Interpersonal problems of alcohol use disorder patients undergoing a physical exercise intervention – a randomised controlled trial

Abstract

The purpose of the present study was to investigate whether the interpersonal relationships of patients suffering from alcohol use disorder (AUD) might be improved as a result of an exercise intervention. Participants were randomised to either treatment as usual (TAU), or TAU in conjunction with participation in one of two intervention arms: exercising individually or supervised group exercise. The Inventory of Interpersonal Problems (IIP-64) was used to assess self-perceived interpersonal problems. The subscales where the AUD patients reported significantly more interpersonal problems were: vindictive, cold, socially inhibited and non-assertive. In all, 116 predominantly male patients participated: an experimental sample of 81, and a control sample of 36. At baseline, on all the IIP subscales, the AUD population achieved higher scores as measured by the IIP than a healthy population. The subscales where the AUD patients reported significantly more interpersonal problems were: vindictive, cold, socially inhibited and non-assertive. The male AUD patients achieved significantly higher scores on the domineering subscale, whereas the female AUD patients achieved significantly higher scores on the non-assertive, exploitable and overly nurturant subscales. On none of the subscales did the intervention group achieve significant changes in IIP scores between pre- and post-treatment. These findings suggest that when working with alcohol patients, it is as important to understand the emotional context of the drinking as it is to target the symptoms of the drinking problems.

Keywords:

• alcohol use disorder (AUD)
• interpersonal problems
• exercise intervention

Introduction

The psychosocial and physical consequences of alcohol use disorder (AUD) are far-reaching (Nutt, King, & Phillips, 2010). Even though AUD can be treated, the costs are high, the outcome modest and there is a high relapse rate during the first year after treatment (Cutler & Fishbain, 2005; Lievens, van der Laenen, & Christiaens, 2014; Miller & Willbourne, 2002).

Consequently, the breakdown of relationships and family problems are significant repercussions of alcohol abuse. For behavioural change to occur, it is important that the individual be capable of forming and maintaining healthy relationships (Riemann, 2008). Therefore, it is relevant to investigate whether relationship patterns might be changed as a result of AUD interventions.

Despite the centrality of interpersonal issues to both theoretical psychological concepts and treatment approaches, relatively few studies have investigated the role and nature of interpersonal problems for alcohol patients (Muller & Clausen, 2015).

Interpersonal problems, measured by the Inventory of Interpersonal Problems (IIP) (Horowitz, Rosenberg, Baer, Ureno, & Villasenor, 1988), have been the focus in the study of other psychiatric disorders. While there are significant individual differences, studies show that people with generalized anxiety disorder mainly experience problems with being too submissive (Salzer et al., 2008), as do bulimia and anorexia patients (Hartmann, Zeeck, & Barrett, 2010). Among patients with depression, people generally experience problems with being too dominant (McFarquhar, Luyten, & Fonagy, 2017). In general, self-perceived interpersonal problems as measured by the IIP appear to be crucial to the understanding of a number of disorders. Patients with substance abuse are also at risk of experiencing significant interpersonal distress, and studies suggest that these patients’ subjective perceptions of their social problems are fundamental to the diagnosis and treatment of alcohol problems. Together with other variables, interpersonal distress can have a significant impact on the individual’s quality of life (Reaney, Martin, & Speight, 2008). Further, such distress may lead to increased alcohol consumption (Lemke, Brennan, Schutte, & Moos, 2007), and play a part in relapse (Leach & Kranzler, 2013; Marlatt & Donovan, 2005). It is a challenging question whether specific types of interpersonal problems are significantly associated with alcohol disorder patients and symptomatic outcome during treatment, and hence worthwhile to investigate whether patients with certain personality types would benefit more (or less) from a group or an individually targeted treatment. Thus, interpersonal distress experienced by the patient should be a focus of treatment (Fjeldstad, Høglend, & Lorentzen, 2015). The current study will examine whether AUD patients do in fact experience more interpersonal problems than the general population, and if so, what types of interpersonal problems are specific to AUD patients.

However, evidence suggests that it would be too simplistic to presume that AUD patients are all alike when it comes to interpersonal problems. Mueller, Degen, Petitjean, Wiesbeck, and Walter (2009) found that gender could account for significant differences in IIP scores in a sample of AUD patients. Thus, this study will also investigate possible gender differences in this sample of AUD patients.

Numerous studies have examined the efficacy of different psychosocial interventions targeting AUD. Matching certain of these interventions to specific groups of patients has not so far improved the effects of these, and in general, different psychotherapies yield the same effect sizes (Imel, Wampold, Miller, & Fleming, 2008). Physical exercise is a new and promising treatment option for mental disorders (Dunn & Jewell, 2010) and for AUD treatment in particular (Brown et al., 2014; Greer et al., 2012; Roessler, Bramsen, Dervisevic, & Bilberg, 2017). The effects of physical exercise on interpersonal aspects of functioning or self-perceived interpersonal distress is a new area of research (Muller & Clausen, 2015; Muller, Skurtveit, & Clausen, 2016). While there are already several studies examining the effects of physical exercise on mental health, they focus on quite different aspects, such as placebo effects (Szabo, 2013), neurobiological effects (Deslandes et al., 2009), or social interactions and capacity for commitment (Roessler, 2011), all of which have potential for improving self-perceived interpersonal problems.

Our study investigates whether the patient’s self-perceived interpersonal difficulties change as a result of engaging in a health intervention based on physical activity as additional treatment. The point of this process is that the physical activity environment is seen as a setting for interpersonal learning, where patients can learn to tackle emotional situations that they had been incapable of handling in the past (Roessler, 2016). For example, the ability to exercise self-regulation, since they are especially vulnerable due to their poor self-control abilities and higher levels of depression (Martinsen, 2008). In order to be helped, the patient needs to undergo emotional experiences that can repair their self-evaluation. A more psychodynamic understanding of addiction identifies the patient’s feelings of helplessness or powerlessness as a key factor behind addiction, as can be shown by the prevalence of non-assertive or self-denying attitudes among AUD patients (Dodes, 2003). The patient displaces these attitudes through e.g. using alcohol to alter their affective state. However, it is not the psychodynamic theory of the origin and emergence of addiction that is the focus of this paper, but the question of whether exercise can have an impact on the self-assessment of interpersonal problems. Regular physical exercise may provide opportunities for patients to challenge beliefs they had previously held about themselves (Roessler, 2011). The experience of exercising and thereby taking control of one’s own physical fitness may reduce underlying feelings of powerlessness and support a sense of being more independent or resilient. Getting more in tune with one’s body and acquiring an increased awareness of physical markers such as pulse or breathing through something as mundane as running or brisk walking may support a sense of independence and self-awareness (Ermalinski, Hanson, Lubin, Thornby, & Nahormek, 1997). Thus, our hypothesis is that through making exercise a part of everyday life, the patient can learn that he or she is indeed neither helpless nor powerless. Changing this underlying perception may feed into other areas of life, and this study specifically investigates whether the individual’s self-perceived social difficulties are affected by the intervention.

The aim of this study is threefold: (a) to compare the self-perceived interpersonal problems of AUD patients with those of a healthy population, (b) to compare the self-perceived interpersonal problems of male and female AUD patients, respectively, and (c) to examine the effects of physical exercise on interpersonal problems in an AUD population.

Our study, providing insights into the possible benefits of exercise on interpersonal problems in an AUD population, is the first of its kind.

Figure 1. Gender differences in self-perceived interpersonal problems.

Methods

This study is comprised of a cross-sectional study and a randomised controlled trial. Participants in the RCT were randomised using urn randomization to either treatment as usual (TAU) or TAU in conjunction with one of two intervention groups: exercising individually or supervised group exercise. TAU consisted of motivational interviewing, cognitive behavioural therapy and family therapy provided by the alcohol treatment centre. The urn randomization was done by a computer using R and initiated at ratios 1:1:1 for the three arms to ensure as equal as possible group sizes during all stages of the sampling process. The randomization was independent from the investigators involved with the participants until a new treatment assignment was requested electronically.

TAU consisted of motivational interviewing, cognitive behavioural therapy and family therapy provided by the alcohol treatment centre. The intervention is described in detail elsewhere (Sari et al., 2013).

Participants and procedures

The sample for the RCT consisted of 116 patients, comprising an experimental sample of 81 patients, and a control sample of 35 patients. Participants were recruited at the alcohol outpatient unit in Odense. A research assistant gave all new patients who met the inclusion criteria written and oral information on the project. They were then offered participation as an add-on to the psychosocial treatment delivered by the treatment facility. All patients were suffering from alcohol use disorder, abuse or dependence, according to the DSM-IV-TR criteria. Inclusion criteria were: age above 18 years, the physical ability to be active and lack of severe psychological problems (see in detail Sari et al., 2013). The study investigated an add-on treatment and participation was voluntary.

The control sample was similar to the experimental sample with regard to age, gender, education and alcohol consumption. Written informed consent was obtained from the participants, and the study was approved by the Regional Scientific Ethical Committee for Southern Denmark (J.nr. S-20130031) and the Danish Data Protection Agency. All procedures in the study were in accordance with the Second Declaration of Helsinki.

If patients chose to participate, they were randomized and, if placed in the exercise intervention, the training would start as soon as possible thereafter. All participants in the RCT received TAU. If the participant was randomized to the solitary training intervention, they were given a session of running instructions by a running instructor and an individual exercise programme. The participants who were randomized to exercise in groups exercised two times per week, together with one or two running instructors.

The cross-sectional study also included a group of healthy controls. The healthy controls were a sample taken from the general population, and were participants in the validation process of the Danish edition of the IIP-64. Data were obtained from the publisher.

Measures

IIP-64: The Inventory of Interpersonal Problems (IIP-64) (Horowitz et al., 1988) was used to assess self-perceived interpersonal problems. The Danish version of the IIP-64 was completed by all participants at baseline testing, after six (T1) and 12 (T2) months. The IIP-64 is a 64-item questionnaire comprised of eight subscales with eight items each. Each subscale describes a particular domain of interpersonal distress, and all are seen as various combinations of the two main domains of dominance and affiliation. The scales are arranged in a circumplex model (Alden, Wiggins, and Pincus, 1990), where dominance and affiliation are orthogonal and negatively correlated, meaning that e.g. high dominance is often accompanied by low affiliation. Dominance is characterized by being too controlling and manipulative and its polar opposite is being non-assertive, i.e. having problems with making one’s own needs known to others. The domain affiliation is characterized at one end by being excessively nurturant, i.e. being too trusting and caring, and trying too hard to please others, while its polar opposite, cold, is characterized by inability to express affection and make long-term commitments.

The other 4 subscales are vindictive, i.e. experiencing problems with distrust and egocentrism, socially inhibited, i.e. experiencing problems approaching others and having feelings of anxiety and embarrassment, exploitable, i.e. experiencing difficulty with feeling and expressing anger, and intrusive, i.e. experiencing problems with being overly self-disclosing and attention-seeking, and having difficulties with being alone (Alden et al., 1990).

Research supports both the construct validity of the IIP (e.g. Gurtman, 1996), test-retest reliability (Horowitz et al., 1988) and internal consistency.

Aims, data analysis and statistics

The primary aim was to compare AUD patients participating in an exercise intervention with a healthy Danish population in respect of their self-reported interpersonal problems.

Secondary aim was to compare the differences between male and female AUD patients, and to measure the proportion of patients that experienced a change in their self-perceived interpersonal problems 26 weeks after treatment start (T1), defined as significant change in the total IIP score.

Descriptive analysis of the participants’ characteristics was conducted using summary statistics. The baseline data are shown in Table 1, where the variables are summarized as mean and standard deviation (SD) for numeric variables, and frequency and percentages for categorical variables.

Table 1. Baseline demographics and clinical characteristics of the study sample (N = 116).

	Intervention N = 81	Control N = 35
Age mean (SD)	43 (11.1)	48,1 (12,1)
Male (%)	71,6	71,4
Education (%)
None	25	23,5
Technical	42,5	41,2
College	32,5	35,3
Employed (%)	57,5	60,6
Family status (%)
Married/cohabiting	27,5	58,8
Divorced/separated	45	17,6
Single/widow/widower	27,5	23,5
Alcohol use^a (IQR)	172 (71.0, 365.0)	195 (46.5, 292.5)

^a Alcohol units consumed in the 30 days prior to treatment entry. A standard unit of alcohol contains 12 grams of pure alcohol.

One-sample t-test was used to compare standardized scales (REF) for AUD patients to the healthy background population, while a linear mixed effect model was used to analyse each of the domains separately in order to evaluate the effect of running on IIP scores and to account for repeated measurements over time taken on each individual. We allowed for effects of age, gender, and interactions between running group and time in order to allow the treatment effect to vary over time for the two groups.

Results

Demographics and alcohol use characteristics

The mean age in the intervention group was 43 (SD 11.1), and 48.1 (SD 12.1) in the control group. The majority of participants in both groups were males: 71.6 and 71.4%, respectively. The level of education for the two groups was similar, as was employment status. A larger proportion of the participants in the intervention group were either divorced or separated; this corresponds to the larger proportion of participants in the control group who were either married or cohabiting. The two groups had comparable alcohol use (see Table 1).

Comparison between interpersonal problems in AUD patients and a healthy population

At baseline, on all the IIP subscales, the AUD population achieved higher scores as measured by the IIP than a healthy population. The AUD patients achieved a significantly higher score on four of the eight subscales. The subscales where the AUD patients were perceived to have significantly more interpersonal problems were: vindictive (p = <0.0001**), cold (p = <0.0001**), socially inhibited (p = <0.0001**) and non-assertive (p = 0.048*) (see Table 2).

Table 2. Baseline differences between an adult (>18) AUD population and a healthy norm population (score 50).

IIP subscale	AUD patients baseline mean IIP scores, compared to healthy population^a	p-value
Domineering	50.59	0.609
Vindictive	56.54	<0.0001**
Cold	57.17	<0.0001**
Socially inhibited	55.32	<0.0001**
Non-assertive	52.52	0.048*
Exploitable	50.59	0.56
Overly nurturant	50.57	0.59
Intrusive	52.24	0.056

^a The healthy population is scored as 50.

Gender differences in self-perceived interpersonal problems

The intervention and control groups consisted of 83 males and 33 females. At baseline the male AUD patients achieved significantly higher scores on the domineering subscale (p = 0.007**), whereas the female AUD patients achieved significantly higher scores on the non-assertive (p = 0.003**), exploitable (p = 0.001**) and overly nurturant (p = 0.037*) subscales (see Figure 1).

Effects of intervention on self-perceived interpersonal problems

On none of the subscales did the intervention group achieve significantly greater change in IIP scores between pre- and post-treatment, after controlling for the two groups’ baseline scores (see Tables 3 and 4).

Table 3. Baseline gender differences between male and female AUD patients.

IIP subscale	Male AUD patients, baseline mean IIP scores (N = 83)	Female AUD patients, baseline mean IIP scores (N = 33)	p-value
Domineering	5.2 (4.4)	3.0 (3.0)	0.007**
Vindictive	6.5 (4.9)	5.1 (4.3)	0.152
Cold	7.5 (6.0)	6.9 (6.3)	0.628
Socially inhibited	9.4 (7.1)	9.9 (7.9)	0.736
Non-assertive	8.9 (5.7)	12.8 (8.1)	0.003**
Exploitable	10.2 (5.3)	13.9 (6.9)	0.001**
Overly nurturant	9.9 (5.6)	12.4 (6.7)	0.037*
Intrusive	7.0 (5.0)	8.0 (5.2)	0.319

Table 4. Difference between exercise (individual + group exercise) and control condition pre- and post-treatment.

IIP subscale	Average difference in IIP score post-intervention^a	Standardized mean difference	p-value
Domineering	0.4742 (95% CI: −0.56; 1.51)	0.17	0.37
Vindictive	0.2984 (95% CI: −0.92; 1.51)	0.099	0.63
Cold	−0.1396 (95% CI: −2.07; 1.79)	−0.046	0.89
Socially inhibited	0.3512 (95% CI: −1.57; 2.27)	0.10	0.72
Non-assertive	1.6969 (95% CI: −0.32; 3.71)	0.39	0.10
Exploitable	0.3327 (95% CI: −1.47; 2.13)	0.084	0.72
Overly nurturant	1.6057 (95% CI: −0.14; 3.35)	0.47	0.075
Intrusive	1.4470 (95% CI: −0.01; 2.90)	0.49	0.056

^a Positive numbers mean that the controls have higher values.

Discussion

The current study revealed significant differences between Danish AUD patients and a healthy Danish population with regard to self-perceived interpersonal problems. AUD patients perceived themselves to be significantly more vindictive, cold, socially inhibited and, to a lesser extent, non-assertive.

Self-perceived problems can be explained as manifestations of underlying feelings of guilt about living a certain lifestyle, with alcohol abuse conceived of as a sign of a generally weak personality (Dörner et al., 2017). This stands in contrast to somatic illness, which is often seen as a matter of fate, and so independent of any particular lifestyle (Kleinman, 1988). Regardless of the connotations surrounding abuse, addiction is often hard to break. One reason might be that benefits accrue from drinking, such as, say, a sense of autonomy. Consequently, when working with alcohol patients, it is as important to understand the emotional context of the drinking as it is to target the symptoms of drinking problems.

Basic human anxiety can account for the use of alcohol to control emotions, and the self-perceived problems of our AUD population resemble those associated with a predominantly depressive personality profile (Giesen, Deimel, & Bloch, 2015). The defining feature of this profile is the anxiety associated with being an individual, with distance from others experienced as abandonment (Riemann, 2008). People with a personality on this part of the spectrum will more frequently sacrifice their own needs, hoping by so doing to achieve and maintain close emotional bonds with other people. Typically, the values they adopt are modesty, unselfishness and a spirit of self-sacrifice. This strategy, however, essentially reflects a fear of autonomy, and might paradoxically result in a situation where the fear of losing relationships is only exacerbated (Riemann, 2008).

A way of countering these feelings could be through alcohol use, which might provide temporary relief from depressive tendencies and strengthen the person’s sense of self. The main self-perceived difficulties in our Danish AUD population were feelings of being cold, vindictive and socially inhibited. These self-perceived problems might reflect the fact that they see themselves as being insufficiently considerate of other people, and that, in their own opinion they should be more self-denying. As the theoretical considerations above would suggest, the interpersonal problems experienced by AUD patients also resemble those of patients suffering from clinical depression (McFarquhar et al., 2017). This is in contrast to e.g. generalized anxiety disorder (Salzer et al., 2008) and some eating disorders (Hartmann et al., 2010), as mentioned above.

The results and theoretical considerations given above indicate possible focal points for future treatment. One clinical implication might be to view physical activity not as a classical treatment, but as an arena, where it is possible for people to express independence. Regular physical exercise can provide an opportunity for the AUD patient to strengthen their sense of self and increase the level of individuation. In Riemann’s (2008) perspective, a confrontation with depressive traits through developing a stronger sense of self will address the underlying causes of abuse and might help the individual towards real maturation.

The implications for psychological research on alcohol abuse more generally include, as in clinical practice, taking the individual’s need for individuation and independence into account. Instead of focusing on possible barriers that might induce people not to turn to a certain desired behaviour (a frequent approach in health sciences), we should perhaps seek to understand people’s individual needs, and open up the possibilities that movement and physical activity can offer.

In this perspective, the practitioner or researcher might consider individual differences regarding interpersonal problems as valuable information, guiding an individualized treatment of alcohol abuse.

Our secondary outcome analysis of the IIP data also revealed significant gender differences in our Danish AUD population. Male AUD patients reported significantly more problems related to the dominance subscale of the IIP, while female patients reported significantly higher levels of interpersonal problems with regard to being non-assertive and overly accommodating. Interestingly, apart from non-assertiveness, there was no overlap with the areas that separated the AUD patients in general from the healthy population. This indicates that particular areas of interpersonal problems: vindictive, cold and socially inhibited are common to AUD patients regardless of gender. Likewise, Mueller and colleagues (2009) compared the interpersonal problems of male and female AUD patients in Switzerland with those of a healthy population, finding that alcohol-dependent men perceived themselves as colder than male controls, while alcohol-dependent women rated themselves as significantly more vindictive, more introverted, more overly accommodating and more intrusive than female controls. They propose a more gender-specific targeted approach, especially for women, to improve the effect of alcohol treatment.

In this study, we set out to examine the effects of a “healthy lifestyle”, here understood as a commitment to regular physical exercise. Our aim was to investigate whether through participation in an intervention consisting of physical exercise two to three times per week, good fitness habits could influence AUD patients’ perceptions of their interpersonal problems. The data indicates that there were no significant effects of the intervention as measured by the IIP. Increasing the sample size and measuring changes on the IIP sooner, and more frequently, than we did in this study might have yielded other results, because on seven out of the eight IIP subscales, the intervention group actually improved their scores more than the control group. But not enough, however, to achieve any significance.

There may be several reasons for this result. Physical exercise, in the form used in this study, may not be suitable for this population; here we have to wait for the final results of the STRIDE (Stimulant Reduction Intervention using Dosed Exercise) study (Greer et al., 2012). The only form of exercise we made available to the group condition was running outdoors, all year round, twice a week at 4 pm In respect of the desire, and the need, for independence and individuation, the fixed and structured trainings that we used in the group condition might not be the best format. We need to remain keenly aware that different groups of patients need different approaches regarding physical activity. This study shows that physical activity as additional treatment for alcohol patients is not a one-size-fits-all prescription. Studies applying more individually targeted approaches and combining different and more flexible forms of physical activity with greater appeal to different gender groups (Muller & Clausen, 2015; Roessler, 2010) would appear to be more promising.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by the Lundbeck Foundation, the Tryg Foundation, and the Region of South Denmark.