European Journal of Clinical Nutrition (2009) 63, S69S74

& 2009 Macmillan Publishers Limited All rights reserved 0954-3007/09 $32.00
www.nature.com/ejcn

ORIGINAL ARTICLE

Dietary assessment in elderly people: experiences gained from studies in the Netherlands
JHM de Vries, LCPGM de Groot and WA van Staveren
Division of Human Nutrition, Wageningen University, Bomenweg 2, Wageningen, The Netherlands

Background/Objectives: In selecting a dietary assessment method, several aspects such as the aim of the study and the characteristics of the target population should be taken into account. In elderly people, diminished functionality and cognitive decline may hamper dietary assessment and require tailored approaches to assess dietary intake. The objective of this paper is to summarize our experience in dietary assessment in a number of different studies in population groups over 65 years of age in the Netherlands, and to discuss this experience in the perspective of other nutrition surveys in the elderly. Methods: In longitudinal studies, we applied a modified dietary history; in clinical nursing home studies, trained staff observed and recorded food consumption; and in a controlled trial in healthy elderly men, we used a food frequency questionnaire (FFQ). Results: For all methods applied in the community-dwelling elderly people, validation studies showed a similar underestimation of intake of 1015% compared with the reference value. In the care-depending elderly, the underestimation was less: 5% according to an observational method. The methods varied widely in the resources required, including burden to the participants, field staff and finances. Conclusions: For effective dietary assessment in older adults, the major challenge will be to distinguish between those elderly who are able to respond correctly to the less intensive methods, such as 24-h recalls or FFQ, and those who are not able to respond to these methods and require adapted techniques, for example, observational records.

European Journal of Clinical Nutrition (2009) 63, S69S74; doi:10.1038/ejcn.2008.68

Keywords: aged; nutrition assessment; observation; diet; methods

Introduction
Elderly people may suffer from many health problems due to chronic diseases that are often related to nutritional intake (Dwyer, 2006). In addition, several changes occur during ageing, which directly or indirectly influence consumption in the elderly. For example, the ability to smell and taste may diminish, or living conditions may change. A better understanding of the changes in intake of elderly people and the relationship with chronic diseases is important to prevent these diseases by nutritional interventions. For this purpose, accurate assessment of intake is key. Several methods are available to assess dietary intake, and each has specific characteristics. For studies in healthy people, food records may be used to assess current intake as compared with 24-h recalls, food frequency questionnaires (FFQ) or dietary history method for assessing past

Correspondence: Dr JHM de Vries, Division of Human Nutrition, Wageningen University, Bomenweg 2, PO Box 8129, 6700 EV Wageningen, The Netherlands. E-mail: Jeanne.devries@wur.nl

intake. For studies in disabled older adults or studies focusing on specific nutrients or other specific purposes, an observed record method, picture sort method or biomarkers may be , 2006). the first choice (van Staveren and Ocke To choose an appropriate dietary assessment method, several aspects have to be taken into consideration (van Staveren et al., 1994). Studies have different purposes requiring the assessment of dietary intake on the individual or group level: studying nutrients, foods or dietary patterns; or investigating associations with the outcome of disease or the adequacy of intake. The best method for the purpose of the study offers the required level of detail about foods or nutrients as well as the appropriate reference period. It is also suitable for the target population, and requires the least resources with regard to time and burden to participants, staff and finances. For the elderly, some additional considerations may play a role in the choice of method. It is difficult to choose one best method for the elderly, because the population is a very heterogeneous group, ranging from the over sixties to centenarians, and from healthy, physically active and independently living people to those who are fully

Dietary assessment in elderly people JHM de Vries et al

S70
dependent on care. Thus, both mental and physical impairments may call for specific adaptations to dietary assessment methods. In the past decades, the Department of Human Nutrition of Wageningen University initiated or became involved in a number of nutritional studies in elderly people over 65 years of age. These studies included epidemiological studies, controlled trials, and nursing home studies. This paper summarizes our experience gained in dietary assessment in elderly people over 65 years of age, which was conducted in order to frame recommendations for valid and efficient studies in different health categories of this population. these minor differences, the data of SENECA and the FINE study were merged and were used for further analysis, including dietary patterns in the HALE study (Knoops et al., 2006). In a validation study, the modified dietary history method as applied in the SENECA study showed acceptable agreement with a 3-day weighed food record in 82 participants in seven centers (Nes et al., 1991), although the data from the dietary history resulted in consistently higher intakes than that from the weighed food record. The differences ranged from 1% (cholesterol) to 28% (mono- and disaccharides) with a median difference of 14% for energy. Pearsons correlation coefficients varied from nearly 0.5 to 0.8 with a lower value for vitamin A; correlations for most nutrients were considered acceptable for ranking intake of individuals. Experiences from this validation study indicated that the weighed record might not be a good reference method for older adults. Therefore, the dietary history was also validated against whole body indirect calorimetry in 12 elderly women. In this study, underreporting appeared to be 12%,

Assessment in community-dwelling populations

SENECA, FINE and HALE studies: the dietary history method One of the major objectives of the SENECA study was to investigate the variation in dietary patterns of elderly people across Europe (de Groot et al., 2004; de Groot and van Staveren, 2000). Information on usual dietary intake of the past month was collected both cross-sectionally and longitudinally during a 10-year follow-up period using a modified diet history method. This method consisted of a combination of an estimated 3-day food record and a mealbased list of foods to check the usual consumption of the earlier month. Portion sizes were based on standard portion sizes or were checked by weighing. A 3-day estimated food record was included to make participants conscious of their food intake (see Table 1). The FINE study was an elderly study in survivors of five cohorts of the Seven Countries Study. In the FINE study, a dietary history method was also used, but it covered only the previous 24 weeks. Another difference was that the inquiry about usual food patterns was on the basis of an oral interview and not on 3-day food records. Accepting

Table 2 Reported energy intake and difference in energy needs, according to dietary history and indirect calorimetry in women (study 1), and according to food frequency questionnaire and energy requirements to maintain stable body weights in a controlled feeding trial in men (study 2) Study 1 women (n 12) Mean Age years Height m Bodyweight kg BMI kg/m2 Reported energy kJ Difference % 74 1.60 65.4 25.6 7166 11.6 s.d. 3 0.05 11.6 4.7 1532 7.4 Study 2 men (n 17) Mean 76 1.76 75.8 24.5 9880 13.2 s.d. 1.5 0.04 7.6 1.9 2.7 17.4

Abbreviation: BMI, body mass index.

Table 1 Overview of the studies Study Community dwelling Seneca Fine Hale Validation studies Seneca Validation DH Balans study Institutions Nursing homes Reference Period Subjects n 2600 2285 3117 Gender Method

de Groot et al. (2004) Menotti et al. (2001) Knoops et al. (2006)

19881999 19902000 19882000

M/F M M/F

DH 3-day estimated FR DH Combination of SENECA and FINE

Nes et al. (1991) Visser et al. (1995) Not published

1990 1994 2008

82 12 17

M/F F M

DH vs 3-day weighed FR DH vs indirect calorimetry FFQ vs energy expenditure

Nijs et al. (2006a, b)

20022003

178

M/F

Observation

Abbreviations: M, male; F, female; DH, dietary history both; FR, food record; FFQ, food frequency questionnaire.

European Journal of Clinical Nutrition

Dietary assessment in elderly people JHM de Vries et al

S71
similar to the dietary history (see Table 2; Visser et al., 1995; van Staveren et al., 2002). Validation of the method against urinary nitrogen showed that protein intake was underestimated by 10%. The extent of underestimation was not related to body weight or body fat but to the level of intake, showing that women with a low intake underreported more. Information from the dietary history method was also used for investigating the relationship of Mediterranean food patterns with morbidity and mortality in the HALE study (Knoops et al., 2004). To apply Mediterranean diet scores, individual food items were combined into food groups using the European classification system, Eurocode (Knoops et al., 2006). The composition of the scores appeared to be successful because three different applications of the Mediterranean diet scores were predictive of beneficial health outcomes, such as survival. We concluded that the modified dietary history applied in the SENECA, FINE and HALE studies appeared to give reliable results in describing the variety in dietary patterns among participating centers and in identifying dietary factors predicting health and survival.

Care-dependent elderly
Observation of intake In several studies in the institutionalized elderly (Manders et al., 2006; Nijs et al., 2006a, b), including the elderly living in long-term care facilities, trained dietitians used an observation and direct measurement method. We chose this method because older residents in these homes often suffer from cognitive decline, which make FFQs, 24-h recalls or other methods relying on memory, less reliable (van Staveren et al., 1994). In these studies, the observational method was applied to investigate the effect of family-style meals on energy intake and to determine the risk of malnutrition in Dutch nursing home residents with a mean age of 77 years, often with cognitive decline but without severe dementia. On average, three trained observers per ward recorded the food and beverage intake of 25 residents during a whole day (Nijs et al., 2006a, b). For the cooked meal, individual menus and recipes were obtained from the kitchen staff. Food consumption of the cooked meal was registered by keeping records of foods and portion sizes served; after the meal, leftovers were weighed and the amount subtracted from the standard portion size. The intake of other meals during the day was observed and recorded in terms of household measures and standard portion sizes. All foods and beverages consumed outside regular mealtimes were also carefully observed. During the night, nurses noted down snacks or beverages consumed. The content of dietary equipment in the wards, such as glasses and spoons, and portion sizes of recipes and meals, were measured thrice. The mean was used as a standard portion. We did not measure on weekends because we expected that the diets of the elderly during these days would be very similar to those during weekdays. Using the observed record method, very low daily energy intakes varying from 5.5 to 6.2 MJ were observed in these institutionalized elderly (Nijs et al., 2006a, b). The low reports might be due to underreporting, but in a validation study this method showed only 5% underestimation of energy intake, (de Jong and Secreve, 1999) because energy expenditure also appeared to be very low in this group. Thus, although an observational record is very burdensome and time-consuming, this method proved its usefulness by showing energy and nutrient deficiencies in the institutionalized elderly.

Assessment in controlled studies: the food frequency questionnaire The FFQ is a popular method of dietary assessment because it is easy to apply and is relatively cheap. We regularly develop new FFQs in a systematic way for different study purposes and age groups (Feunekes et al., 1993; Molag et al., 2007; Verkleij-Hagoort et al., 2007). Foods to be included in the food list of the questionnaire are selected on the basis of their contribution to the absolute intake of the nutrient of interest of the specific age group in the Dutch National Food Consumption Survey. For the AlphaOmega trial, an intervention study investigating the effect of n-3 fatty acids on cardiovascular disease (http://www.clinicaltrials.gov/ct2/ show/NCT00139464), we used this approach to develop a food list for an FFQ to be applied in an elderly population aged 6580 years. The selected foods for the food list were similar to those identified for younger adults, indicating that the food list of the FFQs did not need to be adapted for an elderly population. This was confirmed in a recent controlled feeding trial in apparently healthy men. Just before the trial, we asked the men to fill out an FFQ that was developed for a younger adult population to assess their energy level. We compared the results of the reported energy intake by the FFQ with the energy needs for weight maintenance during the trial. Seventeen elderly male participants with a mean age of 76 years and a body mass index of 24.5 kg/m2 reported on average 87% of their energy needs (see Table 2). The correlation coefficient between reported intake by the FFQ and energy needs during the trial was 0.67 (P 0.003). Body mass index was negatively associated with the difference between reported energy intake and energy needs (R 0.45). Thus, this validation study shows that applying an FFQ to assess the consumption of healthy elderly men might be a good choice.

Discussion
We assessed dietary intake in several studies with varying aims and in different elderly population groups. The applied dietary assessment methods appeared to give reliable results for the different objectives of these studies. For example, the dietary history method in the SENECA study showed an underestimation of energy intake (12%) in elderly women, similar to the underestimation in younger age groups and European Journal of Clinical Nutrition

Dietary assessment in elderly people JHM de Vries et al

S72
similar to the underestimation of 13% by the FFQ in healthy elderly men participating in a controlled feeding study. The observed records in the nursing home studies showed an underestimation of intake of 5%, indicating greater validity than the above methods in a healthier group. In the community-dwelling populations, we used a modified dietary history method, in which attention was given to the meal pattern either by using 3-day food records or through direct questioning by the interviewer. The advantage of this approach is that it helps the participant to recall his or her dietary intake. The dietary history method gave higher estimates than did a weighed dietary food record. Other studies in the elderly also showed that mealbased approaches based on the dietary history method yielded higher estimates of intake than did other methods. (Black et al., 2000; Quandt et al., 2007). On the other hand, the dietary history method still underestimated intake when compared with indirect calorimetry. Thus, although the method has many advantages, low energy reporting was still evident. A dietary history method is not an easy method to apply. It requires well-trained interviewers and a large effort for both interviewers and respondents. In a longitudinal design, cognitive decline is expected to occur in the target population, and therefore a dietary history combined with a record at baseline might be a good choice. An FFQ is much easier to apply than a dietary history, but like 24-h recalls, it is often not used in elderly people because it relies on memory (van Staveren et al., 1994). On the other hand, no strong evidence exists that elderly people provide less valid self-reports using these methods compared with younger populations. Specific problems when using this method have not been reported by researchers (Biro et al., 2002). Our validation study of an FFQ in elderly men showed that this method might be applicable among healthy agers whose memory skills are still intact. Our study also indicated that the food list of the FFQ could be the same as for younger adults. Using the same list is an advantage if intakes of different age groups within one population have to be compared. The results of this validation study may not be generalized to other populations. It was performed in a small number of well-motivated men having a Dutch eating pattern. Moreover, they were, in the main, highly educated and not obese. These factors may have contributed to the small differences found between the FFQ and the reference method applied. Other studies showed, for instance, that men are more likely to provide relatively lower reports on an FFQ than women (Tooze et al., 2007). For community-dwelling populations, the use of other dietary assessment methods than dietary histories or FFQs may also be considered. For example, 24-h recalls could be useful in studies monitoring intake of populations (Brussaard et al., 2002). In general, it is thought that 24-h recalls show less underreporting than do FFQs (Subar et al., 2003), although the results are not consistent (Tooze et al., 2007). According to a validation study in the EPIC (European European Journal of Clinical Nutrition Prospective Investigation into Cancer and Nutrition) study, using two 24-h recalls and comparing them with cutoff limits according to Goldberg, older people were less likely to underestimate energy intake than younger people (Ferrari et al., 2002). On the other hand, it has also been reported that elderly people omit foods by 24-h recalls (Tooze et al., 2007) because of social desirability or memory problems. Underpinning these methods with pictures or the picture sort method (Kumanyika et al., 1997) may improve the quality of the data. It would be worthwhile to investigate whether, in monitoring studies, 24-h recalls could provide accurate information on dietary intake of the healthy elderly population because of the potential to compare results with younger populations. In the institutionalized elderly, an observation and direct measuring method appeared to be a useful but timeconsuming and expensive method. If one is not certain about the abilities of the population of interest, the observation records seem to be a safe choice especially in the care-dependent elderly. However, in clinical daily practice, it will be impossible to apply such an intensive method for all residents. For some specific purposes, other methods may be more practical. If the interest is only energy needs, cheaper and simpler methods such as the HarrisBenedict prediction formula in combination with an estimate for physical activity could be chosen. Some studies (Bardoel et al., 2001), Gaillard et al., 2007) showed that the HarrisBenedict equations can accurately predict resting energy expenditure, adjusted for difference in body weight and fat-free mass of both the successful ageing adult and the older adult with physical impairments. For assessing dehydration, frequently weighing elderly people is a simple but best practice method to monitor changes in water balance. In older residents suffering from edema, this method is however not valid (Schols et al., in press). To assess the exposure to salt or vitamins, such as vitamin D, the use of biological markers in blood or urine could be considered. Finally, screening tools to identify residents at risk for malnutrition may be more efficient than a method by self-report, but as it is beyond the scope of this article, these tools are not discussed further. In future studies, it will be important to obtain a good picture of the nutritional situation of the elderly population as poor intakes are associated with increased risk of poor health, including functional decline. On the other hand, these functional impairments, including mental and physical performance, make it more difficult to get accurate nutritional information from them. Therefore, it is very important that the choice of dietary assessment method should take into account cognitive skills and other characteristics of the elderly population. Memory may already start declining after the age of 50 years (Draaisma, 2008; Gauthier et al., 2006). This means that in an apparently healthy population, the individual elderly might not be able to give good estimates of intake.

Dietary assessment in elderly people JHM de Vries et al

S73
It is suggested that the elderly have more problems with short-term recalls than with long-term recalls, and more with open-ended recalls than with structured recalls using a fixed list with names (Draaisma, 2008). Thus, recalling habitual intake by FFQ using a predefined food list, if necessary underpinned with pictures in which the elderly can recognize the foods they consume, might supply more reliable data than by recalling the current intake of the past 24 h, which relies on their very short-term memory. However, until now, this hypothesis has not been tested in nutrition research. A final challenge is to deal with the decline of functionality.
de Jong S, Secreve A-F (1999). To determine the energy expenditure and intake of patients of a geriatric ward. MSc-thesis, Nijmegen. Draaisma D (2008). De Heimwee Fabriek (in Dutch). Historische uitgeverij: Groningen. Dwyer J (2006). Starting down the right path: nutrition connections with chronic diseases of later life. Am J Clin Nutr 83, 415S420S. Ferrari P, Slimani N, Ciampi A, Trichopoulou A, Naska A, Lauria C et al. (2002). Evaluation of under- and overreporting of energy intake in the 24-hour diet recalls in the European Prospective Investigation into Cancer and Nutrition (EPIC). Public Health Nutr 5, 13291345. Feunekes GI, Van Staveren WA, De Vries JH, Burema J, Hautvast JG (1993). Relative and biomarker-based validity of a food-frequency questionnaire estimating intake of fats and cholesterol. Am J Clin Nutr 58, 489496. Gaillard C, Alix E, Salle A, Berrut G, Ritz P (2007). Energy requirements in frail elderly people: a review of the literature. Clin Nutr 26, 1624. Gauthier S, Reisberg B, Zaudig M, Petersen RC, Ritchie K, Broich K et al. (2006). Mild cognitive impairment. Lancet 367, 12621270. Knoops K, De Groot LCPMG, Fidanza F, Alberti-Fidanza A, van Staveren WA (2006). Comparison of three different dietary scores in relation to 10-year mortality in elderly European subjects: the HALE project. Eur J Clin Nutr 60, 746755. Knoops KT, de Groot LC, Kromhout D, Perrin AE, Moreiras-Varela O, Menotti A et al. (2004). Mediterranean diet, lifestyle factors, and 10-year mortality in elderly European men and women: the HALE project. JAMA 12, 14331439. Kumanyika SK, Tell GS, Shemanski L, Martel J, Chinchilli VM (1997). Dietary assessment using a picture-sort approach. Am J Clin Nutr 65 (Suppl), 1123S1129S. st L, Manders M, de Groot LCPMG, Blauw YH, van Hoekel-Pru Bindels JG, Siebelink E et al. (2006). Effect of a nutrient enriched drink on dietary intake and nutritional status in institutionalized elderly. In: Manders M (ed). Nutritional Care in Old Age, the Effect of Supplementation on Nutritional Status and Performance: Wageningen: University, thesis. Menotti A, Mulder I, Nissinen A, Giampaoli S, Feskens EJM, Kromhout D (2001). Prevalence of morbidity and multimorbidity in elderly male populations and their impact on 10-year all-cause mortality: the FINE study (Finland, Italy, Netherlands, Elderly). J Clin Epidemiol 54, 680686. Molag ML, de Vries JHM, Ocke MC, Dagnelie PC, van den Brandt PA, Jansen MC et al. (2007). Design characteristics of food frequency questionnaires in relation to their validity. Am J Epidemiol 166, 14681478. Nes M, van Staveren WA, Zajkas G, Inelmen EM, Moreiras-Varela O (1991). Validity of the dietary history method in elderly subjects. Euronut SENECA investigators. Eur J Clin Nutr 45 (Suppl 3), 97104. Nijs KA, de Graaf C, Kok FJ, van Staveren WA (2006a). Effect of family style mealtimes on quality of life, physical performance, and body weight of nursing home residents: cluster randomised controlled trial. Br Med J 332, 11801184. Nijs KA, de Graaf C, Siebelink E, Blauw YH, Vanneste V, Kok FJ et al. (2006b). Effect of family-style meals on energy intake and risk of malnutrition in Dutch nursing home residents: a randomized controlled trial. J Gerontol A Biol Sci Med Sci 61, 935942. Quandt SA, Vitolins MZ, Smith SL, Tooze JA, Bell RA, Davis CC et al. (2007). Comparative validation of standard, picture-sort and mealbased food-frequency questionnaires adapted for an elderly population of low socio-economic status. Public Health Nutr 10, 524532. Schols JMGA, de Groot LC, van der Cammen TJM, Olde Rikkert MGM. Preventing and treating dehydration in the elderly during periods of illness and warm weather (in press).

Conclusion
When assessing diets in the elderly, various approaches must be considered in the light of the populations characteristics and abilities. In the healthy elderly, it may not be a problem to use dietary recalls or FFQ, whereas in the infirm elderly, observed records and different measures for nutritional status should be considered. However, the problem is that we are often not able to recognize whether cognitive and other functional skills (vision, hearing, ability to write) of the elderly are sufficient to recall or record their intake. For this, additional research on the relationship between functionality (cognition and other functions) and accurate reporting is urgently needed.

Disclosure
The authors have declared no financial interests.

References
Bardoel EAG, Dicke HC, De Groot LC, Hoefnagels WHL (2001). Schatting van het rustmetabolisme van geriatrische patienten op tisten 56, 172179. basis van predictieformules. Ned Tijdschr Die Biro G, Hulshof KFAM, Ovesen L, Amorim Cruz JA, Efcosum group (2002). Selection of methodology to assess food intake. Eur J Clin Nutr 65, S25S32. Black AE, Welch AA, Bingham SA (2000). Validation of dietary intakes measured by diet history against 24 h urinary nitrogen excretion and energy expenditure measured by the doublylabelled water method in middle-aged women. Br J Nutr 83, 341354. Brussaard JH, Lowik MR, Steingrimsdottir L, Mller A, Kearney J, De Henauw S et al. (2002). A European food consumption survey methodconclusions and recommendations. Eur J Clin Nutr 56 (Suppl 2), S89S94. de Groot LC, van Staveren WA (2000). SENECAs accomplishments and challenges. Nutrition 16, 541543. de Groot LCPMG, Verheijden MW, de Henauw S, Schroll M, van Staveren WA, SENECA investigators (2004). Lifestyle, nutritional status, health, and mortality in elderly people across Europe: a review of the longitudinal results of the SENECA study. J Gerontol A Biol Sci Med Sci 59A, 12771284.

European Journal of Clinical Nutrition

Dietary assessment in elderly people JHM de Vries et al

S74
Subar AF, Kipnis V, Troiano RP, Midthune D, Schoeller DA, Bingham S et al. (2003). Using intake biomarkers to evaluate the extent of dietary misreporting in a large sample of adults: the OPEN study. Am J Epidemiol 158, 113. Tooze JA, Vitolins MZ, Smith SL, Arcury TA, Davis CC, Bell RA et al. (2007). High levels of low energy reporting on 24-hour recalls and three questionnaires in an elderly low-socioeconomic status population. J Nutr 137, 12861293. van Staveren WA, de Groot LC, Haveman-Nies A (2002). The SENECA study: potentials and problems in relating diet to survival over 10 years. Public Health Nutr 5, 901905. van Staveren WA, De Groot LCPMG, Blauw YH, van der Wielen RPJ (1994). Assessing diets of elderly people: problems and approaches. Am J Clin Nutr 59 (Suppl), 221S223S. MC (2006). Estimation of dietary intake. In: van Staveren WA, Ocke Bowman BA, Russell RM (eds). Present Knowledge in Nutrition, vol. 2, ILSI: Washington DC. pp 795807. Verkleij-Hagoort AC, de Vries JH, Stegers MP, Lindemans J, Ursem NT, Steegers-Theunissen RP (2007). Validation of the assessment of folate and vitamin B12 intake in women of reproductive age: the method of triads. Eur J Clin Nutr 61, 610615. Visser M, De Groot LC, Deurenberg P, Van Staveren WA (1995). Validation of dietary history method in a group of elderly women using measurements of total energy expenditure. Br J Nutr 74, 775785.

European Journal of Clinical Nutrition