Posttraumatic Stress Disorder in Older Alcoholic Combat Veterans


SATU at the Veterans Administration Medical Center comprises 102 beds across five inpatient programs and is a part of a large, 1300-bed psychiatric hospital. SATU programs include medical detoxification, 45-day alcohol and 45-day drug therapeutic communities, a 45-day therapeutic community for older veterans or veterans with serious psychiatric disorders, and one 14-day skills-training, relapse prevention program. SATU has been in existence since 1975.

Our patients’ average age is 40 years, with a range of 19-69 years. The racial mix of admitted patients is approximately 65% white, 32% black, and 3% Hispanic. Over a given year, SATU discharges approximately 1200 patients. These patients seek services for alcohol and other substance abuse/ dependence disorders. Patients with a primary diagnosis of alcohol dependence, with other drug use, account for approximately 40% of our case mix. Patients who are principally drug dependent, with or without other alcohol/ drug use, account for approximately 60% of our case mix. Of our admissions, just less than 50% undergo formal detoxification using therapeutic medications (e.g., Librium, methadone).

Typically, when a veteran comes in for substance abuse treatment, he enters the admission/evaluation/detoxification ward for approximately one week. Toward the end of that time, he is assessed by the treatment team and sent to either a 14-day alcohol/drug skills-training program or one of three longer-term rehabilitation programs. A veteran who comes from one of the other Medical Center wards may also enter a rehabilitation program subsequent to evaluation by SATU staff.

Owing to the number of patients potentially at risk for having posttraumatic stress disorder and the effects that untreated posttraumatic stress disorder could have on substance relapse, we instituted procedures of early identification, triage, and treatment for these veterans. Early identification means that all patients on our admission/evaluation wards who score high on both combat- and stress-related symptoms are formally asked about posttraumatic stress disorder during a treatment team meeting. In this meeting, decisions are made regarding transfer to longer-term drug/alcohol rehabilitation care. If that transfer takes place, that rehabilitation program treatment team follows up on any documented evidence of potential PTSD. A handful of core staff have been trained to work clinically with posttraumatic stress disorder patients. These staff assist the treatment teams in the care of the patient. They may suggest referral to a specialized posttraumatic stress disorder unit, offer individual posttraumatic stress disorder specific therapy, or both. The main focus of the staff is to oversee the complex care of these often difficult cases.


The subjects were all veteran patients 49 years of age and older admitted between June 1984 and June 1986. Test data were available on 306 sequential admissions.


Nursing staff handed out the stress package on the evening of admission. Patients were given 1 full day to complete the package, and it was picked up on the morning of their third inpatient day. In cases of poor vision or English illiteracy the package was orally administered. If patients were either too psychiatrically or physically ill to fill out the forms within the time limit, progress was monitored and the material was retrieved as soon as the veteran felt well enough to complete the task. Following hand-scoring, the stress-inventory-package scales were entered into a computer file manager program (VA FileMan program on a DEC PDP 11/24) for storage and retrieval.

The Stress Scale. This scale is made up of 23 items and includes symptoms commonly associated with World War II combat veterans, POW’s from Korea, concentration camp survivors, survivors of a natural disaster, and Vietnam veterans. Items include feelings of dizziness, anxiety, tension, headaches, stomach troubles, asthma, feeling irritable, sleep disturbances, depression, feeling jumpy or easily startled, feeling confused, and so forth. Respondents were asked if they experienced these symptoms during service or within a year after discharge. If they reported “sometimes/ “often,” or “very often/ a score of 1 resulted.

The Combat Scale. This scale is made up of 16 items and is adapted from the original Laufer et al. scale. More recent research caused us to add four items. The original scale included items indicating passive combat (i.e., received incoming fire, flew in aircraft over Vietnam, received sniper fire, and so forth) and active combat (i.e., engaged enemy in combat, saw enemy killed, saw Americans killed, and so forth). The following items were added: saw civilians killed or injured, killed enemy, killed civilians, killed friendly forces. A Likert-type scale (0—never, 1—rarely, 2—sometimes, 3—often, 4—very often) was used to determine level of combat.

The Psychiatric Epidemiology Research Interview (PERI). PERI is comprised of a demoralization scale (with poor self-esteem, hopelessness-helplessness, anxiety, sadness, psychophysiological symptoms subscales) and suicidal thoughts, guilt, angry feelings, active expression of hostility, perceived hostility from others, distrust, sex problems, somatic problems, lack of mastery, and repression of anger scales. Each question is scored on a five-point Likert scale (e.g., in the last 5 months how often have you felt confident? 0—very often; 25—fairly often; 50—sometimes; 75—almost never; 100—never). Scale scores are summed for an overall PERI score.

A Demographic Questionnaire. The questionnaire was developed at this Medical Center to elicit biographical information that might allow us to distinguish high-combat/high-stress veterans from high-combat/low-stress patients.

The Stress Inventory Package. This also included some well-recognized instruments chosen in hope that they might differentiate the high-combat/high-stress and high-combat/low-stress groups. These included the Desirability of Control Scale, the Rotter Questionnaire, and the Sensation Seeking Scale.


Veterans were instructed to indicate specific stress symptoms “during military service” or “within the first year after leaving military service.” This demanded recall of feelings occurring some 30-40 years previously. Of major concern was the possibility that these resulting stress scores were determined by current discomfort and, therefore, did not actually reflect the patient’s stress level during the earlier period. In an unpublished study, we instructed veterans to report stress symptoms during and/or within the first year after service. We also asked them to report stress symptoms for the 12 months before admission. The study results indicated patients’ current feelings had no strong relationship to how they felt during or just after their time in the service. These data are also supported by clinical interview material.

Statistical methods used so far on these data have resulted in disappointing results for the Rotter, Desirability of Control Scale, and the Sensation Seeking Scale. These scale results were divided into low, medium, and high scores, as were the combat and stress scores. Multiple ANOVAs yielded nonsignificant results. A combination of all these scales in a multiple-regression and discriminate-analyses design will hopefully prove more fruitful.

A serendipitous finding that certainly bears mentioning, resulted from an attempt to track down patient files. Seventy percent of the files of patients reporting low stress were on station. Only 30% of the high-stress patient files were available. Files move out of the station record room when patients are in treatment either at our facility or at some other facility. The high percent of low-stress patient files available suggests that these veterans sought treatment on SATU and then exited the VA health-care-delivery system. As for the high-stress group, 70% sought treatment on SATU but were back in the VA system at the time we attempted to recover the files.

Demographic data were analyzed for low stress (mean stress scale score = 0.81) and high stress (mean stress scale score = 13.5) for veterans with combat scores greater than 9. There were no significant differences between the groups. Consequently, results for both low- and high-stress groups are summarized below. For this combined group, 64% either graduated from high school or achieved their general-equivalency diploma and 28% had some college or better. The average number of siblings for both groups was just over three (3.4). Eighty percent of these veterans had moved no more than two times. Ninety percent reported that neither parent had legal problems. Eighty-two percent reported no emotional and/or physical abuse by parents. For this combined group, 90% received honorable with 9% receiving general under honorable discharges. Twenty-three percent felt their experience in the service was average; 73% felt it was good to very good. As a group, these veterans were concerned about possible adjustment problems upon return (70%). For the most part they found friends, family, and peers to be friendly and helpful. Seventy-five percent found employers slightly to very interested in hiring veterans.

Postwar adjustment for these groups was remarkably poor. Only 8% of the total were arrested before entering the service. After service, 70% were arrested. Arrests for both groups increased approximately 500%, number found guilty increased by 700%, and number jailed increased by 500%. Types of crimes committed (e.g., rape, robbery, assault, driving under the influence of alcohol charges) did not vary across the high-stress and low-stress groups. Sixty-one percent of patients in the combined group reported substance use preservice, and 48% felt that usage was excessive. These figures increased to 92% use and 80% excessive use during service and to 95% use and 95% excessive use after discharge. Of the low-stress group, 35% reported alcoholic fathers and 11% reported alcoholic mothers. For the high-stress group, the respective percentages were 40% and 5%.


Before discussing these results it is critical to clarify the limitations of the data. First, the project was forced on us by a patient care crisis. Consequently, it was not the brainchild of an elaborate experimental hypothesis based on theory and observation. As a response to a crisis the project dictated procedures for best possible patient care at the expense of research strategies. As an example, some patients completed the stress inventory package while detoxifying. This affected approximately 50% of our sample and could no doubt bias results. The critical issue, however, was timely collection of the data. Alcohol/drug projects are frequently compromised by the fact that a number of patients seek help but leave early, before they contribute to the database. The major benefit of this early testing procedure was that very few patients escaped being evaluated. Second, projects like this neither involve experimental manipulation nor fall into the conventional longitudinal observation paradigm. Rather they are quasiexperimental, where the dependent variable (in this instance, stress level during and one year after service) is based on the patient recalling how he felt some 30-40 years earlier. Fortunately, analysis of the data supports this being an independent judgment in the sense that current discord did not strongly influence perceived past level of stress. Third, the sample is restricted to male combat veterans. The sample is therefore not representative of the total alcohol-dependent population. The stressor also represents only one very limited dimension of major life events. Thus, the generation of hypotheses about causal effects between major life events and maladaptive outcomes is at best highly speculative. Still, results of the study highlight issues that must become part of our clinical perspective in developing both treatment programs and specific patient treatment plans. Fourth, this is a new area of clinical research. The research area is still in a descriptive phase where criteria are still being enumerated. As a priori choices of significant variables are difficult to make, it is easier to choose a large number of variables in the hope that the important ones will surface. Consequently, some blind alleys may be part of the territory.

The most interesting and critical aspects of the results are the significant difference between the two groups (high combat/high stress versus high combat/low stress) in reported current discord, as measured by the total PERI scores. Combat levels and subjects are fairly well matched. The significant difference in felt psychiatric discord is related to the amount of stress reported during the first year after service. If combat levels of 9 and greater and stress scores of 9 and greater are grounds for a diagnosis of posttraumatic stress disorder in this alcoholic sample, then our clinical impressions have been supported. Therefore, programmatic changes of early testing, triage, and treatment are also justified. Consequently, all VA alcohol/drug treatment programs should seriously consider using similar screening devices and developing the strategies required to address posttraumatic stress disorder in substance-dependent veterans. The implications for the Armed Services are quite clear. Primarily, they can no longer disregard posttraumatic stress disorder as an issue when men have experienced combat, especially when there is a family history of chemical dependency. This project suggests that the costs of doing so are both tragic and staggering.

The generalization of these results to any broader population are highly speculative. However, the patient care and cost issues involved demand that they at least be considered. First, we live in a culture that encourages the use of chemicals as an anodyne for daily stresses, strains, and hassles. Major life events both positive and negative are ritually associated with the use or abuse of alcohol. Health-care professionals must be sensitized to the possible correlation between a major catastrophic stressor (i.e., car accident, sudden death, loss of job, sudden illness) and alcohol abuse. This study strongly suggests that patients experiencing a catastrophic stressor and who have a family history of chemical dependency are at greater risk for abusing alcohol to lessen or alleviate PTSD-like symptoms. Second, the combination of alcoholism and untreated chronic posttraumatic stress disorder may result in a debilitating and demoralizing outcome for the patient. These patients are extremely costly to treat in any health-care-delivery system. Early detection of substance abuse and posttraumatic stress disorder comorbidities should be part of a professional strategy. The cost benefits of an on-the-spot assessment for potential alcoholism and posttraumatic stress disorder symptoms subsequent to a major catastrophe (e.g., combat) far outweigh the costly tragedies that can result from disregarding these important issues.

Clearly, as studied in this project, stress is not a sufficient cause for a failure to adapt. Consequently, the major unanswered question is related to the differential effects of combat on our two groups of men. We administered a number of instruments to our alcohol-dependent veterans hoping the testing would provide a lead. So far these instruments and the statistical methods used have not helped to highlight the mediators of the stress-failure to adapt linkage. We only know that under similar combat durations and intensities, some men develop PTSD-like symptoms while others do not. An answer to the question as to what mediates the stress-failure to adapt link is critical. Especially when we consider the ever-present necessity of sending men into combatlike situations. As of this date, we are not aware of research answering this question. Our plans are to reevaluate the data when the sample is expanded using multiple-regression techniques. We also plan to study high-risk patients through an individual-differences approach.


Selections from the book: “Recent Developments in Alcoholism. Volume 6: Posttraumatic Stress Disorder. The Workplace. Consequences in Women. Markers for Risk.” Edited by Marc Galanter. An Official Publication of the American Medical Society on Alcoholism, the Research Society on Alcoholism, and the National Council on Alcoholism. 1986.