Oup Alcalc Agx061 1..7 ++

Home , Italian Parliament

Prevalence of Polysubstance Abuse and Dual Diagnosis in Patients Admitted to Alcohol Rehabilitation Units for Alcohol-Related Problems in Italy: Changes in 15 Years — A. Zambon, C. Airoldi, G. Corrao, M. Cibin, D. Agostini, F. Aliotta, M. Movalli, F. Biondini, P. Bizzi, G. Zucchi, G. Cerizza, A. Dattola, A. Marmora, G. Vittadini, F. Girardi, L. Liberto, I.M. Hinnenthal, A. Jaretti Sodano, P. Vizzuso, E. Majolino, D. Mioni, L. Pedretti, P. Ranaletti, M. Forghieri, G. Spolaor, I. Giorgi

centroSoranzo Alcohol and Alcoholism, 2017, 1–7 doi: 10.1093/alcalc/agx061 Article

Article Prevalence of Polysubstance Abuse and Dual Diagnosis in Patients Admitted to Alcohol Rehabilitation Units for Alcohol-Related Problems in Italy: Changes in 15 Years A. Zambon1,*, C. Airoldi1, G. Corrao1, M. Cibin2, D. Agostini3, F. Aliotta3, M. Movalli4, F. Biondini4, P. Bizzi5, G. Zucchi5, G. Cerizza6, A. Dattola7, A. Marmora7, G. Vittadini8, F. Girardi9, L. Liberto9, I.M. Hinnenthal10, A. Jaretti Sodano11, P. Vizzuso11, E. Majolino12, D. Mioni13, L. Pedretti14, P. Ranaletti6, M. Forghieri14, G. Spolaor15, and I. Giorgi16

1Department of Statistics and Quantitative Methods, University of Milano-Bicocca, Via Bicocca degli Arcimboldi, 8, 20126 Milano (Milano), Italy, 2Dipartimento per le Dipendenze, Az. Ulss 13 del Veneto, Via Don Giacobbe Sartor, 4, 30035 Mirano (Venezia), Italy, 3Casa di Cura Villa Silvia, Viale Anita Garibaldi, 64, 60019, Senigallia (Ancona), Italy, 4Servizio per le Alcoldipendenze, Ospedale San Raffaele Turro, Via Stamira D’Ancona, 20, 20127 Milano (Milano), Italy, 5Ospedale Privato Accreditato Villa Maria Luigia, Via Montepelato nord, 41, 43022 Monticelli Terme (Parma), Italy, 6U.O. Riabilitazione Alcologica, Ospedale S. Marta, A.O. Ospedale Maggiore, Via Montegrappa, 15, 26027 Rivolta d’Adda (Crema), Italy, 7Centro di Riabilitazione Alcologica—Fondazione Stella Maris Mediterraneo— Chiaromonte (Potenza), Italy, 8Servizio di Psicologia, Istituti Clinici Scientiﬁci Maugeri, IRCCS, Via S. Maugeri, 10, 27100, Pavia (Pavia), Italy, 9U.O. Ospedale San Pancrazio, Istituto di Riabilitazione S. Stefano, Via D. Chiesa, 38062 Arco (Trento), Italy, 10Economia e Gestione Aziende Sanitarie, Università Cattolica del Sacro Cuore, Largo Francesco Vito, 1, 00161, Rome (Roma), Italy; ASL 1, Regione Liguria, Via Giovanni borea, 56, 18038 Sanremo (Imperia), Italy, 11U.O. Alcolfarmacodipendenze Presidio Ospedaliero, Fatebenefratelli, Via Fetebenefratelli, 70, 10077 San maurizio Canavese (Torino), Italy, 12Centro Terapeutico Dipendenze, Ospedale Bad Bachgart Comprensorio Sanitario, Bressanone, Italy, 13Casa di Cura Parco dei Tigli, Via Monticello, 1, 35037 Villa di teolo (Padova), Italy, 14Centro di Riabilitazione Alcologica Villa Rosa, Via F.lli Rosselli, 83, 41125 Modena (Modena), Italy, 15Centro Soranzo (www.centrosoranzo.it), Via Pezzana, 1, 30173 Tessera (Venezia), Italy, and 16U.O. Riabilitazione Alcologica, Istituti Clinici Scientiﬁci Maugeri, IRCCS, Via S. Maugeri, 10, 27100, Pavia (Pavia), Italy

*Corresponding author: Department of Statistics and Quantitative Methods, University of Milano-Bicocca, Via Bicocca degli Arcimboldi, 8, 20126 Milan, Italy. E-mail: [email protected]

Received 5 July 2016; Revised 11 August 2017; Editorial Decision 11 August 2017; Accepted 21 August 2017

ABSTRACT Aims: Alcohol abuse has long been known as a disease with social and economic burden to society. Given the complex nature of alcohol treatment, it is worthwhile to examine the change over time of patients admitted to residential alcohol abuse rehabilitation units. Methods: The data were collected from two Italian projects on alcoholics performed in the mid- 1990s (ASSALT) and in 2009 (CORRAL), respectively. Categorical variables were considered in terms of absolute and relative frequencies. Comparisons of relative frequencies between groups were assessed by means of Fisher’s exact test. Mixed logistic regression models were ﬁtted to CORRAL data to identify the predictors of the probability of being a polysubstance abuser or having a dual diagnosis. The association estimates were reported as adjusted odds ratios and relative 95% conﬁdence intervals.

Downloaded from https://academic.oup.com/alcalc/article-abstract/doi/10.1093/alcalc/agx061/4157313/Prevalence-of-Polysubstance-Abuse-and-Dual by Universita degli studi di Padova- Dipartimento di diritto Pubblico internazionale e Comunitario user on 14 September 2017 2 Alcohol and Alcoholism

Results: Compared to the mid-1990s, in 2009 patients were older (P = 0.0003), with a higher level of education (P = 0.0204), with fewer comorbidities (liver disease except cirrhosis, P < 0.0001; polyneuropathy, P = 0.0001), more frequently polysubstance abusers (P < 0.0001), affected by dual diagnosis (P < 0.0001). In 2009, the probability of being a polysubstance abuser was higher in younger and in patients with dual psychiatric diagnosis. Female gender and polysubstance abuse were positively associated to the probability of being affected by dual psychiatric diagnosis. Conclusions: The increment of patients admitted to residential programs for alcohol dependence with polysubstance abuse and/or dual psychiatric diagnoses suggests the need to pay more attention to both psychological/psychiatric interventions and internal medicine/physical rehabilitation. Short summary: The results of this study suggest that further research is needed to identify the best treatment strategy that is safe and effective for the new population of alcoholics.

INTRODUCTION Services in Northern Italy (Morandi et al.,2015) showed that 10.5% Alcohol abuse has long been known to be a significant disease, of patients had dual diagnosis (with a higher prevalence in women) representing a social and economic burden in societies. A recent and 10.6% used psychoactive substances other than alcohol (polysub- paper showed that in Italy ~5320 deaths (1530 female deaths and stance abuse). Similarly, a recent paper reported high proportions of 3790 male deaths) or 5.9% of all deaths of people aged 15–64 years alcohol users in patients affected by psychiatric diseases varying from were estimated to be alcohol-attributable (Shield et al., 2013). In 65.5% (patients with a diagnosis of schizophrenia) to 88.9% (patients 2001 the Italian Parliament approved a law aimed at the prevention, with a diagnosis of bipolar disorder) (Martinotti et al.,2014). These care and rehabilitation of alcoholics (Official Journal of the Italian patients are particularly problematic because, as suggested in litera- Republic, 2001). Since then, the Italian Ministry of Health has ture, they may be at greater risk of psychosocial and interpersonal issued an annual report on alcohol abuse treatment in Italy problems, treatment non-compliance, alcoholic relapse and attempted (Ministry of Health, 2014). These reports help to understand the and completed suicide (Fein, 2015). extent of these problems and their changing patterns. In 2000, Data from two Italian projects (ASSALT project and CORRAL 34,558 patients were referred to Public Alcoholism Services, com- project) conducted on patients admitted to RAARUs for AD or pared with 69,770 patients in 2012. The doubled number could be abuse was available and included information about both polyabuse the result of organizational improvements and increased sensitivity and psychiatric comorbidities. The first project was carried out in towards alcohol-related problems in these services. The number of the mid-1990s (Corrao et al., 1999) and the second in 2009 (Soares patients treated is nevertheless very small compared to the potential Pinto et al. 2013; Zambon et al., 2013). demand for treatment in the population: a recent survey (Rehm We analyzed the data from the ASSALT and CORRAL projects et al., 2015) found a prevalence of alcohol dependence (AD) of which referred to patients admitted to a RAARU for AD: (a) to ver- 3.4% in the Italian population between 18 and 64 years of age ify the change in the prevalence of polysubstance abusers and of (~37,600,000 people) (Istat, 2014). In other words, there are poten- patients with psychiatric comorbidities over this 15-year period; (b) tially ~1,300,000 people suffering from AD in Italy. In 2012, there to describe the changes in prevalence of other socio-demographical were 454 Public Alcoholism Services in Italy, Services for the treat- and clinical covariates in the same period; and (c) to identify the ment of alcohol problems encompass statutory facilities such as the determinants in recent years associated to the likelihood of admit- Public Ambulatory Services for addictions and Residential Alcohol ting a patient with dual diagnosis or polysubstance abuse to a Abuse Rehabilitation Units (RAARUs). Although they have different RAARU (CORRAL project). organizational structures, these services regard alcoholism as requir- ing a multidisciplinary approach which includes medical, psychological and social aspects (Klingemann et al., 1992). MATERIALS AND METHODS The Italian Ministry of Health report from 2014 showed that the number of polysubstance abusers increased from 7.3 % of AD Data sources patients in 2000 to 9.0 % in 2013. In the same period, the percentage The data for the analysis was taken from two Italian projects carried of hospitalizations completely attributable to alcohol decreased from out on the treatment assessment of AD in Italy in RAARUs. 9.2 to 4.0%. These facts are particularly important because the study In general, RAARUs include a program which provides highly carried out by the National Epidemiologic Survey on Alcohol and structured treatment, such as group therapy, individual therapy and Related Conditions (NESARC) found that consequences due to sub- alcoholism-education sessions. The therapeutic approach is inspired stance use are more severe when alcohol and other drugs are abused by both American experience of the Minnesota Model (Anderson in combination than when abused alone (Hasin and Grant, 2015). As et al., 1999), Croatian psychiatry studies (Hodolin and Corlito, evidence of the potential harm of combining alcohol and other drugs, 1996) and, lastly, German and Austrian ‘psychosomatic clinics’ these findings could be applied to other countries, including Italy. The (Hinnental and Cibin, 2011). Specific features of this model are: same source reported a decrease in the percentage of all the alcohol- shortness of hospitalization (28 days); intensity of intervention related diseases, with liver diseases remaining the most common (53% (physical, psychological and family-related); complexity of interven- in 2011). The percentage of all other diseases was very low (e.g. poly- tion itself (with provision of medical, psychological and educational neuropathy 1.9%). Comorbidity with psychiatric disorders was not services); multiprofessional intervention; elaboration of an aftercare reported. However, a recent survey based on Public Alcoholism project (Giorgi et al., 2015).

The first study sample is selected from the ASSALT project data. psychiatric diagnosis we only considered patients at their first admis- This project was launched in the mid-1990s by the Epidemiological sion to a RAARU from CORRAL data to increase homogeneity Group of the Italian Society of Alcohology (GESIA) with the main between patients. We fitted mixed logistic regression models with aim of identifying prognostic factors associated with alcoholism dual diagnosis or polyabuse condition as a dependent variable and treatment outcomes. Details of the aims, design, phases and instru- clinical/socio-demographic covariates as explanatory ones. We took ments of the ASSALT project are reported elsewhere (Corrao et al., into account the correlation between patients within the RAARU 1999). To summarize in brief, the ASSALT project was based on a assuming a variance–covariance matrix of type compound symmetry multicenter observational prospective design. During 1993–1994, which is particularly appropriate for clustered data with no time 15 Public Alcohol Services extending across the whole of Italy (14 in ordering. The association estimates between each explanatory cov- the north, 2 in the center and 2 in the south) had recruited all the ariate and the probability of being a polysubstance abuser or being patients admitted during a 1-year period for dependence or abuse of affected by dual psychiatric diagnosis were reported as adjusted alcohol. These services identified 680 patients (69% males) with a odds ratios (ORs) with the relative confidence interval (CI) at 95%. mean age of 45 years (range: 16–85). For each patient, data about We performed only two-tailed tests, and P-values of <0.05 were polysubstance abuse and dual diagnosis was evaluated by DSM-IV, considered significant. All statistical analyses were carried out using by the Cognitive Behavioral Assessment (CBA 2.0) and by the SAS software v.9.3. Symptom Distress Checklist (SCL 90) scales. The patients were then followed for a 2-year period. Among the Public Alcohol Services involved in the project there were four RAARUs (three in the north RESULTS of Italy and one in the south). Table 1 shows the distribution of the socio-demographic and clinical The second study sample consists of all the patients admitted to variables of the 120 patients taken from the ASSALT project and of a RAARU in 2009 participating in the CORRAL project, carried the 1779 patients from the CORRAL project hospitalized for AD in out with the aim of investigating the history of alcohol treatment a RAARU. The patients in the CORRAL project were older (P = and the socio-demographic and clinical variables of patients hospita- 0.0003), with a higher level of education (P = 0.0204), less fre- lized in Italian RAARUs for AD or abuse. There were 12 RAARUs quently affected by liver disease except cirrhosis (P < 0.0001) and in this project (10 in the north of Italy, 1 in the center and 1 in the polyneuropathy (P = 0.0001), more frequently abstinent in the south). Patients with multiple hospitalizations during 2009 were month prior to hospitalization (P = 0.0102), with greater support recruited at their first episode of hospitalization. For each patient, from Public Ambulatory Services for Addictions (P < 0.0001). the referent of the RAARU recorded several variables by means of a Moreover, a significant increase of patients with polysubstance 26-item questionnaire including socio-demographic variables (age at abuse (P < 0.0001) and dual diagnosis (P < 0.0001) was observed recovery, education, occupation, residence), clinical variables (date in CORRAL compared to ASSALT data. of recovery, entity/figure sending for hospitalization, date of dis- Tables 2 and 3 refer to patients from the CORRAL project at charge, causes of admission to RAARU, causes of discharge, number their first episode of hospitalization (1040 patients). Table 2 reports of previous recoveries for alcohol problems, duration of AD, poly- the association estimates (adjusted ORs and relative 95% CIs) substance abuse, duration of other dependence/abuse, alcohol intake between patient explanatory covariates and the probability of being in the month before the recovery) and treatment variable (type of a polysubstance abuser. Patients aged 41–50 years and patients old- treatment strategies during recovery). The presence of dual diagnosis er than 50 years were less likely to be polyabusers than patients was based on the variable relative to the causes of admission to younger than 40 years (respectively OR: 0.44 (0.31–0.64) for 41–50 RAARU where all physical and psychiatric comorbidities were years old and OR: 0.17 (0.10–0.29) for >50 years old). Analogously reported by means of DSM-IV or ICD9 codes. Polysubstance abuse the probability of being a polyabuser was lower in patients with was evaluated by DSM-IV. The patients were then followed for a 1- liver diseases except cirrhosis (OR: 0.62 (0.41–0.95)). On the other year period. More details are reported elsewhere (Soares Pinto et al. hand, the presence of other comorbidities excluding pathologies 2013; Zambon et al. 2013). affecting the liver and polyneuropathy and dual psychiatric diagno- The CORRAL and ASSALT projects were approved by the sis increased the probability of being a polysubstance abuser (OR: Scientific Committee of each RAARU involved in the project. 2.67 (1.71–4.15) and OR: 2.06 (1.44–2.94), respectively). The For the analysis we only considered data from admission at the results of the mixed model for dual psychiatric diagnosis are RAARUs involved in the CORRAL and ASSALT projects. We only reported in Table 3. In females the probability of dual diagnosis included patients admitted for AD problems: (a) because alcohol were higher compared to males (OR: 1.83 (1.35–2.48)), while the abusers represent a small proportion of patients admitted to a presence of cirrhosis and other comorbidities excluding pathologies RAARU (7–8%), and (b) to increase the homogeneity of patients affecting the liver and polyneuropathy was associated with a reduc- included in the analysis. tion of the probability of being affected by dual psychiatric diagnosis (OR: 0.49 (0.25–0.97) and OR: 0.68 (0.49–0.94), respectively). Statistical analysis Analysis of ASSALT and CORRAL data All categorical variables were considered in terms of absolute and rela- DISCUSSION tive frequencies. Comparisons of relative frequencies between groups The aim of this study was to describe the changes in prevalence of (ASSALT or CORRAL) were assessed by means of Fisher’s exact test. socio-demographical and clinical covariates, especially polysubstance abuse and dual diagnosis, across a 15-year period comparing Analysis of only CORRAL data two samples of patients admitted to a RAARU in the mid-1990s To identify the association between several covariates and the prob- and in 2009. The main results of this study show that significant ability of being a polysubstance abuser or a patient with dual changes in patient characteristics occurred during this period.

Table 1. Distribution of socio-demographic and clinical variables in patients recruited in the RAARUs participating to ASSALT (mid-1990s) and CORRAL (2009) projects

Variable ASSALT, no. (%) CORRAL, no. (%) Fisher test P-value

Age ≤40 years old 43 (35.83) 518 (29.12) 0.0003 41–50 years old 56 (46.67) 643 (36.14) >50 years old 21 (17.50) 618 (34.74) Gender Male 85 (70.83) 1234 (69.36) 0.8378 Female 35 (29.17) 545 (30.64) Educationa Low 30 (25.00) 267 (15.01) 0.0204 Medium 50 (41.67) 833 (46.82) High 40 (33.33) 679 (38.17) Occupation Unemployed 38 (31.67) 643 (36.14) 0.1522 Employed 68 (56.67) 849 (47.72) Retired 14 (11.67) 287 (16.13) Cirrhosis No 109 (90.83) 1676 (94.21) 0.1598 Yes 11 (9.17) 103 (5.79) Liver disease except cirrhosis No 65 (54.17) 1443 (81.11) <0.0001 Yes 55 (45.83) 336 (18.89) Polyneuropathy No 106 (88.33) 1720 (96.68) 0.0001 Yes 14 (11.67) 59 (3.32) Other comorbiditiesb No 73 (60.83) 1238 (69.59) 0.0523 Yes 47 (39.17) 541 (30.41) History of previous alcohol hospitalizations 0 76 (63.33) 1040 (58.46) 0.3381 ≥1 44 (36.67) 739 (41.54) Alcohol abstinence in the last month prior to hospitalization No 106 (83.33) 1400 (78.70) 0.0102 Yes 14 (11.67) 379 (21.30) Structure/ﬁgure sending patient to hospitalization Public Ambulatory Services for addictions 0 (0.00) 710 (39.91) <0.0001 None 12 (10.00) 95 (5.34) Family 100 (83.33) 146 (8.21) Other 8 (6.67) 828 (46.54) Polysubstance abuse No 111 (92.50) 1273 (71.56) <0.0001 Yes 9 (7.50) 506 (28.44) Dual diagnosis No 85 (70.83) 931 (52.33) <0.0001 Yes 35 (29.17) 848 (47.67)

aEducation: Low, no education or elementary school; medium, junior high; high, high school or graduation. bExcluding polyneuropathy and pathologies affecting the liver.

Firstofall,arelevantincreaseofpolysubstance dependent patients Secondly, a significant increase of individuals with dual psychi- wasobservedinpatientsadmittedtoaRAARU.Thistrendisinline atric diagnosis from mid-1990s to 2009 was observed; this finding is with recent studies which underline the growing interest in research on in line with various international reports and confirmed by the polysubstance use (Chiauzzi et al., 2013; Connor et al., 2014) since it is Italian data (Martinotti et al., 2014; Morandi et al., 2015). associated with increased mental illness, suicidal thoughts and attempts, The significant higher rates of psychiatric disorders and polysub- poor physical health and high risk of being incarcerated (Morley et al., stance abuse in the more recent sample can be explained by the fact 2015). Moreover, the observed positive association between younger that individuals with co-occurring psychiatric or polysubstance dis- age and increased probability of being a polysubstance abuser is con- orders are more likely to seek treatment than individuals with a sin- firmed by literature (Novais et al., 2016). In a sample of 178 alcohol- gle disorder because of the functional impairments and distress they dependent outpatients, a multiple regression analysis showed that age, experience from their co-occurring disorder. With regard to the in addition to educational background, occupational status, years of increase of psychiatric disorders, other possible explanations were: alcoholism and the personality trait of being open to experience were (a) the development of outpatient alcoholism programs as well as predictors of being a multiple substance user. the higher presence of self-help groups in Italy since the 1990s

Table 2. Association estimates (adjusted ORs and relative 95% CI) between patient characteristics and probability to be a polysubstance abusers among patients at ﬁrst hospitalization in a RAARUs for alcohol dependence. CORRAL project 2009

Variable Polyabusers, no. (%) No polyabusers, no. (%) OR 95% CI

Age ≤40 years old 187 (23.88) 146 (56.81) 1 – 41–50 years old 280 (35.76) 81 (31.52) 0.44 (0.31–0.64) >50 years old 316 (40.36) 30 (11.67) 0.17 (0.10–0.29) Gender Male 226 (28.86) 69 (26.85) 1 – Female 557 (71.14) 188 (73.15) 0.93 (0.64–1.34) Educationa Low 144 (18.39) 19 (7.39) 1 – Medium 361 (46.10) 145 (56.42) 1.72 (0.96–3.10) High 278 (35.50) 93 (36.19) 1.57 (0.85–2.89) Occupation Unemployed 265 (33.84) 108 (42.02) 1 – Employed 377 (48.15) 135 (52.53) 1.04 (0.74–1.47) Retired 141 (18.01) 14 (5.45) 0.91 (0.44–1.86) Cirrhosis No 741 (94.64) 245 (95.33) 1 – Yes 42 (5.36) 12 (4.67) 1.48 (0.70–3.12) Liver disease except cirrhosis No 653 (83.40) 202 (78.60) 1 – Yes 130 (16.60) 55 (21.40) 0.62 (0.41–0.95) Polyneuropathy No 755 (96.42) 255 (99.22) 1 – Yes 28 (3.58) 2 (0.78) 0.27 (0.06–1.23) Other comorbiditiesb No 539 (68.84) 215 (83.66) 1 – Yes 244 (31.16) 42 (16.34) 2.67 (1.71–4.15) Alcohol abstinence in the last month prior to hospitalization No 618 (78.93) 204 (79.38) 1 – Yes 165 (21.07) 53 (20.62) 0.94 (0.61–1.45) Structure/ﬁgure sending patient to hospitalization Public Ambulatory Services for addictions 333 (42.53) 113 (43.97) 1 – None 29 (3.70) 18 (7.00) 0.88 (0.40–1.94) Family 57 (7.28) 32 (12.45) 0.93 (0.50–1.73) Other 364 (46.49) 94 (36.58) 0.66 (0.45–0.97) Dual diagnosis No 472 (60.28) 127 (49.42) 1 – Yes 311 (39.72) 130 (50.58) 2.06 (1.44–2.94)

aEducation: Low, no education or elementary school; medium, junior high; high, high school or graduation. bexcluding polyneuropathy and pathologies affecting the liver.

(Ministry of Health, 2014), and (b) the evolution of the general alcohol use disorders are associated with increased likelihood of characteristics of the use of alcohol in Italy, which are undergoing a many psychiatric problems, including depression, post-traumatic change in the style of consumption from ‘Mediterranean’ patterns stress disorder, eating disorders and suicidality (Greenﬁeld et al., (the use, especially of wine, during meals and in a family context) to 2010; Wilsnack et al., 2014; Erol and Karpyak, 2015). ‘northern’ patterns (the use of beer and strong drinks in entertain- Furthermore, the decreased presence of serious alcohol-related ment situations, with a high presence of ‘binge drinking’), especially organic pathologies, the increase of abstinent patients before recovery in young people (Allamani et al., 2014). In a recent survey carried and the increased use of outpatient services can be explained by the out between January 2014 and April 2015, based on a sample of development of Public Ambulatory Services for addictions who take 4275 healthy subjects, aged 18–26 years and residing in different care of patients with alcohol problems before and after the hospital- Italian cities located in the north, center and south of the country, ization and the residential programs. 79.5% reported episodes of binge drinking (Martinotti et al., 2016). Residential programs are an important resource in alcoholism This is an increasing phenomenon. In fact, the number of people treatment (Heather, 1995) and recent studies (Zambon et al., 2013) who had experienced one or more episodes of drunkenness in the showed that the majority of patients admitted for residential treat- 3 months before interview increased to 6.9% of the population ment were referred by the public ambulatory services for addictions. (aged 13 and over) in 2005, compared to 4.1% in 1993 (Allamani These programs initially focused on the treatment of alcohol-related et al., 2010). Moreover, the positive association between female sex problems for patients with a low prevalence of psychiatric problems and the probability of being affected by dual psychiatric diagnosis is and the primary intervention was based, on the one hand, on med- in line with the evidence that women’s heavy-drinking patterns and ical/hospitalization interventions and, on the other hand, on

Table 3. Association estimates (adjusted ORs and relative 95% CI) between patient characteristics and probability to be affected by dual psychiatric diagnosis among patients at ﬁrst hospitalization in a RAARUs for alcohol dependence. CORRAL project 2009

Variable Dual diagnosis, No dual diagnosis, OR CI no. (%) no. (%)

Age ≤40 years old 166 (27.71) 167 (37.87) 1 – 41–50 years old 212 (35.39) 149 (33.79) 0.78 (0.55–1.10) >50 years old 221 (36.89) 125 (28.34) 0.77 (0.51–1.16) Gender Male 141 (23.54) 154 (34.92) 1 – Female 458 (76.46) 287 (65.08) 1.83 (1.35–2.48) Educationa Low 109 (18.20) 54 (12.24) 1 – Medium 310 (51.75) 196 (44.44) 0.92 (0.60–1.42) High 180 (30.05) 191 (43.31) 1.46 (0.94–2.26) Occupation Unemployed 199 (33.22) 174 (39.46) 1 – Employed 297 (49.58) 215 (48.75) 0.86 (0.64–1.16) Retired 103 (17.20) 52 (11.79) 0.84 (0.52–1.37) Cirrhosis No 558 (93.16) 428 (97.05) 1 – Yes 41 (6.84) 13 (2.95) 0.49 (0.25–0.97) Liver disease except cirrhosis No 482 (80.47) 373 (84.58) 1 – Yes 117 (19.53) 68 (15.42) 0.76 (0.52–1.12) Polyneuropathy No 580 (96.83) 430 (97.51) 1 – Yes 19 (3.17) 11 (2.49) 1.12 (0.49–2.56) Other comorbiditiesb No 405 (67.61) 349 (79.14) 1 – Yes 194 (32.39) 92 (20.86) 0.68 (0.49–0.94) Alcohol abstinence in the last month prior to hospitalization No 457 (76.29) 365 (82.77) 1 – Yes 142 (23.71) 76 (17.23) 1.01 (0.70–1.44) Structure/ﬁgure sending patient to hospitalization Public Ambulatory Services for addictions 266 (44.41) 180 (40.82) 1 – None 20 (3.34) 27 (6.12) 1.49 (0.69–3.24) Family 47 (7.85) 42 (9.52) 1.49 (0.85–2.62) Other 266 (44.41) 192 (43.54) 1.25 (0.91–1.71) Polysubstance abusers No 472 (78.80) 311 (70.52) 1 – Yes 127 (21.20) 130 (29.48) 2.09 (1.47–2.97)

aEducation: Low, no education or elementary school; medium, junior high; high, high school or graduation. bExcluding polyneuropathy and pathologies affecting the liver.

educational/motivational interventions facilitating self-help. In the The strength of this study is that data is taken from two of the last decade, the approach has changed (Zambon et al., 2013) and the most important Italian projects conducted on patients admitted to results of this study suggest the need to pay greater attention to the alcohol rehabilitation units in recent years. However, several limita- use of substances other than alcohol and to psychiatric comorbidity, tions of this paper should be underlined. Firstly, although the which entails greater importance of psychological and psychiatric ASSALT and CORRAL projects were designed as natural experi- interventions. Since the CORRAL individuals have more severe ments involving the planned observation of the usual therapeutic psychiatric and substance use problems than the ASSALT indivi- activity of the RAARUs which had participated in the projects, these duals, intervention efforts may be needed in order to keep them results cannot be generalized for the entire population of alcoholics engaged; indeed, the best treatment outcomes are achieved when treated in Italy. Secondly, we compared frequencies from two sam- psychiatric and substance use treatments are integrated and deliv- ples (ASSALT and CORRAL) of unequal sizes. To deal with this ered concurrently (Brunette et al.,2004; Drake et al., 2007). limitation we applied a non-parametric approach which should be Moreover, it is very important to develop ‘shared care’ models, used when the sample sizes of study groups are unequal. In particu- widely used for all chronic diseases, in order to reduce the gap lar, we applied Fisher’s exact test in order to also take into account between the need and the provision of the treatment, which can the problem of low cell expected values (McHugh, 2013). Moreover, only be achieved by investing in a network of comprehensive ser- we used a mixed logistic regression model to take into account the vices that are integrated within healthcare and social structures correlation between patients admitted to the same RAARU because it (Maremmani et al., 2015a, 2015b). was shown that the performance of this model does not depend on

inequality in group size (Heo and Leon, 2005). Thirdly, based on Greenfield SF, Back SE, Lawson K, et al. (2010) Substance abuse in women. CORRAL data, we performed an exploration of the associations Psychiatr Clin North Am 33:339–55. between some socio-demographic and clinical covariates and the Hasin DS, Grant BF. (2015) The National Epidemiologic Survey on Alcohol probability of being polyabusers and/or having psychiatric comorbid- and Related Conditions (NESARC) waves 1 and 2: review and summary of findings. Soc Psychiatry Psychiatr Epidemiol 50:1609–40. ities. We cannot exclude the fact that other covariates not measured Hodolin V, Corlito G. (1996) Psichiatria, Alcologia, Riabilitazione. Trento, in this project could be associated with being a polyabuser and being Italy: Centro Studi Erickson. affected by dual psychiatric diagnosis, allowing residual confounding. Heather N. (1995) Treatment approaches to alcohol problems. WHO Reg fi Finally, we chose to t the mixed logistic regression models to take Publ Eur Ser 65:1–201. into account the correlation among patients within each RAARU. It Heo M, Leon AC. (2005) Performance of a mixed effects logistic regression is known that the ORs estimated from a logistic model are systemat- model for binary outcomes with unequal cluster size. J Biopharm Stat 15: ically more extreme than the corresponding prevalence ratios for 513–26. common outcomes and caution is needed when using these findings Hinnental I, Cibin M. (2011) Il Trattamento Residenziale Breve delle to assess attributable risk, drive cost benefit analyses or shape pol- Dipendenze da Alcol e Cocaina: Il Modello Soranzo. Turin: SEED. icy goals (Lovasi et al. 2012). Because our interest was turner to Klingemann H, Takala JP, Hunt J. (1992) Cure, Care, or Control: Alcoholism Treatment in Sixteen Countries. USA: State University of evaluate potential associations, we performed a sensitivity analysis New York Press. considering a modified Poisson approach using robust error var- ISTAT, Istituto Nazionale di Statistica (2014). Popolazione Residente. www. iances without taking into account the correlation among patients. istat.it. fi fi This analysis identi ed as signi cantthesameassociations Lovasi GS, Underhill LJ, Jack D, et al. (2012) At odds: concerns raised by observed with the mixed logistic regression model. using odds ratios for continuous or common dichotomous outcomes in These findings suggest the need to pay great attention to the use of research on physical activity and obesity. Open Epidemiol J 5:13–7. substances other than alcohol and to psychiatric comorbidity, which Maremmani I, Cibin M, Pani PP, et al. (2015a) Harm reduction as entails the need to pay considerable attention to both psychological/ ‘Continuum Care’ in alcohol abuse disorder. Int J Environ Res Public psychiatric interventions and internal medicine/physical rehabilitation. Health 12:14828–41. doi:10.3390/ijerph121114828. ‘ ’ Further research is needed to identify the best treatment strategy that Maremmani I, Baselice A, BiggiO G, et al. (2015b) Continuum Care in alcohol abuse disorders. A manifesto to bridge the gap in personalisation of is safe and effective for the new population of alcoholics. treatment pathways. J Psychopathol 21:194–209. Martinotti G, Lupi M, Acciavatti T, et al. (2014) Novel psychoactive sub- CONFLICT OF INTEREST STATEMENT stances in young adults with and without psychiatric comorbidities. Biomed Res Int doi:10.1155/2014/815424. fl The authors report no con ict of interest. Martinotti G, Lupi M, Carlucci L, et al. (2016) Alcohol drinking patterns in young people: a survey-based study. J Health Psychol doi: 10.1177/ REFERENCES 1359105316667795 [Epub ahead of print]. McHugh ML. (2013) The Chi-square test of independence. Biochem Med 23: Allamani A, Beccaria F, Voller F. (2010) The puzzle of Italian drinking. 143–9. Trends in alcohol consumption, harms and policy: Italy 1990–2010. Ministry of Health (2014) Report of Health to Parliament With Regard to Nordic Stud Alcohol Drugs 27:465–78. Interventions Carried for Alcohol and Alcohol-Related Problems. Allamani A, Voller F, Pepe P, et al. (2014) Italy between drinking culture and Website of the Italian Ministry of Health. www.salute.governo.it. control policies for alcoholic beverages. Subst Use Misuse 49:1646–64. Morandi G, Periche TE, Pirani M. (2015) Mortality risk in alcoholic patients Anderson DJ, McGovern JP, DuPont RL. (1999) The origins of the in Northern-Italy: comorbidity and treatment retention: effects in a 30 Minnesota model of addiction treatment—a first person account. J Addict year follow up study. Alcohol Alcohol doi:10.1093/alcalc/agv058. Dis 18:107–14. Morley KI, Lynskey LT, Moran P, et al. (2015) Polysubstance use, mental Brunette MF, Mueser KT, Drake RE. (2004) A review of research on residen- health and high-risk behaviours: results from the 2012 Global Drug tial programs for people with severe mental illness and co-occurring sub- Survey. Drug Alcohol Rev 34:427–37. stance use disorders. Drug Alcohol Rev 23:471–81. Novais F, Pombo S, Ismail F. (2016) Predictors of multiple substance use in Chiauzzi E, DasMahapata P, Black RA. (2013) Risk behaviors and drug use: alcohol dependence: the role of personality. J Addict Res Ther doi:10. a latent class analysis of heavy episodic drinking in first-year college stu- 4172/2155-6105.1000258. dents. Psychol Addict Behav 27:974–85. Official Journal of the Italian Republic. Framework law on alcohol and alco- Connor JP, Gullo MJ, White A, et al. (2014) Polysubstance use: diagnostic hol related problems, No. 125 30.03., 2001. challenges, patterns of use and health. Curr Opin Psychiatry 27:269–75. Rehm J, Allamani A, Aubin HJ, et al. (2015) People with alcohol use disor- Corrao G, Bagnardi V, Zambon A, et al. (1999) Outcome variables in the ders in specialized care in eight different European countries. Alcohol evaluation of alcoholics’ treatment: lessons from the Italian Assessment of Alcohol 50:310–8. Alcoholism Treatment (ASSALT) Project. Alcohol Alcohol 34:873–81. Shield KD, Rehm J, Gmel G, et al. (2013) Alcohol consumption, alcohol Drake RE, Mueser KT, Brunette MF. (2007) Management of persons with dependence, and related mortality in Italy in 2004: effects of treatment- co-occurring severe mental illness and substance use disorder: program based interventions on alcohol dependence. Subst Abuse Treat Prev implications. World Psychiatry 6:131–6. Policy 8:21. Erol A, Karpyak VM. (2015) Sex and ated differences in alcohol use and its Soares Pinto SP, Zambon A, Cibin M, et al. (2013) Poliabuso e doppia diag- consequences: contemporary knowledge and future research considera- nosi nelle residenzialità alcologiche ospedaliere: caratteristiche dei pazienti tions. Drug Alcohol Depend 156:1–13. e dei percorsi terapeutici/riabilitativi. Mission 37:19–28. Fein G. (2015) Psychiatric comorbidity in alcohol dependence. Neuropsychol Wilsnack SC, Wilsnack RW, Kantor LW. (2014) Focus on: women and the Rev 25:456–75. costs of alcohol use. Alcohol Res 35:219–28. Giorgi I, Ottonello M, Vittadini G, et al. (2015) Psychological changes in Zambon A, Soares Pinto SP, Agostini D, et al. (2013) Alcohol-addiction inpa- alcohol-dependent patients during a residential rehabilitation program. tients: characteristics of patients and rehabilitation program. Minerva Neuropsychiatr Dis Treat 11:2989–96. Med 104:193–206.