Chapter X
A comparison of psychiatric symptomatology in addicts over the course of treatment in an in-patient setting
Index
10.1. Introduction 10.2. Expectations 10.3. Method 10.4. Results 10.4.i. Level of BSI symptom reporting in males and females 10.4.ii. Reduction of BSI scores after 6 weeks of treatment 10.4.iii. Regression of psychiatric symptomatology on Factor Scores (on entry) 10.4.iv. Regression of psychiatric symptomatology on Factor Scores (after 6 weeks) 10.5. Description and discussion of the results10.1. Introduction
The term co-morbidity was first used by Feinstein (1970) in an appeal to researchers to investigate the occurrence of any co-existing disorders that may impinge on psychological treatment. Since then it has become widely recognised that this phenomenon of co-morbidity affects a large number of disorders including addiction, and poses a perplexing challenge for the comprehension of the etiology, treatment and outcome of addictive disorders. With alcoholics for instance, the most consistently documented co-occurring disorders are other substance use, mood, anxiety and schizophrenic disorders (Kessler, 1995). In relation to the extent of co-morbidity a study investigating the prevalence of co-morbidity in 928 in-patient alcoholics found that 62% of the sample fulfilled inclusive diagnostic criteria for one or more additional psychiatric disorders (Penick, Powell, Nickel & Bingham’ 1994). Studies have also shown that the degree of co-morbidity is similar in alcoholism and drug use disorders in clinical and general populations. (e.g. Hesselbrock, Meyer, & Keener, 1986 (clinical) & Lewis, 1984 (general). In addition to this more than 50% of people with mental disorders have been found to abuse drugs including alcohol (Regier et al, 1990). As this knowledge has developed the term “Dual diagnosis” is frequently used in this field and refers more precisely to the presence of a substance use disorder co-existing with another major psychiatric disorder (Sheehan, 1993). We have seen that within an addicted population the pervasiveness of dual diagnosis is becoming increasingly accepted (Reiger et al, 1990) and the realities of poly-drug use are beginning to be viewed as typical rather then atypical (Carroll, Rounsaville & Bryant, 1993). This obviously increases the complexity of an already difficult to treat disorder. This is because it may be difficult to be confident that existing findings in addiction are attributable to the precise disorder of interest and are not an artefact of co-morbid psychopathology, this obviously including cross addiction. This co-morbidity, especially if not taken into account, may reduce researchers’ ability to unambiguously attribute various factors, such as personality characteristics, to specific disorders. Even though co-morbidity in the addictions has been revealed to be a common occurrence, it can be seen from the above that estimates concerning its prevalence do vary. Regier et al, (1990) found that 37% of alcoholics and over 50% of other drug addicts they studied had a co-morbid mental disorder, the most common being anxiety, affective and anti-social personality disorder. Such statistics suggest that, if such co-morbidity remains unidentified and hence untreated, it might offer an explanation for poor outcome study results, especially if such high co-morbidity occurs in more recently proposed addictions (e.g Internet usage). Also of concern is the lack of in-depth knowledge and understanding in this area. For instance, even though it is known that depression and alcohol dependence are two of the most prevalent psychiatric disorders (Kessler, 1995), relatively little is known about this dual diagnosis population (Cornelius et al, 1997). However, recognition of this complex area does seem to be increasing. Indeed Hall and Farrell (1997) have argued that it is imperative for staff in addiction services to be trained to identify other psychiatric disorders in clients.Prevalence of co-morbidity in relation to gender
In line with other clinical areas, the associated diagnoses for men and women are different. For instance, for alcoholic women, anxiety and affective disorders constitute the largest proportion of lifetime co-occurring cases, while the substance disorders, conduct disorder, and anti-social personality disorder account for the majority of co-occurrence among men (Nestadt et al, 1992). These findings are supported by an investigation which found that depression, anxiety and eating disorders are largely seen in females, whereas alcoholism, aggressive and suicidal behaviour predominate in males (Steiner, Lepate & Dunn, 1997). Even though a number of different disorders are highlighted in this type of research there seems to be a degree of consistency in the nature of the disorders found for men and women which relates to the idea presented earlier (chapter 8) that it is possible that women are more likely to be “internalisers” and men are more likely to be “externalisers” (Luthar et al, 1996) For instance, in one study higher rates of dysthymia and eating disorders were observed in females and higher rates of cluster A personality disorders were observed in males (Grilo, Martino, Walker, & Becker, 1997). Another study of individuals with a primary diagnosis of major depressive disorders indicated that women were significantly more likely than men to meet criteria for co-morbid bulimia nervosa and for simple phobia, while men were significantly more likely to meet criteria for lifetime history of alcohol abuse/dependence and other substance abuse dependence (Fava, Abraham, Alpert & Nierenberg, 1996). Other studies have shown that female alcoholics report more symptoms of depression, anxiety, and global psychopathology than males (Benishek et al, 1992), and that depression in treatment samples show an approximately 2:1 female predominance (Paykel, 1991). Some research has been conducted on the relationship between binge eating disorder (BED) and psychiatric symptomatology e.g. depressive and anxiety symptoms (Telch & Agras, 1994, & Wilson, Nonas & Rosenblum, 1993), with the general conclusion being that BED is associated with high levels of psychiatric symptomatology as well as high rates of co-morbid psychiatric disorders. With regard to psychiatric diagnosis in women with binge eating disorder it was found that they had a higher lifetime prevalence rate for major depression, and axis I or II disorders in comparison to controls (Telch & Agras, 1994).Consequences
When investigating such alarming prevalence figures the potential consequences of such complex diagnoses necessitates a little exploration. At the very least co-morbidity must increase the experienced distress for the sufferer and without a doubt poses a challenge to the health professional in terms of treatment. This is because many of these individuals manifest symptoms which are particularly severe, chronic, and refractory in nature (Regier et al, 1990), and they may not respond as well to treatment as other patients, may have a greater likelihood for relapse and therefore readmission rates. This increased complexity of dual diagnosis is illustrated by one study in which cocaine dependent, and cocaine and alcohol dependent in-patients were compared. The patients who were dependent on both cocaine and alcohol showed significantly more symptoms of depression and anxiety and were more likely to have anti-social and “avoidant personality disorders” (Cunningham, Corrigan, Malow & Smason, 1993). In another study investigating the differences in psychological profile in alcohol dependent men who had either attempted suicide or who had not, it was found that a higher percentage of men who had attempted suicide had histories of either illicit drug use, past psychiatric treatment or both (Roy et al, 1991). The profiles of those who had attempted suicide also showed exaggerated tendencies toward sociopathy, anxiety, depression and hostility. It can be seen that the consequences of dual diagnosis here have potentially far-reaching and serious consequences. In quite a different area, current depression has been shown to be a factor associated with continued smoking, as there is evidence to suggest that smokers with elevated levels of depressive symptoms have an increased risk of failing to stop smoking (Dilsaver, Pariser, Churchill & Larson, 1990). Possible explanations for this include the use of smoking to self-medicate to prevent or reduce depressive episodes, or that smoking and depression share a genetic or emotional basis. These findings if extrapolated to other forms of addiction may suggest that if an addict is also suffering from depressive symptoms this may decrease the likelihood of them successfully reaching or sustaining a successful recovery.Relapse and treatment
Clinical studies suggest that people who have a history of co-morbid mental and addictive disorders, when treated for addiction, tend to relapse at a greater rate than those without a history of mental disorder; and it has been indicated that the combination of substance abuse and depression worsens prognosis (Nunes, Deliyannides, Donovan & McGrath 1996). It is possible that if interventions for mental disorders were implemented in conjunction with the treatment of the addiction that this would reduce the relapse rate of these individuals. The practical recognition of co-morbidity may have prognostic and treatment value, although the direction of causality remains a problem. Addiction may precipitate or exacerbate mental disorders, or self medication for mental disorders may result in drug abuse or addiction. Another area of concern is that the forms of co-morbidity may be affected by a number of different factors including gender, the duration of addiction and individual’s sensitivity to the addiction’s effects. Improving the knowledge of co-morbid disorders and how one may impact the other is important because this type of information may help aid prevention, and enable treatment services to be appropriately configured and designed for clinical populations. Often the almost inevitability of co-morbidity is ignored in treatment contexts as there are different treatment systems for the different disorders and quite often separate programmes that focus on addiction. It seems reasonable to assume whatever the causes of these conditions, both co-morbid conditions need to be treated simultaneously and in an integrated way. However, without the integration of knowledge concerning these areas it is unlikely that treatment methodology will evolve to a point where these diverse problems can be effectively tackled at once.Explanations for these co-occurrences
There have been a number of explanations put forward to explain the rate of addictions co-existing with other psychiatric disorders. Mueslar, Drake & Wallach (1998) note four general models which have been proposed.Common factors model
In this model the high rates of co-morbidity are the result of risk factors that are shared across both substance usage and psychiatric disorders (e.g. a common biological causal pathway or genetic predisposition (Luthar, Merikangas, Rounsaville, 1993), or the sharing of certain neurobiological bases with alcohol dependence (Litten & Allen, 1995). Muesler et al (op cit) state that there is overwhelming evidence which suggests that genetic factors do indeed contribute to the development of not only addiction (especially substance use disorders) but other severe mental illnesses as well such as schizophrenia. It is of interest however, that, research which has addressed the possibility of genetic vulnerability to one disorder increasing the risk to another, provides evidence against this model (op cit).Secondary substance use disorder model
This model proposes that the presence of psychiatric disorder increases the patient’s chance of developing an addiction. One avenue of research was that of the self-medication hypothesis, which was discussed in the previous chapter. A further model, which comes under this rubric of secondary substance use, is that of the alleviation of dysphoria. This is a more general model than the self-medication model in that it indicates that people with severe mental illness are prone anyway to dysphoria which may make them more likely to mood alter (Bartels & Drake, 1988). This is because it has been recognised that the dysphorias are particularly heterogeneous, including anxiety, depression, boredom etc. Research which supports this position comes from Muesler et al (op cit) who states that the literature on self reported reasons for drug use supports the idea that many different types of dysphoria motivate initial alcohol and drug use.Super-sensitivity model
According to this model, “Psychobiological vulnerability determined by a combination of genetic and early environmental events, interacts with environmental stress to either precipitate the onset of a psychiatric disorder or to trigger relapse.” (Muesler et al, op cit, p. 723). In addition to the influence of stress on vulnerability, psychotropic medications can decrease vulnerability, whereas substance abuse may increase it. Because vulnerability is defined in terms of a compromised biological sensitivity to stress it may also apply to the effects of alcohol and drugs. This sensitivity may render patients with severe mental illness more likely to experience negative consequences from using relatively small amount of substances. An interesting implication of this model is that it suggests that negative consequences of substance use, rather than use alone, is what differentiates patients with severe mental illness from the general population. The research seems to support the hypothesis that patients with severe mental illness are prone to experience negative consequences from lower amount of substance use than people in the general population, which could explain some of the excess co-morbidity. For example, Drake and Wallach (1993) looked at the longitudinal course of drinking in two samples of patients with severe mental illness and reported that fewer than 5% were able to sustain symptom free drinking over time without negative consequences, in contrast to approximately 50% of the general population who drink alcohol over time without developing a disorder.Secondary psychiatric disorder model
It still can be argued that even with the previous hypothesis there is the possibility of a reverse explanation which places the substance use first and depression or indeed any other mental illness or symptomatology as being precipitated by the substance use, i.e. addiction may increase the likelihood of, or may exacerbate mental disorders. It is possible that psychiatric disorder develops in addicted individuals who would otherwise not develop these disorders. Clinically it is well known that substances of abuse can actually cause symptoms of both depression and anxiety. Though interestingly these substance related syndromes appear to have a different course and prognosis that uncomplicated, independent anxiety and major depressive disorders have.Bi-directional models
A bi-directional position takes the middle ground in effect and states that either disorder can increase the vulnerability to the other disorder's development.Rationale for study
It is clear that the previously mentioned incidence figures painted a bleak picture. The majority of people with an alcohol disorder have at least one other psychiatric disorder, and this co-occurrence is stronger among women. The implications of these facts are far reaching. In the light of the preceding two chapters, it seems likely that exploring psychiatric symptomatology in relation to addictive orientation may help to elucidate further the complexities of dual diagnosis. Besides amplifying our understanding of the factorial structure, addressing addiction broadly in its association with psychiatric symptomatology, personality variables and the likelihood of the use of other addictive areas creates a fuller picture. This may help not only in overall comprehension but may also prompt a fresh look at appropriate treatment. So, in the next study psychiatric symptomatology is considered in relation to the four factor addictive orientation.10.2. Expectations
10.3. Method
Following admission to PROMIS, in conjunction with the other questionnaires, the Brief Symptom Inventory is completed by all patients. The questionnaire is then administered again six weeks into treatment. This provides two sets of Brief Symptom Iinventory scores: time one (entry into treatment) and time two (six weeks into treatment).Participants
There were 190 participants in total for whom complete data were available, 93 men and 97 women. This is a substantial reduction in the number of participants in comparison to the numbers used in the previous factor analytic studies. This occurred for a variety of reasons, such as patients at the Centre leaving with staff approval before second testing time, or leaving against staff approval before being re-tested, or otherwise leaving at a time that precluded re-testing, with or without intent.Brief Symptom Inventory
The Brief Symptom Inventory (BSI) (Derogatis, 1993) is a 53 item self report symptom inventory designed to reflect the psychological symptom status of psychiatric patients, medical patients and individuals in the community who are not currently patients. It is a shorter version of a longer test developed by Derogatis designed for purposes of diagnosis of psychiatric patients. The BSI contains nine primary symptom sub-scales (described below). The items are listed in Appendix 5. It may be used as a single point in time assessment of an individual's clinical status or it may be used sequentially either to document trends overtime or in pre/post evaluations. General instructions for the BSI are very straightforward and can be found in Appendix 4. Although the Brief Symptom Inventory does not claim to be a diagnostic instrument as such, it was thought that if differences were found between the different Addiction Factors in terms of decrease in Brief Symptom Inventory sub-scale scores, that this would serve to further elucidate the significance of addictive orientations in psychiatric distress. This further exploration utilising these Brief Symptom Inventory variables may in turn have some implications for the comprehension of co-morbidity. This is for a number of reasons as we have already seen e.g. relapse and treatment.Brief descriptions of the nine symptoms assessed in the Brief Symptom Inventory
1. SomatisationReflects distress arising from perceptions of bodily dysfunction.2. Obsessive compulsive
This dimension focuses on thoughts, impulses and actions that are experienced as unremitting and irresistible by the individual, but are of an unwanted nature.3. Interpersonal sensitivity
This centres on feeling of personal inadequacy, inferiority and self-doubt.4. Depression
The symptoms of the depression dimension reflect a representative range of the indicators of clinical depression.5. Anxiety
General signs such as nervousness and tension are included in this dimension, e.g. feelings of apprehension, panic attacks and feelings of terror6. Hostility
This dimension includes thoughts feelings, or actions that are characteristic of the negative affect state of anger.7. Phobic anxiety
Phobic anxiety is defined as a persistence fear response to a specific person, place, object or situation that is irrational disproportionate to the stimulus and leads to avoidance or escape behaviour.8. Paranoid ideation
This dimension represents paranoid behaviour fundamentally as a disordered mode of thinking. The cardinal characteristics of projective thought, hostility, suspiciousness, grandiosity, centrality fear of loss of autonomy and delusions are viewed as primarily aspects of this disorder.9. Psychoticism
This scale was developed to represent the construct of psychoticism as a continuous dimension of human experience. Items indicative of a withdrawn, isolated schizoid lifestyle were included, as were first-rank symptoms of schizophrenia, such as thought control. This scale provided for a graduated continuum from mild interpersonal alienation to dramatic psychosis.
10.4. Results
10.4.i. Level of BSI symptom reporting in males and females
Hypothesis 1
It was anticipated that there would be a high level of symptom reporting by men and women upon entering treatment, and it was predicted that patients on average would score within a clinical range on the Brief Symptom Inventory. Table 10.1. contains the average raw scores for males and females from the BSI. To make the different symptom dimensions comparable these scores are converted to standardised T scores (see bottom of figures 10.1 and 10.2). Standardised scales enable comparisons of the performance of an individual group (here addicts) with that of some relevant reference group. The T score has a mean of 50 and a standard deviation of 10. There are four norm groups for the BSI: adult psychiatric outpatients, adult non-patients psychiatric inpatients and adolescent non-patient and each group has separate norms for males and females. The norm group selected should be the one that best represents the individual or group being examined, so in this case the norm group selected was psychiatric in-patient. Looking at figures 10.1 (male scores) and 10.2 (female scores), it can be seen that in general the scores indicate that they are in a clinical range in that the scores are within one standard deviation of the mean of 50. Though in fact as the scores in most instances are higher than the mean of 50 that this indicates that this population has a particularly high level of psychiatric symptomatology. For the male sample (figure 10.1) it can be seen that Psychoticism and Hostility are the highest scores on intake. For the female sample (figure 10.2.) it can be seen that Psychoticism and Anxiety are the highest scores on intake.
Table 10.1: Male and female BSI mean scores and standard errors at time of intake and after 6 weeks of treatment.
| Female Intake | S.E. | Female 6 weeks | S.E. | Male Intake | S.E. | Male 6 weeks | S.E. | |
| Somatisation | 1.37 | 1.04 | 0.57 | 0.71 | 0.96 | 0.84 | 0.48 | 0.65 |
| Obsessive compulsive | 2.26 | 1.06 | 1.28 | 0.90 | 1.63 | 0.98 | 1.15 | 0.86 |
| Interpersonal sensitivity | 2.45 | 1.04 | 1.49 | 0.94 | 1.53 | 1.13 | 1.07 | 0.95 |
| Depression | 2.53 | 1.02 | 1.28 | 0.98 | 1.86 | 1.03 | 1.01 | 1.03 |
| Anxiety | 2.56 | 0.97 | 1.35 | 0.96 | 1.73 | 1.04 | 1.16 | 0.87 |
| Hostility | 1.03 | 1.01 | 0.88 | 0.82 | 1.02 | 1.03 | 0.79 | 0.77 |
| Phobic anxiety | 1.34 | 1.20 | 0.65 | 0.79 | 0.71 | 0.95 | 0.56 | 0.85 |
| Paranoid ideation | 1.64 | 0.98 | 1.01 | 0.80 | 1.36 | 0.89 | 0.83 | 0.76 |
| Psychoticism | 2.09 | 1.02 | 1.13 | 0.80 | 1.60 | 0.93 | 0.86 | 0.89 |
10.4.ii. Reduction of BSI scores after 6 weeks of treatment
Hypothesis 2
It was further predicted that there would be a reduction of symptomatology scores over a 6-week period in treatment. No predictions were made regarding differences for males and females. To investigate this a t-test was performed. This test checks to see if there are differences between the means of scores from time one and time two. The null hypothesis that is made is that the mean of population A doesn’t differ from the mean of population B. This test is two tailed test because both sides of the distribution are being looked at.Paired t-tests
The t-test can only be conducted as such if the 2 samples are independent. In this case this is not true as the same subjects are being tested before and after a certain period of time. In this situation a new sample is formed given by the difference between the scores of each personality at the admission and after 6 weeks and a one-sample t-test is performed on this newly formed sample. The null hypothesis in this case is that the mean of the difference will be zero against the alternative hypothesis that there is a decrease both at a 5% and 1% level. The results are now presented in tabular form.Table 10.2. Complete sample paired t-test for the addicts difference in BSI scores (Time 1 – Time 2)
| Variable | N | Mean | Std Err | T | Prob >[T] | 95% CI | |
| Somatisation | 190 | 0.64 | 0.07 | 9.57 | 0.001 | 0.51 | 0.77 |
| Obsessive compulsive | 190 | 0.74 | 0.07 | 10.11 | 0.001 | 0.60 | 0.88 |
| Interpersonal sensitivity | 190 | 0.71 | 0.09 | 8.39 | 0.001 | 0.55 | 0.88 |
| Depression | 190 | 1.06 | 0.094 | 11.25 | 0.001 | 0.87 | 1.24 |
| Anxiety | 190 | 0.90 | 0.09 | 10.59 | 0.001 | 0.73 | 1.06 |
| Hostility | 190 | 0.31 | 0.07 | 4.44 | 0.001 | 0.17 | 0.44 |
| Phobia | 190 | 0.43 | 0.08 | 5.48 | 0.001 | 0.27 | 0.58 |
| Paranoid ideation | 190 | 0.58 | 0.07 | 8.72 | 0.001 | 0.45 | 0.71 |
| Psychoticism | 190 | 0.86 | 0.08 | 11.08 | 0.001 | 0.71 | 1.02 |
Table 10.3. Males paired t-test for males difference in BSI scores (Time 1 – Time 2)
| Variable | N | Mean | Std Err | T | Prob >[T] | 95% CI | |
| Somatisation | 93 | 0.48 | 0.10 | 4.90 | 0.001 | 0.29 | 0.67 |
| Obsessive compulsive | 93 | 0.49 | 0.10 | 5.06 | 0.001 | 0.30 | 0.68 |
| Interpersonal sensitivity | 93 | 0.46 | 0.12 | 3.85 | 0.001 | 0.22 | 0.69 |
| Depression | 93 | 0.85 | 0.13 | 6.41 | 0.001 | 0.59 | 1.11 |
| Anxiety | 93 | 0.57 | 0.15 | 4.49 | 0.001 | 0.32 | 0.81 |
| Hostility | 93 | 0.23 | 0.10 | 2.23 | 0.03 | 0.03 | 0.43 |
| Phobia | 93 | 0.14 | 0.09 | 1.69 | 0.09 (ns) | -0.02 | 0.31 |
| Paranoid ideation | 93 | 0.53 | 0.09 | 5.61 | 0.001 | 0.34 | 0.71 |
| Psychoticism | 93 | 0.75 | 0.10 | 7.44 | 0.001 | 0.55 | 0.95 |
Table 10.4. Females Paired t-test for females difference in BSI scores (Time 1 – Time 2)
| Variable | N | Mean | Std Err | T | Prob >[T] | 95% CI | |
| Somatisation | 97 | 0.80 | 0.09 | 8.89 | 0.001 | 0.62 | 0.97 |
| Obsessive compulsive | 97 | 0.98 | 0.10 | 9.42 | 0.001 | 0.77 | 1.18 |
| Interpersonal sensitivity | 97 | 0.96 | 0.12 | 8.22 | 0.001 | 0.73 | 1.18 |
| Depression | 97 | 1.25 | 0.13 | 9.62 | 0.001 | 0.10 | 1.51 |
| Anxiety | 97 | 1.22 | 0.11 | 11.60 | 0.001 | 1.02 | 1.42 |
| Hostility | 97 | 0.38 | 0.09 | 4.09 | 0.001 | 0.20 | 0.57 |
| Phobic anxiety | 97 | 0.70 | 0.12 | 5.67 | 0.001 | 0.46 | 0.94 |
| Paranoid ideation | 97 | 0.63 | 0.09 | 6.69 | 0.001 | 0.44 | 0.81 |
| Psychoticism | 97 | 0.96 | 0.12 | 8.26 | 0.001 | 0.73 | 1.19 |
10.4.iii. Regression of psychiatric symptomatology on Factor Scores (on entry)
Hypothesis 3
It was expected that factors scores would be differentially involved in the prediction of psychiatric symptomatology. Initially the question to be answered was whether the initial BSI scores could be “explained” by the factor scores, and if so, whether there was a statistical model for this relationship, and what was its level of predictive power. The simplest way of determining whether or not the factor scores explain a significant amount of the variation is to perform a linear regression. For this regression the factor scores are regressed against all BSI scores, and the factor variables that explain a significant amount of the reduction are kept in the model and those that don’t, are removed. The most straightforward process for determining which variables to include is to use Efroymson stepwise variable selection technique, which uses a sequence of F-tests to determine which variables to include. It is worth noting that by design the factors are orthogonal, and so are independent of each other, therefore there is no information loss by considering each BSI measure separately. Clearly there are different factor models for the female and male addicts, and so the analysis is done separately for each. Regression analyses were computed to examine whether the factor scores were able to predict the initial Brief Symptom Inventory scores. As the results of this analysis produces a large number of models, the complete table of regression analyses are presented fully in Appendix 6. Contained in Table 10.5 is a summary of the findings.Table 10.5. Summary table: Regression analyses of factor scores predicting initial psychiatric symptomatology scores in males and females
| Male sample: analysis based on scores on intake | Female sample: analysis based on scores on intake | |
| BSI dimensions | Factors which predict BSI scores | Factors which predict BSI scores |
| Somatisation | Drug factor (4) Relationships (3) Food (1) 21%*** |
Drug (1) Food (3) 26%*** |
| Obsessive- compulsive | Food (1) Drug (4) Relationships (3) 19%*** |
Drug (1) Food (3) Relationship (4) Compulsive helping (2) 24%*** |
| Interpersonal-sensitivity | Relationships (3) Food (1) 20%*** |
Food (3) Relationships (4) Drugs (1) 24%*** |
| Depression | Relationship (3) Food (1) Drug (4) 15%*** |
Food (3) Drug (1) Relationships (4) Compulsive helping (2) 25%*** |
| Anxiety | Drug (4) Relationships (3) Food (1) 16%*** |
Drug (1) Food (3) Relationships (4) Compulsive helping (2) 28%*** |
| Hostility | Drug (4) Relationships (3) Food (1) 29%*** |
Relationships (4) 1 drug Food (3) 23%*** |
| Phobic anxiety | Food (1) Drugs (4) Relationships (3) 17%*** |
Food (3) Drug (1) Relationships (4) Compulsive helping (2) 26%*** |
| Paranoid ideation | Relationships (3) Food (1) Drugs (4) 31%*** |
Relationships (4) Compulsive helping (2) Drug (1) Food (3) 24%*** |
| Psychoticism | Relationships (3) Food (1) Drugs (4) 22%*** |
Food (3) Relationships (4) Drug (1) Compulsive helping (2) 28%*** |
| Global Severity Index | Relationships (3) Drugs (4) Food (1) 17%*** |
Drug (1) Food (3) Relationships (4) Compulsive helping (2) 35%*** |
Predicting psychiatric symptomatology scores
Two regression analyses were computed to examine the ability of factor scores to predict psychiatric symptomatology (the BSI scores). From the regression analysis it can be seen that a substantial portion of the variance is explained by the factor scores for all of the BSI dimensions in both males and females, and in all cases the significance of the models is at the 1% level. In all cases a number of the factors in different combinations contribute to the overall R2. As there are a large number of models only the most salient for our discussion will be reported in detail. Looking at the leading predictors for the various symptoms, for men the most relevant factor is the relationship factor (4 out of the 9 symptoms) and for women it is the food factor (4 out of the 9 symptoms). A further notable difference is the difference in the amount of variance explained, with the female sample tending to have more variance explained than the males. For instance with the Depression dimension the male model though significant (F (3) = 14.59, p < 0.001) explained 15% of the variance compared to 25 % of the variance in the female model (F (4) = 19.26, p < 0.001). A similar pattern was observed for the anxiety dimension with 16% of the variance being explained by the male model (F (3) = 16.42, p < 0.001) and 28% of the variance being explained by the female model (F (4) = 22.75, p < 0.001). There are however two notable exceptions these being Hostility and Paranoid ideation. In these cases it is the male sample which has the greater amount of variance explained for Hostility (F (3) = 33.94, p <0.001) 29% of the variance was explained compared to 23% of the variance in the female model (F (3) = 22.18, p< 0.001). For paranoid ideation in the male sample 31% of the variance was explained (F (3) = 37.91, p< 0.001) in comparison to 24% being explained in the female model (F (4) = 17.45, p< 0.001). The greatest amount of variance explained was 35% and this was for the Global severity index in the female sample (F(4) = 31.15, p< 0.001) with the leading predictor being Sensation seeking hedonism (the drug factor) (ß = 0.32, p< 0.001). This compares to only 17% of the variance being explained for the Global severity measure in the male sample (F (3) = 33.3, p< 0.001).10.4.iv. Regression of psychiatric symptomatology on Factor Scores (after 6 weeks)
Hypothesis 3b
It was also predicted that there would be systematic differences between males and females in the relationships found, given that males and females with addictive problems have been found to have different patterns of associated symptomatology. The next stage of the analysis is concerned with determining whether the amount of decrease in BSI scores after undergoing 6 weeks of treatment can be explained by the factor scores. In the same way as with the previous analysis a linear regression was performed for male and female addicts. In the same way as with the previous analysis, as the results produce a large number of models, the complete table of regression analyses are presented fully in Appendix 6. Table 10.6 provides a summary of the findings.| Male sample: analysis based on scores after 6 weeks of treatment | Female sample: analysis based on scores after 6 weeks of treatment | |
| BSI dimensions | Factors which predict BSI scores | Factors which predict BSI scores |
| Somatisation | Drug (4) 15%*** |
Food (3) 6%* |
| Obsessive- compulsive | Relationships (3) 5%*** |
Food (3) Compulsive helping (2) Relationships (4) 28% *** |
| Interpersonal-sensitivity | Relationships (3) 8%* |
|
| Depression | Food (3) 5%* |
|
| Anxiety | Compulsive helping (2) Drug (4) 9%** |
|
| Hostility | Food (1) Drug (4) 16%*** |
Relationships (4) 4%* |
| Phobic anxiety | Food (1) 8%** |
Food (3) 7%** |
| Paranoid ideation | Food (1) Drug (4) 9%** |
Relationships (4) 5%** |
| Psychoticism | Relationships (3) 5%** |
|
| Global Severity Index | Food (1) Relationships (3) Drugs (4) 17%*** |
Food (3) 6%* |
10.5. Description and discussion of the results
1. High level of symptom reporting within clinical range
As predicted, the clinical graphs (figures 10.1 and 10.2.) indicate that in general on entry to treatment both males and females score within the clinical range of the BSI. As the scores in most instances are higher than the mean of 50 that this indicates that this population has a particularly high level of psychiatric symptomatology. This would be expected, as the literature suggests a complex array of potential other problems which concur with addiction (e.g. Kessler, 1995). In addition to this it could be anticipated that on entry to treatment that levels of depression, anxiety etc. may be elevated due to a number of different reasons, such as the added dimension of being in a treatment centre perhaps causing an initial increase in distress. On entry to the treatment centre women in general tend to be particularly high on Somatisation, Obsessive compulsive, Interpersonal sensitivity, Anxiety and Psychoticism. Men on entry tend to score most highly on Interpersonal sensitivity, Hostility and Psychoticism.2. Reduction in scores over six weeks
For the complete sample it was found that all of the BSI measures had decreased over the six-week period beyond the 0.1% significance level. This is a really positive result as it indicates that according to the variables measured the individuals were experiencing significantly fewer troublesome symptoms and behaviours six weeks into treatment. It must be highlighted that the treatment that they receive though specifically aimed at treating addiction on the face of it clearly has a highly positive impact, as it effectively reduces symptomatology.Differences in change between males and females
The most important difference was the level of symptom reporting on intake which was higher in females than in males. After six weeks of treatment interestingly the female’s scores (see figures 10.1 and 10.2) were found to have reduced to a level which is lower than the males 6 week scores. This may indicate that the treatment is particularly helpful and successful for women. Further differences that were found were located in the male sample with the Hostility and Phobic anxiety dimensions. Hostility, though still decreasing significantly, was found to be significant at 5% rather than the .1% level as found with the other dimensions. Phobic anxiety on the other hand though having decreased failed to reach the 5% significance level. For the females the evidence of a decrease in their BSI scores is strong for all of the dimensions at a 0.1% significance level.Regression analyses
The next stage of the analysis aimed to investigate whether the initial BSI could be explained or predicted by factor scores. It was also predicted that there would be differences between males and females. It was found that the results for both males and females indicated that it was possible to identify factors that explained the BSI scores on intake, and for all of the BSI symptom dimensions the level of significance for each regression model was high (p> .001).On intake
The Global severity index, which is a global measure of symptom severity using all of the symptom dimensions, indicates that the key difference on intake is that 35% of the variance is explained for females (F (4) = 31.15, p< 0.001) which is over double for that of males (17%) (F (3) = 33.3, p< 0.001). This may indicate that for women, addiction is more associated with general psychiatric distress than it is for men. This finding seems to be in line with previous findings. For instance, it was mentioned in the introduction that other studies have shown that female alcoholics report more symptoms of depression, anxiety, and global psychopathology than male alcoholics (Benishek et al 1992). In addition to this it can be seen that the Compulsive helping factor (Other orientated nurturance) is involved in the explanation of a large proportion of women’s psychological distress on intake. It appears in 6 out of the 9 primary symptom categories whereas it doesn’t appear as a predictor in any of the symptom categories for men. Therefore it seem to be in some way relevant for women’s distress and not for men’s. It has been mentioned previously that co-dependency (compulsive helping) has been found to be related to self-deprecating personality characteristics (Wells, Glickauf-Hughes & Bruss, 1998), and it has also been mentioned that it is associated with neurotic symptoms Cermark (1986). Perhaps what occurs is that with the general tendency to focus on other’s needs, rather than their own, the development of psychiatric symptomatology may be heightened, as the time devoted to solving personal problems is curtailed by constant concern for others. This in turn may be related to Heimer’s (1996) concept of women predominately internalising their problems. This may occur as a consequence of the expectation that females will look to others in a care-taking capacity, thus leaving little room for an expressive outlet for their distress.Leading predictors
Looking at the leading predictors for the various symptoms, for men the most relevant factor is the Relationship factor, “Power related hedonism” (4 out of the 9 symptoms) and for females it is the Food factor “Self orientated nurturance” (4 out of the 9 symptoms). For instance, in males the dimensions “Interpersonal sensitivity” and “Depression” are best explained by the “Power related hedonism” factor, whilst in females on the other hand it is the “Self orientated nurturance” factor which explains most of the variance in these sub-scales. With respect to the association of “Power related hedonism” with the male symptoms, in the introduction it was noted that for male alcoholics, conduct disorder, and anti-social personality disorder account for a large proportion of co-occurrence among men (Nestadt et al, 1992). It could be argued that conduct disorder, and antisocial personality disorder has at its centre difficulty in relating to and understanding others. Holding this in mind, it is clear that the “Power related hedonism” factor (relationships) is concerned with the exploitative use of relationships for the individual’s own gain and indicates a lack of empathy regarding others involved in the person's life. Perhaps “Power related hedonism” is figural for male addicts in particular, as there is a propensity for them to find it difficult to relate and understand others. Previous research in the field of eating disorders ties in with the Food factor “Self orientated nurturance” being the leading predictors for the various symptoms. For instance research has been conducted looking at the relationship between binge eating disorder and psychiatric symptomatology (Telch & Agras, 1994 & Wilson, Nonas & Rosenblum, 1993), the general conclusion being that binge eating disorder is associated with high levels of psychiatric symptomatology as well as high rates of co-morbid psychiatric disorders. In another study it was found that in conjunction to the high rate of psychiatric symptomatology in women with eating disorders that for anorexia and bulimia nervosa, psychiatric symptoms were significantly more serious in purging than in non-purging groups (Favaro & Santonastaso, 1996). So these studies, which indicate a high level of psychiatric distress in those with eating disorders, seem to help to explain the role of the food factor in this study, as it suggests that for women, eating disorders and this general constellation of behaviours is related to general psychiatric distress. Interestingly for both men and women the drug factor was found to be the leading predictor for Somatisation. This may be to do with the fact that even though patients at PROMIS are medicated where appropriate to minimise withdrawal discomfort, there may still be a number of somatic correlates such as numbness and tingling in the body which may explain this finding.After six weeks
The results for both males and females show that it is not always possible at this level of significance to identify any factor that would explain the decrease of the Brief Symptom Inventory scores. Moreover, even when it is possible to include the factors in the regression models the variability explained is not very high. The percentage of variance explained ranges from 5% explained for the male “Obsessive compulsive”, male “Psychoticism” and female “Depression” dimensions through to 28% of the variability explained for the female Obsessive compulsive dimension. In the male sample the dimensions which had the most variance explained were “Somatisation” 15% and “Hostility” 16%. In the female sample the dimension which had the most variance explained was the “Obsessive compulsive” dimension (28%). Looking at the Global severity index, it can be seen that one key difference between the genders is that 17% of the variance in the decrease in scores is explained for males. This is nearly three times that for females where only 6% of the variance is explained. What is also noteworthy is the fact that for the male sample the food factor is the leading predictor for the decrease in the global severity index and for females it is the only predictor. This may indicate that PROMIS is particularly good at treating people with eating disorders. In conjunction to this, for men, type of addiction is more important in determining reduction in psychiatric symptoms, even though after the six-week period women seem to be scoring lower on most of the dimensions in comparison to men. This on the other hand could indicate that it is something about the treatment at PROMIS which influences such a sharp decline in the females’ scores. It has already been posited as a possibility that females in some way may be at a disadvantage relative to males in terms of overall severity of psychiatric distress. The fact that a significant difference was found between the scores on intake and the scores 6 weeks later for all of the female scores, and that when looking at the involvement of the factors only 6% of the variance for a global measure of severity could be explained, indicates that the “decrease” has something to do with aspects outside that of the particular characteristics of their addictions. Given the evidence it may be the case that it has something to do with being female. If this is the case then given the holistic and relational approach in the treatment setting, then possibly the decrease in scores reflects the suitability of this treatment method for females irrespective of their addiction. In relation to this possibility, in the introduction one of the main areas which was consistently mentioned as a co-morbid factor, was depression (e.g. Roy et al 1991). It is now found that for females, not only is the food factor the leading explanatory factor for the initial score, but it is the only factor which explains part of the variance of the decrease in depressive symptoms six weeks later. For the males, no factor was found to explain the decrease in depression.Criticisms and future research
An inherent problem with this type of research is that when attempting to investigate the possible importance of co-morbid factors it is impossible to know even in terms of symptomatology whether it is overall psychiatric distress which causes the addiction or whether the addiction causes the psychiatric distress. Indeed, another alternative is that they occur together. However, the global finding of male’s psychiatric symptomatology being in some way related to the relationship factor and female’s symptomatology being related to the food factor warrants further attention. Primarily, in the light of previous findings, differences have been revealed in relation to the different types of disorders that tend to dominate in men and women. Here the results from this study have illustrated that on the level of psychiatric symptomatology, differences between men and women have also been found in terms of the amount of variance which can be explained for the various dimensions of psychiatric symptomatology. On the basis of these findings future research may include accurately assessing addict’s co-morbid diagnosis in relation to their addictive orientation, and then investigating whether gender has a role to play in any differences found. This may then further elucidate the potential significance of the idea of women being predominately “Internalisers” and males “Externalisers” of problems (Heimer, 1996). This suggestion may be seen as potentially increasing the complexity of an individual’s diagnosis, and hence treatment, but without a recognition of the fact that individuals often present with a complex array of problems, addictive and otherwise, it is going to be difficult to first understand why a myriad of problems frequently co-occur, and second how to effectively treat such individuals. The fact that some psychological disorders may develop as a consequence of the pharmacological effects of substance abuse, and that forms of co-morbidity may be influenced by the class of drug, the duration of drug use and individual sensitivity to the drug effects of abuse (Mirin, Meyer, & McNamee, 1976), should not be ignored. However, if there is a degree of systematic co-occurrence, further comprehension as to why this is the case would be important as it may shed light on causative factors and hence could enable prevention and treatment services to be appropriately configured. On the face of it, treating addiction has been shown to effectively reduce symptomatology, especially in women, and this finding may be able to shed some light on the validity of the models of co-morbidity mentioned in the introduction (Mueslar, Drake & Wallach, 1998). Even though the purpose of the study was not to evaluate these models it is interesting to note that treatment for addiction does indeed seem to have an impact on the reduction of symptomatology dimensions. In this way it may be the case that the Common factors model would be particularly interesting to concentrate on, as in this model the co-morbidity is thought to be the result of risk factors that are shared across both addiction and psychiatric disorders. In addition we can say that the global finding of males’ psychiatric symptomatology being in some way related to the relationship factor, and females' symptomatology being related to the food factor, indicates that the manifestation of particular clusters of problems is embedded in some facet of gender. A fuller picture may be revealed of the origins of disorders, if psychiatric symptomatology was also investigated in a normal population. Further work conducted in these areas may reveal why different patterns of pathology appear, and perhaps how these behaviours reflect potential differences in the underlying psychology of addiction in men and women. With this may come a more informed and perhaps more holistic and socially relevant way of approaching these complex clients.