health was positively correlated with perceived good health (r = 0·41). As expected, those who perceived themselves to be in good health also saw themselves as having good dental health (r = 0·30) and good mobility (r = 0·41). Living arrangement (alone v. with a partner) was highly negatively correlated with sharing meals with others (r = −0·61) and with the goal of eating in the company of others (r = −0·35). In terms of perceived resource levels, cooking skills was positively correlated with food knowledge (r = 0·43) and good income was positively correlated with having access to high-quality products/brands (r = 0·33) and negatively correlated with the goal of keeping expenditure low (r = −0·41). Moreover, the resource of having access to convenient products was positively correlated with the perceived resource level of low prices (r = 0·31), access to new products (r = 0·34) and access to high-quality products/brands (r = 0·32).
The influence of material resources (income, physical health, mental health, access to a car and living arrangement) on varied eating was investigated by hierarchical multiple regression analysis where the regression of the summed measure of varied eating v. monthly income, PCS8, MCS8, living alone/with partner and having a car was performed, in the first step. In order to see whether perceived resources could add to the prediction of varied eating, the twenty-two perceived levels of resources were added as a second step. To analyse whether people’s goals contribute to their eating habits over and above the influences of perceived and actual resources, the eleven individual goals were entered into the hierarchical analysis as a third step.
Actual resources and their influence on variety of diet
Multiple regression of varied eating v. the five objective resource levels (adjusted R 2 = 0·07, F(5, 1479) = 22·39, P<0·001) revealed significant independent effects for living arrangement (β = −0·09, P < 0·001), car availability (β = –0·10, P = 0·01), physical health (β = 0·10, P < 0·001), mental health (β = 0·07, P < 0·001) and monthly income (β = 0·09, P < 0·001). This suggests that those who live with a partner eat a more varied diet than those who live alone. Also, those who have access to a car have a more varied diet. In addition, mental and physical health affects variety of diet such that those who are in better health have a more varied diet. Finally, income affects variety of diet such that the higher the income the more the variety in the person’s diet. As the β coefficients of all the predictor variables are similar in value, this suggests that the sizes of the effects of living arrangements, car availability, physical health, mental health and monthly income on the variety of the diet consumed are roughly equal.
Perceived resources and their influence on variety of diet
Regression of varied eating v. the twenty-two perceived resource levels as the second step revealed that the increase in explained variance (change in R 2 = 13 %) was significant (F change (22, 1457) = 10·74, P < 0·001). Good appetite (β = 0·10, P < 0·001), food knowledge (β = 0·20, P < 0·001), access to convenient food products (β = −0·08, P < 0·01), access to a good food service provider (β = 0·15, P < 0·001), kitchen appliances (β = 0·08, P < 0·01), short distance to the shops (β = 0·05, P < 0·05), access to high-quality products (β = 0·09, P < 0·01) and support from friends and neighbours (β = 0·06, P < 0·05) were shown to be significant predictors while good health, good dental health, cooking skills, mobility, ability to smell/taste, good income, low prices, access to new products, access to food storage, convenient transport, access to organic products, sharing meals with others, formal support and support from family were not seen as influencing a varied diet. This suggests that those who have a good appetite, a good knowledge of different foods and less access to convenient foods, but good access to a good service provider, have a more varied diet than those whose appetite is poor, whose food knowledge is low, have good access to convenient foods but do not have good access to a good service provider. In addition, those who have good kitchen appliances, good access to high-quality products, have a short distance to go to the shops and have the support of friends and neighbours have a more varied diet than those who think they are low on these resources. It is important to note that some of the perceived resources like income, access to transport, sharing meals and health may be insignificant here because the effects of these were already captured by related objective measures in the first step. Food knowledge, access to a good service provider and good appetite appear to be stronger predictors than the other five significant predictors.
Influence of goals on variety of diet
Regression of varied eating v. the eleven goals in the third step found that the increase in explained variance of 3 % was significant (F change (11, 1446) = 4·31, P < 0·001). Goals such as controlling weight (β = 0·08, P < 0·01), having a variety of foods on the menu (β = 0·13, P < 0·001) and cooking for others (β = 0·06, P < 0·05) were significant predictors, whereas keeping expenditure low, eating a healthy diet, choosing convenient food, enjoying food and meals, eating in other’s company, not receiving help, maintaining food culture and eating in a nice surrounding were found not to influence a varied food choice. This suggests that those who want to eat a varied diet or control their diet or cook for others eat a more varied diet than those who do not strive to achieve these goals. Of these goals, wanting to have a variety of foods on the menu was the best predictor of having a varied diet.
Discussion
Results showed that the actual resources older people have affect the quality of their diet in terms of how many different foods they eat. This is not surprising as we would expect people with more money to be able to afford a better quality and variety of food. These results confirm previous findings(Reference Sharpe, Huston and Finke13, Reference Blaylock, Smallwood, Kassel, Variyam and Aldrich23, Reference Nord42, Reference Quinn, Johnson, Poon, Martin and Nickols-Richardson43). For example, Nord(Reference Nord42) found that three-quarters of food-insecure elderly households obtained enough food to avoid hunger by eating a less varied diet. Banister and Bowling(Reference Banister and Bowling19) found that those older people able to travel by car to shops were able to buy and consume a wider range of food.
The impact of physical and mental health on dietary variety is also plausible, as those who are not in pain, have more energy and without emotional problems are more likely to eat a varied diet. This is in line with previous findings(Reference Hendy, Nelson and Greco14) showing health to affect food-related behaviours. The finding that those who live alone eat a narrower diet than those living with a partner fits well with previous research(Reference Sahyoun and Zhang20). Loneliness and lack of incentive to cook for one may contribute to why those who live alone eat a less varied diet. Studies have shown that eating with a television on or eating in a communal setting (e.g. community centres) widens the diet of those who live alone(Reference De Castro44, Reference Stroebele and de Castro45).
In addition to the actual resources, perceived resources were found to influence older people’s diets. Those who think they have a good appetite and know a lot about food were found to eat a more varied diet than those who said their appetite was poor or didn’t know much about food. This suggests that increasing older people’s food knowledge, by means such as information provision and cooking classes, may be a possible way to increase dietary variety. Increasing older people’s access to a good service provider and improving their kitchen facilities are alternative means of increasing dietary variety. Sharpe et al.(Reference Sharpe, Huston and Finke13) also found that lack of access to services such as meals on wheels served as a barrier to healthy eating for rural elderly.
Increasing access to high-quality meals was also found to increase older people’s varied food intake, although access to organic foods or new products did not have significant effects. This suggests that older people’s diets could be improved by providing good-quality foods that are familiar rather than new products and not necessarily organically grown.
In terms of social resources, having help from friends and neighbours rather than family increased the variety of older people’s diets. This suggests that friends’ and neighbours’ help may be related to food insomuch as they may help with the shopping or taking the person to the shops, whereas help from family may be needed when there is illness or on more serious matters. Interestingly, while living with a partner did affect the quality of the diet, sharing meals with others did not contribute to a varied diet, suggesting that those who share meals with others eat the same range of foods as those who do not share meals with others.
The study also found that older people’s food-related goals had an independent influence on their quality of diet over and above their actual and perceived resources. Unsurprisingly, those whose food-related goals include eating a varied diet ate a greater variety of foods than those who did not have this as a goal. Thus one way to change a narrow diet would be to find ways to influence older people to have this as one of their food-related goals, perhaps through information expounding the merits of a varied diet as well as examples of how to increase variety in a diet. In addition, those who wanted to cook for others also had a more varied diet than those who did not have this as a goal. Here too providing recipes and information to those who want to cook for others on how to increase the variety of foods used in their cooking can help. Surprisingly, those who wanted to control their weight also were found to have a more varied diet; similar results were demonstrated by Oaks and Slotterback(Reference Oaks and Slotterback32) where people’s judgement of healthiness of food was influenced by whether or not they had dieting as a food-related goal. It could be argued that when trying to lose weight one needs to think more about what one eats and plan the meals carefully. This may give people the opportunity to think about the variety of foods they can eat, influencing their food intake. It could also be that when wanting to lose weight people have to learn more about foods in general and the increase in food knowledge may help to increase the types of foods they eat, as discussed earlier.
Taken together, the present results suggest that older people’s variety of food intake does depend on resources such as their monthly income, whether they have access to a car, their living arrangement and their physical and mental health. However, in addition to these variables, the way older people perceive other resources such as their level of appetite, their food knowledge, their perception of the distance to the shops, access to high-quality products, having better kitchen facilities, access to good service providers and support from their friends and neighbours all contribute to how varied a diet they eat. Further, older people’s goals such as wanting to eat a varied diet, wanting to cook for others and wanting to control their weight also influence the variety in their diet.
The findings that perceived levels of resources and food-related goals impact on dietary variety gives additional means of possible interventions that can be used to encourage and influence older people’s variety of diet. The results suggest that by encouraging older people to adopt certain food-related goals, such as wanting to eat a varied diet, wanting to cook for others or watching their weight, it would be possible to change their eating habits. Similarly, by increasing older people’s food and nutrition knowledge it may be possible to increase dietary variety. A varied diet could also be fostered by making high-quality products accessible to older people.
The study has a number of limitations. Food variety was measured using the weekly food variety score as it was easy to use across different countries. There are many other ways of measuring variety(Reference Arvaniti and Panagiotakos46). Future studies should compare how the different food variety measures square up against this one. There were only a limited number of indicators available for objective resources, and most resources were measured only as perceived by respondents. A more complete set of objective resource indicators could not only shed more light on the relationship between objective and perceived resources, but also aid in deriving implications for how older people’s endowment with relevant resources can be improved. We should also note that responses to an inventory of self-reported resource items may be subject to halo effects. Future research should investigate the mechanisms through which these above-mentioned possible interventions influence dietary change. Further, additional food-related goals that may increase variety in older people’s diet should be elicited and their effects studied in order to develop further intervention strategies and policies, resulting in a healthier elderly population.
Acknowledgements
Sources of funding: The study was carried out with financial support from the Commission of the European Communities, specific RTD programme ‘Quality of Life and Management of Living Resources’, QLK1-2002-02447, ‘Choosing foods, eating meals: sustaining independence and quality of life in old age’. It does not necessarily reflect its views and in no way anticipates the Commission’s future policy in this area. Preparation of this paper benefited also from a grant from the Danish Research Council, Programme Committee for Food and Health. Conflict of interest declaration: The authors have no conflicts of interest. Author contributions: M.D. – analysis and writing; M.R. and K.G.G. – design and editing; M.L. – design. Acknowledgements: The Food in Later Life Project Team comprises research scientists, interviewers, technicians, administrative staff and managers who continue to make the study possible. We are extremely grateful to all the representatives from day-care centres, local authorities, industry and professional bodies who took part.
