readiness to change and risk drinking women

Journal of Substance Abuse Treatment 40 (2011) 230–240

Regular article

Readiness to change and risk drinking women

Sonia Matwin, (Ph.D.)a, Grace Chang, (M.D., M.P.H.)a,b,⁎

aHarvard Medical School, Department of Psychiatry, Boston, MA 02115, USAbBrigham and Women's Hospital, Department of Psychiatry, Boston, MA 02115, USA

Received 28 May 2010; received in revised form 21 October 2010; accepted 8 November 2010

Abstract

The predictive value of the Readiness to Change Questionnaire (RTCQ) for subsequent drinking was evaluated in 499 women. Thesewomen had medical problems potentially exacerbated by alcohol use and were enrolled in an intervention study. Correlates and predictors ofstage of change were analyzed. Results indicated that the categorical application of the RTCQ did not predict drinking in the 6–12 monthsafter enrollment. Preliminary findings support rescoring the RTCQ into a continuous measure. Following this conversion, situational risksfactors for drinking were examined as potential mediators of readiness to change (RTC). Heightened risk for alcohol consumption duringargument or boredom was found to attenuate the association between one's RTC and later drinking. Finally, medical condition moderatedthe association of RTC on later drinking; women with diabetes, infertility, or osteoporosis drank the most in the contemplation stage. Incontrast, hypertensive women drank more when action oriented to change. The implications for treating risk drinking in women arediscussed. © 2011 Elsevier Inc. All rights reserved.

Keywords: Risk drinking; RTCQ; Women; Medical problems

1. Introduction

Understanding the processes that drive change in alcoholconsumption is an area of active inquiry (Huebner &Tonigan, 2007). Although the processes may be psycholo-gical, social, or neuropsychological, the shift in emphasisfrom whether people change to how people change reflectsadvances in alcohol treatment and the desire to improveoutcomes (Willenbring, 2007).

Readiness to change (RTC) has been conceptualized as acombination of a patient's perceived importance of aproblem and confidence in his or her ability to change, andis a central tenet of the Transtheoretical Model of intentional

This study was supported by grants from NIAAA and the Office ofResearch on Women's Health (R01 AA 014678 and K24 AA 00289, bothG. C.).

⁎ Corresponding author. Brigham and Women's Hospital, Departmentof Psychiatry, 221 Longwood Avenue, Boston, MA 02115. Tel.: +1 617 7326775; fax: +1 617 264 6370.

E-mail address: [email protected] (G. Chang).

0740-5472/10/$ – see front matter © 2011 Elsevier Inc. All rights reserved.doi:10.1016/j.jsat.2010.11.004

change (DiClemente, Schlundt, & Gemmell, 2004). Thisstage of change model addresses the underlying motivationalprocesses that drive people to modify their behavior andproposes that the process of change is based on progressionthrough distinct qualitative stages of change. RTC can bemeasured using the brief Readiness to Change Questionnaire(RTCQ), where 12 items are rated on a 5-point scale, rangingfrom strongly agree to strongly disagree (Heather, Gold, &Rollnick, 1991). This measure consists of three subscales(four items each) that correspond to the precontemplation(e.g., “I don't think I drink too much”), contemplation (e.g.,“I enjoy my drinking but sometimes drink too much”), andaction (e.g., “I am trying to drink less than I used to”) stagesof change. The RTCQ assigns the individual to one of thethree stages of change, using either the quick (assigning theindividual to the stage having the highest raw score) orrefined (performing profile analysis of scores across the threescales) method of allocation (Hannover et al., 2002). Themeasure was developed primarily for use with hazardousdrinkers who are not necessarily aware that they have analcohol problem (i.e., non-treatment-seeking) and provides a

http://dx.doi.org/10.1016/j.jsat.2010.11.004



mailto:[email protected]


231S. Matwin, G. Chang / Journal of Substance Abuse Treatment 40 (2011) 230–240

means of assessing a drinker's RTC quantity and frequencyof drinking. Readiness ranges from less readiness, with noproblem recognition, to greater readiness, as manifested byconsidering, or actually making, behavior changes (Forsberg,Halldin, & Wennberg, 2003).

RTC has been studied as a predictor of later drinkingreduction (Borsari, Murphy, & Carey, 2009), as a mecha-nism of drinking change (Freyer et al., 2005), and as amediator of treatment (Stein et al., 2009). Several investi-gators have examined the utility of RTC stages as a predictorof drinking, by women in particular. Although women havehad lower rates of alcohol use disorders than men, the gendergap for drinking problems is narrowing (Hasin, Stinson,Ogburn, & Grant, 2007; Keyes, Brant, & Hasin, 2008).Women are more vulnerable to the negative medicalconsequences of alcohol because of metabolic differencesand as a result experience serious negative consequencesfrom alcohol consumption earlier, and to a greater degree,than men (Nolen-Hoeksema, 2004).

One of the first studies to focus on women drinkersclassified 119 treatment-seeking alcohol-dependent womeninto contemplation or action stages at study enrollment,before treatment. Those who had greater RTC perceivedmore advantages and fewer disadvantages of changing theirdrinking behavior (Share, McCrady, & Epstein, 2004). Alater study of 301 pregnant women who were not seekingalcohol treatment found that stage of change did not predictsubsequent prenatal alcohol use. However, stage of changedesignation was associated with different patterns of drinkingbefore pregnancy in that sample. In other words, those in theprecontemplation stage drank more per episode and moreoften than those in the action stage before pregnancy (Chang,McNamara, Wilkins-Haug, & Orav, 2007). More recently, astudy of 285 (non-treatment seeking) female college studentssought to determine under what circumstances RTC variedwithin individuals (Kaysen, Lee, LaBrie, & Tollison, 2009).Findings revealed that changes in RTC were not consistentlypredictive of subsequent drinking behaviors.

Budd and Rollnick (1996) have suggested that the RTCQlacks discriminant validity and have shown that the threestages of change are not related in a simplex structure. Thisstatistical finding is corroborated by the observation thatindividuals typically do not progress through the stages ofchange in the invariant sequence the model hypothesizes(Sutton, 1996). As such, it has been suggested thatconstruing RTC as a continuous variable (rather than as aseries of discrete stages) may be a more parsimoniousformation of this construct. In fact, Budd and Rollnick(1996) have shown that the RTCQ items can be rescored toform a valid and reliable continuous measure of RTC.

Similarly, when we consider that the RTCQ is not alwayspredictive of consumption (Gavin, Sobell, & Sobell, 1998), itis possible that the effects of RTC as a predictor ofsubsequent drinking behavior may be mediated or moderatedby other factors. For example, impediments to self-efficacyor outcome efficacy may adversely impact attempts to reduce

drinking. This is supported by the fact that measures ofcraving have been demonstrated to be a useful means ofpredicting drinking (Flannery, Poole, Gallop, & Volpicelli,2003). Therefore, if cravings for alcohol are higher, then littlebehavior change might be expected, as the barriers to changeare greater. Similarly, sensitivity to situational risk factorsmay lead to reduced behavior change, as the individualencounters more frequent obstacles to initiating and thenmaintaining behavior changes. In addition, it is possible thatpersonal variables interact with one's RTC, such that thesefactors may systematically increase or reduce the predictivevalue of the RTCQ. For example, it is possible that the effectof RTC on follow-up drinking is dependent on the medicalcondition for which participants were seeking treatment. Thisis especially plausible when we consider that these medicalproblems (i.e., diabetes, hypertension, infertility, or osteopo-rosis) are potentially exacerbated by excessive alcohol use.

The results of studies focusing on women drinkers aresimilar to others of more general populations, which havefound that stage of change status neither predicts participa-tion in treatment nor drinking behavior change outcomes(DiClemente et al., 2004). Hence, additional studies aboutthe stages of change and, specifically, its measurement incomparison to continuous readiness have been recom-mended, given the enduring popularity of the Transtheore-tical Model of intentional behavior change and the continuedneed to understand the role of motivation in drinkingbehavior change.

1.1. Study purpose

The purpose of this study is to evaluate the efficacy ofthe RTCQ as a predictor of drinking change in a sample of499 adult women with medical problems potentiallyexacerbated by excessive alcohol use. In addition, wesought to analyze the correlates and predictors of stage ofchange in this non-treatment-seeking sample. Our specifichypotheses were as follows:

H1. RTCQ stage of change designations will not predictdrinking behavior in the 6- to 12-month period followingstudy enrollment among non-treatment-seeking adultwomen with medical problems potentially exacerbatedby alcohol consumption.H2. Rescoring the RTCQ (by combining items to form ahomogenous scale) to treat RTC as a continuousvariable will increase the predictive value of thisconstruct on drinking behavior in the 6- to 12-monthperiod following study enrollment among non-treatment-seeking adult women.H3. Situational risk factors (e.g., increased risk fordrinking when bored; increased risk for drinkingfollowing an argument) for drinking and levels ofweekly craving (as measured on a visual analogue scale)will mediate the predictive value of one's (continuous)RTC score.

232 S. Matwin, G. Chang / Journal of Substance Abuse Treatment 40 (2011) 230–240

H4. Medical diagnoses (i.e., medical problems—diabe-tes, hypertension, infertility, or osteoporosis—potentiallyexacerbated by excessive alcohol use) will moderate thepredictive value of the RTCQ.

2. Method

2.1. Participants

Participants were 499 of 511 (97%) women enrolled in arandomized trial of brief intervention for risk drinking; 12 ofthe 511 did not complete the RTCQ and were thereforeexcluded from analyses. Overall, 96% completed the study,which included follow-up interviews 12 months afterenrollment. Participants were randomized to receive eithera single session brief intervention or treatment as usualfollowing the enrollment interview. There were no diffe-rences in drinking outcomes by treatment group assignment(Chang, 2009).

Participants were women receiving outpatient treatmentfor diabetes, hypertension, osteoporosis, or infertility, butthey were not seeking alcohol treatment. They met eligibilitycriteria for the clinical trial, which included (a) medicaldiagnosis confirmed by a physician (viz., diabetes, hyper-tension, infertility, or osteoporosis), (b) sufficient commandof English to comprehend and complete study question-naires, (c) no current treatment for alcohol or drug usedisorders, (d) no current physical dependence on alcohol orother substances, and (e) not currently pregnant or nursing.In addition, participants needed to consume alcohol at levelsexceeding the National Institute on Alcohol Abuse andAlcoholism (NIAAA) sensible drinking limits in the past6 months or to be T-ACE (Tolerance, Annoyed, Cut Down,Eye Opener) alcohol screen positive. The NIAAA sensibledrinking limits for adult women up to age 65 are no morethan seven standard drinks per week and no more than threestandard drinks per day (U.S. Department of Health andHuman Services, National Institutes of Health, & NationalInstitute on Alcohol Abuse and Alcoholism, 2007). TheT-ACE asks four questions (T = How many drinks does ittake for you to feel high [tolerance]?; A = Have peopleannoyed you by talking about your drinking; C = Have youtried to cut down on your drinking; and E = Have you everhad a drink first thing in the morning to steady your nerves orget rid of a hangover [eye-opener]?) and was scored usingthe standard cutoff of 2 (or more) points (Sokol, Martier, &Ager, 1989). Additional details about the study are availableelsewhere (Chang, Fisher, Hornstein, Jones, & Orav, 2010).

2.2. Measures

Participants completed a series of measures at enrollment,during the brief intervention, and at study follow-up.Enrollment measures included the following: (a) the alcoholand drug abuse modules from the Structured Clinical

Interview for DSM-IV (SCID; First, Spitzer, Gibbon, &Williams, 2002)—to obtain current and lifetime alcohol anddrug disorder diagnoses; (b) the alcohol timeline followback(TLFB; Sobell & Sobell, 1992)—to obtain estimates of dailydrinking for the 6 months prior to study enrollment; (c) theRTCQ (Rollnick, Heather, Gold, & Hall, 1992)—12 items toassess an individual's RTC drinking by using the quickmethod of assignment to one of three stages of change(precontemplation [no interest in change], contemplation[considering a change], and action [making a change];Heather, Rollnick, & Bell, 1993); (d) Craving Scales (visualanalogue 100-mm line on which participants indicate theintensity of their desire to drink)—to estimate the intensity ofcravings for alcohol, currently (NOW) and in the past week(PAST); and (e) the Perceived Stress Scale (PSS; Cohen &Williamson, 1988)—to measure the perception and expe-rience of stress in one's life during the past month.

Following enrollment, women in the brief interventiongroup were asked to name a drinking goal (e.g., no change,reduce, abstain), to identify situations that would put them atrisk for alcohol consumption (i.e., during an argument,feeling bored, others are drinking, feeling depressed, feelinglonely, feeling anxious, want to have fun), and to listbehaviors they could pursue instead of drinking (e.g., callinga friend, having a snack, exercising, watching television).The women's responses were documented on the briefintervention summary sheet and then tabulated.

At the 12-month follow-up, participants completed theTLFB for drinking history of the preceding 6 months (i.e.,months 6–12 after enrollment), the PSS, and the CravingScales (already described). From the TLFB self-reportmeasure of drinking history, four drinking outcomes werecalculated: (a) drinks per drinking day (D/DD), (b) percentdrinking days, (c) number of binge episodes defined as fouror more drinks per occasion, and (d) number of weeksexceeding NIAAA sensible drinking limits for women.Baseline alcohol use was calculated using the TLFB data forthe 6 months prior to enrollment. Drinking at follow-up wascalculated using the TLFB data for the previous 6 months.

This study was reviewed and approved by the PartnersInstitutional Review Board (2004-p-000687). Participantsreceived an honorarium of $50.00 for the enrollment andfollow-up interviews. A federal Certificate of Confidentialitywas obtained (AA-30-2005) as well. The Clinicaltrials.govIdentifier for the umbrella study is NCT00846638.

2.3. Data analysis

All analyses were carried out using the SPSS statisticalpackage (version 14.0). Simple descriptive statistics werecalculated and are reported as percentages, means, andstandard deviations, as appropriate.

2.3.1. Descriptive outcome measuresDifferences in alcohol use at baseline and follow-up were

examined with paired sample t tests.


2.3.2. Analyzing H1: Predictive value of RTCQThe RTCQ was evaluated as a measure of discrete stages

of RTC. That is, participant responses on the RTCQ werescored according to the procedures outlined by the RTCQuser's manual (Heather et al., 1991), including the instruc-tions to reverse-score the precontemplation scale items.Based on these scores, participants were assigned toprecontemplation, contemplation, or action stages withregard to their RTC drinking habits (Allen & Columbus,1995). Associations between the three RTCQ subscales wereexamined with correlations. Confirmatory factor analyseswere employed to verify the internal consistency andreliability of the three RTCQ subcategories in the currentsample. To evaluate the RTCQ, a series of one-way analysesof variance (ANOVAs) were conducted to examine therelationship between stage of change and the four drinkingoutcomes at follow-up (i.e., D/DD, percent drinking days,number of binge episodes, and number of weeks exceedingNIAAA sensible drinking limits for women) based onalcohol consumption in the 6 to 12 months after enrollment.Furthermore, this analysis allowed us to determine whetherthe three RTC categories differed with regard to follow-updrinking behaviors.

2.3.3. Analyzing H2: RTCQ as a measure of continuousreadiness of change

The RTCQ was further evaluated after a continuum ofsummed scores was used in place of the ordinal categoriesof RTC. Standard scoring procedures (Heather et al., 1991)were maintained (i.e., precontemplation items were reversecoded); however, whereas previously the individual wasassigned to that stage that had the highest raw score,participant scores were now summed across subscales tocreate a single continuous measure of RTC. This method ofcombining all 12 RTCQ items to form a single continuousmeasure of RTC has been shown to be a reliable and validpredictor of later drinking behavior (Sutton, 1996).Following this conversion to continuous scores (whereascending values indicated movement toward actionorientation), predictors of RTCQ were evaluated withstepwise regressions.

2.3.4. Analyzing H3 and H4: examining mediators andmoderators of the RTCQ

Situational risk factors and self-reported alcohol cravingswere examined as potential mediators of RTC (see Fig. 1)using Baron and Kenny's (1986) regression procedures. This

Proposed medi1. Increased urge when bored; 2. Increased urge during argument; 3. Increased alcohover past week

IV: Readiness to Change Score

Fig. 1. Predicted mediators of RTC an

method of analysis posits several requirements, which are (1)the total effect of the independent variable (i.e., RTC scorecollected at enrollment) on the dependent variable (i.e., 6- to12-month follow-up drinking behaviors) must be significant;(2) the path from the independent variable to the mediator(i.e., risk factors for drinking, collected at enrollment) mustbe significant; and (3) the path from the mediator to thedependent variable must be significant (i.e., must lead to areduction in coefficient magnitude while controlling for theindependent variable). In addition, there is a simultaneousfourth step, which is required to provide evidence forcomplete mediation. (4) This step stipulates that theindependent variable must no longer have any effect on thedependent variable when the mediator has been controlled(i.e., the effect of RTC on follow-up drinking is reduced tozero when controlling for the proposed mediator). It shouldbe noted that the effects in both Steps 3 and 4 are estimated inthe same equation. Using these procedures, the meditationalmodels depicted in Fig. 1 were examined.

Finally, we examined the impact of four medicaldiagnoses (diabetes, hypertension, infertility, or osteoporo-sis) on stages of RTC. These diagnoses were selectedbecause they are medical problems exacerbated by alcoholconsumption and may therefore affect the strength or form ofthe relationship between RTC stage designation and follow-up (6–12 months) drinking behaviors. Moderation wasexamined with standard ANOVA procedures in a four(medical condition: diabetes, hypertension, infertility, orosteoporosis) by three (original categorical RTC designa-tions: precontemplation, contemplation, action) analysis.These findings were replicated with multiple regressionprocedures by creating the necessary interaction term (i.e.,Medical diagnosis × RTC continuous variable) and dummyvariables (i.e., for the moderator and causal variables).

3. Results

3.1. Demographic and clinical characteristics

The average participant age was 45 years (SD = 13.6years), with a range from 19 to 78 years. More than half weremarried (53%), completed at least 4 years of college (62%),and worked outside of the home (65%). Less than half (48%)had children at home. The women were diverse, with 73%being Caucasian, 22% African American, and 5% Hispanic.They had diabetes (19%), hypertension (31.5%), infertility

ators: to drink

to drink ol cravings

DVs: 12-month follow-up drinking 1. # of weeks exceeding NIAAA limits; 2. # of binge episodes

d follow-up drinking outcomes.


(32%), and osteoporosis (17.5%). Almost half (49%)were overweight with an average body mass index of 27.3(SD = 7.5).

3.2. Alcohol and substance use

Current and lifetime alcohol and drug disorder diagnoseswere obtained from the alcohol and drug abuse modules ofthe SCID. Lifetime (43%) and current (9.3%) alcohol usedisorders were the most common substance use diagnosesamong participants. Lifetime and current marijuana (13%and 2.3%, respectively) and cocaine (7% and 1.4%,respectively) use disorders were less frequent. Whereas5.6% of the participants had lifetime sedative–hypnotic,stimulant, and hallucinogen diagnoses, only 1% met criteriafor current disorders. No participant satisfied diagnosticcriteria for current opiate use disorders, but 2% had alifetime history.

For the 6-month period prior to enrollment, the womenaveraged two or more D/DD on 43 of 180 (24%) days.During this same 6-month period, they had about one bingeepisode (four or more drinks) per month (or six binges in all)and had 4 weeks that exceeded any NIAAA sensibledrinking limits. Drinking patterns subsequent to enrollmentwere significantly lower, with an average of less than twoD/DD, t(489) = 4.06, p = .00, on 37 of 180 days (20%),t(489) = 3.27, p = .00. The total number of binge episodeswas reduced to 4.5, t(489) = 1.93, p = .05, and the numberof weeks that exceeded NIAAA sensible drinking limitsdropped to three in the final 6 months of the study, t(489) =2.85, p = .00.

3.2.1. RTC stagesMost (64%) of the women were classified to be in the

precontemplation stage; 24% were in the action stage, and12% were in the contemplation stage. Mean scores on eachsubscale were the following: precontemplation, 0.9 (SD =3.6); contemplation, 3.3 (SD = 4.1); and action, 1.4 (SD =4.5). Subscale correlations were the following: precontem-plation and contemplation, r = .55; precontemplation andaction, r = .29; and contemplation and action, r = .56.Notably, all three subscales were significantly correlatedwith one another (p b .01), indicating that the RTCQ is notmeasuring three independent factors. Nonetheless, follow-ing standard practice, the three unique underlying subscalefactors (one corresponding to each subscale) of the RTCQwere evaluated and verified using confirmatory factoranalysis. Reliability analysis of the three subscales resultedin Cronbach's alpha coefficients as follows: precontempla-tion, .67; contemplation, .83, and action, .84, therebyindicating sufficient internal consistency among itemswithin each scale. When all items were treated as a single12-item scale, the alpha coefficient was .72, with anaverage inter-item correlation of r = .46 (all ps b .01),clearly indicating that these items combine to form ahomogenous scale.

The relationship between one's stage of change anddrinking at 12-month follow-up was evaluated (see Table 1for control variables of follow-up drinking behaviors). Stageof change did not predict alcohol consumption on any of thefour measures of drinking at 12 months: (a) D/DD, F(2, 480)= 0.13, p = .88; (b) percent drinking days, F(2, 480) = 0.69,p = .52; (c) number of binge episodes, F(2, 480) = 0.28, p =.39; or (d) number of weeks exceeding sensible drinkinglimits, F(2, 480) = 0.65, p = .27. At baseline, those in theprecontemplation stage (mean D/DD = 1.83, SD = 1.03)drank significantly less, F(3, 498) = 24.41, p = .00, thanthose in either the contemplation (mean D/DD = 2.68, SD =1.23) and action (mean D/DD = 2.61, SD = 1.99) stages.Sixteen percent of the women assigned to the precontempla-tion stage drank than less than one drink per drinking day.

3.2.2. RTC as a continuous scoreAs described previously, the three stages of change

designations were converted to continuous scores for furtheranalysis. The women's drinking goals were classifiedinto one of four categories: (a) reduce drinking, (b) maintain(no change) current drinking, (c) abstain from drinking,and (d) unknown, for the women who did not receive thebrief intervention.

3.2.3. Mediators of RTCSituational risk factors for drinking were obtained (shortly

after enrollment) from the brief intervention summary. These“risk” variables were considered for mediation becauseheightened risk situations were thought to account for thevariability in the predictive value of RTC scores. Sevenrisk situations were assessed (viz., heightened risk fordrinking when bored, depressed, lonely, anxious, arguing,seeking fun, others are drinking) and tested as potentialmediators of RTC.

First, regression analyses were run to examine whetherparticipants' continuous RTC scores did in fact significantlypredict follow-up drinking behaviors. This effect wassignificant for the number of binge episodes (R2 = .023,p = .00, b = 0.153), number of weeks exceeding NIAAAsensible drinking limits (R2 = .018, p = .00, b = 0.136), andnumber of D/DD (R2 = .038, p = .00, b = 0.194) in the 6–12 months following study enrollment. These findingsindicate that as continuous RTC scores moved towardaction, certain drinking behaviors increased at follow-up. Itshould be noted that continuous RTC scores did notsignificantly predict follow-up percentage of drinking days(R2 = .001, p = .476, b = 0.033); hence, this dependentvariable was excluded from further mediational analyses.

To test whether the path from the independent variable tothe mediator was significant, the subject's risky situations (ormoods) for drinking were regressed onto continuous RTCscores, which revealed that “risk during argument” (R2 =.040, p = .00, b = 0.199) and “risk when bored” (R2 = .023,p = .02, b = 0.150) were significantly predicted by higherRTC scores (i.e., movement toward action orientation). To

Table 1Regression results for alcohol consumption at 12-month follow-up

Predictors

D/DDNo. of weeksexceeding limits No. of binge episodes % of days drinking

b (SE) p b (SE) p b (SE) p b (SE) p

Constant 1.49 (0.44) .00 ⁎ 2.44 (1.77) .17 4.76 (4.84) .33 0.16 (0.07) .01 ⁎

Stage of change (dummy coded categories) 0.03 (0.15) .95 −0.03 (0.18) .46 0.031 (1.68) .50 0.06 (0.02) .17Drinking goal of abstinence −0.21 (0.16) .18 −1.72 (0.64) .01 ⁎ −1.10 (1.75) .53 −0.10 (0.02) .00 ⁎

Risk of drinking whenAnxious −0.04 (0.54) .95 −2.43 (2.19) .27 −8.05 (5.99) .18 −0.05 (0.08) .57Others drinking 0.11 (0.29) .64 0.39 (1.00) .70 0.56 (2.73) .84 −0.03 (0.04) .47Bored 0.61 (0.57) .29 0.20 (0.29) .00 ⁎ 0.22 (0.28) .00 ⁎ 0.06 (0.088) .52Depressed −0.18 (0.49) .72 1.66 (1.97) .40 −0.19 (5.40) .97 −0.04 (0.07) .57Lonely 1.10 (0.89) .22 6.29 (3.58) .08 21.90 (9.81) .067 0.18 (0.14) .20Arguing −1.21 (0.75) .11 0.17 (1.02) .01 ⁎ 0.21 (8.28) .00 ⁎ −0.03 (0.11) .77Want fun −0.19 (0.25) .45 −1.92 (1.02) .06 −4.06 (2.81) .15 −0.10 (0.04) .01 ⁎

Cope with drink desire byNonalcoholic drinks −0.32 (0.26) .21 −0.49 (1.03) .63 1.94 (2.82) .49 −0.04 (0.04) .28Calling a friend 0.13 (0.64) .84 0.36 (2.59) .89 2.12 (7.09) .77 −0.04 (0.10) .67Eating 0.47 (0.33) .15 −2.21 (1.31) .09 −2.97 (3.59) .41 −0.07 (0.05) .17Exercising 0.33 (0.34) .34 −0.56 (1.38) .69 1.21 (3.78) .75 −0.05 (0.05) .36Watching television −1.81 (0.91) .05 ⁎ −2.07 (3.67) .57 4.30 (10.05) .67 −0.19 (0.14) .17Listening to music −0.19 (1.33) .89 −0.88 (5.35) .87 −0.92 (14.65) .95 −0.03 (0.20) .87Lifetime abuse/dependence onAlcohol 0.63 (0.28) .03 ⁎ 2.26 (1.11) .04 ⁎ 5.30 (3.05) .08 0.02 (0.04) .71Opiates 1.78 (0.77) .02 ⁎ −1.77 (3.08) .57 −8.32 (8.44) .32 0.04 (0.12) .73Marijuana 0.77 (0.40) .06 0.36 (1.63) .83 −2.73 (4.46) .54 0.02 (0.06) .73Cocaine 0.97 (0.54) .07 −0.81 (2.17) .71 11.33 (5.93) .06 −0.02 (0.08) .76Hallucinogens −1.39 (0.57) .02 ⁎ −3.57 (2.30) .12 −9.94 (6.31) .12 −0.10 (0.09) .27NOW alcohol craving 0.068 (0.01) .13 0.06 (0.04) .12 0.08 (0.11) .43 0.00 (0.00) .25PAST week alcohol craving −0.01 (0.01) .33 0.23 (0.03) .00 ⁎ 0.17 (0.07) .00 ⁎ 0.00 (0.00) .14Medical condition −0.21 (0.07) .00 ⁎ −0.09 (0.29) .04 ⁎ −0.16 (0.66) .00 ⁎ 0.06 (0.01) .15

⁎ p b .05.


test whether the path from the mediator to the dependentvariable was significant, follow-up drinking was regressedon risky situations for drinking (controlling for RTC). Thisanalysis revealed a significant effect, such that identifyingarguments as a risky situation for drinking was associatedwith greater number of weeks exceeding NIAAA sensibledrinking limits (R2 = .029, p = .01, b = 0.171) and number ofbinge episodes (R2 = .042, p = .00, b = 0.206) at follow-up.Similarly, identifying boredom as a risky situation fordrinking lead to a greater number of weeks exceedingNIAAA sensible drinking limits (R2 = .039, p = .00, b =0.197) and number of binge episodes (R2 = .051, p = .00, b =0.225) at follow-up. The number of D/DD at follow-up wasnot significantly related to being at increased risk fordrinking due to argument (R2 = .002, p = .49, b = 0.045) orboredom (R2 = .002, p = .30, b = 0.143).

Next, both continuous RTC scores and target risksituations for drinking were regressed on follow-up drinkingbehaviors (i.e., controlling for continuous RTC in themediation to outcome path). These analyses confirmedcomplete mediation, as the effect of RTC was attenuatedwhen “risk during argument” (R2 = .038, p = .15, b = 0.095,outcome variable = weeks exceeding drinking limits atfollow-up; R2 = .059, p = .08, b = 0.116, outcome variable =number of binges at follow-up) and “risk when bored”(outcome = weeks exceeding drinking limits at follow-up:

R2 = .049, p = .13, b = 0.098; outcome = number of binges atfollow-up: R2 = .069, p = .06, b = 0.122) were controlledfor, but the effect of these risk factors on binges andexceeding drink limits remained significant (see Fig. 2).

In addition, the role of alcohol craving was examined, asit was predicted that increased cravings, in the week beforefollow-up measurement was obtained, might reduce thepredictive value of the RTCQ. Baron and Kenny's analyticprocedures were followed. First, it was established thatcontinuous RTC scores significantly predicted follow-upbinge episodes, D/DD, and number of weeks exceedingNIAAA sensible drinking limits. The percentage of daysdrinking at follow-up was not significantly related tocontinuous RTC scores. Second, PAST week alcoholcravings were regressed onto continuous RTC scores, R2 =.030, p = .00, b = 0.172. Third, follow-up drinking wasregressed on weekly alcohol cravings. This analysis revealeda significant effect, such that higher PAST week alcoholcravings were associated with greater number of weeksexceeding drinking limits (R2 = .052, p = .00, b = 0.229) andnumber of binge episodes (R2 = .029, p = .00, b = 0.169) atfollow-up (see Fig. 3). Because the effect of RTC scores onfollow-up drinking was not reduced to zero when controllingfor PAST craving, partial mediation is indicated. It should benoted that PAST week alcohol cravings were not signifi-cantly associated with the number of D/DD (R2 = .008,

Continuous Readiness to Change Score

At risk for

drinking when bored

# weeks exceeding

NIAAA limits at follow-up


At risk for

drinking during argument

# weeks exceeding



At risk for

drinking during argument

# of binge episodes

at follow-up


At risk for

drinking when bored

# of binge episodes

at follow-up

R2=.02, p = .02 R2=.04, p = .00

R2 = .02, p = .00, direct path: RTC to follow-up drinkingR2 =.05, p = .13, controlling for risk

R2=.02, p = .02 R2=.05, p = .00


R2=.04, p = .00 R2=.03, p = .01


R2=.04, p = .00 R2=.04, p = .00


Fig. 2. Complete mediation of situational risk on the relationship between RTC and follow-up drinking outcomes.


p = .06, b = 0.090) at follow-up. In addition, these results,including the simple direct effect on follow-up alcoholconsumption: D/DD (R2 = .005, p = .13, b = −0.070);percent drinking days (R2 = .007, p = .07, b = 0.083);


PAST cravin(i.e., self-reporte

intensity of desiredrink during previ

week)


PAST cravin(i.e., self-reporte

intensity of desiredrink during previ

week)

R2=.03, p = .00

R2 = .02, p = .00, direct path: RTC to followR2 =.06, p = .03, controlling for PAST crav

R2=.03, p = .00

R2 = .02, p = .00, direct path: RTC to followR2 =.04, p = .00, controlling for PAST crav

Fig. 3. Partial mediation of PAST craving on the relationsh

number of binge episodes (R2 = .000, p = .87, b = −0.007);or number of weeks exceeding sensible drinking limits (R2 =.000, p = .13, b = 0.068) could not be replicated with NOWalcohol craving measures.

g d to ous

# weeks exceeding


g d to ous

# of binge episodes

at follow-up

R2=.05, p = .00

-up drinkinging

R2=.03, p = .00

-up drinkinging

ip between RTC and follow-up drinking outcomes.


3.2.4. Moderators of RTCQDrinks per drinking day at the 12-month follow-up

varied significantly by participants' medical condition (R2 =.044, p = .00, b = −0.210). Those with hypertension (M =3.33, SD = .44) or diabetes (M = 2.31, SD = .29) drankmore than those with either infertility (M = 1.72, SD = .27)or osteoporosis (M = 1.42, SD = .30). To determine if theeffect of RTC on follow-up drinking behaviors wasmoderated by participants' medical condition (i.e., hyper-tension, diabetes, infertility, or osteoporosis), an interactionterm (continuous RTC × Condition) was created, tested,and found to be significant (R2 change = .02, p = .00,cumulative R2 = .10). Thus, the effect of (continuous) RTCon follow-up D/DD was dependent on the medicalcondition for which participants were seeking treatment(R2 = .034, p = .00, b = −0.185). These findings werereplicated with the original stage of change (categorical)designations in a four (medical condition: diabetes, hyper-tension, infertility, or osteoporosis) by three (originalcategorical RTC designations: precontemplation, contempla-tion, action) ANOVA.

This analysis confirmed that the effect of RTC stages wassignificant, F(3.490) = 6.395, p = .00. Post hoc analysesusing the Scheffé post hoc criterion for significance indicatedthat participants drank significantly more when they werecontemplating a change to their drinking behaviors (M =2.54; vs. precontemplative, M = 1.61; or action oriented,M = 2.11). Interestingly, there was also a main effect formedical condition, F(3.490) = 4.961, p = .002, such thatwomen with hypertension drank significantly more than theircounterparts (M = 3.33; vs. osteoporosis, M = 1.42;infertility, M = 1.72; or diabetes, M = 2.31). Finally, theinteraction, F(9, 474) = 2.343, p = .01, revealed that thosewomen with a diagnosis of hypertension drank the mostwhen they were action oriented for change (see Table 2 fora comparison of means). Thus, it would seem that theRTCQ has reduced predictive value within this uniquemedical population.

Table 2Mean number of D/DD at follow-up for RTCQ categories × Medical condition in

Medical condition RTCQ M SE Mean difference

Hypertension Precontemplation 1.67 0.17Contemplation 2.76 0.43 −1.09Action 3.28 0.25 −1.61

Diabetes Precontemplation 2.16 0.23Contemplation 3.15 0.42 −0.99Action 2.25 0.38 −0.09

Infertility Precontemplation 1.29 0.17Contemplation 2.43 0.39 −1.14Action a 1.40 0.28 −0.11

Osteoporosis Precontemplation 1.32 0.22Contemplation 1.83 0.49 −0.51Action 1.49 0.51 −0.17

a Of the participants with a diagnosis of infertility, those in who indicated thecontemplation stage of change (M difference = −1.02, p = .00). None of the other po

⁎ p = .05.

4. Discussion

Stage of change designation did not predict drinkingbehavior in the 6- to 12-month period subsequent toenrollment in this study. That is, for these 499 adultwomen with medical problems potentially exacerbated bydrinking (who were not actively seeking alcohol treatment),the RTCQ failed to predict the expected changes (i.e.,reductions) in drinking outcomes. Stage of change designa-tion, however, was significantly associated with D/DD atbaseline, whereby those in the precontemplation stage (64%)had the least D/DD, and those in the contemplation (12%)and action (24%) stages drank more. These results supportthe conclusions of others, who have found that patients whodrink more and have more severe alcohol-related problemsactually have more recognition or interest in changing theirdrinking behaviors (Williams et al., 2006).

Interestingly, when RTC scores were interpreted in acontinuous fashion (cf., Sutton, 1996), the measure signi-ficantly predicted drinking outcomes, such that as partici-pants' motivation to change drinking behaviors increased(e.g., moved from precontemplation to contemplation andaction), their drinking also increased. This suggests thathigher rates of drinking are associated with greatermotivation to reduce drinking in this non-treatment-seeking,medically at-risk sample. Moreover, these findings suggestthat the traditional, stage-based interpretation of the RTCQmay have less value in predicting one's motivation to changehigh-risk drinking. In light of these findings, it appears thatRTC is a construct that requires further study. For instance,this “readiness” may be a reflection of the severity of one'sdrinking problem, the accumulated consequences that lead tochange (e.g., increased awareness of the problematic natureof baseline drinking behaviors), or may indicate a shift inintentions to change. Hence, it remains unclear whatmechanism drives the change process and whether it trulyoccurs in a series of discrete changes or, as suggested by thecurrent data, in a continuous fashion.

teraction

(from precontemplation w/in medical condition) p n %

99 20.10 15 3.00 ⁎ 45 9

55 11.04 ⁎ 16 3.83 23 5

105 21.00 ⁎ 18 4.57 37 7

61 12.07 12 2.55 13 3

y were in the action stage of change differed significantly from those in thest hoc comparisons between contemplation and action stages were significant.


In addition, several preliminary findings about themediators and moderators of RTCQ, when used as acontinuous variable, were identified. For example, certainrisk situations (during an argument or boredom) for drinkingappeared to attenuate the association between an individual'sRTC and later drinking behaviors.

This finding is consistent with reported gender diffe-rences in motives for alcohol use, whereby women are morelikely than men to drink alcohol in response to stress andnegative emotions (Greenfield, Back, Lawson, & Brady,2010). As such, women drinking in this context mayexperience diminished motivation to reduce drinking,regardless of their RTC. For example, it is possible thatwomen who lack effective coping responses drink inresponse to stressful situations (e.g., argument) or unpleasantcircumstances (e.g., boredom), such that even if their RTC isstrong, their ability to change is greatly reduced by the lackof alternative coping skills. Hence, development of active,problem-focused coping strategies instead of avoidant,emotion-focused coping (cf., Folkman & Lazarus, 1980)may help such women reduce their alcohol use and mayserve as the basis of future research efforts.

Past week craving for alcohol was significantly associatedwith greater number of weeks exceeding drinking limits andnumber of binge episodes at follow-up. Moreover, takingpast weekly craving for alcohol into account improved thepredictive value of the RTCQ. Past week craving for alcoholwas measured using a simple Likert scale. Social desirabilitymay be a possible explanation as to why reports of PASTweek alcohol craving partially mediated the impact of RTCon future use, as compared to reports about current (NOW)craving, which did not. Some women may be reluctant orembarrassed to acknowledge current alcohol craving. Otherstudies have demonstrated that measures of craving can be auseful means of predicting drinking, although the standardinterpretation of craving still needs to be conceptualized(Flannery et al. 2003; Mezinskis, Honos-Webb, Kropp, &Somoza, 2001).

The association of RTC on later drinking was alsomoderated by the individual's medical condition, wherebywomen with diabetes or hypertension drank more than thosewith osteoporosis or infertility. A possible explanation maybe that the delayed, adverse consequences of diabetes orhypertension moderate a women's motivation (or possiblyperceptions of self-efficacy) to change her use of alcohol.Women with hypertension, in particular, appear to warrantspecial attention because their RTC is inversely related totheir drinking behavior. That is, for these women, actionorientation to change leads to greater follow-up drinking (ascompared to all other study participants). The finding thatmedical conditions may moderate RTC problem drinkingmay be important as medical providers initiate screening andintervention efforts for their risky drinking patients. A “one-size-fits-all” approach may not be helpful when the medicalproblems of women are considered. For example, medicalproviders implementing screening tools should carefully

consider how a diagnosis of illness might impact efforts atchange (e.g., motivation) and adherence to interventions.Moreover, future research should examine the mechanismsby which different illnesses impact efforts at change, so thatinterventions may effectively tailor to the individual'sunique (medical) needs.

In addition, it is of interest to note that those womendiagnosed with osteoporosis, infertility, and diabetes alldisplayed the same pattern of drinking in relation to theirRTC. That is, among these women, those in the contempla-tion stage drank significantly more than those in theprecontemplation and action stages (which did not differfrom one another). In contrast, those women diagnosed withhypertension displayed a different pattern of drinkingoutcomes. For these women, the action stage representednot only increased RTC but also greater drinking risk, aswomen diagnosed with hypertension who reported actionorientation to change their alcohol consumption drank themost (at follow-up). In fact, for women with a diagnosis ofhypertension, one's RTC was inversely related to theirdrinking outcomes. That is, those in the action stagedrank significantly more than their non-action-orientedhypertensive peers. In addition, hypertensive women in thecontemplation stage drank more than those who wereprecontemplating a change. This finding has importantimplications for understanding RTC, as well as fordeveloping interventions that target problem-drinkingwomen. Moreover, this finding suggests that those womenwho are especially vulnerable to the negative medicalconsequences of alcohol may experience an especially strongdisparity between their motivation and ability to reducedrinking. Future research may want to further examine thispossibility, as well as potential factors that may mediate therelationship between RTC and drinking outcomes for thisgrowing high-risk population.

Potential limitations to study findings include concernsabout reliance on self-reports of drinking (cf., Midanik,1982), measurement of stage of change, interpretation ofmediational findings, and conversion of the RTCQ to acontinuous score. Inaccurate self-reports of consumption arepossible (particularly when we consider the retrospectivenature of the TLFB measure), as well as reactivity to researchprotocols and regression to the mean (Clifford & Maisto,2000; Cunningham, 2006; Finney, 2007). Ceiling effects forRTC may be associated with the self-reported levels ofalcohol consumed by the women. In addition, the media-tional findings should be interpreted with caution when weconsider that the PAST craving measure was a self-reportretrospective account collected in tandem with the NOWcraving assessment (during the 12-month follow-up).Moreover, there is some difficulty in interpreting themediational findings due the fact that the same variablemay serve as both a mediator and a moderator. For instance,risk situations for drinking might serve both roles. First, as amediator, RTC might lead to higher drinking outcomesbecause it facilitates greater drinking during boredom and/or


argument, which subsequently obstructs change. Second, asa moderator, RTC might be related to higher drinkingoutcomes especially for women who drink when bored and/or arguing because this tendency makes their readinessdifficult to realize. Several assumptions underlie thedemonstrated tests of mediation, including assumptions ofthe correct specification of causal ordering and causaldirection. For these analyses, we assumed both temporalstability (i.e., the value of an observation does not depend onwhen the treatment is delivered) and causal transience (i.e.,the effect of a treatment or of a prior measurement does notpersist over time). Furthermore, the fact that the measure-ments were taken at times that reflect causal action (i.e., RTCwas obtained prior to proposed mediators, which weremeasured before the follow-up drinking outcomes) providesadditional support to our mediational findings. Still, theseinterpretations are based primarily on nonstatistical ratio-nales (due to the fact that RTC is observed rather thanrandomly assigned) and therefore difficult to defend. Assuch, these preliminary findings may be temporallyconfounded and should be interpreted with caution. Relatedto this is the fact that the information about drinking goals,risk situations for drinking, and behaviors to cope withdrinking was collected only for the women receiving thebrief intervention. Hence, this information was unavailablefor the women receiving treatment as usual. Studyparticipants may have been more motivated than averagepatients with similar health problems, although those withdiabetes and hypertension had the least reductions in overalldrinking behaviors. Those with infertility may have becomepregnant and so became abstinent for the sake of the fetus.Finally, stage of change was not measured at follow-up, sostage progression could not be ascertained. These potentiallimitations notwithstanding, the results of this studyenhance our understanding of RTC. For example, possiblemediators and moderators of change stage were identified,which may help to explain why RTC is not necessarilyitself a mechanism of behavior change (Borsari et al.,2009). Past week craving for alcohol, the risk of drinkingduring arguments or periods of boredom, and a person'smedical problems all appeared to be important factorsinfluencing alcohol consumption in this sample of 499women. These findings may point to future directions forinventions with problem-drinking women who are vulne-rable to the negative medical consequences of alcohol andwhose numbers are increasing (Baraona et al., 2001; Keyeset al., 2008).

Moreover, the findings presented here suggest that theRTCQ is not measuring discrete stages of change, andtherefore, converting RTCQ stages to continuous scoresmay increase the predictive value of the measure. Inaddition, it is possible that behavior change itself is acontinuous process rather than a progression throughdiscrete stages. Therefore, future work might examine therelationship between a continuous model of RTC andexisting models of behavior change (e.g., the Theory of

Planned Behavior; Ajzen, 1985), as this may provide agreater understanding of those factors that determinemotivation to change drinking behaviors.

References

Allen, J. P., & Columbus, M. (1995). Assessing alcohol problems: A guidefor clinicians and researchers. NIAAA treatment handbook series 4.Bethesda, MD: U.S. Department of Health and Human Services, PublicHealth Service, National Institutes of Health, National Institute onAlcohol Abuse and Alcoholism (NIH publication no. 95-3745).

Ajzen, I. (1985). From intentions to actions: A theory of planned behavior.In J. Kuhl, & J. Beckman (Eds.), Action-control: From cognition tobehavior (pp. 11−39). Heidelberg: Springer.

Baraona, E., Abittan, C. S., Dohmen, K., Moretti, M., Pozzato, G., Chayes,Z. W., et al. (2001). Gender differences in pharmacokinetics of alcohol.Alcoholism: Clinical and Experimental Research, 25, 502−507.

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variabledistinction in social psychological research: Conceptual, strategic, andstatistical considerations. Journal of Perssonality and Social Psychology,51, 1173−1182.

Borsari, B., Murphy, J. G., & Carey, K. B. (2009). Readiness to change inbrief motivational interventions: A requisite condition for drinkingreductions. Addictive Behaviors, 34, 232−235.

Budd, R. J., & Rollnick, S. (1996). The structure of the Readiness to ChangeQuestionnaire: A test of Prochaska and DiClemente's TranstheoreticalModel. British Journal of Health Psychology, 1, 365−376.

Chang, G. (2009). Screening and brief intervention for risk drinking women.In G. Chang (Ed.), Treating women with alcohol use disorders.Symposium conducted at the American Psychiatric AssociationMeeting, San Francisco, CA.

Chang, G., Fisher, N. D. L., Hornstein, M. D., Jones, J. A., & Orav, E. J.(2010). Identification of risk drinking women: T-ACE or the medicalrecord. Journal of Women's Health, 19, 1933−1939.

Chang, G., McNamara, T., Wilkins-Haug, L., & Orav, E. J. (2007). Stages ofchange and prenatal alcohol use. Journal of Substance Abuse Treatment,32, 105−109.

Clifford, P. R., &Maisto, S. A. (2000). Subject reactivity effects and alcoholtreatment outcome research. Journal of Studies on Alcohol, 61,787−793.

Cohen, S., & Williamson, G. (1988). Perceived Stress in a probabilitysample of the United States. In S. Spacapan, & S. Oskamp (Eds.), Thesocial psychology of health (pp. 31−67). Newbury Park, CA: Sage.

Cunningham, J. A. (2006). Regression to the mean: What does it mean?Alcohol, 41, 580.

DiClemente, C. C., Schlundt, D., & Gemmell, L. (2004). Readiness andstages of change in addiction treatment. American Journal onAddictions, 13, 103−119.

Finney, J. W. (2007). Regression to the mean in substance use disordertreatment research. Addiction, 103, 42−52.

First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (2002).Structured clinical interview for DSM-IV TR Axis I Disorders—Non-Patient Edition (SCID-I/NP 11/2002 Revision). New York StatePsychiatric Institute.: Biometrics Research Department.

Flannery, B. A., Poole, S. A., Gallop, R. J., & Volpicelli, J. R. (2003).Alcohol craving predicts drinking during treatment: An analysis of threeassessment instruments. Journal of Studies on Alcohol and Drugs, 64,120−126.

Folkman, S., & Lazarus, R. S. (1980). An analysis of coping in a middle-aged community sample. Journal of Health and Social Behavior, 21,219−239.

Forsberg, L., Halldin, J., & Wennberg, P. (2003). Psychometric propertiesand factor structure of the Readiness to Change Questionnaire. Alcoholand Alcoholism, 38, 276−280.

Freyer, J., Tonigan, J. S., Keller, S., Rumpf, H. J., John, U., & Hapke, U.(2005). Readiness for change and readiness for help-seeking: A


composite assessment of client motivation. Alcohol and Alcoholism, 40,540−544.

Gavin, D. R., Sobell, L. C., & Sobell, M. B. (1998). Evaluation of theReadiness to Change Questionnaire with problem drinkers in treatment.Journal of Substance Abuse, 10, 53−58.

Greenfield, S. F., Back, S. E., Lawson, K., & Brady, K. T. (2010). Substanceabuse in women. Psychiatric Clinics in North America, 33, 339−355.

Hannover, W., Thurian, J. R., Hapke, U., Rumpf, J., Meyer, C., & John, U.(2002). The Readiness to Change Questionnaire in subjects withhazardous alcohol consumption, alcohol misuse, and dependence in ageneral population survey. Alcohol and Alcoholism, 37, 362−369.

Hasin, D. S., Stinson, F. S., Ogburn, E., & Grant, B. F. (2007). Prevalence,correlates, disability, and comorbidity of DSM-IV alcohol abuse anddependence in the United States. Archives of General Psychiatry, 64,830−842.

Heather, N., Gold, R., & Rollnick, S. (1991). Readiness to ChangeQuestionnaire: User's manual. Technical Report 15. Kensington,Australia: National Drug and Alcohol Research Centre, University ofNew South Wales.

Heather, N., Rollnick, S., & Bell, A. (1993). Predictive validity of theReadiness to Change Questionnaire. Addiction, 88, 1667−1677.

Huebner, R. B., & Tonigan, J. S. (2007). The search for mechanisms ofbehavior change in evidence-based behavioral treatments for alcohol usedisorders: Overview. Alcoholism: Clinical and Experimental Research,31, 1−3.

Kaysen, D. L., Lee, C. M., LaBrie, J. W., & Tollison, S. J. (2009). Readinessto change drinking behavior in female college students. Journal ofStudies on Alcohol and Drugs, Supplement, 16, 106−114.

Keyes, K. M., Brant, D. F., & Hasin, D. F. (2008). Evidence for a closinggender gap in alcohol use, abuse, and dependence in the United Statespopulation. Drug and Alcohol Dependence, 93, 21−29.

Mezinskis, J. P., Honos-Webb, L., Kropp, F., & Somoza, E. (2001). Themeasurement of craving. Journal of Addictive Diseases, 20, 67−85.

Midanik, L. (1982). The validity of self-reported alcohol consumption andalcohol problems: A literature review. British Journal of Addiction, 77,357−382.

Nolen-Hoeksema, S. (2004). Gender difference in risk factors andconsequences for alcohol use and problems. Clinical PsychologyReview, 24, 981−1010.

Rollnick, S., Heather, N., Gold, R., & Hall, W. (1992). Development of ashort Readiness to Change Questionnaire for use in brief opportunisticinterventions. British Journal of Addiction, 87, 743−754.

Share, D., McCrady, B., & Epstein, E. (2004). Stage of change anddecisional balance for women seeking alcohol treatment. AddictiveBehaviors, 29, 525−535.

Sobell, L. C., & Sobell, M. B. (1992). Timeline Followback: A technique forassessing self-reported ethanol consumption. In J. Allen, & R. Z. Litten(Eds.), Measuring alcohol consumption: Psychosocial and biologicalmethods (pp. 41−72). Totowa, NJ: Humana Press.

Sokol, R. J., Martier, S. S., & Ager, J. W. (1989). The T-ACE questions:Practical prenatal detection of risk-drinking. American Journal ofObstetrics and Gynecology, 160, 863−871.

Stein, L. A. R., Minugh, P. A., Longabaugh, R. L., Wirtz, P., Baird, J.,Nirenberg, T. D., et al. (2009). Readiness to change as a mediator of theeffect of a brief motivational intervention on post-treatment alcohol-related consequences of injured emergency department hazardousdrinkers. Psychology of Addictive Behaviors, 23, 185−195.

Sutton, S. (1996). Can “stages of change” provide guidance in the treatmentof addictions? A critical examination of Prochaska and DiClemente'smodel. In G. Edwards, & C. Dare (Eds.), Psychotherapy, psychologicaltreatments, and the addictions (pp. 189−205). Cambridge: CambridgeUniv. Press.

U.S. Department of Health and Human Services, National Institutes ofHealth, & National Institute on Alcohol Abuse and Alcoholism. 2007(Reprinted May 2007, Updated 2005 Edition). Helping patients whodrink too much: A clinician's guide (NIH Publication No. 07-3769).

Willenbring, M. L. (2007). A broader view of change in drinking behavior.Alcoholism: Clinical and Experimental Research, 31, 84−86.

Williams, E. C., Kivlahan, D. R., Saitz, R., Merrill, J. O., Achtmeyer, C. E.,McCormick, K. A., et al. (2006). Readiness to change in primary carepatients who screened positive for alcohol misuse. Annals of FamilyMedicine, 4, 213−220.

readiness to change and risk drinking women

Documents