In this study, we found that 65% of people with SUD have symptoms of mental health disorders and psychological distress. Mental health symptoms were particularly prevalent among females, people with frequent use of cannabis, non-OAT opioids, and benzodiazepines compared to men and people with no or less frequent use of these substances. Interestingly, there were no clear associations between substance use patterns and change in mental health symptoms over time. This could suggest that the differences observed were indicating self-medication to larger degree than medication-related decline in mental health.

People with SUD are a heterogeneous population; fifteen and 35 % reported lower mean SCL-10 item scores compared to the general population and the standard reference score for symptoms of mental health disorders, respectively. Despite vast intra-individual variations in SCL-10 score from baseline to first follow-up, going in both directions, there were no time trends indicating change over time for the total study sample. This indicates that mental health disorders and psychological distress persist over time for this group and we are not able to explain the huge shift, positive and negative, in mental status of many individuals.

The mean SCL-10 for our cohort was 2.2, which is considerable lower compared to the general Norwegian population at 1.4, estimated to be around 11% of the population [31]. Around two-thirds of the total study sample reported symptoms of mental health disorders. This was somewhat higher symptom burden compared to cohort among people with SUD in Sweden [35], however, lower compared to a study among people entering SUD treatment in Norway, which found that over 80% had a level of mental distress above the 1.85 cut-off for SCL-10 at admission [36]. This could reflect that initiating SUD treatment, often combined with strict detoxification, is a very stressful event, whereas most of the patients included in our cohort were long-term OAT patients with a mean treatment time of almost eight years [13]. Correspondingly, follow-up studies have shown that there may be a significant reduction in SCL-10 symptoms when these individuals are discharged from inpatient treatment, however, presence of mental health disorders and severity of substance use seem to be independent predictors of considerable symptoms of mental health disorders in the long-term [37, 38]. We found that mental health symptoms at baseline were associated with a worrying debt situation, unstable living conditions and a frequent use of some of the substances. Severe debt has been found to correlate with poor mental health in a systematic review summarizing a number of studies [39]. There are also several studies suggesting a strong relationships between substance use and psychological distress, despite hardship to establish exact causality [40–42]. In the above study among people entering SUD treatment, severity of substance use, although stratified into alcohol use, illicit drug use and number of substances used– but not the actual substances used; was the most significant predictor of symptoms of mental health disorders [36]. However, again the question arises whether these symptoms are the direct result of the substance use or symptoms of mental distress presenting upon treatment admission [36].

In our study, use of cannabis, non-OAT opioids and benzodiazepines were co-occurring with mental health distress at baseline, while the opposite was seen for stimulants. There were no changes in time trends between use of substances and mental health symptoms. One hypothesis for these findings could be that the associations at baseline might be due to reverse causality, i.e. that participants with substantial mental health symptoms use substances to self-medicate symptoms [43]. It is also possible that there is a “flattening effect” and that potential negative impact of substances are more substantial at an earlier phase and that the change in later phases are less pronounced. Other research indicate that high doses of benzodiazepines reduce social functioning, and that it may also increase psychological distress and worsen mental health [16, 44], and misuse of benzodiazepines is seen among both SUD and psychiatric populations alike [45]. Similarly, the use of stimulants, in particular methamphetamine, has been associated with poor mental health outcomes [23]. Self-medication of attention deficit hyperactivity disorder (ADHD) with stimulants could be one explanation for these findings. Yet one study found that high ADHD symptom burden was associated with higher mental distress and use of simulants among OAT patients [46]. It is estimated that up to a third of patients in OAT have ADHD and previously we have found that coverage of central acting stimulants in this patient group is very low [12, 47, 48]. An alternative explanation could be that stimulants have a direct positive impact on mental health symptoms among these patients. However, the time trend analyses does not support these hypotheses.

Although prevalence of mental disorders and SUD comorbidity has been found to vary among European countries; research consistently shows a high total prevalence of around 50%, with depression, anxiety disorders and personality disorders being the most frequent [9]. However, some facility based studies indicate an even higher comorbidity prevalence as people with severe symptoms are more likely to seek support; 70% for personality disorders [3] and a lifetime substance-independent mental disorder was found in nine out of ten patients enrolled in treatment facilities [49]. Comorbid mental health disorders and SUD have been found to be associated with poor treatment outcomes and show a higher psychopathological severity compared to people with a single disorder [50–52], and this underlies the importance of assessing mental health status in clinical settings among people with SUD. We endorse that evaluation of mental health and linkage to mental health care services should be included in OAT programs and low-threshold SUD clinics; be gender-sensitive and follow and integrated treatment approach, which have been found superior compared to separate treatment plans [53–55].

The major strength of this study is the relatively large sample size of a “hard-to-reach” population of people with SUD as well as a cohort design. However, there are some limitations. Firstly, only a minority contributed to the prospective analyses (268/707). To reduce the potential for selection bias between the sub-group with follow-up SCL-10 measurements presented in Fig. 2 and the baseline cohort, we conducted an inverse probability weighted analysis. Our study sample is also mainly relevant for people with opioid dependence being enrolled in OAT treatment as most were in this group. Thus, our research might not be generalized to other groups with SUD. Moreover, both in the OAT and low-threshold SUD clinics, patient- and system delays contributed to non-accurate annual health assessments, which could in turn affect both answers and results. Thirdly, the SCL-10 has limitations. It is not a diagnostic tool for mental health disorders and is no replacement for clinical interviews and more comprehensive psychiatric instruments among people with SUD. Literature also suggests that the SCL-10 predicts depression and anxiety better than other diagnosis, and that some 50–60% of the patients identified with symptoms of mental disorders qualify for at least one or more mental disorders when assessed clinically [31, 56, 57].