PSYCHOMETRIC PROPERTIES AND FACTOR STRUCTURE OF THE READINESS TO CHANGE QUESTIONNAIRE

Lars Forsberg*, Jan Halldin1 and Peter Wennberg

Department of Clinical Neuroscience, Section of Clinical Alcohol and Drug Addiction Research and
1 Department of Public Health Sciences, Division of Social Medicine, Karolinska Institutet, Stockholm, Sweden

Received 24 May 2002; in revised form 23 December 2002; accepted 9 January 2003


    ABSTRACT
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Aims: This study is aimed at describing the psychometric properties and factor structure of the Swedish version of the Readiness to Change Questionnaire (RTCQ). Originally, the questionnaire included three factors (Precontemplation, Contemplation and Action) in a model that intends to measure stages of change. Methods: Patients at an emergency ward were screened for hazardous alcohol consumption and patients who met risk criteria and responded to the RTCQ were included in the study (n = 165). A confirmatory factor analysis proved to be slightly more suitable for a three-factor solution, compared to the one-factor solution. Results: The three scales reflecting the three factors showed satisfactory reliability. Conclusions: The presented data give no support for abandoning the original use of the test.


    INTRODUCTION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
The stages of change model (Prochaska and DiClemente, 1986Go; DiClemente and Prochaska, 1998Go) has been helpful in conceptualizing motivation to change to clinicians working in different areas of risky behaviours, such as smoking and hazardous and harmful alcohol consumption (Davidson, 1998Go). According to the model, an individual is more or less receptive to different kinds of action depending on his or her stage of change. The first of five stages of change is labelled Precontemplation and is characterized by lack of problem recognition. During the next stage — Contemplation — the individual starts to consider a change. Progressing to the stage of Preparation, a decision to change is taken, and plans are made on how to succeed. The following stage is Action, which implies that the individual is experimenting with alternative behavioural patterns. The fifth stage, Maintenance, includes long-term reinforcement of the individual’s new healthier behaviours in order to stabilize these behaviours.

The Readiness to Change Questionnaire (RTCQ) (Rollnick et al., 1992Go) was designed for clients with alcohol problems who might not be aware of having an alcohol problem. Items were initially chosen to represent a specific stage of change according to the Prochaska and DiClemente model (Prochaska and DiClemente, 1986Go). The Maintenance stage is not included, because some individuals endorsed both Precontemplation and Maintenance items. Hence, the RTCQ aims at assessing three of the stages of change — Precontemplation, Contemplation and Action. The test has been shown to have a three-factor structure (Rollnick et al., 1992Go), corresponding to the three stages of change. However, the three factors are correlated, which indicates that the test measures a common dimension besides reflecting the three stages of change. The three-factor structure has been replicated with the Spanish version (Rodriguez-Martos et al., 2000Go) and with the German version of the RTCQ (Hapke et al., 1998Go) in samples of hazardous and harmful drinkers coming into contact with health care. However, in a German sample of hazardous drinkers, or alcohol-dependent subjects in the general population, the three-factor structure could not be replicated (Hannöver et al., 2002Go). Five of 12 items loaded high on both the factors that emerged. In a Dutch study (Defuentes-Merillas et al., 2002Go) on excessive drinkers who were admitted to an addiction treatment centre, a two-factor structure was found, with action items forming one factor and a bi-polar factor with Precontemplation and Contemplation items.

In the original study (Rollnick et al., 1992Go) the internal consistency reliability of the three scales was found to be satisfactory, as was the case in the Spanish (Rodriguez-Martos et al., 2000Go) and the German (Hapke et al., 1998Go) studies. The test–retest reliability, calculated as a Pearson correlation coefficient, was reported in the original study (Rollnick et al., 1992Go) and the Spanish study (Rodriguez-Martos et al., 2000Go), and ranged from 0.78 to 0.87 for the three scales. However, an American study (Gavin et al., 1998Go) reported a low internal reliability (Cronbach’s alpha) for the Precontemplation and the Contemplation subscales. The latter authors suggested that this was due to the sample used, which was composed of alcohol-dependent patients applying for treatment. In contrast, in the Dutch study (Defuentes-Merillas et al., 2002Go) on patients applying for alcohol treatment, the internal consistency in the three scales was acceptable.

The underlying model of stages of change has been questioned by the constructor of the RTCQ (Budd and Rollnick, 1996Go; Rollnick, 1998Go), and a subsequent reanalysis of the original data (Budd and Rollnick, 1996Go) produced a global second-order factor termed ‘readiness to change’, indicating that one single factor fits the data better than the three-factor solution. The continuous scale shows good internal consistency reliability. Furthermore, in a comprehensive review (Carey et al., 1999Go) of the psychometric properties of tests assessing stages of change, including the RTCQ, Carey was in favour of a one-factor continuous readiness to change construct. Thus, there are mainly two ways to conceptualize the RTCQ, which is the comprising of one factor or three factors.

The aim of the present study was to examine psychometric properties and factor structure of the Swedish version of the RTCQ and to test the original three-factor concept to a one-factor model of the same test.


    SUBJECTS AND METHODS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
The present study was carried out as part of a larger randomized controlled study concerning brief interventions (Forsberg et al., 2000Go). The study was approved by the local ethics committee at Karolinska Institutet (no. 92-327).

Subjects
The original cohort consisted of patients (n = 563) recruited consecutively at a surgical ward at Danderyd Hospital in Stockholm, Sweden. Patients were recruited to the study if they were between 16 and 73 years of age and excluded if they were terminally ill, had insufficient language skills or used some form of illicit drug. The patients were informed that their results would not be registered in the hospital records and that they could withdraw from the study at any time.

Patients who gave their informed consent were interviewed using the screening questionnaires Mm-MAST (Kristenson and Trell, 1982Go), CAGE (Ewing, 1984Go), and the Trauma Scale (Skinner et al., 1984Go). Cut-off limits for hazardous alcohol habits were two affirmative answers in any of the questionnaires. Positive cases were further assessed with regard to alcohol consumption, and those showing risky consumption according to either of two criteria were subject to further analysis (Forsberg et al., 2000Go). The first criterion of hazardous consumption was regular drinking [>=162 g of absolute alcohol per week on average for men and >=82 g for women (Ashley et al., 1994Go)]. The second criterion was peak consumption during the last 12-month period [for males >=1.05 g of absolute alcohol per kg body weight and for females >=0.90 g (Babor and Grant, 1991Go)]. In total, 165 patients out of 563 (29%) were classified as risky consumers and were included in the study (see sample characteristics in Table 1Go). The preliminary medical diagnoses of the risky consumers at the surgical ward were classified in 11 categories and the most frequent diagnoses were appendicitis (27%), ‘abdominal observation’ (22%) and trauma (21%).


View this table:
[in this window]
[in a new window]
 
Table 1. Sample characteristics
 
The questionnaire
The RTCQ was translated into Swedish and back-translated into English in order to identify difficult phrases. Items with double-negative phrases were difficult to understand in Swedish and were in one case reformulated into a positive phrase. Revised versions of the test were repeatedly tested on staff and patients until the readability and comprehension of the test was clearly established. The patients responded to each item of the test with one of the following five alternatives: totally agree (+2), partly agree (+1), unsure (0), partly disagree (-1) and totally disagree (-2) (Rollnick et al., 1992Go). The items representing each of the stages of change were regarded as scales measuring the extent to which the subjects endorsed that stage of change. In the three-stage model, allocation of subjects to one of the stages was based on the highest score obtained either on the Precontemplation, Contemplation or Action scales. In the event of a tie between two scale scores, the most advanced stage was chosen (Rollnick et al., 1992Go; Heather et al., 1993Go). In the one-factor model, a readiness score was calculated by combining all 12 items to form a continuous Likert scale. The signs of the Precontemplation items were altered to represent readiness to change.

Analysis of data
The original three-factor model was tested against three criteria: (1) the model should show satisfactory fit-indices in a confirmatory factor analysis; (2) the correlations between adjacent stages/factors should be higher than for distanced stages/factors; (3) the internal reliability and test–retest reliability of the test should be satisfactory.

The factor structure of the test was examined using a confirmatory factor analysis [LISREL 8.12a (Jöreskog and Sörbom, 1996Go)]. Fit-indices of one- and three-factor solutions of the test were compared. Internal consistency reliability of the test was provided when treating the test as one scale or three scales. In order to investigate the test–retest reliability of the test, 18 patients responded to the questionnaire twice. These patients were chosen, because they were in-patients and were available for the second follow-up 1 or 2 days after the first occasion.


    RESULTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
The factor structure of the questionnaire
In order to compare how well a one- or a three-factor solution of the test fits the data, confirmatory factor analyses were conducted (summarized in Table 2Go). The model with the best fit for data was the three-factor model, in which the three factors were correlated. The root mean square error of approximation (RMSEA) was 0.06, compared to the other solutions with RMSEA ranging between 0.08 and 0.16. The other fit-indices (goodness of fit index and comparative fit index) went in the same direction. This gives some support for the original idea of a three-factor model of the test. The three-factor solution with factor loadings is shown in Table 3Go.


View this table:
[in this window]
[in a new window]
 
Table 2. Confirmatory factor analysis according to four different models (n = 165)
 

View this table:
[in this window]
[in a new window]
 
Table 3. Items with factor loadings in the Swedish Readiness to Change Questionnaire showing the hypothesized three stages of change: Precontemplation (P), Contemplation (C) and Action (A) (n = 165)
 
Furthermore, factors reflecting two stages near to each other had higher correlations than the two factors reflecting stages far from each other in the stages of change model. The strong positive correlations between the three factors show that a high score in one scale is often accompanied by high scores in the other scales, which then might indicate that the test also measures another construct, besides reflecting the three stages of change. The one-dimensional construct might be the level of problem severity.

Reliability of the test
The internal consistencies for each of the three scales of the test were reported as well as for one continuous 12-item scale (Table 4Go). Deleting items from the continuous (one-dimensional) scale did not increase reliability. Test–retest reliability was calculated as a Pearson correlation between the two times. Furthermore, allocations to stage of change (or readiness score) were compared between the two times. As a three-factor test, 16 out of 18 patients were allocated to the same stage of change both times, and as a one-factor test, when treating the continuous variable as a trichotomous discrete variable, 15 patients were in the same third of the variable both times.


View this table:
[in this window]
[in a new window]
 
Table 4. Internal consistency (n = 165) and test–retest reliability (n = 18) of the RTCQ
 

    DISCUSSION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Three criteria were set to test whether the original three-factor model holds, and according to our data the original model seems to hold against all three. Conceptually, stages and dimensions should not be mixed up. While a stage reflects a position in a developmental process, different dimensions or factors only reflect that a test measures different types of qualities. Hence, the fact that this paper supports the notion of a three-dimensional test does not imply that the idea of the three stages also holds.

The factor structure of RTCQ has now been studied in at least five countries, in five different languages and cultures (Rollnick et al., 1992Go; Budd and Rollnick, 1996Go; Hapke et al., 1998Go; Rodriguez-Martos et al., 2000Go; Defuentes-Merillas et al., 2002Go; Hannöver et al., 2002Go) and the results, which are somewhat diverse, favour not only the notion that the RTCQ factor structure is consistent with the stages of change model with a three-factor structure, but also that the test measures a common construct representing severity of alcohol problems.

The test–retest reliability was satisfactory for each of the three scales corresponding to the stages of the change model, and just as good for the continuous readiness scale. The internal consistency reliability was higher in the continuous scale, due to more items in this scale. The results correspond to earlier studies of RTCQ, which, with one exception, have shown the test to be reliable.

Yet, future research on the psychometrics of the RTCQ is needed. Budd and Rollnick (1996)Go proposed that the three-factor model and the one-factor model are examined further by contrasting the scale scores derived from each model. With the stage model, RTCQ is scored as a measure of stage of change, and as a continuous variable, the RTCQ score is the sum of the items in the questionnaire. A forthcoming study will investigate the proposal of Budd and Rollnick (1996)Go to further examine the two alternative models with regard to validity, contrasting their relative merits.

The patients in the present study were representative of the target population that the test was developed for; however, patients who were alcohol-dependent were not excluded. Hence, the sample was heterogeneous with respect to the severity of alcohol problems and readiness to change the alcohol behaviour. As a consequence, the sample includes sufficient variation in responses to the RTCQ.


    ACKNOWLEDGEMENTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Dr Sören Ekman, Department of Surgery at Danderyd Hospital, together with the first author, translated the RTCQ and significantly contributed to the project about brief interventions. Professor Hans Bergman contributed valuable points of view. We are also indebted to the staff of the surgical admission ward (Ward 61 at Danderyd Hospital) under the leadership of Chief Nurse Mildred Rutström as well as to the staff on the surgical unit and surgical wards. The Public Health and Medical Services’ Committee in the County of Stockholm provided the project with financial support.


    FOOTNOTES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
* Author to whom correspondence should be addressed at: Karolinska Institutet, Department of Clinical Neuroscience, Section of Clinical Alcohol and Drug Addiction Research, Magnus Huss Clinic, M4:04 Karolinska Hospital, SE-171 76 Stockholm, Sweden. Back


    REFERENCES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Ashley, M. J., Ferrence, R., Room, R., Rankin, J. and Single, E. (1994) Moderate drinking and health: report of an international symposium. Canadian Medical Association Journal 151, 1–16.[Abstract]

Babor, T. F. and Grant, M. (1991) Project on Identification and Management of Alcohol-Related Problems. Report on Phase II: A Randomized Clinical Trial of Brief Intervention in Primary Health Care. World Health Organization, Geneva.

Budd, R. and Rollnick, S. (1996) The structure of the readiness to change questionnaire: a test of Prochaska & DiClemente’s transtheoretical model. British Journal of Health Psychology 1, 365–376.[ISI]

Carey, K., Purmine, D., Maisto, S. and Carey, M. (1999) Assessing readiness to change substance abuse. Clinical Psychology in Science and Practice 6, 245–266.[Abstract]

Davidson, R. (1998) The transtheoretical model: a critical overview. In Treating Addictive Behaviors, 2nd edn, Miller, W. and Heather, N. eds, pp. 25–38. Plenum Press, London.

Defuentes-Merillas, L., Dejong, C. and Schippers, G. (2002) Reliability and validity of the Dutch version of readiness to change questionnaire. Alcohol and Alcoholism 37, 93–99.[Abstract/Free Full Text]

DiClemente, C. and Prochaska, J. (1998) Toward a comprehensive, transtheoretical model of change. In Treating Addictive Behaviors, 2nd edn, Miller, W. and Heather, N. eds, pp. 3–24. Plenum Press, New York, NY.

Ewing, J. A. (1984) Detecting Alcoholism. Journal of the American Medical Association 252, 1905–1907.[Abstract]

Forsberg, L., Ekman, S., Halldin, J. and Rönnberg, S. (2000) Brief interventions for risk consumption of alcohol at an emergency surgical ward. Addictive Behaviors 25, 471–475.[CrossRef][ISI][Medline]

Gavin, D., Sobell, L. and Sobell, M. (1998) Evaluation of the readiness to change questionnaire with problem drinkers in treatment. Journal of Substance Abuse 10, 53–58.[CrossRef][ISI][Medline]

Hannöver, W., Thyrian, J., Hapke, U., Rumpf, H.-J., Meyer, C. and John, U. (2002) The readiness to change questionnaire in subjects with hazardous alcohol consumption, alcohol misuse and dependence in a general population survey. Alcohol and Alcoholism 37, 362–369.[Abstract/Free Full Text]

Hapke, U., Rumpf, H. and John, U. (1998) Differences between hospital patients with alcohol problems referred for counselling by physicians’ routine clinical practice versus screening questionnaires. Addiction 93, 1777–1785.[CrossRef][ISI][Medline]

Heather, N., Rollnick, S. and Bell, A. (1993) Predictive validity of the readiness to change questionnaire. Addiction 88, 1667–1677.[ISI][Medline]

Jöreskog, K. and Sörbom, D. (1996) LISREL 8: Chicago, IL: Scientific Software International, Incorporated.

Kristenson, H. and Trell, E. (1982) Indicators of alcohol consumption: comparisons between a questionnaire (Mm-Mast), interviews and serum {gamma}-glutamyl transferase (GGT) in a health survey of middle-aged males. British Journal of Addiction 77, 297–304.[ISI][Medline]

Prochaska, J. and DiClemente, C. (1986) Toward a comprehensive model of change. In Treating Addictive Behaviours: Processes of Change, Miller, W. and Heather, N. eds, pp. 4–27. Plenum Press, New York, NY.

Rodriguez-Martos, A., Rubio, G., Auba, G., Santo-Domingo, J., Rorralba, L. and Campillo, M. (2000) Readiness to change questionnaire: reliability study of its Spanish version. Alcohol and Alcoholism 35, 270–275.[Abstract/Free Full Text]

Rollnick, S. (1998) Readiness, importance and confidence: critical conditions of change in treatment. In Treating Addictive Behaviors, 2nd edn, Miller, W. and Heather, N. eds, pp. 49–60. Plenum Press, London.

Rollnick, S., Heather, N., Gold, R. and Hall, W. (1992) Development of a short ‘readiness to change’ questionnaire for use in brief, opportunistic interventions among excessive drinkers. British Journal of Addiction 87, 743–754.[ISI][Medline]

Skinner, H. A., Holt, S., Schuller, R., Roy, J. and Israel, Y. (1984) Identification of alcohol abuse using laboratory tests and a history of trauma. Annals of Internal Medicine 101, 847–851.[ISI][Medline]





This Article
Abstract
FREE Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Request Permissions
Google Scholar
Articles by Forsberg, L.
Articles by Wennberg, P.
PubMed
PubMed Citation
Articles by Forsberg, L.
Articles by Wennberg, P.