Background
Persons in inpatient treatment for substance use disorders (SUD) have extensive service needs [
1] linked to a complexity of symptoms [
2,
3], including psychological and psychiatric burden [
4‐
6]. In particular, there is a high prevalence of co- occurring psychiatric disorders (COD) (i.e. the co-occurrence of SUD with another psychiatric disorder) in SUD inpatient treatment settings, with rates around 50% or above [
7‐
9]. The combination of substance use and psychiatric disorders may complicate shift in substance use behavior [
10] and worsen treatment prognosis [
11]. This combination is also associated with more severe features, such as more severe psychiatric symptoms, substance use, and relapse [
12‐
15]. Despite this, little is known regarding the demographic, psychological and clinical characteristics (e.g. types of SUD diagnoses and psychiatric diagnoses) of inpatients with COD, and the extent to which these factors have an impact on their treatment outcome.
A review of studies reporting the prevalence of COD within diverse SUD treatment settings in Australia [
16] concluded that most studies did not analyze demographic characteristics among patients with COD. Regarding inpatient SUD treatment settings, the few studies that compared the demographic characteristics of patients with - and without COD, revealed a predominance of females among the former group of inpatients [
7,
8]. It was also indicated that the age distribution was roughly the same across the two patient groups [
7,
12]. Results regarding education level were mixed. For example, one study found that the education level of inpatients with and without COD was comparable (7), while another study (12) reported that patients with COD had more education than those without COD.
Studies investigating psychological characteristics among patients with and without COD consistently showed that those with COD reported higher levels of general mental distress, [
12‐
14,
17]. Anxiety and depression are the two most frequently co-occurring psychiatric disorders among SUD patients [
16]. One study showed that these disorders occurred particularly frequently among female inpatients with COD [
8]. An unaddressed research question in the literature is whether the prevalence of COD varies with the specific type of SUD [
16]. As far as we know, the study by Bergman et al. (7), suggesting that patients with COD were more likely diagnosed with alcohol use disorders and polysubstance use, has been the only study to investigate types of drugs involved in COD within an inpatient SUD treatment setting. However, their study was restricted to young adult patients (18–25 years), and did not use objective diagnostic criteria, limiting the generalizability of findings.
Our prior research among a general SUD inpatient sample showed that having a co-occurring psychiatric disorder was associated with an increased relapse risk [
15]. However, the specific relapse predictors among inpatients with COD has not been investigated previously. Previous research on general SUD treatment samples has identified both demographic and psychological factors that may have an impact on SUD treatment outcome (including treatment dropout and relapse). These variables included younger age [
15,
18,
19], lower education, and unstable living arrangements [
20], whereas previous research on putative gender differences in SUD treatment outcomes revealed mixed results [
21‐
23]. Among psychological variables, both baseline motivation [
24‐
26] and baseline mental distress level (i.e. symptoms of anxiety/depression) [
24,
27‐
29] have been found to be associated with SUD treatment outcomes.
Whether different types of SUD diagnoses influence the treatment outcome of patients with COD has not been sufficiently investigated. While analyses of data from different SUD treatment programs in USA [
30] suggested higher risk of treatment non-completion among patients with COD admitted for alcohol and cannabis use, previous research did not examine whether the risk of post-treatment relapse among COD inpatients varied with particular types of SUD diagnoses.
There is a need for updated and extended knowledge about the characteristics of SUD inpatients with COD, and the factors associated with their treatment outcome. This study will advance existing research by investigating such factors among a non-selected inpatient SUD treatment sample. The aims were: (1) To examine demographic, psychological, and diagnostic characteristics of COD inpatients; (2) To investigate the risk for post-treatment relapse rates according to types of SUD diagnoses; (3) To identify patient characteristics associated with post-treatment relapse to substance use.
Based on previous research from general SUD treatment samples, we hypothesized that patients with COD would be more marginalized in terms of mental distress and post-treatment relapse rate. We also hypothesized that higher levels of intrinsic motivation for changing substance use behavior would have a moderating effect on relapse risk.
Discussion
The current study has provided updated and extended data on the characteristics of SUD patients with co-occurring psychiatric disorders, by investigating a large non-selected inpatient SUD treatment sample. The study also adds a through comparison of post-treatment relapse risk and associated factors among inpatients with and without COD.
The high prevalence of co-occurring psychiatric disorders found in our sample corresponds to previous research comprising inpatient SUD treatment populations [
7‐
9]. Also as expected [
8,
16], anxiety and depression were the most frequently occurring disorders.
As hypothesized, our data showed that patients with COD were younger, and appeared to be more socially and socio-economically deprived. The fact that patients with COD had longer inpatient treatment stays compared to patients without COD probably reflects a more challenging symptom profile and management in these patients (10, 11).
Previous research among SUD inpatients has not explored the socio-demographic characteristics of patients with COD, however the current finding correspond to reports from mental health treatment settings [
36]. We did not find a higher prevalence of COD among females as was suggested in previous work [
8,
30,
37]. It is noteworthy, however, that females had twice the risk of multiple co-occurring psychiatric diagnoses compared to males, and that particularly anxiety disorders occurred more frequently in females. The current results accords with research suggesting that females with substance dependence may have more complex mental health issues than males [
38,
39].
In line with prior research including SUD inpatient treatment samples [
12,
13,
17] patients with COD had indications of persistently poorer psychological functioning in terms of higher levels of mental distress, as compared to patients without COD. This finding underscores the need for SUD treatment providers to attend to mental health problems among patients with COD, and also alludes to the importance of available and accessible community mental health services upon discharge from the inpatient stay. An important area for future research is whether the treatment provided for patients with COD is in accordance with the recommendations in national treatment guidelines [
40], which includes mapping and assessment of mental health, and the use of evidence-based interventions, such a cognitive behavioral therapy [
41] and motivational interviewing [
42]. In this context, it would also be interesting to examine the effect of integrated dual disorder treatment (IDDT) [
43] which is suggested to be a promising approach for the treatment of patients with co-occurring SUD and psychiatric disorders [
42].
Little research has been conducted to investigate whether the prevalence of COD among inpatients in SUD treatment varies with the type of SUD diagnosis. The current study showed that patients with COD had a distinct drug use profile. Contrary to Bergmann et al. (7), who reported that patients with COD were characterized by having an alcohol use disorder, we found that the proportion of patients who were dependent upon illicit drugs, particularly cannabis use disorders, was higher among patients with COD relative to patients without COD. The divergent results may be attributed to methodological differences between the studies. For example, while Bergmans’s study (7) was restricted to young adult SUD inpatients, our study included a relatively large and heterogeneous COD inpatient sample in terms of age and SUD diagnoses.
In accordance with our previous research conducted among a general SUD treatment sample of illicit drug users (15), the current study indicated a higher relapse rate among patients with COD as compared to those without COD. The present data showed that the relapse rate was particularly high among those with a cannabis use disorder, among whom more than 50% of the patients with COD reported a relapse at follow-up.
The multivariate analysis indicated that among the demographic variables included in the model, age and gender emerged as uniquely associated with relapse in patients with COD. Whereas previous studies did not evaluate predictors for relapse in patients with COD specifically, the current finding aligns with previous research within the general SUD inpatient population demonstrating an association between younger age and adverse SUD treatment outcomes, including treatment dropout (18) and relapse (15). Knowledge about the characteristics of the most vulnerable SUD patient groups is essential for the development of targeted preventive measures for relapse. Thus, future research among patients with COD may want to focus specifically on the youngest patient group, in an effort to reveal the extent to which they receive adequate treatment, including accommodating their needs for psychiatric and psychological treatment. We found that the female patients with COD were less prone to relapse than their male counterparts. The psychosocial and social implications of substance use differ among males and females (23), and may account for the mixed findings among previous studies that considered gender difference in relapse rate (e.g. 21, 22). Further studies, with sufficient statistical power, should examine factors that may moderate the association between gender and treatment outcome among patients with COD. For example, further research could investigate whether the lower risk of relapse in females could be explained by better access to protective factors such as social networks and community health services.
Our finding concerning the reduced relapse risk associated with higher baseline motivation was expected, and it is in accordance with a previous study reporting a positive association between high initial readiness for substance use change and treatment outcome (i.e. retention) among inpatients with COD [
26]. Our results suggest that baseline motivation may be a particularly relevant risk factor of relapse among these patients. The association between baseline motivation and relapse among patients with COD may be due to underlying moderating characteristics, such as these patients’ tendency of poor SUD treatment attendance and engagement [
10,
11]. Nonetheless, the current data suggest a potential of motivational enhancement interventions in contributing to reduced relapse risk among patients with COD.
Higher mental distress level (i.e. level of anxiety/depression symptoms) did not influence the relapse risk among patients with COD. This finding suggests that problems associated with having a psychiatric diagnosis, such as adverse social [
36] and socio-economic characteristics [
44], rather than the associated psychological symptoms per se, make these patients more prone to relapse. Studies of general SUD treatment samples have found that higher mental distress (i.e. anxiety/depression symptoms) is associated with adverse treatment outcomes [
24,
27‐
29] as was also found in the current study among the patient group without COD. The higher predictive value associated with mental distress among patients without COD may be associated with untreated anxiety/depression symptoms, and/or that health care providers are insufficiently attentive to their mental health care needs following an inpatient treatment stay.
A unique relapse risk associated with cannabis use disorder was equally high in patients with COD, as in those without COD. Previous research has shown that among SUD patients, those with cannabis use disorders are typically below 25 years [
27,
45]. These young adult SUD patients may have a distinct profile [
27,
46], including the presence of more adverse psychosocial risk factors [
47]. Thus, associated unmeasured background factors may have contributed to the increased relapse risk found among patients with cannabis use disorders in the present study. Patients with cannabis use disorders make up a large proportion of the SUD inpatient sample. It is possible that cannabis holds a lower perceived severity profile among patients, at least compared to other substances, which may contribute to a higher relapse risk as the potential severity of consequences might be perceived as less harmful. Studies to come could investigate the role of substance risk perception on relapse rates. Moreover, there is a need for quality research studies to identify the factors that contribute to the increased relapse risk among these patients.
Strengths and limitations
Strengths of this study include the prospective study design and the relatively large SUD inpatient sample size. The sample size enabled a comparison of patients with and without COD with respect to several clinical and demographic factors and follow-up data on relapse rates and associated risk factors. Although the use of objective diagnostic criteria (ICD-10) for both SUD and psychiatric disorders represents a methodological strength of this study, as this is the most relevant source of diagnostic information available for clinicians, we cannot rule out the possibility of both over - and under-diagnostics [
48]. The study has some other limitations that should be considered. A limited number of sociodemographic variables were included. Other sociodemographic variables, potentially predictive of SUD treatment outcome, such as for example employment status [
49,
50] and previous criminal problems [
51] were not available in this study. We used self-report data of relapse as an outcome variable. The accuracy of self-report data on relapse may be questioned. However, several studies have found satisfactory test-retest reliability regarding substance use [
52‐
54]. Nonetheless, the use of laboratory tests to show relapse to substance use could have provided more accurate data. About 30% of the COD sample were lost for follow-up assessments. This rate was just as high for patients without COD, and was otherwise in line with other SUD follow-up studies [
55]. Those who were lost for follow-up were more likely than those who retained to manifest social marginalization (i.e. low education, unstable housing situation). Thus, the current patient sample who were assessed for follow-up may be somewhat biased towards the less marginalized, and we cannot rule out that other relapse predictors could have occurred if also the more marginalized patients responded to the follow-up interview.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.