Demographics, health-related behaviors, eating habits and knowledge associated with vegetable intake in Japanese adults

Abstract

Objectives:

To analyze demographic, health-related behaviors, eating habit and knowledge associated with vegetable intake.

Methods:

Secondary analyses using the dataset from the National Health and Nutrition Survey 2003. Food intake data measured by the food-weighing method in one-day and a questionnaire assessed the dietary intake and health-related behaviors, eating habit and knowledge. This study was made in Japan. The data of 1742 men and 2519 nonpregnant/nonlactating women, aged 20–69 years, energy intake between percentiles 1 and 99 were included. Vegetable intake was analyzed according to the Japanese vegetable recommendation (⩾350 g/day) after age adjustment.

Results:

Average of VI was 307 g/day in men and 297 g/day in women. Only 35% of men and 31% of women met the recommended amount of vegetable intake. Japanese from city areas, aged 60–69 years, had the highest vegetable intake and subjects from metropolitan areas had the lowest vegetable intake. Depending on the age groups, risks for low vegetable intake in Japanese were found in subjects with skipping meals, alcohol intake and history of smoking.

Conclusions:

To increase vegetable intake, it is necessary to provide more nutritional education and lifestyle-related diseases education.

Introduction

There is strong and consistent epidemiological evidence that vegetable intake (VI) is beneficial to health, decreasing the risk for a range of chronic diseases and many cancers (Mozaffarian et al., 2003; Sauvaget et al., 2003; World Health Organization, 2004; Pomerleau et al., 2006). As a consequence, in many countries, dietary guidelines include recommendations for vegetable consumption. In Japan, a VI of ⩾350 g/day is recommended. However, most Japanese consume less than this amount. The National Health and Nutrition Survey of 2003 (NHNS, 2003) showed that the average daily VI was 278 g. The highest intake was among the Japanese aged between 60 and 69 years, though it was still below the recommended amount (Ministry of Health, Labour and Welfare, 2005). Therefore, health authorities are making an effort to develop dietary programs, such as the Food Guide Spinning Top (Yoshiike et al., 2007), to encourage people to eat a balanced diet by choosing enough servings of vegetable dishes. Many research studies have been carried out in a number of countries to ascertain the determinants of fruit and vegetable consumption. The majority of the studies have focused on household income (Kirkpatrick and Tarasuk, 2003; Laaksonen et al., 2003), regional differences (Pollard et al., 2001; Papadaki and Scott, 2002) and understanding the psychosocial and sociodemographic determinants of fruit and VI. These include knowledge, perception of benefits and barriers, food preparation skills, gender and social status (Satia et al., 2002; Friel et al., 2004; Larson et al., 2006). Furthermore, a higher VI is associated with eating homegrown vegetables (Billson et al., 1999), eating vegetables at lunch or dinner, eating salads (Satia et al., 2002) and skill in preparing vegetables (Larson et al., 2006; Crawford et al., 2007). In addition, other studies showed that distress was associated with unhealthy habits, such as low VI (Unusan, 2006). In Japan, very little research has been conducted on these topics. In terms of initiatives to promote increased consumption, a more thorough understanding of the behavioral correlates of VI is likely to be important; thus, the aim of this paper is to analyze demographics, health-related behaviors, eating habits and the knowledge associated with VI among Japanese adults, because such information could be beneficial for developing effective dietary interventions.

Methods

NHNS 2003 data

NHNS 2003 data is a cross-sectional survey of a nationally representative sample of the noninstitutionalized population of Japan. It includes (1) a physical examination (anthropometry measurements, blood pressure, blood test, a questionnaire on medication, smoking status, alcohol intake, exercise and number of steps measured by a pedometer), (2) a dietary survey that involves weighing the amount of food consumed over a day by a household and individual household members and (3) a questionnaire on health-related behaviors, eating habits and knowledge (Iwaoka et al., 2001).

The design and protocol of the national survey conducted by the Ministry of Health, Labour and Welfare of Japan (MHLWJ) were thoroughly reviewed by the technical committee on the survey in the ministry, and also approved by the Ministry of Internal Affairs and Communications in the government office, including the ethical issues. Detailed explanation was made to the households selected as the survey sample by the dietitians of the local public health centers and informed consent was obtained from the households participating in the survey.

Subjects

We performed a secondary analysis of the dataset from NHNS 2003 with the permission of the MHLWJ. From 11 105 subjects, 1742 men and 2519 women were selected according to the following inclusion criteria: age (between 20 and 69 years), energy intake (between percentiles 1 and 99 according to sex cutoff (910–4015 kcal for men and 726–3079 kcal for women)) and those who completed the health-related behaviors, eating habits and knowledge questionnaire. The inclusion criteria for energy were restricted to percentiles 1–99 to exclude outlying data for the analyses. The exclusion criteria were pregnant/lactating women.

Vegetable classification

Green-yellow vegetables, light color vegetables, pickles made from vegetables, vegetable juice. Mushrooms, seaweeds and plants foods that contain variable amounts of starch, such as potato, were not included in vegetable groups.

Demographics, health-related behaviors, eating habits and knowledge

A questionnaire about demographic factors, health-related behaviors, eating habits and knowledge were carried out. It was assessed as follow:

• Demographic factors: age, sex, region of residence (12 regions: Hokkaido, Tohoku, Kanto I, Kanto II, Hokuriku, Tokai, Kinki I, Kinki II, Chugoku, Shikoku, Northern Kyushu and Southern Kyushu), area of residence (metropolitan: more than 150 000 people, city: between 50 000 and 150 000 people and town/rural: less than 50 000 people).

• Health-related behaviors: exercise habit (yes, no), perception of overall health status (very good, good, not good), average hours of sleep (less than 6 h or more than 9 h, 7–9 h), feeling dissatisfaction, distress or burden and others (very often, sometimes, never), history of smoking (yes, no), frequency of alcohol intake (every day, 6 days/week to 1 day/month, gave up/rarely).

• Eating habits: skipping meals (yes, no), eating snacks (yes, no).

• Knowledge: knowledge about ‘Health Japan 21’ (yes, no), ‘lifestyle-related diseases (LSRD)’ (I know the content, I have heard about it, I do not know what it is), attendance at health-related education programs (yes, no).

Statistical analysis

All analyses were performed using the Statistical Package for Social Science (version 11; SPSS Inc., Chicago, IL, USA). Data were expressed as mean and 95% confidence interval (CI) according to sex and age groups (20–39, 40–59 and 60–69 years).

Two approaches were used to assess the relationship between vegetable consumption and other factors. First, we used a general linear model (GLM) with VI as the dependent variable, and all other independent factors as categorical variables: demographics, health-related behaviors, eating habits and knowledge factors. We determined age-adjusted means of VI across factors and correlations with P values <0.05 to be significant.

In the second approach, we computed the proportion of subjects who ate a total of ⩾350 g/day of vegetables, the intake recommended for adults by ‘Health Japan 21’. Factors correlated with meeting the recommendation of VI were assessed the odd ratio (OR), with 95% CI, estimated from a multiple logistic regression after age adjustment.

Results

All the analyses were carried out in the subset of 4261 subjects, 1742 men and 2519 women, who met all the inclusion criteria.

In Table 1 the body mass index (BMI) and nutrient intakes among Japanese adults is shown. BMI was slightly higher in men aged 40–59 years, and in women aged 60–69 years.

Table 1 BMI and macronutrient intake according to group of age

Energy intake was significantly higher in women aged 40–59 years. For both men and women, protein intake tends to increase with age whereas fat intake decreases. Mean VI was 307 g for men and 297 g for women. The proportion of those who met the recommendation of ⩾350 g of VI per day was 35% in men and 31% in women (Table 2).

Table 2 Vegetable intake in Japanese adults

For men and women aged 20–39 years, mean of VI was 265 and 243 g, respectively. For those aged 30–49 years, it was 299 and 302 g, and for the oldest group, it was 359 and 343 g, indicating a trend for VI to increase with age.

By region of residence, the regions with the highest VI were Hokuriku (central northeast Japan) for men and Kanto II (central Japan) for women. The lowest VI was in Shikoku for both men and women (P<0.05; Table 2).

Men and women from the city areas aged 60–69 years had the highest VI, whereas men and women from metropolitan areas trend to the lowest intake. For young men, VI was higher when the number of family was 3. For middle-aged and elderly men, VI was higher when the number of family members was 2. On the other hand, men living alone had the lowest intake of vegetables (P<0.05; Table 2).

When health-related behaviors, eating habits and knowledge were analyzed (Tables 3 and 4), we found that in men and women who had breakfast and lunch at home had a higher VI.

Table 3 Vegetable consumption according to the health and lifestyle behaviors

Table 4 Vegetable consumption according to health-related behaviors, eating habits and knowledge factors

In men aged 20–39 years, those who did not skip meals, drink alcohol everyday, and attended health education groups had a higher VI. Men aged 40–59 years who did not skip meals, did not feel dissatisfaction, distress or burden, drank alcohol 6 days/week to 1 day/month and had never smoked consumed more vegetables. In the oldest age group, people who perceived their status as good, did not skip meals, did not feel dissatisfaction, distress or burden and drank alcohol 6 days/week to 1 day/month had a higher intake of vegetables.

For the youngest women, those who did not skip meals, did not eat snacks, had a lack of knowledge of LSRD and had never smoked had a higher VI. In women aged 40–59 years, those who had exercise habit, perceived their health status as very good, did not skip meals, had knowledge of LSRD, attended health education groups and had never smoked had a higher VI. Women aged 60–69 years, who did not skip meals, attended health education groups and had never smoked had the highest VI.

In logistic regression, men aged 20–39 years who skipped breakfast, had breakfast away from home, perceived their status as good, skipped meals and had snacks were associated with lower VI (Table 5). In subjects aged 40–59 years who skipped breakfast, had breakfast away from home, did not perceive their health status as good, skipped meals, felt distress or dissatisfaction very often or often and had ever smoked showed association with low VI (Table 6), whereas those drinking alcohol 6 days/week to 1 day/month had a higher VI. In men aged 60–69 years who had breakfast and lunch away from home and perceived their health status as good, the VI were higher.

Table 5 Risk for low vegetable intake in Japanese adults according to exercise habits and place for breakfast, lunch and dinner

Table 6 Risk for low vegetable intake in Japanese adults, according to health-related behaviors, eating habits and knowledge factors

Women aged 20–39 years who had breakfast away from home, heard about LSRD, but did not know what LSRD is had a higher risk for low VI. Women who did not know about LSRD had a higher VI. Women aged 40–59 years who skipped breakfast, had breakfast and lunch away from home, did not do exercise, skipped meals, drank alcohol every day, had heard about LSRD and had a history of smoking were related to a low VI. In women aged between 60–69 years who ate breakfast or lunch away from home, did not exercise, did not know and/or had not heard about LSRD, did not attend health-education group and had ever smoked, the risk for low VI was higher.

Discussion

Most of the associations analyzed in this study have been reported previously in Western countries; however, there are no nationally representative data from Japan. Our aim was to examine the relationships of demographic factors, health-related behaviors, eating habits and knowledge of VI among Japanese adults.

In this study, a large number of subjects did not meet the current ‘Health Japan 21’ recommendation for VI, and this finding was positively correlated with age (MHLWJ, 2005). We found that about 65% of men and 69% of women ate less than 350 g of vegetables in a selected typical day. However, the average amount of VI in both men and women approached the recommended dietary target for vegetables of ⩾350 g in a day.

We demonstrated that household size was related to VI as previously reported by (Papadaki and Scott, 2002). Men living with someone had a higher intake than those living alone. We also found that residence regions and area were associated with VI: low intake was particularly marked in those regions separated from the main Japanese island (Honshu), that is, Shikoku and southern Kyushu. This finding could be related to higher vegetable prices because of the natural conditions, such as low temperature and snowfall damaging agricultural crops. The Annual Report on Food, Agriculture and Rural Areas in Japan, Fiscal Year 2003 reported that the total field crop damage from low temperatures for vegetables nationwide was 7%. The field crops regions damage from low temperature was 15% for Hokkaido and 8% for Kyushu (The Ministry of Agriculture, Forestry and Fisheries of Japan, 2003).

On the other hand, subjects from city areas had a higher VI compared with those living in metropolitan or town-rural areas. Furthermore, men and women from metropolitan areas had the lowest consumption, indicating that transportation of vegetables to markets, availability and price could be the main factors for their low intake.

When we analyzed the mean intake using GLM, we divided total VI into green and yellow vegetables, light color vegetables, vegetable juice and pickles (data not shown). After dividing vegetables into these groups, we found that vegetable juice was consumed in large quantities by young men living alone, men aged 40–59 years who had never smoked and women aged 20–39 years who had knowledge of LSRD. Furthermore, women aged 40–59 years who perceived their health status as very good and women aged 60–69 years living alone had a high consumption of vegetable juice. The results suggest that these subjects with healthier behaviors or those living alone have more vegetable juice intake. In addition, young women aged 20–39 years who had breakfast out also had high intakes of vegetable juice. This may be attributable to subjects being aware of their low VI and attempting to compensate for this by consuming vegetable juice. This finding is interesting because there are no previous Japanese studies identifying factors to explain vegetable juice intake.

In men and women aged 60–69 years, having snacks was associated with pickle intake (data not shown). In Japanese culture, many old people have snacks before going to bed, particularly pickles, even though there is a campaign to reduce salt intake (MHLWJ, 2002).

Logistic regression analyses showed that risk factors associated with low VI were different by age group in both men and women. Our findings suggest that in the youngest and oldest age groups, especially for women, educational initiatives that aim to increase VI should include such strategies as increasing availability of vegetables in cafeterias and restaurants where they usually have lunch, furthermore, food price must be taken into consideration. Other authors found that subjects in their twenties reported low VI because of lack of time to cook vegetables or living alone for the first time and not having the skills to prepare vegetables dishes (Papadaki and Scott, 2002; Larson et al., 2006). We also found other unhealthy behaviors that were negatively correlated with VI in each age group. For example, more health-related behaviors and eating habits factors affecting VI, such as eating snacks, were observed among the youngest group, whereas behaviors related to distress, such as drinking alcohol and a history of smoking, were common in the middle-aged group. In the oldest age group, lack of health-related knowledge and not attending health-education groups were factors negatively associated with VI, especially in women. Our finding is consistent with previous studies showing that unhealthier behaviors, such as drinking alcohol, are associated not only with distress (Hiro et al., 2007) but also with low VI (Agudo et al., 1999; Unusan, 2006). Women in the middle and in the oldest age groups, lack of knowledge or having only heard about LSRD was associated with lower VI, indicating that merely having heard about LSRD has no effect on behavior; therefore, women need to learn the content of LSRD to improve VI. The differences in health-related behaviors and eating habits across age groups could be because the subjects in the youngest group are likely to live alone and eat out, whereas middle-aged Japanese (aged 40–59 years) may be busy with work and have family obligations.

After analyzing the responses to the question ‘during the last month, on average, how many hours did you sleep’ it was found that this factor was not significant either when the mean intake were calculated using GLM or in the logistical regression analysis (P>0.05; data not shown). As the number of subjects skipping dinner was small, we could not analyze the relation between dinner and VI. On the other hand, number of subjects who had knowledge of Health Japan 21 was few, however, they clearly trended to have higher intake of vegetable (Table 4). Thus, we suggest that it is important to spread more information of Health Japan 21 among Japanese to raise awareness about healthy lifestyle to carry out more effective dietary interventions.

Finally, this study had a large number of subjects and the data considered to be representative for Japanese adults. Although the consumption of energy and nutrient were closed to other surveys done in Japan adults (Nakamura et al., 2002; Zhou et al., 2003), this study has several limitations. First, the dietary intake was assessed for one day in November; thus, it may not be a fair representation of the typical dietary consumption of individual subjects and, not only because of under/over reporting of VI, but also because the day-to-day food intake may vary and there may be seasonal variations as well. Second, some portion of the findings could likely be due to chance because we dealt with many variables in statistical models to cover important groups of vegetables and various indicators for lifestyles. Third, we were not able to access information on household income or educational level to analyze whether those factors are also associated with VI as has been reported in previous studies (Agudo et al., 1999; Kirkpatrick and Tarasuk, 2003; Laaksonen et al., 2003); and fourth, due to the cross-sectional design, causal hypotheses cannot be supported.

Conclusions

These analyses clearly demonstrate that VI among Japanese adults is associated with demographic factors such as area of residence, healthier behaviors and eating habits such as having meals away from home. We concluded that to increase VI, it is necessary to provide more nutritional education, such as how to cook vegetables, for Japanese aged 20–39 years and to focus on education about LSRD for subjects aged 40–69 years.

Conflict of interest

The authors declare no conflict of interest.