Smokers and drinkers in a health maintenance organization population: Lifestyles and health status

Smokers and drinkers in a health maintenance organization population: Lifestyles and health status

PREVENTIVE MEDICINE Smokers and Drinkers Population: DAVID Center for 16, 783-795 (1987) Health in a Health Maintenance Organization Lifestyl...

863KB Sizes 7 Downloads 22 Views

PREVENTIVE

MEDICINE

Smokers

and Drinkers Population:

DAVID

Center

for

16, 783-795 (1987)

Health

in a Health Maintenance Organization Lifestyles and Health Status

C. PEARSON, PH.D. ,l LOUIS C. GROTHAUS, M.S., ROBERT S.THOMPSON, M.D.,M.I?H.,AND EDWARD H. WAGNER, M.D.,M.P.H. Promotion,

Group Health Cooperative of Puget Seattle, Washington 98121

Sound,

521 Wall Street,

This study examines, among a large health maintenance organization population, the prevalence of two high-risk lifestyle practices (smoking and problem drinking), their interrelationships, and their relationships with other lifestyle practices, sociodemographic characteristics, and health status measures. Results, based on a random sample of 1,133 adults, showed that smoking and problem drinking are strongly correlated. Individuals with no drinking problems had an age-, sex-, and education-adjusted smoking prevalence of approximately 20%, while problem drinkers smoked at about twice that rate. In addition, reporting one type of problem drinking hehavior (hinge, chronic, or drinking and driving) at least doubled, and in one instance increased by sixfold, the likelihood of reporting another type of problem drinking behavior. Smokers and problem drinkers were more likely to be younger than age 65, to be irregular seat belt users (smokers and binge drinkers only), and not to belong to voluntary organizations. Results of the analysis suggest that detection, prevention, and treatment of drug use, in general, might prove more beneficial than only focusing on smoking and problem drinking. In addition, because binge drinking and drinking and driving were so widespread among younger age groups, it might prove more beneficial to consider preventive strategies that change the sale and distribution of alcohol and make the environment safer in which to drink, such as providing transportation to get drinkers back home. 0 1987 Academic Precc. Inc

INTRODUCTION

Mortality and morbidity related to smoking and abusive drinking have been investigated extensively (19, 26, 29, 32). Cigarette smoking is the single most important preventable cause of death in the United States (27). The causative role of tobacco in relation to a variety of cancers, premature deaths from heart disease, and respiratory problems is no longer in doubt. Similarly, alcohol abuse is known to play a substantial role in preventable deaths, illnesses, and disabilities (13, 14, 23, 24). It has been estimated that alcohol plays an integral part in at least 10% of all U.S. deaths (27). In addition, smoking and drinking together have been shown to increase the risk of developing certain cancers above that expected from smoking or drinking alone (12, 25). Some researchers have examined how smoking and drinking are correlated; but, with the exception of one recent study (6), results have not been applicable to the general population, because the studies have been conducted with alcoholics in treatment (15, 18, 20). The analysis of smoking and drinking practices and their joint occurrence in more representative populations is an important step i To whom

requests

for reprints

should

be addressed. 783 0091-7435187

$3.00

Copyright 0 1987 by Academic Press, Inc. All rights of reproduction in any form reserved.

784

PEARSON

ET AL.

toward increasing the understanding of the development and maintenance of high-risk lifestyle practices in general. This study examines, among a large health maintenance organization (HMO) population, the prevalence of two risky lifestyle practices (smoking and problem drinking), their interrelationships, and their relationships with other lifestyle practices, sociodemographic characteristics, and health status measures. Implications of these results are also discussed with respect to preventing and reducing the consequences of smoking and drinking in an HMO population. METHODS

The study was conducted at the Group Health Cooperative (GHC) of Puget Sound, a large (approximately 320,000 members), closed-panel, not-for-profit HMO located in western Washington State. A random sample stratified by age and sex of 1,432 GHC adult enrollees was drawn. Enrollees over the age of 65 were oversampled to increase the precision of estimates for this age group. Data were collected using a mailed questionnaire with telephone follow-up interviews of nonrespondents. Completed surveys were obtained from 79% of the original sample. After subjects who either were no longer enrolled in GHC or had died ,X,P~P

t-~mn~raTI

frnm

the

rlonnminatnr

the

recnnn~~

rate

inn-c=acr-CI

tn

521%

BP-

SMOKERS

AND

DRINKERS

IN AN

HMO

POPULATION

785

tween smoking/drinking and health status and other lifestyle behaviors, adjusting for age, sex, and education. Adjustment for education was done in an effort to control for the effects of socioeconomic status. Marital status was not adjusted for in the final stage of the analysis because it had little impact on computed odds ratios. Age-, sex-, and education-adjusted rates were derived using the MantelHaenszel procedure, as described above. In cases where the association of smoking and drinking measures with an independent variable varied significantly between men and women, results are reported separately by sex. RESULTS Sociodemographic

Characteristics

of Sample

In Table 1 the GHC enrollee population is compared with the Seattle Standard Metropolitan Statistical Area (SMSA) population. The GHC enrollee population differs primarily by having more women and being more highly educated than the SMSA population. Because the majority of GHC’s population is enrolled as a result of their or a family member’s employment, young adults are underrepresented. Sociodemographic

Characteristics

of Smokers

and Drinkers

About one-quarter of all GHC adults currently smoke cigarettes. There was no significant difference in the overall smoking rates between men (26%) and women (24%). However, age-specific smoking rates were significantly different for men and women (log-linear analysis of variance; x2 (3) = 8.0, P = 0.04). Men over age 25 years were 50% more likely to smoke than women (Fig. 1A). In contrast, women ages 18 to 24 years smoked at twice the rate of men. Smoking prevalence was greatest among those between ages 45 and 64, where the rate was 31%. Among men, those ages 25 to 64 were significantly more likely to smoke than men younger or older. Smoking rates for women followed the same pattern as for men above age 25. Twenty percent of the adult GHC population reported binge drinking (had five or more drinks at one sitting) during the preceding month. Men were twice as likely to be binge drinkers as women (28% vs 13%, x2 (1) = 15.4, P = 0.001). Binge drinking rates varied significantly with age (x2 (3) = 40.7, P = 0.001); however, the effect of age on binge drinking rates was the same for men and women (Fig. IB). About 11% of GHC adults reported chronic drinking (at least two drinks per day). Men were twice as likely to be chronic drinkers as women (15% vs 7%, x2 (1) = 13.2, P = 0.001). In contrast to binge drinking, there was no significant decline in prevalence with increasing age (Fig. 1C). Thirty percent of the adult population reported that at least once in the prior year they had driven an automobile after having had more than two drinks. Drinking and driving rates were twice as great among men as among women (39% vs 21%). While drinking and driving rates generally decreased with age, age-specitic rates were different for men and women (x2 (3) = 10.2, P = 0.001) (Fig. 1D). Men age 25 and over were much more likely to be drinking drivers than women, while 18 to 24-year-olds of both sexes had similar rates. The rate of drinking and

786

PEARSON

ET AL.

TABLE 1 COMPARISON OF THE SOCIODEMOGRAPHIC CHARACTERISTICS OF THE GROUP HEALTH COOPERATIVE OF PUGET SOUND (GHC) POPULATION VERSUS THE LOCAL STANDARD METROPOLITAN STATISTICAL AREA (SMSA) (% REPORTING)

Age (years) 18-24 25-44 45-64 65+ Sex Male Female Race Caucasian Black Asian/Pacific islander Other Education (persons 25 years and older) 11 years or less 12 years (high school grad) 13-15 years 16 years (college grad) Income’ Less than $10,000 $10,000-$14,999 $15,000-$24,999 $25,000-$34,999 $35,000-$49,999 $50,000 and over Marital stat& Married or living as married Never married Divorced Widowed Separated

GHC”

SMSAb

13 48 26 13

22 39 26 13

45 55

49 51

91 3 4 2

90 4 4 2

9 24 33 34

18 35 23 24

10 10 24 23 20 13

15 8 18 19 21 19

67 17 10 5 1

56 27 9 6 2

a Based on random sample of 1,133 adults ages 18 and over. b Source: Seattle-Everett, Washington SMSA. 1980 Census of Population and Housing, U.S. Dept. of Commerce, Bureau of Census. c The reported level for the SMSA was adjusted upward from 1979 to 1984 based on a 43% increase in the U.S. Consumer Price Index for that period. d For GHC population, marital status is reported for those 18 years and over; for SMSA, marital status is reported for those 15 years and over.

driving remained high (35% or more) up to age 65, while for women the rate declined considerably by age 45. For both men and women, the rates were lowest among those 65 and over. Educational level was significantly related to smoking rates (Table 2). The ageand sex-adjusted smoking prevalence was 31% for those with no college, while

SMOKERS

18.24

AND DRINKERS

25 44

4544

787

IN AN HMO POPULATION

6%

18~24

2544

4564

65+

Age Groups FIG. 1. Prevalence of smoking and problem drinking, by age and sex, among Group Health Cooperative adults. (A) Smoker, current smoker; (B) binge drinker, five or more drinks on at least one occasion in last 30 days; (C) chronic drinker, averages at least two drinks/day; (D) drinking driver, driven automobile after more than two drinks in last 12 months.

for those with at least some college the smoking rate was 21% (P < 0.01). Smoking rates were also correlated with marital status (Table 2). The direction of the effect, however, was different for men and women (data not shown). Unmarried women were much more likely to be smokers than their married counterparts (P < O.OOl), while married men were slightly (but not significantly) more likely to be smokers than unmarried men. Educational achievement was not associated with drinking prevalence rates to the same degree as with smoking and the rates were in the opposite direction (Table 2). All drinking problems were reported more frequently by individuals with a college education than by those with less education, but the differences reached statistical significance only for drinking and driving. Although marital status was not significantly associated with chronic drinking (Table 2), both men and women not currently married reported prevalence rates of binge drinking and drinking and driving twice those of men and women currently married. Lifestyle

Characteristics

of Smokers and Drinkers

Table 3 shows the relationships between smoking and drinking and other lifestyle practices or factors. Smoking was only modestly associated with the other risky lifestyle practices (excluding drinking practices). Irregular seat-belt users

788

PEARSON ET AL.

TABLE 2 AGE- AND SEX-ADJUSTED SMOKING AND PROBLEM DRINKING PREVALENCE, BY EDUCATION AND MARITAL STATUS Variable

N”

% Smoke

Education High school grad or less Some college (OR)b

276 588

31% 21% (1.6)** [1.2, 2.31

KII’ Marital status Not Marriedd Married (OR)

[CII

281 581

26% 20% (1.4)% [l.O, 2.11

% Binge drinker

% Drink and drive

17% 21% [.5, 1.21

24% 32% (0.7)* [S, 1.01

22% 135 (1.8)** [1.2, 2.71

35% 17% (2.7)** H.9, 3.91

(0.8)

% Chronic drinker 8% 12%

(0.6) 10.4, 1.11 12% (14; [0.8, 2.31

Note. Significance of association with smoking or drinking: ***P < 0.001, **P < 0.01, *P < 0.05. r?N, effective sample size after weighting for oversampling in over-65 age groups. b Odds ratio (OR) adjusted for age and sex using Mantel-Haenszel procedure. c Ninety-five percent confidence interval (test-based) for Mantel-Haenszel odds ratio. d Includes never married, divorced, and widowed.

were more likely to smoke than those who regularly wore their seat belts (P < 0.01) (Table 3). In general, problem drinking measures were not associated with exercise practices, obesity, or seat-belt use (Table 3). However, it is of public health concern that binge drinking was significantly related to seat-belt use (P < 0.05). Health

Status Characteristics

After adjustment for age, sex, and education, those who perceived their health status unfavorably (poor or fair) were much more likely to be smokers (Table 4). The smoking prevalence for those who viewed their health status as “poor or fair” was 37%, while the smoking prevalence for those who viewed their health status as “good or excellent” was 22%. This was especially evident for men between ages 45 and 64, where the smoking prevalence for those reporting unfavorable health status was 77%, while for those reporting favorable health status it was 28% (data not shown). There were no differences in smoking rates between those who reported having chronic medical conditions (using the Alameda Health Status Scale (5)-e.g., arthritis, hypertension, diabetes) and those who did not (data not shown). However, smokers were more likely to report being worn out at the end of the day (P < 0.05) but less likely to report being tired after only 4 or 5 hours of sleep (P < 0.01). Additionally, smokers were much more likely to report “shortness of breath” (P < 0.001) and emphysema (P < 0.05) than nonsmokers. With respect to the indicators of social health status, smoking rates were lower among those who belonged to voluntary groups and organizations than among those who did not. In contrast, there was no significant association between

SMOKERS

AND DRINKERS

IN AN HMO POPULATION

789

TABLE 3 AGE-,SEX-,ANDEDUCATION-ADJUSTEDSMOKINGANDPROBLEMDRINKINGPREVALENCE,BY LIFESTYLECHARACTERISTICSORFACTORS % Smoke

% Binge drinker

% Drink and drive

% Chronic drinker

406 452

29% 20% (1.6)** [l.l, 2.21

24% 16% (1.6)* [l.l, 2.31

32% 27% (1.3) [0.9, 1.91

11% 11% (1.1) [0.6, 1.91

501 30.5

26% 21% (1.3) [0.9, 2.01

20% 20% (1.0) [0.7, 1.41

29% 32% (0.9) [0.6, 1.31

11% 10% (1.1) [0.5, 2.31

150 666

24% 24% (1.0) [0.4, 2.11

21% 19% (1.1) [0.5, 2.61

34% 28% (1.3) [0.8, 2.11

7% 12% (0.5) 10.3, 1.11

Variable Seat belt useb Infrequent Always (OR)’

[W Exercise’ Inactive Active (OR)

[CII Obesityf Overweight Not overweight (OR) cc11

Note. Significance of association with smoking or drinking: ***P < 0.001, **P < 0.01, *P < 0.05. n N, effective sample size after weighting for oversampling in over-65 age groups. b Infrequent: never, seldom, sometimes. Always: nearly always or always. c Odds ratio (OR) adjusted for age, sex, and education using Mantel-Haenszel procedure. d Ninety-five percent confidence interval (test-based) for Mantel-Haenszel odds ratio. e Inactive, less than 60 min active exercise per week. f Overweight, 120% or more of recommended weight using CDC Height/Weight Tables (1983).

smoking and our measure of social support. There was no association between smoking and depression or positive affect. Physical health status and self-perceived health status were not significantly associated with any of the problem drinking variables. Social participation was inversely related to all types of problem drinking behavior (P < 0.01). For example, chronic drinking prevalence rates were 5% for social participants and 15% for nonparticipants. Reported social support was not associated with any of the problem drinking indicators. Overall, there were no significant relationships between the two measures of mental health status (depression and positive affect) and chronic drinking. However, binge drinking and drinking and driving prevalences were significantly higher among those who reported lower levels of positive affect. In addition, drinking and driving practices were significantly higher among those who were depressed. Smoking and Problem Drinking Interrelationships The prevalence of smoking rates was significantly higher among those reporting each problem drinking behavior-binge drinking, chronic drinking, and drinking and driving (Table 5). Individuals with no drinking problems had an age-, sex-, and education-adjusted smoking prevalence of approximately 20%, while problem drinkers smoked at about twice that rate (P < 0.001).

790

PEARSON ET AL.

TABLE 4 AGE-, SEX-, AND EDUCATION-ADJUSTED SMOKING AND PROBLEM DRINKING PREVALENCE, BY HEALTH STATUS CHARACTERISTICS Variable

N”

% Smoke

% Binge drinker

% Drink and drive

Physical health status (Alameda)b Low High WV

350 514

25% 23% (1.1) [0.8, 1.51

19% 20% (0.9) [0.6, 1.41

31% 28% (1.1) [0.8, 1.61

37% 22% (2.1)** [l.l, 3.71

25% 19% (1.4) fO.6, 3.41

28% 30% (0.9) [0.3, 2.51

28% 24%

20% 20% (1.0) LO.7, 1.51

43% 28% (1.8)* [l.l, 3.21

14% 11% (1.4) [0.5, 3.61

36% 25% (1.7)** (1.2, 2.31

12% 10%

[0.8, I.71

24% 17% (1.5)* [l.O, 2.21

479 368

28% 18% (1.9)** [1.3, 2.71

25% 11% (2.6)** L1.7, 4.01

33% 24% (1.6)** [l.l, 2.21

15% 5% (3.5)** [2.1, 6.01

143 676

24% 24% (0.95) LO.5, 1.91

21% 19% (1.1) 10.5, 2.21

25% 29%

11% 11% (1.0) [0.6, I.71

[W

Self-perceived health status Poor-Fair Good-Excellent (ORI

71 782

KU Depression’ Present Absent (ON

71 765

(1.2) [0.6, 2.41

[CII Positive affectf Low High (ON

348 463

(1.2)

[CII

Social participationg Low High (OR)

[CII Social supporth Low High (OR)

[CII

22% 20%

(0.8) fO.5, 1.31

% Chronic drinker 9% 11%

(0.8) [0.5, I.21 9% 11%

(0.8) [0.2, 2.91

(1.2) [0.8, 2.11

Note. Significance of association with smoking or drinking: ***P < 0.001, **P < 0.01, *P < 0.05. a N, effective sample size after weighting for oversampling in over-65 age groups. b Low, with chronic conditions; high, without chronic conditions. c Odds ratio (OR) adjusted for age, sex, and education using Mantel-Haenszel procedure. d Ninety-five percent confidence interval (test-based) for Mantel-Haenszel odds ratio. e Depression present: reports being moody, low in spirits, and depressed a good bit of the time. f Positive affect low: reports being relaxed, cheerful, and interested in life less than a good bit of the time. g Social participation low: reports not belonging to voluntary groups or organizations. h Social support low: reports not having friends or family to talk with about personal problems.

Reporting one type of problem drinking behavior at least doubled, and in one instance increased by sixfold, the likelihood of reporting another type of problem drinking behavior (Table 5). This was most evident with chronic and binge drinking, where 82% of chronic drinkers were also binge drinkers, as compared with only 13% of those who drank fewer than two drinks daily (P < 0.001).

SMOKERS

AND DRINKERS

791

IN AN HMO POPULATION

TABLE 5 AGE-, SEX-, AND EDUCATION-ADJUSTED SMOKING AND PROBLEM DRINKING PREVALENCE, BY SMOKING AND DRINKING STATUS Response variable % Binge drinker

% Drink and drive

% Chronic drinker

NAb NA

39% 14% (4.1)*** [2.7, 6.01

45% 25% (2.4)*** [1.7, 3.51

19% 9% (2.4)*** [1.5, 3.81

169 691

47% 18% (4.1)*** [2.7, 6.01

NA NA

56% 23% (4.4)*** [3.1, 6.21

32% 2% (30.9)*** [18.5, 51.51

249 601

36% 19% (2.4)*** 11.7, 3.51

38% 12% (4.4)*** [3.1, 6.21

NA NA

27% 9% (3.8)*** [2.4, 6.01

94 774

41% 22% (2.4)*** [IS, 3.81

82% 13% (30.9)*** [18.5, 51.51

59% 26% (3.8)*** [2.4, 6.01

NA NA

Subgroup

N”

% Smoke

Current smoker Yes No (OR)’

204 862

[W Binge drinker Yes No (ON

Ul Drink and drive Yes No (ON

[CII

Chronic drinker Yes ER)

tcu

Note. Significance of association with smoking or drinking: ***P < 0.001, **P < 0.01, *P < 0.05. a N, effective sample size after weighting for oversampling in over-65 age groups. b NA, not applicable. c Odds ratio (OR) adjusted for age, sex, and education using Mantel-Haenszel procedure. d Ninety-five percent confidence interval (test-based) for Mantel-Haenszel odds ratio.

Interest

in Health

Promotion

Both smokers and problem drinkers expressed interest in receiving assistance to stop smoking or to deal with alcohol use. Although our definitions of problem drinking may seem severe, respondents in these categories sensed a need for help also. While only 7-8% of nonproblem drinkers expressed interest in receiving help from GHC regarding alcohol use, 27% of chronic drinkers, 21% of binge drinkers, and 14% of drinking drivers indicated they were interested in help from GHC regarding alcohol use (data not shown). DISCUSSION

Study data were collected exclusively by self-report which, for various reasons, may yield inaccurate information. The use of a self-report approach to collect information about drinking causes special concern because of the social stigma attached to excessive drinking. Previous studies have suggested that the reliability and validity of the frequency of drinking is satisfactory for both alcohol treatment populations and general population groups (2, 31). Additionally, the

792

PEARSON

ET AL.

validity of the reported amount of drinking may be adequate for the alcoholic treatment population, but in the general population, it appears to be less than satisfactory. Despite its shortcomings, the use of quantity-frequency measures to identify potentially dangerous drinking remains one of the best ways to monitor the drinking habits of a general population. In addition to the difficulties in measurement, the survey permits us only to estimate the prevalence of specific lifestyle practices and the strength of associations among various lifestyle practices and sociodemographic and health status variables at a single point in time. Our cross-sectional data demonstrate that smoking and drinking are correlated, but do not allow us to decipher the temporal relationship between the two. Our data show that approximately 25% of adult HMO enrollees smoked, despite the favorable predisposing characteristics of our population (better educated, predominantly white) against smoking (29) in contrast to national statistics. Smoking prevalence remained highest among the middle-age groups (45-64) and especially high among men of this age with a high school diploma or less education. Young women (18-24) had an alarmingly high rate of smoking-more than double that of their male counterparts. This difference in smoking prevalence between young men and women is similar to that reported in a recent study (22). The overall smoking prevalence in our HMO enrollee population is lower than other national survey data adjusting for differences in age, sex, and education (25% vs 30%) (1). It is unclear why our smoking prevalence is lower. We can only hypothesize that it may be due to HMO enrollees smoking less than the general population, west coast residents smoking less than the general population (3), or residual confounding. Comparison of our results with those of the Behavioral Risk Survey (22) showed there were virtually no differences in problem drinking rates. Nationally, 23% of the adult population reported having a binge drinking episode in the last 30 days, which is identical to the GHC figure. Also, 9% of the national population were chronic drinkers, while 11% of our population were chronic drinkers. Rates of problem drinking across both study populations were approximately twice as high for men as for women. Unfortunately because of differences in the way the drinking and driving questions were asked, it was not possible to make comparisons as to this behavior. Consistent with other studies (7, 8), we found that the persons who were at especially high risk for having alcohol abuse problems were younger, unmarried men and young women. Furthermore, our analysis confirms findings from a recent report (6) that suggests smoking and problem drinking do cluster together in population-based studies. Smokers and problem drinkers shared two common characteristics: a marked reduction in their prevalence by age 65, and a tendency not to belong to voluntary groups or organizations. With respect to the first, Cahalan (8) has suggested that most young or middle-aged problem drinkers “mature out” of their abusive drinking practices later in life; with respect to smoking, our data show that among seniors 49% are ever smokers, while only 11% are current smokers. Thus, the lower prevalence of smoking and drinking among seniors may be due to higher quit rates. Second, smokers and drinkers have higher mortality rates and are less

SMOKERS

AND

DRINKERS

IN AN

HMO

POPULATION

793

likely to survive past age 65. With respect to social participation, it seems apparent that smokers and drinkers, for whatever reason, are less likely to belong to or participate in voluntary groups or organizations. Whatever the explanation for this association, it suggests that smokers and drinkers may be less likely to participate in health promotion programming as well. Being married was significantly associated with a lower prevalence of binge drinking and drinking and driving, and, for women only, with smoking. It is possible that being married may be a protective or moderating factor against the development and/or maintenance of problem drinking and smoking. This, in part, is supported by a recent report (17) which suggests that divorced or separated women are at much higher risk for experiencing alcohol-related accidents. Education was negatively correlated with smoking, positively correlated with drinking and driving, and not correlated with binge drinking and chronic drinking. The inverse relationship between smoking and education is consistent with recent research which suggests that smoking is rapidly becoming concentrated among the working class and poor in the United States (9). The lack of association between binge and chronic drinking and education has also been reported by others (8). It has been suggested by several researchers that problem drinking in general does not have an education gradient, while specific subsets of problem drinkers do. For example, heavy beer drinkers come predominantly from blue-collar, lesseducated subgroups of the general population (21). Smokers were more likely to assess their health as poor or fair and to report the presence of emphysema but were no more likely to report having other chronic medical conditions, such as diabetes and hypertension. The extent to which the poorer self-perceived health status of smokers can be accounted for by smokingrelated cardiac or respiratory diseases cannot be estimated from these data. In contrast, there were no differences in the way problem drinkers and nonproblem drinkers viewed and reported their health. The clustering of smoking and drinking did not, for the most part, extend to other potentially risky lifestyle practices or factors. While binge drinkers and smokers were somewhat more likely (relative risk about 1.4) not to wear seat belts regularly, chronic drinkers and drinking drivers were no less likely to wear their seat belts regularly. Smokers and problem drinkers reported prevalence rates of sedentary behavior and obesity similar to those of their counterparts. The lack of association between smoking and being sedentary was somewhat surprising. Because of their reported reduced health status we expected that smokers would not be as likely to have a regular exercise program as nonsmokers. On the other hand, the lack of association between problem drinking and exercise was not surprising, since our definitions of problem drinking probably included a large number of drinkers whose alcohol consumption levels would not be expected to limit their exercise practices. The association between binge drinking or drinking and driving with depression and less positive affect is consistent with reports that alcohol use is greater with some forms of psychopathology. The direction of the association cannot be claritied by these cross-sectional data. Finally, binge drinking and drinking and driving are widespread among the

794

PEARSON

ET AL.

younger age groups. Forty-five percent of 18- to 44-year-old men and 30% of women in this age group reported a drinking and driving episode in the last 30 days. Similarly, among 18- to 24-year-olds, 50% of men and 35% of the women reported binge drinking in the last 30 days. The 18- to 24-year-old group was also the group showing the strongest association between binge drinking and nonseat-belt use: 67% of men and 46% of women who were non-seat-belt users were also binge drinkers. These data again confirm the importance of aggressive approaches to reducing the consequences of alcohol abuse for younger adults. Because binge drinking is so widespread and because it is considered by many to be “normal” drinking, other strategies might be considered in reducing the consequences of alcohol use. These include (a) changing the sale and distribution of alcohol by affecting when, where, and how much people drink-for example, increasing the tax on alcohol may reduce alcohol consumption, highway crash fatalities, and deaths due to cirrhosis of the liver; (b) teaching safe and appropriate drinking patterns, such as not drinking more than a specified amount per day, eating while drinking, and substituting nonalcoholic beverages for alcoholic beverages; and (c) making the environment safer in which to drink-examples include the installation of passive seat belts or air bags in all cars and the adoption of “Dial-a-Ride” programs for people who have had too much to drink. ACKNOWLEDGMENT We gratefully acknowledge Robin Dunlap for her support in preparing the manuscript.

REFERENCES 1. Advance Data, Vital Health Statistics of the National Center for Health Statistics, No. 113, November 15, 1985. 2. Armor, D. .I., Polch, J. M., and Stanbul, H. B. “Alcoholism and Treatment.” Rand Corp., Santa Monica, CA, 1976. 3. Bachman, J., et al. Smoking, drinking, and drug use among American high school students: Correlates and trends, 1975-1979. Amer. J. Public Health 71, 59-69 (1981). 4. Beery, W., et al. “Health Status Study of Adult Members of the Aid Association for Lutherans.” Univ. of North Caroiina Press, Chapel Hill, 1983. 5. Belloc, N. B., Breslow, L., and Hochstim, D. J. Measurement of physical health in a general population survey. Amer. J. Epidemiol. 193, 328-336 (1971). 6. Bradstock, K. M., et al. The behavioral risk factor surveys. III. Chronic heavy alcohol use in the United States. Amer. J. Prev. Med. 1, 15-20 (1985). 7. Cahalan, D., Cisin, I., and Crossley, H. “American Drinking Practices: A National Study of Drinking Behavior and Attitudes.” Rutgers Center of Alcohol Studies, New Brunswick, NJ, 1969. 8. Cahalan, D., and Room, R. “Problem Drinking among American Men.” Rutgers Center of Alcohol Studies, New Brunswick, NJ, 1974. 9. Cooper, R., and Simmons, B. E. Cigarette smoking and ill health among black Americans. N. Y. State J. Med. 85, 344-349 (1985). 10. Danchik, K. M., Schoenborn, C. A., and Elinson, J. “National Center for Health Statistics: Highlights from Wave 1 of the National Survey of Personal Health Practices and Consequences, United States, 1979.” Vital Health Statistics, Series 15, No. 1, DSSH Publ. No. (PHS) 81-1162, Public Health Service. U.S. Govt. Printing Office, Washington, DC, 1981. 11. Feinberg, S. E. “The Analysis of Cross-Classified Categorical Data,” 2nd ed. MIT Press, Cambridge, MA, 1980.

SMOKERS

AND DRINKERS

IN AN HMO POPULATION

795

12. Friedman, G. D., Loring, G. D., and Ury, H. K. Mortality in middle-aged smokers and nonsmokers. New Engl. J. Med. 300, 213-217 (1979). 13. Morbidity and Mortality Weekly 32, 573-576. “Alcohol and Fatal Injuries-Fulton County, Georgia, 1982.” Deaths-United States, 14. Morbidity and Mortality Weekly 32, 649-651. “Alcohol-Related 1968-1978.” 15. Maletzky, B. M., and Klotter, J. Smoking and alcoholism. Amer. J. Psychiatry 131, 445-447 (1974). 16. Mantel, N., and Haenszel, W. Statistical aspects of the analysis of data from retrospective studies of disease. .I. Nat/. Cancer Inst. 22, 719-748 (1959). 17. McCormack, A. Risk for alcohol-related accidents in divorced and separated women. J. Stud. Alcohol 46, 240-243 (1985). 18. Miller, W. R., Hedrick, K. E., and Taylor, C. A. Addictive behaviors and life problems before and after behavioral treatment of problem drinkers. Addict. Behav. 8, 403-412 (1983). 19. Noble, E. P. (Ed.). “Third Special Report to the U.S. Congress on Alcohol and Health from the Secretary of Health, Education, and Welfare.” National Institute of Alcohol Abuse and Alcoholism, Rockville, MD, 1978. 20. O’Farrell, T. J., Conners, G. J., and Upper, D. Addictive behaviors among hospitalized psychiatric patients. Addict. Behav. 8, 329-333 (1983). 21. Pittman, D. J., and Snyder, C. R. “Society, Culture and Drinking Patterns.” Wiley, New York, NY, 1962. 22. Remington, P. L., Forman, M. R., Gentry, E. M., et al. Current smoking trends in the United States: The 1981-1983 behavioral risk factor surveys. JAMA 253, 2975-2978 (1985). 23. Schmidt, W., and Popham, R. G. Heavy alcohol consumption and physical health problems: A review of the epidemiological evidence. Drug Alcohol Depend. 1, 27-50 (1976). 24. Schucket, M. A. Prevelance of affective disorder in a sample of young men. Amer. J. Psychiatry 139, 1431-1436 (1982). 25. Stason, W. B., Neff, R. K., Miettinen, 0. S., et al. Alcohol consumption and nonfatal myocardial infarction. Amer. J. Epidemiol. 104, 603-608 (1976). 26. Tuyns, A. J. Alcohol, in ‘Cancer Epidemiology and Prevention” (D. Schottenfeld and J. Fraumeni, Eds.), pp. 293-303. Saunders, Philadelphia, PA, 1982. 27. U.S. Dept. of Health, Education, and Welfare. “Smoking and Health: A report to the Surgeon General.” DHEW Publ. No. (PHS) 79-50066. U.S. Govt. Printing Office, Washington, DC, 1979. 28. U.S. Dept. of Health, Education, and Welfare. “Healthy People: The Surgeon General’s Report on Health Promotion and Disease Prevention.” DHEW (PHS) Publ. No. 79-55071. U.S. Govt. Printing Office, Washington, DC, 1979. 29. U.S. Dept. of Health and Human Services. “Health-United States, 1982.” DHHs Publ. No. (PHS) 83-1232. U.S. Govt. Printing Office, Washington, DC, 1982. 30. Wald, N. J. Mortality from lung cancer and coronary heart disease in relation to changes in smoking habits. Lancet 1, 136-138 (1976). 31. Ware, J. E., Johnston, S. A., Davies-Avery, A., and Brook, R. H. “Conceptualization and Measurement of Mental Health for Adults in the Health Insurance Study (Vol. 3 Mental Health R-1987/3 HEW).” U.S. Govt. Printing Office, Washington, DC, 1979. 32. Williams, G., et al. Reliability of self-reported alcohol consumption in a general population survey. J. Stud. Alcohol 46, 223-227 (1985). 33. Wynder, E. L., Mabuchi, K., and Beattie, E. J. The epidemiology of lung cancer: Recent trends. JAMA 213, 2221-2228 (1970). 34. Wynder, E. L., and Stellman, S. D. The comparative epidemiology of tobacco-related cancers. Cancer Res. 31, 4608-4622 (1977).