Tobacco Documents Online

I I I I I I I I I I I I I I I I I I ,4ddtctwn (1993) 88, 1573-1579 NEW DATA NOTE SERIES-20 Severity of dependence: data from the DSM-IV field trials GEORGE E. WOODY,' LINDA B. COTTLERZ & JOHN CACCIOLA1 'Department of Psychiatry, Washington University School of Medicine, 4940 Children's Place, St Louis, MO 63110 & ZSubstance Abuse Treatment Unit, Philadelphia VA Medical Center, 39th and Woodland Ave S, Philadelphia, PA 19104, USA The Data Notes Series is edited by Dr Bridget Grant, Chief of Biometry of the National Institute on Alcohol Abuse and Alcoholism. Abstract Tiu concept of a dependence syndrome with graded levels of severity was originally derirxd fiom work with alcoholics. The applicability and clinical utility of the dependence syrtdmrne across a wider range of substances was examined as part of the DSM-IV feld trials. When using a criterioh count method to astess set.writy, it was found that persons cluster at diffennt sevniry levelr according to the drug on which they are dependent. Across all drug classes, severiry correlated reasonably well with measures of quantity and frequency of use and with associated problems. Tlu relationship between severity ratings and outcome was not tested, however data from other studies indicate that severity is only one of many factors that can :nfluence outcome. Introduction The concept that dependence exists in gnded levels of severity was part of the original descrip- tion of the alcohol dependence syndrome, known as the ADS (Edwards & Gross, 1976; Edwards, Arif & Hodgson, 1981; Edwards, 1986). The ADS has been found to have internal consistency and external validity, and alcohol- specific instruments have been developed to measure its elements (Stockwell er a1, 1979; Hodgson, Stockwell & Ranldn, 1978; Skinner, 1981; Skinner & Allen, 1982; Hesselbrock, Babor & Hesselbrock, 1983; Davidson, Bunting & Raistrick, 1989). Tbis work .vaa supported by a Srant from the Jotm and Catherine D. MacArthur Foundation; by NIMH Grant MH 47200-03; and by NIDA Grma DA-05186, DA-05593, and DA-05585. Some of the studies that have examined the validity of the ADS have related severity of de- pendence to potential external validators such as quantity and frequency of use, or alcohol-related physical, psychiatric, and social problems (Drummond, 1990). Others have examined severity as it relates to outcome, and some have found that it is a reasonably good predictor (Orford, Oppenheimer & Edwards, 1976; Hodg- son, Rankin & Stockwell, 1979; Skinner, 1981; Polich, Armor & Braiker, 1981; Hesselbrock a aL, 1983; Babor, Cooney & Lauerman, 1987). Related studies have been done with nicotine dependence (Breslau, Kilbey & Andreski, 1991). With few exceptions (Doherty & Webb, 1989) studies that have examined the relationships between dependence severity, alcohol con- sumption, associated problems, and outcome 1573 I

I I I I I I I I I I I I I I I I I I 1574 George E. Zoodv et al. have not focused on the three DSM-III-R sever- irv levels (mild, moderate and severe). Rather, they have usually examined the two ends of the severity connnuum by comparing subjects with severe vs. mild/moderate dependence (Hodgson et al., 1979). In addition these studies have typi- cally used substance-specific rating scales such as the Seventy of Alcohol Dependence Scale (SADQ) or the Faegerstrom (for nicotine), rather than counts of DSM or ICD diagnostic criteria. Though developed from work with alcoholics and later applied to nicotine, the Edwards con- cepts were generalized and applied to all substances in DSM-III-R and ICD-9. Even at the time DSM-III-R was published, very little work had been done to examine how well the ADS and the DSM-III-R criteria for severity applied to this wider range of substances. Subsequently, more work has been done in these areas including efforts to develop and test an instrument to measure the severity of opiate dependence. This instrument is modeled after the SADQ and given the related title: Severity of Opiate Dependence Questionnaire (SODQ). Preliminary results suggest that severity of opiate dependence may behave differently than severity of alcohol dependence, as all items on the SODQ were found to load on two factors: with- drawal symptoms, both physical and affective; and withdrawal relief (Phillips er al., 1987). In another study, (Babor et al., 1987) the SADQ and SODQ were used to measure severity of dependence in treated alcoholics and opiate ad- dicts and it was found that severity predicted outcome in alcoholics but not in opiate addicts. Kosten and colleagues (1987) examined the ap- plicabiliry of the ADS to a wide range of drugs including opiates, cocaine, alcohol, hallucino- gens, sedatives, marijuana, cocaine and other stimulants. They found that the severity concept appeared to be more applicable to some sub- stances than to others. A study that was done as part of the DSM-IV field trials (Cacciola & Woody, 1993) used a secondary analysis of data from two treatment outcome studies to examine severity among pa- tients seeking treatment for cocaine problems. The Diagnostic Interview Schedule (DIS) had been used in each study, and items from it were selected to approximate DSM-III-R dependence criteria and to then construct three severity groupings based on criterion counts. Patients were also divided into low, mid, and high levels of cocaine use, and into low, mid, and high levels of cocaine-associated problems. All groups were then compared at intake, and at one and three month followup. Dependence seventy was found to correlate with frequency and quantity of use and with cocaine-associated problems at intake, but the differences were clearly detectable only at the extremes. This finding resembled that of (Hodg- son et al., 1979) who also found only two clearly-identifiable severity levels among alco- holics. Unlike many of the studies with alcohol and nicotine, severity of dependence at intake did not predict outcome. In fact, there was some evidence that patients with severe dependence had better outcomes than those with mild or moderate dependence. This latter finding was not seen, however, in persons with antisocial personality disorder (ASPD). These patients were generally more severe than those without ASPD, and higher dependence severity was asso- ciated with poorer outcome within this subgroup. Though these studies had methodological lim- itations, when taken together they suggest that severity varies in its ability to predict outcome as a function of: (1) the drug under study; (2) the method used for rating it (i.e. drug-specific mea- sures vs a DSM symptom count); and (3) the nature and extent of associated problems (psy- chiatric, family/social, medical, etc.). These relationships may be complex, as pointed out by Hasin, Grant & Endicott (1988) and by the work of Cloninger, Sigvardsson & Bohman (1988). Though the relationship between severity and outcome is not always detectable, these and other studies have usually found that severity of dependence is well-correlated with severity of substance use and severity of associated prob- lems (Skinner, 1981; Drummond, 1990). Thus it appears that severity of dependence is a useful way of describing a patient's status at a given point in time, even though severity may not always be a major determinant of outcome. The DSM-IV Field trial for substance use disorders provided a unique opportunity to ex- tend this work to other drug classes because it used a standardized instrument to evaluate per- sons with a wide range of patterns of use. The methods used in the field trials have been de- scribed elsewhere (Cottler, 1993) and will only be summarized here. Approximately 1100 sub-

I I I I I I I I I I I I I I I I I I I Sevenr}• oi dependenue 1575 Table 1. Sez,entv of nerotn dependence b}numoer of crsurea and dose/tiequenc~ compos:rr scores Dose/frequencv(composite score Low (0-32 percentile) Intermediate ;33-66 percentile) High (67-100 percentile) Total Cnterion count n= 45 n=-12 n= 39 = 126 0-2 (No dependence) 16 (36%) 3 (7%) 4 ( 10°'0 l 23 ! 18°• o; 3-4 (Mild) 2 (4%) 4 (9%) 5 (13%) 11 (9%) 5-6 (Moderate) 5 (11%) 4 (9%) 2 (5%) 11 (9%) 7-9 (Severe) 22 (49%) 31 (74%) 28 (72%) 81 (64%) Table 2. Scvertry of opuzu dependence by number of crireria and doseJfrequnuy composue scores Dose/frequency/composite score Criterion count Low (0-32 percentile) n = 26 Intermediate (33-66 percentile) n = 25 High (67-100 percentile) n = 24 Total n= 75 0-2 (No dependence) 5 (19%) 5 (50%) 3 (13%) 13 (17%) 3-4 (Mild) 1 (4%) 2(81%) 3 (13%) 6 (8%) 5-6 (Moderate) 4 (15%) 2 (8%) 2 (8%) 8 (11%) 7-9 (Severe) 16 (61%) 16 (64%) 16 (67%) 48 (64%) jects were interviewed using a modified version of the Composite International Diagnosric Inter- view-Substance Abuse Module, known as the CIDI-SAM (Robins et aL, 1988, 1989). Those selected for study were persons who had used alcohol, nicotine or illicit drugs on at least six occasions. The ratio of treatment to non-treat- ment subjects was approximately 65/35. Most subjects were recruited from five sites in the US: Burlington, Vermont; Denver, Colorado; Philadelphia, Pennsylvania; San Diego, Califor- nia; and St Louis, Missouri. The data obtained permitted an examination of severity groupings according to counts of de- pendence criteria across drug classes. It was then possible to relate severity to independent valida- tors such as quantify and frequency of use, and substance-associated problems. Those chosen for presentation here are quantity/frequency of drug use. Methods Items 1,2,3,4,5,6,7,9, and 11 from the DSM-IV Options Book (1991) were used to assess depen- dence. These items are similar to those in DSM-III-R and are the criteria that the DSM-IV Substance Use Disorders Work Group was using during the Spring of 1992 to assess dependence (later modifications were made for the final version of DSM-IV). Severity of dependence on the seven most commonly used substances was examined. These substances were: heroin, opi- ates, cocaine, amphetamines, alcohol, cannabis, and nicotine. Heroin and opiates were examined separately though they are similar pharmacologi- cally. The rationale was that heroin is usually associated with illicit use whereas opiates are typically obtained by prescription. Data on other substances such as sedatives, hallucinogens, phencyclidine, and inhalants was available, but the sample sizes were too small to use for these analyses. A composite score measuring the intensity and frequency of use was developed for each drug using data from the CIDI-SAM. The specific formulas used for calculating the composite scores will not be described here, but they can be obtained on request from the first author. Scores reflecting severity of drug use were calculated for all subjects who had used each drug six or more times on a lifetime basis. The scores were divided into thirds, with the lowest third representing 'mild' levels of drug use, etc. Severity ratings were then developed accord- ing to a count of dependence criterion items. There were four groupings: 0-2 criteria = no de- pendence; 3-4 criteria = mild dependence; 5-6 criteria = moderate dependence; and 7-9 crite- ria = severe dependence. These groupings were

I I I I I I I I I I I 1 I I I I I I I 2576 George E. T'oodv er al. Table 3. Seventy of coca:ne dependence oy number of crnrersa and dos%frequency composite scores Dosdfrequency/composue score Criterion count Low Intermediate High (0-32 percentzle) (33-66 percentile) (67-100 percentile) Total n= 110 n=102 n=105 n=317 0-2 (No dependence) 32 (29%) 7 (7%) 6 (6%) 45 (14%) 3-4 (Mild) 16 (15%) 11 (11%) 5 (5%) 32 (10%) 5-6 (Moderate) 22 (20%) 21 (21%) 17 (16%) 60 (19%) 7-9 (Severe) 40 (36%) 63 (62%) 77 (73%) 180 (57°/.) Table 4. Severuy of asrepiutamrrse dependence by number of crueria and dosdfrequnrcy composue scores Dose/frequency/composite score Criterion count Low (0-32 percentile) n = 94 Intermediate (33-66 percentile) n=94 High (67-100 percentile) n=93 Total n=281 0-2 (No dependence) 79 (84%) 51 (54%) 29 (31%) 159 (57%) 3-4 (Mild) 12 (13%) 26 (28%) 12 (13%) 50 (18%) 5-6 (Moderate) 3 (3%) 5 (5'/.) 23 (25%) 31 (11%) 7-9 (Severe) 0 (0%) 12 (13'/.) 29 (31%) 41 (15%) then cross tabulated with ratings for the intensity of drug use. The data are seen in Tables 1-7. Results Three general findings emerged, as seen in Tables 1-7. First, a large proportion of persons with high levels of drug use met criteria for dependence; low levels of use were usually asso- ciated with not meeting criteria for dependence. This finding was seen in all drug classes includ- ing heroin; it was weakest for cannabis. However there were exceptions; some persons with high levels of drug use did not meet dependence criteria and vice versa. Second, the proportion of persons having mild, moderate or severe dependence among those who met dependence criteria differed across drug classes. Two-thirds or more of those who met dependence criteria for heroin, opiates, or cocaine had seven or more dependence crite- ria and thus were labeled as high severity. Persons meeting dependence criteria for alcohol were equally distributed between low/moderate, and high severity. In contrast, two thirds of those who met dependence criteria for amphetamines, cannabis or nicotine were found in the low/mod- erate severity range. Third, the proportion of those who had used 6 or more times and did not meet dependence criteria varied greatly across drug classes (Table 8). Only 13.3% of those who had used tobacco, 14.2% of those using cocaine, 17.3% of those using opiates and 18.3% of those using heroin did not meet dependence criteria. In contnst, 46.7% of those who had used alcohol, 56.5% of those using amphetamines and 58.7% of those using cannabis did not meet dependence criteria. Stated another way, roughly 80-90% of those who had used the former three substances met dependence criteria, while only 45-6A% of those using the latter three had become dependent. Discussion These findings are generally consistent with the concept of abuse liability. Heroin, opiates, and cocaine are considered to have a high abuse liability (the Drug Enforcement Administration classifies them as Schedule I or II drugs). Over 80% of those who used these drugs six or more times met criteria for dependence, and two thirds or more of those who met dependence criteria had severe dependence. Thus, persons who use heroin, opiates or cocaine six or more times were at considerable risk for becoming dependent, and if they become dependent they were likely to have severe dependence. In contrast, amphetamines, alcohol and can- nabis were much less likely to be associated with ,

I I I I I I I I I I I I I I I I I I I Sez•enn, of dependence 1577 Table 5. Sevenn• of alcohol dependence by number of cnrena and dosdfsequencv composite scores Dose/frdquency/composue score Critenon count Low (0-32 percentile) n = 285 Intermediate (33-66 percentile) n=286 High (67-100 perentile) n=295 Total n=866 0-2 (No dependence) 250 (88°io) 129 (45%) 25 (8°/o) 404 (47"0) 3-4 (Mild) 22 (8%) 64 (22%) 34 (12%) 120 (14%) 5-6 (Moderate) 9 (3%) 43 (15%) 55 (19%) 107 (12%) 7-9 (Severe) 4 (1%) 50 (17%) 181 (61%) 235 (27%) Table 6. Severery of cannabu dependenu -by numb" of criuria and doseJfrequencv composue scons Dose/frequency/composite score Criterion count Low (0-32 percentile) n = 152 Intermediate (33-66 percentile) n=145 High (67-100 percentile) n=136 Total n=433 0-2 (No dependence) 129 (85%) 77 (53%) 48 (35%) 254 (59%) 3-4 (Mild) 16 (11%) 31 (21%) 33 (24%) 80 (18%) 5-6 (Moderate) 3 (2%) 20 (14%) 32 (23%) 55 (13%) 7-9 (Severe) 4 (3%) 17 (12%) 23 (17%) 44 (1(r%) dependence. Among those who had become de- pendent on these drugs, one half to two thirds had low/moderate severity. The relatively low dependence severity that was seen among those using amphetamines is somewhat at odds with clinical experience, where amphetamines are considered to have a relatively higb. abuse liabil- ity (greater than alcohol but less than cocaine and heroin). However, this finding also demon- strates the usefulness of collecting information in a systematic way using a diagnostic instrament that does not allow for preconceived ideas to influence the results. Tobacco is a particularly interesting case. Over 80% of those who had used tobacco 6 or more times met dependence criteria, yet among those who were dependent 78% had mild or moderate dependence severity. It thus appears that to- bacco readily produces dependence (perhaps more so than most other substances), yet it does not progress to severe levels of dependence as readily as cocaine, heroin and most other drugs. Though it easily causes compulsive use, toler- ance, and withdrawal, tobacco may be less likely to get 'out of control' and progress to severe dependence than most other substances. Alternatively, it is possible that some persons with severe tobacco dependence are not well- identified by the DSM criteria or by their operationalization in the CIDI-SAM. The crite- ria were derived from work with alcohol and they may be less applicable to drugs that produce less adverse psychological symptoms. For example, tobacco has few or no sedative properties, espe- cially when compared to alcohol and narcotics; tobacco differs from cocaine and amphetamines Table 7. Seturuy of robacca depaidsres by ntanber of csiterra and doulJnquency coneposits uorsr Doteffrequency/composite score Criterion couat Low (0-32 perceatile) n= 243 Intermediate (33-66 percentile) n= 222 High (67-100 percentile) n= 180 Total n= 645 0-2 (No dependence) 45 (18%) 34 (14'/.) 9 (5%) 86 (13%) 3-4 (Mild) 69 (28%) 67 (30%) 39 (22%) 175 (27%) 5-6 (Moderate) 82 (34%) 86 (39°/.) 91 (51%) 259 (40%) 7-9 (Severe) 47 (19%) 37 (1T/.) 41 (23%) 125 (2(r)

I I I I I I I I I I I 1 I I I I 1578 George E. W'ooav et al. Table 8. Percent oi those rcreny srx or more times who dud not meet depenaence critena Tobacco 13.3% Cocaine 14.2% Opiates 17.3% Heroin 18.3% Alcohol -16.7% Amphetamines 56.5°!0 Cannabis 58.7% by its relative lack of stimulant properties and its inability to produce paranoid states and the other severe organic mental syndromes that are associ- ated with dependence on these drugs. Taken together, these findings suggest at least one reason that might account for variability in the relationship between severity and outcome across drug classes. Namely, that severity is less likely to exert an important influence on out- come if its range is restricted. Persons with dependence on heroin, other opiates, and co- caine have a restricted range-most are severe. Thus, it is not surprising that factors such as psychiatric symptoms (McLellan et aL, 1983; Rounsaville ct aL, 1986; Woody & McLellan, 1984) or andsocial petsonaliry disorder (Hessel- brock, Hesselbrock & Workman-Daniels, 1986; Woody et aL, 1985) will emerge as more impor- tant predictors of outcome than severity of dependence. Whether or not severity predicts outcome it probably serves as a reasonably good descriptor of patient status a given point in time. When measured by either an item count or a drug- specific scale, most studies, and the data pre- sented here in Tables 1-7, show that severity of dependence relates fairly well to measures of drug consumption. Other data, mentioned here but not reviewed in detail, have shown that severiry also relates to the presence of associated problems; persons with higher levels of depen- dence usually have more associated problems. Severity assessments may thus be useful in traatment planning. For example, patients with low dependence severity may be better candidates for outpatient or less intensive therapies than persons with high severity. Since severity of dependence is usually correlates with associated problems, it may also indicate which patients need a wide range of services and which will progress with only drug-focused treatment. References B.kBOR. T. F.. Coo\-Ev, N. L. & L-~UER.MtIN. R. J. (1987) The dependence syndrome as a psychological theory of relapse behavior: an empirical evaluation of alcoholic and opiate addicts, Bruuh Yournal of Addu- non, 82, pp. 393-405. BRESLAt:, N.. 1CILBEY, M. & A.\DREsKI. P. ;1991` Nicotine dependence. mator depression, and anxset}• in young adults, Archives of Generai Prvchtatry, 48, pp. 1069-1074. CACCto[a, J. & WooDY, G. E. (1993) Cocaine abuse vs dependence and levels of severity: a secondary analysis for DSM-IV, in: WIDIGER, T., FRANCES, A. & PtNCUS, H. (Eds) DSM-IV Sourubook (Washing- ton, DC, American Psychiatric Association). CLONnaGER, C. R, SIGVARDSSON, S. & BOHMAN, M. (1988) Childhood personality predicts alcohol abuse in young adults, Akoholunc: Clinual and 8spenmettal Ressarch, 12, pp. 494-505. CoTTlEtt, L B. (1993) Comparing DSM-III-R and ICD-10 substance use disorders, Addiction, 88, pp. 689-696. DAVIDSON, R, BtNmxG, B. & RvsTiucx, D. (1989) The homogeneiry of the alcohol dependence syn- drome: a factorial analysis of the SADD quaaonaaire, Brauk Journal of Addiction, 84, pp. 907J914. DoHSlrrt. B. & WEee, M. (1989) The distribunon of alcohol dependence severity among in-patlent prob- lem drinkets, British lournal oir Addiction, 84, pp. 917-922. DttvluMoND, D. C. (1990) The relationship between alcohol dependence and alcohol-related problems in a clinical population, Bniisk Journal of Addiction, 85, pp. 357-366. FiuNCEs, A., FtxsT, M. B., PtNcus, H. A. & WrntGFx, T. A. (Eds) (1991) DSM-IV Ophow Book.• Work in Progras (Washington, DC, American Psyctuatnc As- sociation). EDwARos, G. (1986) The alcohol dependence syn- drome: a concept as stimulus to enquiry, Brinak ,3onrnal of Addiction, 81, pp. 171-183. EDwAttDS, G. & Gxoss, M. M. (1976) Alcohol depen- dence: provisional description of a clinical syndrome, Britrah Jounml of Medicine, 1, pp. 1058- 1061. EDvARDs, G., ARti:, A. & HODGSON, R. (1981) Nomenclature and classification of drug and alcohol related problems, BtrUstve of the WHO, 59, pp. 225- 242. HASna, D. S., GtuN-t, B. & Ermtcorr, J. (1988) Sever- iry of alcohol dependence and social/occupational problems: relationship to clinical and familial his- tory, Atcoholum• Cliaical and E xperrrnextal Ressarrk, 12, pp. 660-664. HmsmBttocx, M., BwRaox, T. F. & HESSS.exocx, V. (1983) Never believe an alcoholic? On the validisy of self-report measures of alcohol dependence and re- lated consnvcts, Inurrta>yonal Jow+wi of AddieaoR, 18, pp. 593-609. HEssm4ocx, V. M., HESMsROCx, M. N. & Wowc- ntAv-DArttt+ss, K L(1986) Effect of maior depression and antisocial personality on alcoholism: course and motivational pattetns. HoncsoN, R, STocKwEL[, T. & RAxxnN, H. (1978)

I I I I I I I I I I I I I I I I I I Alcohol dependence: the concept, its uulitti• and measurement, Br:ttsh Journal of Addiction, 73, pp. 339-342. HoDGsoN, R, RAr11cu1, H. & STOCxavEU., T. (1979) Alcohol dependence and the ptimmg effect, Behavior Research & Therapy, 17, pp. 379-387. Kos7zx. T. R.. Rot.\sAw1.E, B. J.. B.soR, T. F.. SPITZER, R. L. & WtLU.4,NIs, J. B. a. !1987) Sub- stance-use disorders in DSM-III-R: evidence for the dependence syndrome across different psychoactive substances, British 3ournal of Psych=atry, 151, pp. 834-843. McLEUa.ti, A. T., LusoxsKY, L., WOODY, G. E., DRU- tfY, K. A. & O-BttrE.*t, C. P. (1983) Predicting response to alcohol and drug abuse treatatents: Role of psychiattic severity, Archives of General Prychsarry, 40 pp. 620-625. ORFORD, J., OPPE?JHEIMER, E. & EDWARDS, G. (1976) Abstinence or control: the outcome for excessive drinkers two years after consultation, Behavior Research & T herapy, 14, pp. 409-418. PHu.uPs, G. T., Gossop, M. R, EDWARDS, G., cz ai. (1987) The application of the SODQ to the mea- surement of the severity of opiate dependence in a British sample, British 3ournal of Addiction, 82, pp. 691-699. PoucH, J. M., ARntoR, D. J. & BRntxER, H. B. (1981) The Course of Alcoholism: Four Years After Treatment (New York, Wiley). Roenas, L N., Wnvc, J., WrrrcKD4, H.-U., ct ai. (1988) The Composite International Diagnostic In- terview: an epidemiologic instrument suitable for use Seventv o,f dependence I 5-9 :n coniuncuon with different d:a¢nost:c systems and :n different cultures, Archmes of General Pscch:atn. 45, pp. 1069-1077. ROBTNS, L N., Cor=, L B., Rosnvs, L N. & HEtzER, J. E. (1989) The reliability of the Com- posite International Diagnostic Interview Substance Abuse Module-.'CIDI-SA.M)-a comprehensive substance abuse interview. British 3ournai of Addu- tran, 84, pp. 801-814. Rotr.rsnvtt.tE, B. J., KosTE.*:, T. R, WEtssrinN, M. M. & KizsER, H. D. (1986) Prognostic significance of psychopathology in treated opiate addicts, Archicv,s of General Asychuury, 43, pp. 739-745. StaxxER, H. A. (1981) Primary syndromes of alcohol abuse: their measurement and correlates, Brursh Journal of Addicrion, 76, pp. 63-76. SxivrrEit, H. A. & At,ti-N, B. A. (1982) Alcohol dependence syndrome: measurement and valid- ation, Journal of AbrsorMal PsycholoV, 91, pp. 199-209. STOCKWELL, T., HoDGSON, R., EDWl1RDS, G., TAYLOR, C. & RArtKN, H. (1979) The development of a questionnaire to measure alcohol dependence, Bruish fournal of Addiction, 74, pp. 79-87. WOODY, G. E. & McLfu.Ala, A. T. (1984) Psychiatric seventy as a predictor of benefits from psychother- apy: the Penn-VA study, Arsencen: 3ourxal of Psychuury, 45, pp. 9-13. WOODY, G. E., McLELLArt, A. T., LusottsKY, L & O'BRiEta, C. P. (1985) Sociopathy and psychother- apy outcome, Archives of General Psycfttatry, 42, pp. 1081-1086.