Clinical Epidemiology of Single Versus Multiple Substance Use Disorders: Polysubstance Use Disorder : Medical Care

Secondary Logo

Journal Logo

Original Research

Clinical Epidemiology of Single Versus Multiple Substance Use Disorders

Polysubstance Use Disorder

Bhalla, Ish P. MD*; Stefanovics, Elina A. PhD*,†; Rosenheck, Robert A. MD*,†,‡

Author Information
Medical Care 55():p S24-S32, September 2017. | DOI: 10.1097/MLR.0000000000000731
  • Free


Substance use and addiction have long been recognized as major public health and clinical problems with concern about opiate use becoming a focus of special urgency in recent years.1 Although most research on addiction has focused on one or another substance of abuse, many people use more than 1 substance.2 Empirical research on people who have problems with multiple substances has been limited, although some evidence suggests that polysubstance use may be increasing.3 Diagnostic and statistical manual (DSM)-IV included a specific diagnostic category of polysubstance dependence to identify people who abuse multiple drugs, but this diagnosis has not been operationally defined and has not been the subject of research attention.

Demographically, polysubstance use has been associated with young age,4–6 childhood maltreatment,7 low education,6 socioeconomic disadvantage,8 and male sex.6,7 The few studies that have investigated the clinical impact of polysubstance use report associations with higher rates of lifetime suicide attempts,9 infection,4 incarceration,4 deviant behaviors,4 arrests,10 medical, financial, and legal problems,10 and depression.5 It has been reported that use of multiple substances results in more severe medical11 and psychiatric12 comorbidities as well as more risky social behaviors.13

However, these studies have not been based on any standardized definition and have been limited by relatively small sample sizes, reliance on self-reported data, and limited focus on selected populations at especially high risk. We sought to consider whether patients with multiple diagnosed SUDs ought to be identified as a distinctive group and studied systematically under the name of polysubstance use disorder (PSUD).

As the largest integrated health care provider in United States, the Veterans Health Administration (VHA) treats over 5.5 million US veterans annually. Over 8% are diagnosed with substance abuse or dependence under DSM-IV, now termed substance use disorders (SUDs) under DSM-5. We use the latter designation in this study, although the data were gathered while DSM-IV terminology was in use in VHA. VHA addresses substance use as well as both psychiatric and general medical problems and thus may provide unique information on the prevalence, comorbidities and service use correlates of PSUD. Even though US veterans have similar rates of SUDs as the general population,14 VHA substance abuse treatment programs have been reported to provide more specialized substance abuse treatment than their non-VHA counterparts.15

In this study, we use national VHA data from fiscal year (FY) 2012 to compare veterans who carried a diagnosis of 1 SUD to veterans diagnosed with 2 or 3 SUDs, and with those diagnosed with more than 3 such disorders on sociodemographic characteristics, comorbid psychiatric and medical diagnoses, and measures of health service and psychotropic medication use. We thus hope to identify and describe the basic clinical epidemiology of a previously undefined and understudied group of potentially highly vulnerable patients who may require a special array of coordinated psychiatric, medical and social services.



Of the 5,774,903 veterans who used VHA services in FY 2012, 8.2% (n=472,642) had a diagnosis of at least 1 DSM-IV SUD and 2.2% (n=126,313) had PSUD including abuse or dependence on alcohol [International Classification of Diseases (ICD)-9 303.xx or 305.00], opiates (ICD-9 304.0x or 305.5), cannabis (ICD-9 304.3x or 305.2), cocaine (ICD-9 304.2x or 305.6), barbiturates (ICD-9 304.1x), amphetamines (ICD-9 304.4x or 305.7), or hallucinogens (ICD-9 304.5x or 305.3). In the current study, populations were classified based on of the number of unique substances for which they had received a diagnosis: 73.3% (n=346,329) had 1 SUD; 24.0% (n=113,598) 2 or 3 SUDS, more than 3 SUDs 2.7% (n=12,715). We chose this grouping on the basis of the frequencies in the data and their intuitive clinical meaning: 73.3% had 1 substance diagnoses, 17.1% had 2, 6.9% 3 and only 2.7% had 4 or more. The choice of 3 groups allowed us to observe monotonic trends across the groups, but preserved a degree of parsimony and respected the frequencies in the data.


Data on sociodemographic characteristics, medical and psychiatric diagnoses (identified with ICD-9 codes), service use (medical and psychiatric), and psychotropic medication prescription fills were obtained from a preconstructed dataset from the Northeastern Evaluation Center created for analytic purposes based on records from the VHA Computerized Personal Record System.

Sociodemographic characteristics included age, sex, race, geographic location (rural or urban residence), Veterans Affairs (VA) pension status, service-connected disability status, and homelessness in the past year. Geographic location data were obtained through zip codes based on Rural-Urban Commuting Area classification.16 Homelessness was defined as having an ICD-9 code of V60.0 and/or use of specialized VA homeless program services in FY 2012.

Medical diagnoses were chosen based on the Charlson comorbidity index.17 The Charlson is an aggregate measure of medical comorbidity that weights and sums diagnoses together to predict 1-year mortality. Several common medical diagnoses included in the Charlson known to be highly prevalent among VHA patients were identified using ICD-9 codes. These diagnoses included myocardial infarction, peripheral vascular disease, cerebrovascular accident, chronic obstructive airway disease, diabetes mellitus, and cancer. In addition, diagnoses known to be associated with substance use including hepatic disease and human immunodeficiency virus (HIV) were also included in the analysis, as well as the Charlson index itself. A summary measure representing “any pain” was also created and is described elsewhere.18

Psychiatric diagnoses included major depressive (ICD-9 296.2-296.39) and other depressive disorders (ICD-9 300.4x, 296.9x, 301.10-301.19, 311.x), bipolar disorder (ICD-9 296.0x, 296.1x, 296.40-296.89), schizophrenia (ICD-9 295.x), posttraumatic stress disorder (PTSD) (ICD-9 309.81), and anxiety disorders (300.xx excluding 300.4). Also included in the analysis were 7 specific drug use disorders (alcohol, cannabis, cocaine, opioid, barbiturate, amphetamine, hallucinogen) and personality disorders (ICD-9 301.x).

Measures of VHA health service use were derived from clinic stop codes for different types of services. Routine service use included the number of general psychiatric, substance abuse, primary care, specialty medical, and emergency department visits. Visits were counted 1 per day in which a given stop was used. Summary variables were created to represent the mean number of outpatient encounters.

Individual codes were used to identify use of diverse subspecialty mental health services in FY 2012. These include specialized programs for homeless veterans, veterans with criminal justice system involvement, and Assertive Community Treatment referred to as Mental Health Intensive Case Management in VHA. Also included was a dichotomous variable that represented use of any mental health inpatient treatment and another that concerned use of any residential rehabilitative or halfway house treatment.

Psychotropic medications were classified into several groups including antidepressants, antipsychotics, anxiolytic/sedative/hypnotics, stimulants, and anticonvulsant/mood stabilizers. Lithium was included as a separate class because it has a different mechanism of action than other anticonvulsants that provide mood stabilization and is regarded by clinicians as fundamentally different. Summary variables were used to represent the number of all psychotropic medication fills during FY 2012 and the percentage of subjects who had any psychotropic prescription fills.


First, bivariate analyses were performed to compare veterans diagnosed with a single SUD to those having 2–3 SUDs or 4 or more SUDs on sociodemographic characteristics, medical/psychiatric diagnosis, measures of service use, and psychotropic prescription fills. In addition, patients with more than 3 SUDs were compared with those with 2 or 3 SUDs.

As the sample examined in this study is very large (including all VHA patients diagnosed with substance abuse or dependence in FY 2012) the usual significance testing would likely be uninformative, as very small differences with little clinical importance could have low P-values and thus be misleadingly identified as “significantly” different. For this reason, effect sizes were adopted to identify substantial differences between groups instead. The Cohen d was computed for continuous variables and risk ratios for categorical variables. Cohen d, calculated as the mean difference between groups divided by the pooled standard deviation has been classified into small, medium and large effect sizes19 with values of >0.2 or <−0.20 reflecting at least small differences. For dichotomous variables, risk ratios of >1.5 or <0.67 as the cutoff for substantial differences.20

Next, a single multinomial logistic regression was fit to identify the most parsimonious set of measures that independently differentiated the groups. We included a fixed effects term representing each medical center with the one seeing the greatest number of veterans as the reference group. Subjects with 1 SUD were compared with those with 2 or 3 SUDs and to those with more than 3 SUDs in 1 model. A second logistic regression compared those with >3 diagnoses to those with 2 or 3. Variables identified as substantially different across groups in bivariate comparisons were included in these analyses in addition to other variables that were included due to their conceptual relevance (eg, age). The analysis included sociodemographic and diagnostic measures (Table 3), and measures reflecting service use and psychotropic medication fills (Table 4). As both dichotomous and continuous measures were included in the analysis, standardized regression coefficients were calculated to allow comparison of effect sizes across measures.

The study was approved by the Institutional Review Board committee of the VA Connecticut Healthcare System. A waiver of informed consent was obtained as the study used administrative data and there were no patient identifiers included.

All analyses were conducted using SAS statistical software (version 9.2; SAS Institute Inc., Cary, NC).


The sample included all 472,642 veterans with at least 1 diagnosed SUD. Of these, 346,329 (73.2%) had 1 SUD; 80,828 (17.1%) had 2 SUDs and 32,770 (6.9%) had 3 SUDs for a total of 113,598 (24.0%) in this second group; and 12,715 (2.7%) had more than 3 SUDs. Altogether, 95.4% of the total sample was male, which is higher than the 93.6% of the VHA population as a whole. A total of 69.7% self-identified as white, and 26.3% as black. Of those who had only 1 SUD, a majority (77.2%) had alcohol dependence followed in frequency by opiate dependence (10.3%) and cannabis dependence (7.2%). In total, 376,520 (79.7%) had alcohol use disorders, 82,661 (17.5%) had cannabis use disorders, 99,394 (21.0%) had opioid use disorders, and 77,364 (16.4%) had cocaine use disorders.

Bivariate analysis showed younger age, black race, receipt of VA pension benefits, homelessness, and not living in isolated rural areas were associated with more SUDs (Table 1).

Bivariate Analysis of Demographics and Medical/Psychiatric Diagnoses [n (%), Mean (SD)] by Level of Polysubstance Use Disorder

Veterans diagnosed with PSUD had lower rates of peripheral vascular disease, but higher rates of liver disease and HIV. The highest level of PSUD (more than 3 SUDs) was also associated with insomnia (Table 1).

Most psychiatric and SUDs were substantially associated with both levels of PSUD on bivariate analysis. Specific SUDs associated with PSUD included all the drug use disorders except alcohol use disorder because of high base rates in all 3 groups.

Use of each of the several types of specialized mental health programs was strongly associated with both levels of PSUD, especially intensive substance abuse day programs, residential treatment, vocational rehabilitation, criminal justice system, homeless services, and mental health inpatient treatment. Veterans diagnosed with PSUD were also prescribed substantially more antidepressants and antipsychotics than those with 1 SUD (Table 2).

Bivariate Analysis of Service Utilization (n, %; Mean, SD) by Level of Polysubstance Use Disorder

Multivariate analyses highlighted the association of PSUD with black race and homelessness among the sociodemographic factors; hepatic disease among medical diagnoses; and schizophrenia, bipolar disorder, depressive disorders, and personality disorder among the psychiatric diagnoses (Table 3). Standardized regression coefficients highlight the frequent use by veterans with PSUD of inpatient and residential psychiatric treatment, as well as substance abuse and emergency department services, and antidepressant and antipsychotic medications (Table 4).

Multivariate Analysis of the Association of Demographics and Medical/Psychiatric Diagnosis With Level of Polysubstance Use Disorder
Multivariate Analysis of the Association of Service Use and Psychotropic Medication Fills With Level of Polysubstance Use Disorder


This national study of VHA service users compared veterans diagnosed with PSUD who had 2–3 and 3 or more diagnosed SUDs. We found substantial differences between these groups suggesting more numerous and serious comorbidities and greater service use among veterans with PSUD, especially at high levels of comorbidity. Black race and homelessness were among the sociodemographic factors most strongly associated with PSUD. Patients with PSUD had more medical problems that may have been aggravated by severe psychiatric diagnoses (most notably the more severe schizophrenia and bipolar disorder diagnoses) and personality disorders. In addition, they used more specialized mental health services, especially services targeting substance abuse and serious social dysfunction, and more psychotropic medications. In virtually all cases, these indicators became more pronounced with higher levels of PSUD. Although many of these results are consistent with previous studies of people who use multiple substances, to our knowledge no previous study has examined multiple SUDs.

Many of the observed relationships to SUDs have been noted in previous studies, but they have not been considered together within a common PSUD conceptual framework. For example, the association of black race and homelessness with PSUD has been observed by many other researchers in both the general population21 and in a veteran cohort.22 One study specifically showed that veterans who used alcohol and another substance concurrently were at a higher risk for homelessness than those with alcohol problems alone.22 Also consistent with previous studies are results suggesting that PSUD is associated with more severe medical11 and psychiatric comorbidities.12 Intoxication with cocaine can lead to seizures,23 and use of alcohol is strongly associated with hepatic disease.24 Intravenous drug use is a well-known risk factor for bloodborne infections such as HIV and viral hepatitis.25 Processes associating the use of multiple substances with medical illness are well documented.26 What is new here, again, is the broadened conceptual context.

The relationship between PSUD and psychiatric disorders can reflect causality in 2 directions. On one hand, patients may use substances chronically to self-medicate symptoms from an underlying mental illness.27 On the other hand, extensive use of substances can mimic symptoms28 and possibly even lead to the development of psychiatric disorders. Studies have found specific links between the use of opiates and depression29 and between use of cannabis with psychosis.30 Patients who abuse multiple substances may even use one of them to address symptoms of withdrawal from another. For example, patients withdrawing from alcohol may be erroneously diagnosed with an anxiety disorder and be treated with an anxiolytic or antidepressant. In contrast, severe anxiety disorder can also lead to self-mediation with alcohol or sedatives. The presence of several addictive disorders seems to be associated with more numerous psychiatric disorders suggesting an intensification of this bidirectional process. Further complicating the picture is the fact that people abusing multiple substances are less compliant with their prescribed psychotropic medications, potentially further exacerbating their acute psychiatric symptoms.31

Physicians are typically trained to diagnose and treat diseases individually, and most clinical trials are conducted on samples with little or no comorbidity. However, there is a growing awareness, that began within geriatrics, that multimorbidity is the norm in clinical practice and represents more than a simple cooccurrence of multiple medical or psychiatric problems.32,33 There has been recent recognition that multimorbidity needs to be understood as including SUDs, serious mental illness, and the psychosocial problems that arise from these comorbidities.34 However, there have been few empirical studies of this clinical phenomenon from the integrated perspective as suggested by the PSUD concept and little consideration of the treatment of psychiatric and medical disorders in the presence of PSUD. One example of this complexity is found in a recent investigation showing that although prazosin is an effective treatment for nightmares associated with PTSD,35 its utility is quite limited in those who are dually diagnosed with PTSD and alcohol use disorders.36 The data presented here suggest an urgent need for research on treatment of the kind of multimorbidity as many ineffective treatments are probably provided to patients with PSUD based on research conducted on uncomplicated and thus only partially relevant to multimorbid cases.

Veterans with PSUD use substantially more mental health services than those with a single SUD across all types of services including inpatient and residential care, and a wide range of specialized and subspecialized rehabilitative mental health services and psychotropic medications. Previous studies found that substance abuse is a considerable risk factor for suicide37 and violence,38 and those high risks may account for the increased use of residential and inpatient treatment. Treating each SUD, its medical and psychiatric correlates, and social dysfunction in isolation through a fragmented pattern of service delivery, however, may fail to adequately address these challenging clinical situations in their full complexity.39 On one hand, substance abuse screening and treatment may be crowded out by attention to other conditions.40,41 On the other hand, the portrait of the PSUD patient painted here suggests that these patients are prescribed substantially more antidepressants, antipsychotics, and opiates than others most likely by multiple providers with limited communication with each other. These medications pose a high risk for adverse drug-drug interactions and in combination with illicit substances can have hazardous sedating effects, increasing the risk of injury, falls, and traffic accidents. Coordinating and integrating care among all of the involved clinics and services may be more effective in managing the multiple concomitant treatments, and setting priorities for which conditions should be treated at what stage.42,43

There is some evidence to support the use of mental health case management to provide integrated care for patients with many psychiatric and medical problems,44 as well as in providing integrated medical care for those with comorbid medical problems.45 There has been some interest in “1-stop shopping” to integrate mental health and medical treatment,46 with some evidence of its benefit.47 Considering the immense psychosocial problems associated with PSUD, coordinated delivery of these services in a single setting may be advantageous. One way to provide such coordinated and informed care is to co-locate not only mental health and primary care,48,49 but also have a 1-stop shop for related psychosocial services as well. Co-location of the full range of these services might lead to greater utilization in a “bottom-up” fashion, where case managers work with PSUD patients provide additional services more readily accessed if they are located “under the same roof.”50

Several methodological limitations require comment. First, this study relied entirely on administrative diagnoses that have not been independently validated with formal research tools. Although these data could be biased by either over or under diagnosis, they do reflect the perceptions of clinicians about the patients they are treating and thus have important face validity. Second, while the data presented in this study are from FY 2012 and may seem dated, we would not expect there to be substantial differences if a more recent dataset were used. It must also be further acknowledged that many of the relevant symptoms and functional consequences of PSUD may be undocumented by administrative data and require further study through primary data collection. At this point, it is difficult to determine whether PSUD is a distinct emergent entity with a fundamentally unique constellation of medical, psychiatric and functional impairments or if it represents an additive syndrome, whereby people are worse off because of their cumulative exposure to harmful substances.

PSUD is seen in only 2.2% of the VHA patient population, and clinicians may therefore not be able to recognize it as a unique phenomenon unless the data are aggregated over many facilities, as we have done here. Having a name for this phenomenon may aid in its identification and study. Finally, although the VHA data are national in scope, women and younger people are underrepresented in comparison with the US population. Hence, generalizability of results is unknown.

Veterans with PSUD have exceptionally serious psychiatric and medical comorbidities, have multiple social dysfunctions, fill more psychotropic medication prescriptions and receive more specialized VA services than other veterans with SUD. We argue that PSUD and its resulting social dysfunctions should be seen through the evolving framework of multimorbid chronic. This study has newly identified a not uncommon phenomenon with very serious adverse consequences. Much needed research on its treatment has only just begun.


1. Kolodny A, Courtwright DT, Hwang CS, et al. The prescription opioid and heroin crisis: a public health approach to an epidemic of addiction. Annu Rev Public Health. 2015;36:559–574.
2. Morley KI, Ferris J, Winstock AR, et al. Polysubstance use and misuse or abuse of prescription opioid analgesics: a multi-level analysis of international data. Pain. 2017. [Epub ahead of print].
3. Dart RC, Surratt HL, Cicero TJ, et al. Trends in opioid analgesic abuse and mortality in the United States. N Engl J Med. 2015;372:241–248.
4. Hedden SL, Martins SS, Malcolm RJ, et al. Patterns of illegal drug use among an adult alcohol dependent population: results from the National Survey on Drug Use and Health. Drug Alcohol Depend. 2010;106:119–125.
5. Midanik LT, Tam TW, Weisner C. Concurrent and simultaneous drug and alcohol use: results of the 2000 National Alcohol Survey. Drug Alcohol Depend. 2007;90:72–80.
6. Quek L-H, Chan GCK, White A, et al. Concurrent and simultaneous polydrug use: latent class analysis of an Australian nationally representative sample of young adults. Front Public Health. 2013;61:1–9.
7. Armour C, Shorter GW, Elhai JD, et al. Polydrug use typologies and childhood maltreatment in a nationally representative survey of Danish young adults. J Stud Alcohol Drugs. 2014;75:170–178.
8. Redonnet B, Chollet A, Fombonne E, et al. Tobacco, alcohol, cannabis and other illegal drug use among young adults: the socioeconomic context. Drug Alcohol Depend. 2012;121:231–239.
9. Smith GW, Farrell M, Bunting BP, et al. Patterns of polydrug use in Great Britain: findings from a national household population survey. Drug Alcohol Depend. 2011;113:222–228.
10. Herbeck DM, Brecht M-L, Lovinger K, et al. Poly-drug and marijuana use among adults who primarily used methamphetamine. J Psychoactive Drugs. 2013;45:132–140.
11. De Alba I, Samet JH, Saitz R. Burden of medical illness in drug- and alcohol-dependent persons without primary care. Am J Addict. 2004;13:33–45.
12. Mertens JR, Lu YW, Parthasarathy S, et al. Medical and psychiatric conditions of alcohol and drug treatment patients in an HMO: comparison with matched controls. Arch Intern Med. 2003;163:2511–2517.
13. Andrade LF, Carroll KM, Petry NM. Marijuana use is associated with risky sexual behaviors in treatment-seeking polysubstance abusers. Am J Drug Alcohol Abuse. 2013;39:266–271.
14. Golub A, Vazan P, Bennett AS, et al. Unmet need for treatment of substance use disorders and serious psychological distress among veterans: a nationwide analysis using the NSDUH. Mil Med. 2013;178:107–114.
15. National Survey of Substance Abuse Treatment Services (N-SSATS) [database online]. Rockville, MD: Substance Abuse and Mental Health Data Archive; 2014.
16. Morrill R, Cromartie J, Hart G. Metropolitan, urban, and rural commuting areas: toward a better depiction of the United States settlement system. Urban Geogr. 1999;20:727–748.
17. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383.
18. Barry DT, Sofuoglu M, Kerns RD, et al. Prevalence and correlates of co-prescribing psychotropic medications with long-term opioid use nationally in the Veterans Health Administration. Psychiatry Res. 2015;227:324–332.
19. Cohen J. Statistical Power Analysis for the Behavioral Sciences. Routledge: Lawrence Erlbaum Associates; 1988.
20. Ferguson CJ. An effect size primer: a guide for clinicians and researchers. Prof Psychol Res Pract. 2009;40:532–538.
21. Greenberg GA, Rosenheck RA. Mental health correlates of past homelessness in the National Comorbidity Study Replication. J Health Care Poor Underserved. 2010;21:1234–1249.
22. Tsai J, Kasprow WJ, Rosenheck RA. Alcohol and drug use disorders among homeless veterans: prevalence and association with supported housing outcomes. Addict Behav. 2014;39:455–460.
23. Pascual-Leone A, Dhuna A, Altafullah I, et al. Cocaine-induced seizures. Neurology. 1990;40:404–407.
24. Ghabril M, Chalasani N, Björnsson E. Drug-induced liver injury: a clinical update. Curr Opin Gastroenterol. 2010;26:222–226.
25. Garfein RS, Vlahov D, Galai N, et al. Viral infections in short-term injection drug users: the prevalence of the hepatitis C, hepatitis B, human immunodeficiency, and human T-lymphotropic viruses. Am J Public Health. 1996;86:655–661.
26. Palepu A, Tyndall M, Yip B, et al. Impaired virologic response to highly active antiretroviral therapy associated with ongoing injection drug use. J Acquir Immune Defic Syndr. 2003;32:522–526.
27. Harris KM, Edlund MJ. Self-medication of mental health problems: new evidence from a national survey. Health Serv Res. 2005;40:117–134.
28. Quello SB, Brady KT, Sonne SC. Mood disorders and substance use disorder: a complex comorbidity. Sci Pract Perspect. 2005;3:13–21.
29. Scherrer JF, Svrakic DM, Freedland KE, et al. Prescription opioid analgesics increase the risk of depression. J Gen Intern Med. 2014;29:491–499.
30. Caspi A, Moffitt TE, Cannon M, et al. Moderation of the effect of adolescent-onset cannabis use on adult psychosis by a functional polymorphism in the catechol-O-methyltransferase gene: longitudinal evidence of a gene X environment interaction. Biol Psychiatry. 2005;57:1117–1127.
31. Tucker JS, Burnam MA, Sherbourne CD, et al. Substance use and mental health correlates of nonadherence to antiretroviral medications in a sample of patients with human immunodeficiency virus infection. Am J Med. 2003;114:573–580.
32. Boyd CM, Martin Fortin MD. Future of multimorbidity research: how should understanding of multimorbidity inform health system design? Public Health Rev. 2010;32:1.
33. Tinetti ME, Basu J. Research on multiple chronic conditions: where we are and where we need to go. Med Care. 2014;52(suppl 3):S3–S6.
34. North CS, Brown ES, Pollio DE. Expanded conceptualization of multimorbidity to encompass substance use disorders and other psychiatric illness. Ann Clin Psychiatry. 2016;28:182–188.
35. Raskind MA, Peskind ER, Kanter ED, et al. Reduction of nightmares and other PTSD symptoms in combat veterans by prazosin: a placebo-controlled study. Am J Psychiatry. 2003;160:371–373.
36. Petrakis IL, Desai N, Gueorguieva R, et al. Prazosin for veterans with posttraumatic stress disorder and comorbid alcohol dependence: a clinical trial. Alcohol Clin Exp Res. 2016;40:178–186.
37. Hall RC, Platt DE, Hall RC. Suicide risk assessment: a review of risk factors for suicide in 100 patients who made severe suicide attempts. Evaluation of suicide risk in a time of managed care. Psychosomatics. 1999;40:18–27.
38. Swartz MS, Swanson JW, Hiday VA, et al. Violence and severe mental illness: the effects of substance abuse and nonadherence to medication. Am J Psychiatry. 1998;155:226–231.
39. Watkins KE, Burnam A, Kung FY, et al. A national survey of care for persons with co-occurring mental and substance use disorders. Psychiatr Serv. 2001;52:1062–1068.
40. Edlund MJ, Unützer J, Wells KB. Clinician screening and treatment of alcohol, drug, and mental problems in primary care: results from healthcare for communities. Med Care. 2004;42:1158–1166.
41. Friedmann PD, McCullough D, Chin MH, et al. Screening and intervention for alcohol problems. A national survey of primary care physicians and psychiatrists. J Gen Intern Med. 2000;15:84–91.
42. Druss BG, Rohrbaugh RM, Levinson CM, et al. Integrated medical care for patients with serious psychiatric illness: a randomized trial. Arch Gen Psychiatry. 2001;58:861–868.
43. Gilmer TP, Henwood BF, Goode M, et al. Implementation of integrated health homes and health outcomes for persons with serious mental illness in Los Angeles county. Psychiatr Serv. 2016;67:1062–1067.
44. Ziguras SJ, Stuart GW. A meta-analysis of the effectiveness of mental health case management over 20 years. Psychiatr Serv. 2000;51:1410–1421.
45. Druss BG, Esenwein von SA, Compton MT, et al. A randomized trial of medical care management for community mental health settings: the Primary Care Access, Referral, and Evaluation (PCARE) study. Am J Psychiatry. 2010;167:151–159.
46. Pourat N, Hadler MW, Dixon B, et al. One-stop shopping: efforts to integrate physical and behavioral health care in five California community health centers. Policy Brief UCLA Cent Health Policy Res. 2015;PB2015-1:1–11.
47. Wells K, Sherbourne C, Schoenbaum M, et al. Five-year impact of quality improvement for depression: results of a group-level randomized controlled trial. Arch Gen Psychiatry. 2004;61:378–386.
48. Guerrero EG, Heslin KC, Chang E, et al. Organizational correlates of implementation of colocation of mental health and primary care in the Veterans Health Administration. Adm Policy Ment Health. 2015;42:420–428.
49. McGuire J, Gelberg L, Blue-Howells J, et al. Access to primary care for homeless veterans with serious mental illness or substance abuse: a follow-up evaluation of co-located primary care and homeless social services. Adm Policy Ment Health. 2009;36:255–264.
50. McGuire J, Rosenheck R, Burnette C. Expanding service delivery: does it improve relationships among agencies serving homeless people with mental illness? Adm Policy Ment Health. 2002;29:243–256.

substance use disorder; multimorbidity; multiple chronic conditions

Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.