The Shorter PROMIS Questionnaire

Abstract

There is both a theoretical and clinical need to develop a questionnaire that assesses a range of addictive behaviours. The Shorter PROMIS Questionnaire (SPQ) is a 16-scale self-report instrument assessing the use of nicotine, recreational drugs, prescription drugs, gambling, sex, caffeine, food bingeing, food starving, exercise, shopping, work, relationships dominant and submissive, and compulsive helping dominant and submissive. Clinical cut-off scores using the 90th percentile were derived from a normative group of 508 individuals. These cut-offs correctly identified 78–100% of cases within clinical criterion groups of specific disorders. The clinical sample also completed other validated scales assessing gambling, eating, alcohol, and drug use. Correlations were typically .7 with relevant SPQ scales. The SPQ food, drug, and alcohol scales were at least equivalent to validated comparison scales in the strength of their relationship to relevant clinical criterion groups. Internal consistency was high for all scales, and test-retest reliability was generally good. This clinically useful instrument provides a broad assessment of addictive problems, thereby benefiting both the treatment provider and the client.

Introduction

Functional similarities between substance-related and other addictive behaviours have been observed at the biological, psychological, and social levels of analysis Donovan, 1988, Orford, 1985. It is also the case that different addictions tend to co-occur in predictable ways (Stephenson, Maggi, Lefever, & Morojele, 1995). Moreover, as McKay and McLellan's (1998) recent review would suggest, there is a clinical, not to say financial, imperative to deal simultaneously with the related difficulties of “polyproblem individuals.” These various considerations provide a compelling incentive to develop an instrument that assesses a broad range of excessive behaviours that could be deemed as being addictive.

It has been observed that behaviours not involving the use of psychoactive substances can still produce physiological arousal. The heart rates of pathological gamblers increase during a gambling session Coventry & Constable, 1999, Coventry & Norman, 1998, and altered dopaminergic and serotonergic functions have also been observed Bergh et al., 1997, De-Caria et al., 1998. Similarly, cortical arousal is a discriminating factor in the identification of individuals with a dependency on exercise (Beh, Mathers, & Holden, 1996). Other processes usually associated with addictive psychoactive substance use have been observed in excessive use of sex (Roth, 1992), work, and gambling (Orford, 1985); food (Cummings, Gordon, & Marlatt, 1980); tobacco and caffeine (Aubin, Laureaux, Tilikete, & Barrucand, 1999); shopping (Christensen, Farber, & DeZwaan, 1994); exercise (Furst & Germone, 1993); and the playing of video games (Phillips, Rolls, Rouse, & Griffiths, 1995).

Other evidence suggests that addictive behaviours consistently covary. It is common for a high proportion of alcoholics in treatment to report previous drug use (Sokolow, Welte, Hynes, & Lyons, 1981). Conversely, alcohol use is considered a complicating factor in the treatment of drug users (Miller, Belkin, & Gold, 1990). Wiederman and Pryor (1996) reported that a third of adolescents with bulimia also drank alcohol, used nicotine, and smoked marijuana. A study of high school students classified as problem drinkers found that 35% also had eating problems (Peluso, Ricciardelli, & Williams, 1999). There have been similar rates of comorbidity reported in inpatient populations. For example, approximately 40% of individuals with a diagnosis of anorexia or bulimia also abused drugs and alcohol (Zerbe, Marsh, & Coyne, 1990). In addition, 22% of those with an eating disorder reported using cocaine in an attempt to alleviate their symptoms (Gold, Gold, Sweeney, & Potash, 1987).

High levels of comorbidity have been found in a survey of gamblers in treatment, with 47% also abusing drugs and alcohol (Ramirez, McCormick, Russo, & Taber, 1984). Conversely, 14% of substance abusers in treatment also met the criteria for pathological gambling, with a further 14% identified as experiencing problems with gambling behaviour (Lesieur & Heineman, 1988). These relationships also exist in nonclinical populations. Griffiths and Sutherland (1998) discovered that among adolescents, self-reported gamblers were also significantly more likely than nongamblers to report that they drank alcohol, took drugs, and smoked tobacco. In a study investigating ‘overlapping’ addictions, Greenberg, Lewis, and Dodd (1999) found high correlations between college students' use of alcohol, nicotine, caffeine, chocolate, exercise, and gambling; with a clear tendency to become addicted to more than one substance or activity.

The above evidence suggests that treatments targeting single behaviours may not be as effective as those that are wider in scope. For example, cocaine addicts commonly return to using cocaine through drinking alcohol (Smith, 1986). Donovan (1988) suggested that potentially addictive behaviours are interdependent with a decrease in the target behaviour accompanied by an increase in an associated behaviour. For example, in a study investigating nicotine and caffeine use in alcohol-dependent individuals in treatment, it was found that both caffeine and nicotine intake significantly increased following abstinence from alcohol (Aubin et al., 1999). According to Donovan, associated behaviours may cause abstinent individuals to return to their target behaviour, or develop a new addiction to the associated behaviour, or develop two individual but related addictions.

To address the above issues, the original PROMIS Addiction Questionnaire (Lefever, 1988) was developed to assess addictive behaviours in patients admitted to a residential facility for the treatment of alcohol and drug dependence, eating disorders, and other forms of addiction. The questionnaire consisted of 16 scales, each containing 30 items contributing a score of ‘1’ to the scales. The scale items each reflected seven common characteristics of addictive behaviour: preoccupation, use alone, use for effect, use as a medicine, protection of supply, using more than planned, and increased capacity or tolerance (Lefever, 1988).

Preliminary evidence supporting the validity of the PROMIS Questionnaire scales was provided in an archival study conducted by Stephenson et al. (1995). Systematic questionnaire data, which had been collected from 471 patients admitted consecutively to treatment between 1988 and 1993, were used in a factor analytical study of addictive orientations. The scale scores had a reliability coefficient (Cochran's Q statistic) in the region of .9 for each scale. They also found a statistically significant relationship between diagnosis and mean scale scores. Based on the results of these and additional factor analyses, the highest loading items were selected, and in some cases combined and rewritten, to produce a smaller set of items that had equivalent scale reliability. These formed the basis of a shortened version of the questionnaire, the Shorter PROMIS Questionnaire (SPQ, see Appendix A).

The purpose of this study was threefold. Firstly, to standardise the SPQ using comparison and cut-off scores derived from a nonclinical community sample population. Secondly, to validate SPQ scales by using the cut-offs with a clinical sample population encompassing multiple addictive behaviours. Also, convergent and divergent validity were examined using subsets of the clinical sample that have completed other relevant validated scales. Finally, internal consistency and test–retest reliability were assessed using clinical samples.