In view of decreasing lifestyle-related disease rates and increasing life expectancy, primary disease prevention for health promotion and secondary disease prevention for early detection of diseases have become more and more important [1] . The key for such approaches are general and preventive health checks; screenings and interventions which include assessment of an individual’s lifestyle risk factors and current health by physical examination and clinical laboratory tests. Despite the importance and benefit of such health checks, their understanding is not sufficient and the attendance rate in Japan is not high enough [2] .

The Japanese social security system is based on equality. Health care services are equally offered to everyone. The health insurance system is a combination of the system for employees and their dependents and the community-based system for farmers, self-employed, pensioners and their dependents [3] . The latter system is called National Health Insurance (NHI) and covers almost 30% of the population in Japan [3] . Health insurers of NHI are obligated to implement health checks by the Japanese National Health Insurance Act and Act on Assurance of Medical Care for Elderly People. These health checks and the followed medical counseling are particularly aimed at decreasing lifestyle-related diseases rates by the early detection of metabolic syndromes and associated factors [4] .

Numerous studies investigated the associations between social status and health check attendance [5] - [19] . Lifestyles, which are related to a variety of lifestyle-related diseases [20] - [25] , were studied concerning health check attendance in the previous studies [5] [6] [7] [12] [18] [19] [26] [27] . These studies have suggested decreased health check attendance rates among groups with lower social status or unhealthy lifestyles, such as lower income, lower education level, being unemployed, being single, physical inactivity, unhealthy diets, smoking, obesity and poor self-rated health status compared with those groups with higher social status or healthier lifestyles. These findings underline the prevention approach for specific vulnerable groups, although the studies have not reached the generally accepted theory. The success of health checks depends heavily on good understanding of changes needed and lifestyle modification after that. It is thus important to understand what affects a person’s decision to attend health checks and their motivation to change their lifestyle following the health checks. Even if it is difficult to compare attendance rates at health checks globally because health checks are defined differently in various health care systems by sex and age, understanding the determinants in health check attendance featured by sex and age is important in order to improve the preventive health strategy. To the best of our knowledge, studies on associations of social status and health-re- lated behaviors with health check attendance using one study population in all age groups are very few. Therefore, we aimed to examine the sex- and age-spe- cific association between social status and health-related behaviors, including lifestyle and mental health status, with health check attendance in the model areas of Kanazawa city where we could undertake the complete survey.

Data were derived from the cross-sectional Kanazawa Study 2011, which investigated the residents in two model areas of Kanazawa, Ishikawa, Japan. Information on social status, lifestyle, mental health status, and use of health services was gathered in the study.

Questionnaires were distributed to 12,871 people, all residents aged 12 years or older in the model areas in November 2011. A total of 12,253 respondents completed the questionnaire. The response rate was 95.2%. To make us well understand the relationship of social status, health-related behaviors with health check attendance, we excluded participants who are obligated to attend health checks by the Japanese Industrial Safety and Health Act or the Japanese School Health and Safety Act. People with employee’s health insurance and students are obligated by these Acts. On the other hand, people with National Health Insurance are not obligated to attend health checks, so they have the choice of whether to attend or not. Exclusion criteria are shown on Figure 1. On these analyses, we included participants with National Health Insurance aged 23 years or older (n = 4920).

The outcome measures of the analyses were self-reported attendances at health checks. Participants were asked whether they had attended the health checks recommended by the National Health Insurers in the past year. We created a binary variable indicating health check attendance in the past year (yes = 1, no = 0).

Age was stratified in three intervals, 1) 23 to 39 years (young), 2) 40 to 64 years

(middle-aged) and 3) 65 years or older (elderly). We used social status and health-related behaviors as predictor variables. A questionnaire was used to ascertain predictor variables. We classified self-rated health status as 1) poor or very poor and 2) very good or good. Working was classified as 1) non-working and 2) working. We categorized household size as 1) one person and 2) more than one person. Physical activity was ascertained by questions on the number of days per week of physical activity including walking and gymnastics. We classified physical activity as 1) not active at all (no weekly physical activity), 2) slightly active (less than 7 days per week), and 3) active (every day). Participants were asked how many days per week they consume vegetables. We categorized the number of days as 1) low (never or 1 to 3 days per week), 2) medium (4 to 6 days per week) and 3) high (everyday). Alcohol consumption was ascertained by questions on the number of days per week when alcohol was consumed. We categorized it as 1) almost every day, 2) sometimes (1 to 5 days per week) and 3) never. We assessed participants smoking status by including a question for participants to categorize themselves as 1) smokers, 2) ex-smokers or 3) non- smokers. Referring to the World Health Organization (WHO) definition [28] , we categorized Body mass index (BMI) in three groups, 1) preobese or obese (25 or over), 2) normal weight (18.5 to 25), 3) underweight (less than 18.5). BMI was calculated using self-reported height and weight. Mental health status was measured using the GHQ-12. The General Health Questionnaire (GHQ) is a self-administered screening questionnaire, designed for use in counseling aimed at detecting individuals with diagnosable psychiatric disorders [29] . In its original version, it had 60 items (GHQ-60), which were reduced to 30 (GHQ-30), 28 (GHQ-28) and 12 items (GHQ-12). The 12-Item General Health Questionnaire is the most usually used screening instrument for common mental disorders, in addition to being a general measure of psychiatric well-being [30] . Numerous studies have assessed the validity of the GHQ for use with adults [31] . Our categories were 1) poor (scored 3 or over), 2) slightly good (scored 1 to 3) and 3) good (scored 0).

We excluded the bias due to sex or age as much as possible. Firstly, we analyzed the data by sex. After that we applied stratified analyses by age groups for each sex. The bivariate analyses were conducted by means of cross-tabs. We calculated health check attendance rates by each variable. We used Pearson’s χ²-test to examine statistically significant differences. We fitted logistic regression models to estimate adjusted odds ratios (ORs) of attendance in the past one year. We computed ORs in logistic regression model containing all variables described above (Working, Household size, Physical activity, Vegetable consumption, Alcohol consumption, Smoking, BMI, Self-rated health status, Mental health). We considered p-value of <0.05 to indicate statistical significance; all tests were two-tailed. All statistical analyses were performed using SPSS version 19 for Windows.

Study participants characteristics are shown on Table 1.

The one year health check attendance rate was 61.6% in men and 59.9% in women. Working, physical activity, vegetable consumption, smoking, self-rated health status, and mental health status was associated with health check attendance in particular sex and age groups (Table 2).

Workingmen had significantly increased health check attendance rate (Wor- king = 63.9%, p < 0.01) compared with non-working persons. However, the significance was not observed in working women. Working men and women aged 23 to 39 years (Working men = 69.9%, p < 0.001; Working women = 50.5%, p < 0.001), and aged 40 to 64 years (Working men = 58.7%, p < 0.001; Working women = 57.3%, p < 0.05) had significantly increased attendance rates. There was no significance in other working persons.

Men with more physical activity had significantly increased attendance rates (Not active at all = 53.9%, Slightly active = 66.2%, Active = 68.9%, p < 0.001), although women with more physical activity did not. Men aged 23 to 39 years (Not active at all = 51.8%, Slightly active = 64.6%, Active = 86.2%, p < 0.001) and men aged 65 years or older (Not active at all = 59.8%, Slightly active = 66.9%, Active = 86.2%, p < 0.05) with more physical activity had significantly increased attendance rates, although persons in other groups with that did not.

Women who consumed more vegetables had significantly increased attendance rates (Low = 49.7%, Medium = 52.9%, High = 62.1%, p < 0.001), although men with more vegetable consumption did not. Women aged 40 to 64 years (Low = 43.8%, Medium = 44.4%, High = 59.3%, p < 0.001) and women aged 65 years or older (Low = 56.4%, Medium = 60.0%, High = 67.4%, p < 0.05) with more vegetable consumption had significantly increased attendance rates.