Peter W. Vik
San Diego Veterans Affairs Medical Center
John R. McQuaid
San Diego Veterans Affairs Medical Center
Thomas L. Patterson
San Diego Veterans Affairs Medical Center
Michael R. Irwin
San Diego Veterans Affairs Medical Center
San Diego Veterans Affairs Medical Center
Journal of Abnormal Psychology, November 1990 Vol. 99, No. 4, 344-348
We examined the relation between stressful life events and drinking outcome among 129 male alcoholics who had completed an alcohol treatment program. Life events were assessed for the year prior to treatment and for the 3 months after treatment and were rated on the Psychiatric Epidemiology Research Interview and the Contextual Rating System. Approximately 40% of the pretreatment stressors were found to be directly or indirectly related to alcohol use. When stressors related to drinking were excluded from consideration, we found that men who returned to drinking after treatment experienced more severe or highly threatening stress before their relapse than men who remained abstinent during the follow-up period. These data suggest that although less severe stress may not increase risk for relapse, acute severe stressors and highly threatening chronic difficulties may be associated with elevated relapse risk.
The relation between stress and alcohol abuse is well documented, albeit not well understood. Alcoholics report high levels of life change (e.g., Mules, Hague, & Dudley, 1977 ), and abusers who relapse report stressful experiences before relapse drinking ( Hore, 1971a , 1971b ) and more negative life events than recovered alcoholics ( Billings & Moos, 1983 ; Miller, Hedrick, & Taylor, 1983 ; Rosenberg, 1983 ). In general, these studies have demonstrated an association between stress and alcohol relapse, despite wide differences in the measurement of stress, outcome, and the length of time between treatment and follow-up.
Although stress is often portrayed as a causal factor in alcohol relapse, the causal relations between posthospitalization experiences and treatment outcome are dynamic and reciprocal ( Finney, Moos, & Mewborn, 1980 ). Heavy or abusive use of alcohol can generate its own stress, including job difficulties, family and marital disturbances, and legal or medical problems. In fact, such major life problems are often used to define the severity of alcohol dependence (e.g., American Psychiatric Association, 1987 ).
Unfortunately, studies that relate psychosocial stress and alcohol abuse often fail to consider the extent to which the stressful experience occurs independently of alcohol use and the degree to which stressors are related to or a consequence of alcohol consumption. As Allan and Cooke (1985) noted, the reciprocal relation between stress and substance abuse confounds these measures and often limits interpretation. To demonstrate that stressful life experiences increase the risk of relapse as is hypothesized in many relapse models (e.g., Brownell, Marlatt, Lichtenstein, & Wilson, 1986 ), stressful events that clearly occur independently of alcohol use must be examined separately.
Studies of alcohol relapse typically limit their focus to posttreatment stress as a factor in returning to drinking (e.g., Billings & Moos, 1983 ), even though stressful life events such as job loss or divorce that occur before treatment may continue to have an impact on the alcoholic during posttreatment periods. In this study we introduce a method of classifying stressful life events and ongoing difficulties according to their association with alcohol, and we assess whether alcoholics who return to drinking experience more threatening and chronic stressors that are unrelated to their drinking than those who abstain after treatment.
The purpose of our study was to examine several aspects of the stressful life event-alcohol relapse association. It was hypothesized that psychosocial stress that maximally taxes one's coping repertoire by virtue of stress severity or chronicity may increase the risk for relapse drinking after treatment. Essentially, we examined severe stress as a potential causal agent in relapse after treatment for alcoholism. We first addressed the confounded nature of stress and alcohol abuse by examining separately stressful life events that were independent of alcohol use and those related to alcohol use. Thus, we introduced a method for evaluating the degree of association between a stressful event and alcohol use. Second, we examined whether more threatening stress of an acute or chronic nature was related to poorer drinking outcome after treatment.
One hundred twenty-nine male inpatients in an Alcohol Treatment Program (ATP) at the San Diego Veterans Administration Medical Center were included in this study. The subjects were participants in the Veterans Administration Clinical Alcohol Research Center (ARC), which included those ATP patients who met Diagnostic and Statistical Manual of Mental Disorders (rev. 3rd ed.; DSM-III-R ; American Psychiatric Association, 1987 ) criteria for alcohol dependence, had no preexisting major psychopathology (e.g., affective disorder or antisocial personality disorder), and had a resource person (family member or close friend) who participated in an independent interview to provide corroborative information. Because the clinical course of alcoholism may vary as a function of primary disorder ( Schuckit, 1985 ), ATP participants with a major psychiatric disorder or drug dependence that predated onset of alcohol dependence were excluded from the study.
The men ranged in age from 22 to 70 years ( M = 45.8, SD = 10.6). Two thirds (66%) of the sample were currently unmarried, with 47% divorced, 17% single, and 2% widowed. The subjects were predominantly White (85%), with 8.5% Black, 4% Mexican-American, and 2% Native American. Most men expressed a religious preference (83%), yet only 36% reported currently practicing their religion. Sixty percent of the men had completed high school or the General Equivalency Degree, and an additional 17% reported technical schooling, 11% graduated from college, and 6% had obtained an advanced degree. This sample consumed alcohol an average of 25.2 days per month ( SD = 7.6) and 16.4 drinks ( SD = 10.9) per occasion during the 3 months before treatment. Although 61% of the sample reported use of some other drug during their lifetime, only 24% reported any drug use during the 3 months before treatment, with an average last use of 33 days before ATP admission.
Within 48 hours of admission, ATP participants completed a structured diagnostic interview ( Schuckit, Irwin, Howard, & Smith, 1988 ) that covered the patient's social, family, and educational background and his alcohol, drug, medical, and psychiatric history. Additionally, a resource person (family member or close friend, typically living with the alcoholic) completed the same interview to corroborate the data gathered. Patient, resource person, and medical chart information was reviewed, and those ATP participants who met DSM-III-R criteria for alcohol dependence with no prior history of other drug dependence or psychiatric disorders were accepted into the ARC, in which 97% agreed to participate.
The Psychiatric Epidemiology Research Interview-Modified (PERI-M; Hirschfield et al., 1977 ) was conducted by a trained technician during the 3rd week of the 4-week ATP. Assessment at this time allowed for an average of 29.4 days of abstinence, which, it was hoped, minimized the effects of acute alcohol withdrawal on memory. The subjects and resource persons were again independently interviewed with the ARC follow-up interview ( Schuckit et al., 1988 ) 3 months after discharge from the program, and the subjects completed the PERI-M. Although the details of the follow-up interview have been reported elsewhere (e.g., Irwin, Baird, Smith, & Schuckit, 1988 ; Schuckit, Irwin, & Brown, 1990 ), the procedure is a structured clinical interview separately completed by a patient and a resource person and designed to ascertain information about alcohol and drug use during the follow-up period as well as functioning in major life domains (e.g., work, family, health, etc). With a variety of procedures to enhance follow-up, 86% (n = 111) of the original sample completed the full ARC assessment battery during treatment and at follow-up. For outcome classification purposes any alcohol or drug use during the follow-up period was considered a relapse.
Stressful life events and difficulties were evaluated on the Psychiatric Epidemiology Research Interview as modified by the National Institute of Mental Health Clinical Research Branch ( Hirschfield et al., 1977 ). The PERI-M is a structured, confidential interview based on the original work of Dohrenwend and Dohrenwend (1974) . Our study also incorporated additional instructions and probes that enhance reliability ( Grant, McDonald, Patterson, & Trimble, 1989 ). In a structured format the interviewer inquired about the occurrence of 133 events related to school, work, love, health (personal and of important others), crime, money, childbirth, family, residence, personal habits, and death that had occurred within the 12 months before admission to the ATP and at follow-up for the 3 months after discharge. The subjects were also asked about any other events they had experienced that did not appear on the PERI-M list. In the case of the occurrence of any event, participants provided a detailed description of the experience including precursors, consequences, and proximity to drinking.
After the interview three raters, trained to reliability ( r > .90), reviewed all interview information and resource person data to score stressful life experiences according to the Brown and Harris Contextual Rating System (CRS; G. W. Brown & Harris, 1982 ; Seligman & Meyer, 1970 ). The CRS was used to assign by consensus several standardized ratings to each event: (a) short-term threat, (b) long-term threat, (c) focus, and (d) category. Whenever questions about specifics of events (e.g., when an event occurred relative to drinking) arose, other data were examined (i.e., resource person interview, diagnostic and alcohol-use interviews, or patient's medical file).
Of particular interest to this report is the personal-threat rating of the CRS assigned to items experienced during the evaluation period. In the CRS, stressors that represent an acute threat to the participant are classified as stressful events (e.g., accidents), whereas ongoing stressors with a significant personal threat are considered ongoing difficulties (e.g., chronic marital discord). CRS events receive both a short-term (immediate) and a long-term threat rating, but ongoing difficulties receive a single chronic threat rating. Event ratings range from 1 ( severe stress ) to 4 ( mild or minimal ) and ongoing difficulties range from 1 ( severe stress ) to 7 ( trivial ). An example of an event with high short-term threat but little long-term threat is a man receives news that his spouse was taken to the hospital in an ambulance with chest pain (short-term threat level = 1) only to discover that the spouse was fine and had suffered only from an anxiety attack (long-term threat level = 4). In contrast, if it was discovered that his spouse had experienced a fatal myocardial infarction, a more severe long-term threat is evident (long-term threat level = 1). Short-term and long-term ratings as well as chronic threat assessments have been shown to have high (> .90) interrater reliabilities (e.g., G. W. Brown & Harris, 1978 ; Parry, Shapiro, & Davis, 1981 ; Tennant, Smith, Bebbington, & Hurry, 1979 ).
In order to examine highly threatening stress that may increase the risk for relapse separately, stressful experiences were classified as severe or nonsevere. With a categorization system established for persons with chronic disorders ( Grant, Brown et al., 1989 ; Grant, McDonald et al., 1989 ), those events that received a short-term rating of 1 and a long-term rating of 1 or 2 and ongoing difficulties that received ratings of 1, 2, or 3 were classified as severe stressors. This classification of more severe stressors and less severe stressors has proven to be a useful distinction in the examination of the clinical course of chronic disorders such as depression ( G. W. Brown, Bifulco, & Harris, 1987 ).
In addition to the standard scoring systems for the PERI-M, each stressor was evaluated according to its relatedness to or independence from the subject's drinking and given an alcohol independence rating (AIR). Items were classified into the following categories: (a) occurrences independent of alcohol use, or alcohol independent (AI; e.g., home was burglarized, or spouse was hospitalized for cerebral vascular accident), or (b) occurrences possibly or probably related to the subject's use of alcohol, or alcohol related (AR; e.g., job loss because of drinking, or hospitalization for injuries sustained while drinking). With the time of occurrence (i.e., date and if event occurred during drinking or nondrinking time periods) and whether alcohol was directly or indirectly involved in the occurrence of the event (e.g., while drinking, during withdrawal, or while attempting to obtain alcohol), AIR categorical ratings were assigned by the interviewer at the time of the interview. Information obtained from the structured follow-up interview with the resource person and medical records were used to validate these ratings. In the uncommon circumstance that items were difficult to rate, information from all sources was discussed with other researchers and assigned a consensus rating. A conservative approach was taken in that when contradictory information was present, a worst-case-scenario was recorded (i.e., presumed to be possibly or probably alcohol related). Interrater reliability of this system was assessed with two raters who independently rated 100 items with a concordance of 97%.
Of the 111 men who completed the follow-up, 76 (68%) remained abstinent from alcohol and other drugs for the entire 3 months. Of the remaining 35 men who used alcohol or drugs during the follow-up period, 34% drank during only 1 month, 32% drank during 2 months, and 20% drank during all 3 months, with an average of 8.30 drinks per drinking day ( SD = 8.07). Five relapsed subjects (14%) used drugs other than alcohol. For the relapsers the mean length of initial abstinence was 52.02 days ( SD = 22.21), and the average number of drinking days was 26.54 ( SD = 30.49), with a range of 1-90 days of drinking. Because we examined the hypothesis that severe stress increases risk for relapse (any alcohol or drug use), all posttreatment relapsers were combined in the analyses. Drinking and abstaining groups were not significantly different ( p < .05) on age, marital status, educational level, or pretreatment drinking variables such as quantity of consumption, frequency of drinking, severity of withdrawal symptoms, or number of alcohol-related life problems.
The primary alcoholics in this study reported an average of 9.11 ( SD = 4.45) PERI-M events for the 1 year before admission to the ATP. With the AIR system, 42% of these events ( M = 3.82, SD = 2.90) were classified as AR. Thus, alcoholics reported an average of 5.16 ( SD = 3.37) PERI-M events that were AI in the 1 year before treatment.
Abstainers and relapsers do not differ during the pretreatment period on the more general stress measures of total number of PERI-M items, F (1, 128) = 0.55, p = .46, number of AR stressful life events, F (1, 128) = 0.59, p = .45, or AI stressors, F (1, 128) = 0.07, p = .79. Both groups report a comparable number of highly threatening life events and ongoing difficulties. Of note, such severe stress constituted a small portion of pretreatment stress.
Similarly, abstaining and relapsing alcoholics reported comparable numbers of total PERI-M items during the 3-month follow-up period, F (1, 110) = 0.94, p = .33. Alcoholics who relapsed within the first 3 months after treatment tended to report more AR stressful life experiences than abstaining alcoholics, F (1, 110) = 2.83, p = .10, although even abstainers reported life stressors secondary to pretreatment alcohol use ( M = 1.14, SD = 1.25). Outcome groups did not significantly differ on the total number of life AI events, F (1, 110) = 0.74, p = .39.
In order to examine whether severe stress was a precipitant to relapse after treatment, abstaining and relapsing men were compared on the number of highly threatening events and number of severe ongoing difficulties during the follow-up period. First,a greater number of highly threatening life events and severe ongoing difficulties were identified for relapsers as compared with abstainers, F (1, 109) = 4.19, p < .05, and F (1, 109) = 8.21, p < .005, respectively.
As a second, more stringent assessment of the severe stress-relapse association, only severe stress that occurred independently of alcohol use and before relapse was considered. A severe stress composite was calculated by adding the number of highly threatening events and severe ongoing difficulties that were rated as AI for each assessment period (before and after treatment). Alcoholics who returned to drinking after treatment were not significantly different from abstainers on the number of AI severe stressors that occurred before treatment ( M = 0.51, SD = 0.74, and M = 0.38, SD = 0.72, respectively); however, there was a trend ( p < .10) for relapsers to report more severe AI ongoing difficulties than abstainers ( M = 0.31, SD = 0.33, and M = 0.17, SD = 0.31, respectively).
Those alcoholics who returned to drinking within 3 months after treatment experienced significantly more posttreatment AI severe stress that occurred before initial relapse ( M = 52.02 days, SD = 22.21) than abstainers did during the entire 3-month period, F (1, 110) = 5.53, p < .05. Thirty percent of the relapsers experienced at least one AI severe stressor before their relapse date as compared with 16% of abstainers who reported such a stressor during the 90 posttreatment days. The alcoholics who drank during the follow-up period experienced both more marked AI ongoing difficulties ( p < .05) and tended to have more severe AI events ( p < .10) than did abstainers.
Stress and outcome were next examined in two continuous posttreatment drinking measures: length of initial abstinence and Drinking Days × Average Number of Drinks per Drinking Day (severity of relapse). The total number of PERI-M items, the number of severe stress experiences, and AI severe stress at both assessment times (pretreatment and posttreatment) were not correlated with length of initial abstinence after treatment ( r s = -.09-.00). In contrast, although pretreatment measures of stress were not correlated with the severity of relapse, posttreatment measures of stress including number of PERI-M items, severe stress, and severe AI stress were significantly ( p < .05) associated with this outcome measure ( r s = .18, .43, and .21, respectively).
The results of this study indicate that the severity of stressful life experiences and the relation of such events to alcohol consumption are important dimensions to be considered when examining the life event and alcohol relapse association. A significant portion of stressful life experiences that alcoholics reported for the year prior to treatment and one third of the posttreatment stress reported by relapsing alcoholics appears to be related to their alcohol use in that the stressful experience occurred during a drinking episode or was a consequence of alcohol use. Thus, whereas previous studies (e.g., Mules et al., 1977 ; Rosenberg, 1983 ) have reported elevated levels of stressful life events among alcoholics, our findings provide support for a bidirectional relation between alcohol use and psychosocial stress as has been previously hypothesized (e.g., Allan & Cooke, 1985 ; Niaura et al., 1988 ).
The second major finding of our study was that severe stress, defined as highly threatening life experiences, was associated with poorer drinking outcome. Alcoholics who relapsed after treatment experienced more severe stress before their initial posttreatment relapse than abstaining men reported for the entire 3 months after treatment. Because the base rate for severe life events is relatively low and because not all relapsers had experienced a highly threatening stressful life event, severe stress is obviously not the only precursor to relapse. Abstainers and relapsers reported comparable numbers of events that involved alcohol use before treatment and similar pretreatment drinking habits, which suggests comparable severity of dependence. When considered in total, this pattern of results suggests that models of relapse that include stress as a risk factor (e.g., Brownell et al., 1986 ; Crutchfield & Grove, 1984 ) must attend to stress severity and chronicity in the theoretical formulation of relapse risk.
The elevated incidence of severe stress before relapse provides a starting point for future studies to incorporate longer follow-up periods, more refined evaluations of the relapse process ( S. A. Brown, Vik, & Creamer, 1989 ), and possible moderators that may account for abstinence despite the experience of severe stress.
Allan, C. A. & Cooke, D. J. (1985). Stressful life events and alcohol misuse in women: A critical review. Journal of Studies on Alcohol, 46, 147-152.
American Psychiatric Association.(1987). Diagnostic and statistical manual of mental disorders (Rev. 3rd ed.). Washington, DC: Author.
Billings, A. G. & Moos, R. H. (1983). Psychosocial processes of recovery among alcoholics and their families: Implications for clinicians and program evaluators. Addictive Behaviors, 8, 205-218.
Brown, G. W. & Harris, T. O. (1978). Social origins of depression: A study of psychiatric disorder in women. London: Tavistock.
Brown, G. W. & Harris, T. O. (1982). Life events and psychiatric disorders II. Psychosocial Medicine, 3, 159-176.
Brown, G. W., Bifulco, A. & Harris, T. O. (1987). Life events, vulnerability and onset of depression: Some refinements. British Journal of Psychiatry, 150, 30-42.
Brown, S. A., Vik, P. W. & Creamer, V. A. (1989). Characteristics of relapse following adolescent substance abuse treatment. Addictive Behaviors, 14, 291-300.
Brownell, K. D., Marlatt, G. A., Lichtenstein, E. & Wilson, G. T. (1986). Understanding and preventing relapse. American Psychologist, 41, 765-782.
Crutchfield, R. D. & Grove, W. R. (1984). Determinants of drug use: A test of the coping hypothesis. Social Science Medicine, 18, 503-509.
Dohrenwend, B. S. & Dohrenwend, B. P. (1974). Stressful life events: Their nature and effects. New York: Wiley.
Finney, J. W., Moos, R. H. & Mewborn, C. R. (1980). Posttreatment experiences and treatment outcome of alcoholic patients six months and two years after hospitalization. Journal of Consulting and Clinical Psychology, 48, 17-29.
Grant, I., Brown, G. W., Harris, T., McDonald, W. I., Patterson, T. & Trimble, M. R. (1989). Severely threatening events and marked life difficulties preceding onset or exacerbation of multiple sclerosis. Journal of Neurology, Neurosurgery, and Psychiatry, 52, 8-13.
Grant, I., McDonald, W. I., Patterson, T. & Trimble, M. R. (1989). Life events and multiple sclerosis. In G. W. Brown & T. O. Harris (Eds.), Life events and illness (pp. 295-311). New York: Guilford Press.
Hirschfield, R. M. A., Lerman, G. L., Schless, A. P., Endicott, J., Lichtenstaeder, S. & Clayton, P. J. (1977). Modified Life Events sections of the Psychiatric Epidemiology Research Interview (PERI-M). Bethesda, MD: National Institute of Mental Health.
Hore, B. D. (1971a). Life events and alcohol relapse. British Journal of Addiction, 66, 83-88.
Hore, B. D. (1971b). Factors in alcoholic relapse. British Journal of Addiction, 66, 89-96.
Irwin, M., Baird, S., Smith, T. L. & Schuckit, M. (1988). Use of laboratory tests to monitor heavy drinking by alcoholic men discharged from a treatment program. American Journal of Psychiatry, 145, 595-599.
Miller, W. R., Hedrick, K. E. & Taylor, C. A. (1983). Addictive behaviors and life problems before and after behavioral treatment of problem drinkers. Addictive Behaviors, 8, 403-412.
Mules, J. E., Hague, W. H. & Dudley, D. L. (1977). Life change, its perception and alcohol addiction. Journal of Studies on Alcohol, 38, 487-493.
Niaura, R. S., Rohsenow, D. J., Binkoff, J. A., Monti, P. M., Pedraza, M. & Abrams, D. B. (1988). Relevance of cue reactivity to understanding alcohol and smoking relapse. Journal of Abnormal Psychology, 97, 133-152.
Parry, G., Shapiro, D. A. & Davis, L. (1981). Reliability of life event ratings: An independent replication. British Journal of Clinical Psychology, 20, 133-134.
Rosenberg, H. (1983). Relapsed versus non-relapsed alcohol abusers: Coping skills, life events, and social support. Addictive Behaviors, 8, 183-186.
Schuckit, M. A. (1985). The clinical implications of primary diagnostic groups among alcoholics. Archives of General Psychiatry, 42, 1043-1049.
Schuckit, M. A., Irwin, M. & Brown, S. A. (1990). The history of anxiety symptoms among 171 primary alcoholics. Journal of Studies on Alcohol, 51, 34-41.
Schuckit, M. A., Irwin, M., Howard, T. & Smith, T. (1988). A structured diagnostic interview for identification of primary alcoholism: A preliminary evaluation. Journal of Studies on Alcohol, 49, 93-99.
Seligman, M. E. P. & Meyer, B. (1970). Chronic fears and ulcers as a function of unpredictability of safety. Journal of Comparative and Physiological Psychology, 73, 202-211.
Tennant, C., Smith, A., Bebbington, P. & Hurry, J. (1979). The contextual threat of life events: The concept and its reliability. Psychological Medicine, 9, 525-528.
Vallant, G. E. & Milofsky, E. S. (1982). Natural history of male alcoholism. Archives of General Psychiatry, 39, 127-133.
Watson, C. G. & Purcell, J. (1985). Consistency of posttreatment alcoholics' drinking patterns. Journal of Consulting and Clinical Psychology, 53, 679-683.
Copyright © 2000, Southern Development Group,
S.A. All Rights Reserved.
A Private Non-Profit Agency for the good of all, published in the UK & Honduras