The Effect of Age on the Relationships between Work-related Factors and Heavy Drinking

Abstract

Aims: The aim of this study was to investigate age-related differences in the relationship between work-related factors and heavy drinking. Methods: This study in 3,398 male workers at a factory in Japan examined data on heavy drinking, defined as an alcohol consumption >40 g/day, and work-related factors including occupation, shift work, and job stress evaluated using the Japanese version of the Karasek's Job Content Questionnaire (JCQ). In the present cross-sectional study, alcohol consumption was assessed using a self-administered diet history questionnaire (DHQ). Results: The prevalence of heavy drinkers in the study population was 15.5% and rose with increasing age. An increase in the odds ratio (OR) for heavy drinking was observed consistently in blue-collar workers compared with white-collar workers in all age groups. In subjects aged 20–29 years, shift work had also increased the OR for heavy drinking. In subjects aged 40–49 years, the two groups with a lower decision latitude had an increased OR compared with the highest group. In subjects aged 20–29 years, the age adjusted OR for individuals who received the lowest level of social support in the workplace was increased significantly compared with the highest group (4.22 [95%CI, 1.07–16.62] ). On the other hand, social support showed a positive association with heavy drinking in subjects aged 40–49 and 50–59 years. Job demand was not related to heavy drinking in any of the age groups. Conclusions: Our findings suggest that occupation and work schedule are related to alcohol use more apparently in a younger age group and that psychosocial factors are related to enhancement or prevention of alcohol use.

(J Occup Health 2014; 56: 141-149)

Introduction

Excessive intake of alcohol has been linked to major morbidities, including neuropsychiatric disorders, gastrointestinal illness, cancer, and cardiovascular disease, as well as both intentional and unintentional injury¹⁾. The WHO²⁾ warned that alcohol is the leading risk factor for death in males aged 15–59 years, which represents the working population. Numerous studies have investigated the factors related to excessive drinking or problematic drinking and have shown that socioeconomic status such as educational attainment, family history of alcohol dependence, anxiety traits and cultural norms for alcohol use may contribute to drinking problems^3-9).

For work-related factors, heavy drinking or problem drinking behavior was more prevalent in manual workers and lower occupational classes^{6, 10-12)}. Although the effects of shift work on alcohol consumption have not been established, we have reported previously that shift work increased the prevalence of heavy drinking in middle-aged male workers when accompanied by sleep problems¹³⁾. Studies that have examined the relationship between psychosocial work characteristics and alcohol consumption have generally produced equivocal findings. Some studies investigated the risk of high job demands or high-strain work on heavy drinking¹⁴⁾, whereas others examined whether passive work was a risk factor for heavy drinking¹⁵⁾. Several studies failed to show a significant relationship between job stress and heavy drinking^16-18). Social support in the workplace may both encourage and supress heavy drinking¹⁹⁾. The effects of effort-reward imbalance have also been shown to be equivocal^{16, 17)}.

Such inconsistencies may be caused by differences in the age distribution of the study populations, as it is known that drinking patterns change with increasing age²⁰⁾. Some studies have reported that younger adults show higher enhancement, social and coping tendencies associated with anxiety motives for drinking compared with older adults^{21, 22)}. To date, however, only a few relevant cross-sectional studies, one from Japan and the other from Finland, have reported age-related differences in the effects of work characteristics on alcohol consumption^{17, 23)}. Although the measures evaluating work characteristics and the relationships between job stress and heavy drinking were different between two studies, both studies concluded that job stressors were not consistently associated with heavy drinking.

For epidemiological study, alcohol consumption must be determined as accurately as possible. However, it is suggested that people drink more alcohol than indicated by self-reported survey^{24, 25)}. Underestimation of alcohol consumption may affect the association of heavy drinking and work-related factors. To minimize the bias, it is important to use questionnaires that have been validated with respect to measurement of alcohol intake. However, no studies on age-related differences in the effects of work characteristics on alcohol consumption have used such validated questionnaires.

In the present cross-sectional study, alcohol consumption was assessed using a self-administered diet history questionnaire (DHQ)²⁶⁾, which has been validated with respect to the accuracy of measurement of alcohol consumption²⁷⁾. We investigated the associations of work-related factors including occupation, work schedule and job strain on heavy drinking in male factory workers in Japan stratified according to age. We investigated only male subjects, as the prevalence of heavy drinking in female workers was too small for the purposes of the study. Job strain was evaluated using the Japanese version of the Karasek's Job Content Questionnaire (JCQ)²⁸⁾.

Methods

The study was approved by the Ethics Committee for Epidemiologic Research at the Kanazawa Medical University (Ishikawa, Japan).

Study subjects

The study population consisted of male workers employed at a factory that produces light-metal products. In 2002, a self-administered survey on job stress, depressive score, occupational category and work schedule was carried out on 7,271 employees (4,703 males, 2,568 females). One year later, a self-administered survey on alcohol consumption, other health-related behaviors, and history of medical treatment was also carried out. The data of both surveys were merged and used for analysis. Of the 4,703 subjects, 3,398 (72.3%) provided valid responses in the two surveys. The mean age (SD) of these 3,398 subjects was 41.8 (10.6) years. After excluding 610 nondrinkers, 2,788 subjects were included in the analysis. Subjects who did not drink were excluded because they may have had specific reasons for not consuming alcohol, such as severe illness or constitutional factors.

Occupation, work schedule, psychosocial work characteristics and other social variables

Occupational status was categorized into three groups: white-collar, blue-collar engaged in fixed-day work, and blue-collar engaged in shift work. White-collar workers included administrators, managers, clerical workers and professional workers, while blue-collar work included operation of machinery, processing or construction of aluminium products, and other manual work. Shift work included two- and three-shift rosters.

Individual work characteristics including job demands, job control, and worksite support were determined using the Japanese version of the Karasek's JCQ which is based on a job demand-control model²⁶⁾. The depression score was assessed using a self-rated scale, the Center for Epidemiological Studies Depression Scale (CESD)²⁹⁾.

Marital status was divided into “married” and “not married”.

Alcohol consumption, smoking, and medical treatment

In accordance with the Japanese Ministry of Health, Labor, and Welfare guidelines, “21^st Century Measures for National Health Promotion (2^nd)”, heavy drinking was defined as an alcohol intake exceeding 40 g/day³⁰⁾. Alcohol intake during the preceding month was assessed using DHQ²⁶⁾. The DHQ was developed originally to estimate the dietary intakes of macronutrients and micronutrients in epidemiological studies in Japan. The DHQ includes questions about the nature of alcohol intake, including weekly or monthly frequency, and the amount and type of alcohol consumed on each occasion. The mean alcohol intake per day was calculated over a one-month period using an ad hoc computer algorithm developed for the DHQ, which was based on the Standard Tables of Food Composition in Japan. The validity of alcohol consumption derived from the DHQ was confirmed using 16-day semi-weighed dietary records as a reference²⁷⁾. The subjects were classified as current-smokers, non-smokers or ex-smokers based on their smoking habits. Medical treatment history for hypertension, dyslipidemia and diabetes mellitus was also recorded.

Statistical analysis

The data of the four age groups (20–29, 30–39, 40–49 and 50–59 years) were analyzed separately. A multiple logistic regression model was used to evaluate the effects of work-related factors on heavy drinking. The factors included in the model were occupational category (white-collar work/blue-collar work engaged in fixed-day work/blue-collar work engaged in shift work) and psychosocial work characteristics (job demand, decision latitude, and social support in the workplace). Potential confounding factors in the model included age (continuous variable), marital status (single/married), current smoking (no/yes), depression score, and the use of medication for hypertension, diabetes or dyslipidemia (no/yes). Analysis of the data of subjects aged 20–29 years did not include the use of medication in the model, as only four workers were taking these drugs. The data on work characteristics, job demand, decision latitude, social support in the workplace and depression scores were grouped into tertiles.

Statistical analyses were performed using IBM SPSS 19.0 (IBM Corp., Armonk, NY, USA). All probability values were two-tailed with the level of statistical significance defined as p<0.05.

Results

Table 1 summarizes the distribution of sociodemographic factors, work-related factors, and the amount of alcohol consumed, grouped according to age. The prevalence of heavy drinkers was 15.5% (n=525) in the study population and rose with increasing age: 20–29 years, 4.9% (n=29); 30–39 years, 10.4% (n=80); 40–49 years, 19.5% (n=197); 50–59 years, 21.3% (n=219).

Table 2-1 and Table 2-2 present the findings of a multiple logistic regression analysis that assessed the correlation between heavy drinking and sociodemographic factors, depression score and work-related factors after adjustment for age. Current smoking was related significantly to heavy drinking in all the age groups. Medical treatment, marital status and depression scale (CESD) were not related to heavy drinking in any of the age groups. Compared with the white-collar group, the OR for heavy drinking was increased in the two blue-collar groups in all age groups. In the group aged 20–29 years, the OR ratio for blue-collar workers engaged in shift work was increased significantly (OR 4.79 [95%CI, 1.49–15.44]) compared with white-collar workers. Among the other three age groups, there were significant differences in ORs between white-collar and blue-collar groups, although shift work did not influence heavy drinking.

The relationship between psychosocial work characteristics and heavy drinking was different in the four age groups. In subjects aged 40–49 years, lower decision latitude increased the frequency of heavy drinking, with the OR of the intermediate group (OR 1.90 [95%CI, 1.24–2.92]) and the lowest group (OR 1.82 [95%CI, 1.17–2.84]) being increased significantly. In subjects aged 20–29 years, low social support in the workplace was related to heavy drinking (OR 5.91 [95%CI, 1.28–27.32]). On the other hand, social support encouraged heavy drinking in subjects aged 50–59 years.

Table 1. Distribution of sociodemographic factors, work-related factors and amount of alcohol consumed according to age in 2003

	All	20–29 years	30–39 years	40–49 years	50–59 years
Number	3,398	592	768	1,011	1,027
Age in years, (mean ± sd)	41.8 ± 10.6	25.6 ± 2.8	34.4 ± 2.9	44.7 ± 2.7	54.0 ± 2.6
Occupationa), n (%)
White-collar	1,274 (37.5)	186 (31.4)	344 (44.8)	391 (38.7)	362 (35.2)
Blue-collar	2,115 (62.2)	406 (68.6)	424 (55.2)	620 (61.3)	665 (64.8)
Shift worka), n (%)
Day work	2,554 (75.2)	374 (63.2)	575 (74.9)	787 (77.8)	818 (79.6)
Shift work	844 (24.8)	218 (36.8)	193 (25.1)	224 (22.2)	209 (20.4)
Score of JCQa), (mean ± sd)
Decision latitude	65.9 ± 9.9	65.3 ± 10.4	67.3 ± 9.3	65.6 ± 9.7	65.4 ± 10.3
Job demand	33.0 ± 5.1	33.3 ± 5.5	34.3 ± 5.4	33.0 ± 4.8	31.8 ± 4.7
Social support	22.6 ± 3.3	23.9 ± 3.5	22.7 ± 3.3	22.1 ± 3.2	22.3 ± 3.1
Depression score (0–51)a), (mean ± sd)	14.9 ± 7.7	15.0 ± 7.7	14.9 ± 8.5	15.2 ± 7.7	14.4 ± 7.0
Marital status, n (%)
Single	29 (29.0)	468 (79.1)	256 (33.3)	162 (16.0)	100 (9.7)
Married	986 (71.0)	124 (20.9)	512 (66.7)	849 (84.0)	927 (90.3)
Smoking, n (%)
Never	1,152 (33.9)	255 (43.1)	300 (39.1)	295 (29.2)	302 (29.4)
Current smoker	1,751 (51.5)	298 (50.3)	380 (49.5)	559 (55.3)	514 (50.0)
Ex-smoker	495 (14.6)	39 (6.6)	88 (11.5)	157 (15.5)	211 (20.5)
Medical treatment, n (%)
No	3,057 (90.0)	588 (99.3)	737 (96.0)	920 (91.0)	812 (79.1)
Yes	341 (10.0)	4 (0.7)	31 (4.0)	91 (9.0)	215 (20.9)
Alcohol consumption per day, n (%)
0 g/day	610 (18.0)	137 (23.1)	136 (17.7)	165 (16.3)	172 (16.7)
<20 g/day	1,612 (47.4)	371 (62.7)	419 (54.6)	436 (43.1)	386 (37.6)
<40 g/day	651 (19.2)	55 (9.3)	133 (17.3)	213 (21.1)	250 (24.3)
≥40 g/day (heavy drinking)	525 (15.5)	29 (4.9)	80 (10.4)	197 (19.5)	219 (21.3)

Nondrinkers were included. Medical treatment: the use of medication for hypertension, diabetes or dyslipidemia. ^a): Data of job stress, depressive score, occupational category and work schedule were collecte in 2002.

Table 2-1. Age-adjusted odds ratios for heavy drinking related to sociodemographic factors, work-related factors, and depression score in workers aged 20–39 years

		20–29 years					30–39 years
Variables	(Range)	n	Cases	(%)	OR	95%CI	n	Cases	(%)	OR	95%CI
Marital status
Single		351	19	(5.4)	1.00		207	20	(9.7)	1.00
Married		104	10	(9.6)	1.50	(0.58–3.87)	425	60	(14.1)	1.36	(0.80–2.30)
Current smoking
No		209	7	(3.3)	1.00		320	32	(10.0)	1.00
Yes		246	22	(8.9)	2.92	(1.13–7.46)*	312	48	(15.4)	1.64	(1.02–2.62)*
Medical treatment
No		452	29	(6.4)	—	— —	608	74	(12.2)	1.00
Yes		3	0	(0.0)	—	— —	24	6	(25.0)	2.24	(0.85–5.89)
CESD
Low	(0–11)	170	14	(8.2)	1.00		255	29	(11.4)	1.00
Intermediate	(12–17)	147	10	(6.8)	0.99	(0.36–2.48)	189	22	(11.6)	1.01	(0.56–1.79)
High	(18–51)	138	5	(3.6)	0.43	(0.13–1.38)	188	29	(15.4)	1.32	(0.76–2.28)
Type of occupation
White collar		150	5	(3.3)	1.00		296	24	(8.1)	1.00
Blue collar, day work		145	7	(4.8)	1.99	(0.54–7.35)	190	31	(16.3)	2.14	(1.21–3.79)**
Blue collar, shift work		160	17	(10.6)	4.79	(1.49–15.44)**	146	25	(17.1)	2.61	(1.45–4.70)**
Job demand
Low	(12–31)	172	11	(6.4)	1.00		184	23	(12.5)	1.00
Intermediate	(32–35)	120	9	(7.5)	1.32	(0.49–3.54)	163	29	(17.8)	1.32	(0.74–2.38)
High	(36–48)	163	9	(5.5)	0.73	(0.26–2.03)	285	28	(9.8)	0.71	(0.40–1.26)
Decision latitude
High	(71–96)	135	6	(4.4)	1.00		223	28	(12.6)	1.00
Intermediate	(63–70)	155	8	(5.2)	0.90	(0.26–3.20)	241	35	(14.5)	1.19	(0.70–2.02)
Low	(24–62)	165	15	(9.1)	2.76	(0.96–8.00)	168	17	(10.1)	0.77	(0.41–1.45)
Social support
High	(25–32)	140	3	(2.1)	1.00		95	10	(10.5)	1.00
Intermediate	(23–24)	183	13	(7.1)	4.12	(0.90–18.96)	321	41	(12.8)	1.24	(0.59–2.58)
Low	(8–22)	132	13	(9.8)	5.91	(1.28–27.32)*	216	29	(13.4)	1.31	(0.61–2.81)

Nondrinkers were excluded. Definition of heavy drinking: (≥40 g/day). Medical treatment: the use of medication for hypertension, diabetes or dyslipidemia. OR (95% confidence interval (CI)): age-adjusted odds ratios calculated using a logistic regression model. Age and each variable were included in the model. ^* p<0.05; ^** p<0.001. CESD: Center for Epidemiological Studies Depression Scale.

Table 2-2. Age-adjusted odds ratios for heavy drinking related to sociodemographic factors, work-related factors, and depression score in workers aged 40–59 years

		40–49 years					50–59 years
Variables	(Range)	n	Cases	(%)	OR	95%CI	n	Cases	(%)	OR	95%CI
Marital status
Single		117	31	(26.5)	1.00		73	22	(30.1)	1.00
Married		729	166	(22.8)	0.77	(0.49–1.20)	782	197	(25.2)	0.77	(0.45–1.29)
Current smoking
No		371	72	(19.4)	1.00		433	86	(19.9)	1.00
Yes		475	125	(26.3)	1.56	(1.12–2.17)**	422	133	(31.5)	1.86	(1.36–2.54)*
Medical treatment
No		769	179	(23.3)	1.00		681	173	(25.4)	1.00
Yes		77	18	(23.4)	1.04	(0.59–1.82)	174	46	(26.4)	1.02	(0.70–1.48)
CESD
Low	(0–11)	300	61	(20.3)	1.00		322	92	(28.6)	1.00
Intermediate	(12–17)	287	71	(24.7)	1.25	(0.85–1.85)	299	65	(21.7)	0.70	(0.48–1.00)
High	(18–51)	259	65	(25.1)	1.30	(0.87–1.94)	234	62	(26.5)	0.91	(0.63–1.33)
Type of occupation
White collar		350	61	(17.4)	1.00		322	64	(19.9)	1.00
Blue collar, day work		315	88	(27.9)	1.82	(1.25–2.63)**	369	105	(28.5)	1.61	(1.13–2.29)**
Blue collar, shift work		181	48	(26.5)	1.73	(1.12–2.66)*	164	50	(30.5)	1.80	(1.17–2.77)**
Job demand
Low	(12–31)	307	77	(25.1)	1.00		398	105	(26.4)	1.00
Intermediate	(32–35)	252	63	(25.0)	0.95	(0.64–1.40)	258	62	(24.0)	0.94	(0.66–1.35)
High	(36–48)	287	57	(19.9)	0.76	(0.52–1.13)	199	52	(26.1)	0.99	(0.67–1.54)
Decision latitude
High	(71–96)	243	38	(15.6)	1.00		234	60	(25.6)	1.00
Intermediate	(63–70)	325	87	(26.8)	1.90	(1.24–2.92)**	313	74	(23.6)	0.95	(0.65–1.41)
Low	(24–62)	278	72	(25.9)	1.82	(1.17–2.84)**	308	85	(27.6)	1.13	(0.77–1.66)
Social support
High	(25–32)	86	19	(22.1)	1.00		75	27	(36.0)	1.00
Intermediate	(23–24)	383	87	(22.7)	1.12	(0.64–1.97)	427	104	(24.4)	0.56	(0.33–0.93)*
Low	(8–22)	377	91	(24.1)	1.16	(0.66–2.04)	353	88	(24.9)	0.58	(0.34–0.97)*

Nondrinkers were excluded. Definition of heavy drinking: (≥40 g/day). Medical treatment: the use of medication for hypertension, diabetes or dyslipidemia. OR (95% confidence interval (CI)): age-adjusted odds ratios calculated using a logistic regression model. Age and each variable were included in the model. ^* p<0.05; ^** p<0.001. CESD: Center for Epidemiological Studies Depression Scale.

Table 3 summarizes the relationships between occupational factors and heavy drinking after adjustment for age, marital status, medical treatment, smoking, and all the variables listed in the table. An increased OR was observed consistently in blue-collar workers compared with white-collar workers in all age groups. In subjects aged 20–29, shift work showed an additional relationship with heavy drinking. For psychosocial work characteristics, decision latitude correlated significantly with heavy drinking in subjects aged 30–39 or 40–49 years, although these two age groups showed opposite direction. In subjects aged 30–39 years, the two groups with lower decision latitude had decreased ORs compared with the highest group. In contrast, in subjects aged 40–49 years, the two groups with lower decision latitude had increased ORs compared with the highest group. The association between social support and heavy drinking changed from a negative to a positive relationship with increasing age. In subjects aged 20–29 years, the OR of the group receiving the lowest level of social support in the workplace was increased significantly (OR 4.70 [95%CI, 1.15–19.28]) compared with the highest group. On the other hand, social support showed a positive association with heavy drinking in subjects aged 40–49 or 50–59 years. Job demand was not related with heavy drinking in any of the age groups.

Table 3. Relationship between work-related factors and heavy drinking in 2,788 subjects excluding 610 abstainers after adjustment for all confounding factors determined by multiple logistic regression

	20–29		30–39		40–49		50–59
Variables	OR	95%CI	OR	95%CI	OR	95%CI	OR	95%CI
Type of occupation
White collar	1.00		1.00		1.00		1.00
Blue collar, day work	1.57	(0.43–5.66)	2.27	(1.23–4.19)**	1.60	(1.07–2.39)*	1.62	(1.09–2.40)*
Blue collar, shift work	4.68	(1.36–16.04)*	2.80	(1.44–5.48)**	1.48	(0.91–2.40)	1.95	(1.19–3.18)**
Job demand
Low	1.00		1.00		1.00		1.00
Intermediate	1.14	(0.41–3.12)	1.40	(0.75–2.61)	1.01	(0.68–1.50)	0.93	(0.64–1.35)
High	0.93	(0.32–2.69)	0.72	(0.38–1.34)	0.79	(0.53–1.19)	1.04	(0.70–1.56)
Decision latitude
High	1.00		1.00		1.00		1.00
Intermediate	0.54	(0.15–1.88)	0.83	(0.46–0.15)	1.70	(1.07–2.69)*	0.88	(0.57–1.36)
Low	1.23	(0.36–4.25)	0.42	(0.20–0.90)*	1.47	(0.87–2.49)	0.93	(0.58–1.47)
Social support
High	1.00		1.00		1.00		1.00
Intermediate	2.66	(0.68–10.35)	1.32	(0.61–2.87)	0.80	(0.44–1.44)	0.62	(0.35–1.08)
Low	4.70	(1.15–19.28)*	1.49	(0.64–3.46)	0.80	(0.43–1.48)	0.61	(0.34–1.08)

Nondrinkers were excluded. Definition of heavy drinking: (≥40 g/day). Medical treatment: the use of medication for hypertension, diabetes, or dyslipidemia. OR (95% confidence interval (CI)) calculated using logistic regression after adjustment for variables including age, marital status, smoking, medical treatment (the use of medication for hypertension, diabetes or dyslipidemia) and all the items listed in the table. ^* p<0.05; ^** p<0.001.

Discussion

We carried out a cross-sectional study to investigate the effect of age on the association between work-related factors and alcohol consumption, and found age-related differences in the relationship between work-related factors and heavy drinking. Occupation was related to alcohol use in all age groups, particularly in the younger age group, and shift work was related to alcohol use more apparently in the younger age group. However, work-related psychosocial factors were related to either encouragement or prevention of alcohol use, and the directions of the relationships were different according to age.

The frequency of heavy drinking among the all subjects, 15.5%, was similar to the 19.7% calculated from the data for the same year of the National Health and Nutrition Survey in Japan³¹⁾, although the questionnaire used for estimation of alcohol intake was different from that used by us. Our study showed that the frequency of heavy drinking increased with age. An international comparative study reported that Japanese men consumed more alcohol and had a higher proportion of middle-aged men were heavy drinkers, whereas men in many European populations and in the US population reduced their alcohol consumption as they got older²⁰⁾. It is possible these different trends are a consequence of the social drinking norms in Japan including high alcohol consumption in job-related drinking, permissive attitudes towards drinking by middle-aged men and the relatively greater expense of alcoholic beverages for young men.

We found an increased OR for heavy drinking in blue-collar workers compared with white-collar workers in all age groups, particularly in the younger age group. This finding is in agreement with those of former studies. A cross-sectional study from the US by Harford et al.⁶⁾ showed that the percentage of drinkers in white-collar occupations in men was higher than in blue-collar occupations, while the men in blue-collar who drink have a higher average daily consumption than drinkers in white-collar occupations. We also reported a cross-sectional study of male workers in smaller sized enterprises that showed that the frequency of heavy drinkers was higher in manual workers and transportation workers compared with white-collar workers¹¹⁾. Some studies have also reported that highly hazardous physical working conditions are associated positively with heavy and binge drinking^{18, 23)}. A higher frequency of alcohol dependence in blue-collar workers compared with white-collar workers has also been reported^{10, 12)}. These differences in drinking habits between occupational groups may reflect differences in cultural norms for drinking and other socioeconomic factors including educational attainment.

This study showed that shift work was associated with a higher likelihood of heavy drinking in subjects aged 20–29 years. To date, the effects of shift work on alcohol consumption or heavy drinking have not been established. We have reported previously that shift work increased the prevalence of heavy drinking in middle-aged male workers when accompanied by sleep problems, and suggested that shift workers may try to cope with sleep problems by using alcohol¹³⁾. Shift workers in a young age group may also use alcohol for adaptation to irregular working schedules.

We found that a decrease in decision latitude increased the frequency of heavy drinking in subjects aged 40–49 years. However, the opposite relationship was found in subjects aged 30–39 years. The positive relationship between decision latitude and heavy drinking in subjects aged 30–39 years was also observed in the same age group in the study by Hiro et al.²³⁾. These different associations between decision latitude and drinking habits according to age may be explained by the motivational model of alcohol use proposed by Cox et al.³²⁾. According to the model, the motives for drinking are categorized into four groups: social (external, positive reinforcement), enhancement (internal, positive enhancement), conformity (external, negative reinforcement) and coping (internal, negative reinforcement). Subjects aged 40–49 years with a lower decision latitude may use alcohol to cope with job stress (negative reinforcement). However, subjects aged 30–39 years with higher decision latitude may use alcohol for social motives (positive reinforcement).

In the youngest age group, a decrease in social support in the workplace increased the frequency of heavy drinking, whereas a positive association was found in subjects aged 40 years or older. These findings suggested that this young age group may have used alcohol to cope with job stress, whereas older age group may have used alcohol according to social positive enhancement. However, in the population investigated by Hiro et al.²³⁾, social support increased the likelihood of heavy drinking in both the younger and older age groups. Social support therefore appears to have disparate effects of either enhancing or preventing heavy drinking, particularly in a young age group. However, the prevalence of heavy drinking in each age group of subjects in that study was half that of our study subjects. The differences of prevalence in heavy drinking may contribute to the differences in the association of heavy drinking and work-related factors between the two studies.

While the strength of the present study was its investigation of the relationship between work-related factors and heavy drinking according to age by using a validated questionnaire for measurement of alcohol consumption, it also had several limitations. First, the study was conducted in a single factory in Japan, with the majority of subjects living in the district around the factory. This means that they shared similar drinking norms, which may have influenced them to acquire drinking habits^{9, 33)}. Generalization of our results should therefore be carried out with caution. Second, due to the cross-sectional design of the study, we were unable to determine causal relationships. There is evidence that socioeconomic status (SES) in early adolescence affects drinking habits³⁴⁾. SES may also affect the selection of jobs in youth. As we did not collect information about educational attainment, which is a better indicator of SES in adolescents, we were unable to exclude the effects of selection bias. Third, there was a one-year time lag between investigating work characteristics and depression score and evaluating alcohol consumption. It is therefore possible that work characteristics and depression score may have changed over this one-year period.

In conclusion, our findings suggest that there are age-related differences in the relationship between work-related factors and heavy drinking. Occupation and work schedule related to alcohol use more apparently in the younger age group. Psychosocial factors were related to either enhancement or prevention of alcohol use, and the directions of the relationships were different according to age. Therefore, approaches in consideration of age-specific characteristics in the relationships between work-related factors and drinking behavior would be necessary for prevention of heavy drinking among the working population.

Acknowledgments: The present study was supported by Health and Labor Sciences Research Grants from the Ministry of Health, Labor and Welfare of Japan (H18-Junkankitou [Seishuu]-Ippan-012, H19-Junkankitou [Seishuu]-Ippan-012, H19-Junkankitou [Seishuu]-Ippan-021, H20-Junkankitou [Seishuu]-Ippan-013, H22-Junkankitou [Seishuu]-Ippan-005, H23-Junkankitou [Seishuu]-Ippan-005), a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan for Scientific Research (B) (20390188) and the Japan Arteriosclerosis Prevention Fund.

References

• 1)   Rehm  J,  Baliunas  D,  Borges  GL, et al. The relation between different dimensions of alcohol consumption and burden of disease: an overview. Addiction 2010; 105: 817-43.
• 2)  WHO. Global Status Report on Alcohol and Health 2011. 2011.
• 3)   Casswell  S,  Pledger  M,  Hooper  R. Socioeconomic status and drinking patterns in young adults. Addiction 2003; 98: 601-10.
• 4)   Batty  GD,  Bhaskar  A,  Emslie  C, et al. Association of life course socioeconomic disadvantage with future problem drinking and heavy drinking: gender differentials in the west of Scotland. Int J Public Health 2012; 57: 119-26.
• 5)   Kuendig  H,  Plant  ML,  Plant  MA, et al. Beyond drinking: differential effects of demographic and socioeconomic factors on alcohol-related adverse consequences across European countries. Eur Addict Res 2008; 14: 150-60.
• 6)   Harford  TC,  Parker  DA,  Grant  BF,  Dawson  DA. Alcohol use and dependence among employed men and women in the United States in 1988. Alcohol Clin Exp Res 1992; 16: 146-8.
• 7)   Kuntsche  E,  Knibbe  R,  Gmel  G,  Engels  R. Why do young people drink? A review of drinking motives. Clin Psychol Rev 2005; 25: 841-61.
• 8)   Kuntsche  E,  Knibbe  R,  Gmel  G,  Engels  R. Who drinks and why? A review of socio-demographic, personality, and contextual issues behind the drinking motives in young people. Addict Behav 2006; 31: 1844-57.
• 9)   Marchand  A,  Parent-Lamarche  A,  Blanc  ME. Work and high-risk alcohol consumption in the Canadian workforce. Int J Environ Res Public Health 2011; 8: 2692-705.
• 10)   Cheng  WJ,  Cheng  Y,  Huang  MC,  Chen  CJ. Alcohol dependence, consumption of alcoholic energy drinks and associated work characteristics in the Taiwan working population. Alcohol Alcohol 2012; 47: 372-9.
• 11)   Morikawa  Y,  Tabata  M,  Kido  T,  Koyama  Y. Occupational class inequalities in behavioral and biological risk factors for cardiovascular disease among workers in medium- and small-scale enterprises. Ind Health 2012; 50: 529-39.
• 12)   Mandell  W,  Eaton  WW,  Anthony  JC,  Garrison  R. Alcoholism and occupations: a review and analysis of 104 occupations. Alcohol Clin Exp Res 1992; 16: 734-46.
• 13)   Morikawa  Y,  Sakurai  M,  Nakamura  K, et al. Correlation between shift-work-related sleep problems and heavy drinking in Japanese male factory workers. Alcohol Alcohol 2013: 48; 202-6.
• 14)   Saade  SL,  Marchand  A. Work organisation conditions, alcohol misuse: the moderating role of personality traits. Work 2012; 44: 191-200.
• 15)   Gimeno  D,  Amick  BC, 3rd,  Barrientos-Gutierrez  T,  Mangione  TW. Work organization and drinking: an epidemiological comparison of two psychosocial work exposure models. Int Arch Occup Environ Health 2009; 82: 305-17.
• 16)   Head  J,  Stansfeld  SA,  Siegrist  J. The psychosocial work environment and alcohol dependence: a prospective study. Occup Environ Med 2004; 61: 219-24.
• 17)   Kouvonen  A,  Kivimaki  M,  Cox  SJ,  Poikolainen  K,  Cox  T,  Vahtera  J. Job strain, effort-reward imbalance, and heavy drinking: a study in 40,851 employees. J Occup Environ Med 2005; 47: 503-13.
• 18)   San Jose  B,  van de Mheen  H,  van Oers  JA,  Mackenbach  JP,  Garretsen  HF. Adverse working conditions and alcohol use in men and women. Alcohol Clin Exp Res 2000; 24: 1207-13.
• 19)   Hagihara  A,  Tarumi  K,  Nobutomo  K. Positive and negative effects of social support on the relationship between work stress and alcohol consumption. J Stud Alcohol 2003; 64: 874-83.
• 20)   Higuchi  S,  Parrish  KM,  Dufour  MC,  Towle  LH,  Harford  TC. Relationship between age and drinking patterns and drinking problems among Japanese, Japanese-Americans, and Caucasians. Alcohol Clin Exp Res 1994; 18: 305-10.
• 21)   Kim  O,  Jeon  HO. Relations of drinking motives and alcohol consumption in Korean male office workers. Psychol Rep 2012; 111: 963-70.
• 22)   Mezquita  L,  Stewart  SH,  Ibanez  MI, et al. Drinking motives in clinical and general populations. Eur Addict Res 2011; 17: 250-61.
• 23)   Hiro  H,  Kawakami  N,  Tanaka  K,  Nakamura  K. Association between job stressors and heavy drinking: age differences in male Japanese workers. Ind Health 2007; 45: 415-25.
• 24)   Dawson  DA. Methodological issues in measuring alcohl use. Alcohol Res Health 2003; 27: 18-29.
• 25)   Stockwell  T,  Donath  S,  Cooper-Stanbury  M,  Chikritzhs  T,  Catalano  P,  Meteo  C. Addiction 2004; 99: 1024-33.
• 26)   Sasaki  S,  Yanagibori  R,  Amano  K. Self-administered diet history questionnaire developed for health education: a relative validation of the test-version by comparison with 3-day diet record in women. J Epidemiol 1998; 8: 203-15.
• 27)   Kobayashi  S,  Honda  S,  Murakami  K, et al. Both comprehensive and brief self-administered diet history questionnaires satisfactory rank nutrient intakes in Japanese adults. J Epi 2012; 22: 151-9.
• 28)   Ministry of Health, Labour and Welfare, Japan. 2003 National Health and Nutrition Survey. [Online]. 2003 [cited 2013 Nov 19]; Available from: URL: http://www.mhlw.go.jp/bunya/kenkou/eiyou-chosa2-01/pdf/06a.pdf.
• 29)   Kawakami  N,  Kobayashi  F,  Araki  S,  Haratani  T,  Furui  H. Assessment of job stress dimensions based on the job demands-control model of employees of telecommunication and electric power companies in Japan: reliability and validity of the Japanese version of the Job Content Questionnaire. Int J Behav Med 1995; 2: 358-75.
• 30)   Shima  S,  Kano  T,  Kitamura  T. Seishin Igaku 1985; 27: 717-23 (in Japanese).
• 31)  Ministry of Health Labor and Welfare. 21st Century Measures for National Health Promotion. 2012.
• 32)   Cox  WM,  Klinger  E. A motivational model of alcohol use. J Abnorm Psychol 1988; 97: 168-80.
• 33)   Biron  M,  Bamberger  PA,  Noyman  T. Work-related risk factors and employee substance use: insights from a sample of Israeli blue-collar workers. J Occup Health Psychol 2011; 16: 247-63.
• 34)   Caldwell  TM,  Rodgers  B,  Clark  C,  Jefferis  BJ,  Stansfeld  SA,  Power  C. Lifecourse socioeconomic predictors of midlife drinking patterns, problems and abstention: findings from the 1958 British Birth Cohort Study. Drug Alcohol Depend 2008; 95: 269-78.