Introduction

Mental disorders pose a significant global health burden, accounting for 7–16% of the total disability-adjusted life years worldwide [1,2]. Severe mental illnesses (SMI) refers to chronic and disabling mental health conditions that include schizophrenia and bipolar disorder which profoundly contribute to this burden by impacting individuals’ thoughts, emotions, and behaviors, and often impairing their ability to function in daily life and relationships [3]. This burden is even greater in settings like Ethiopia, where mental disorders still receive limited attention from policymakers, despite accounting for more than 11% of the years lived with disability in the country [4]. However, the actual causes and risk factors of SMIs are not well understood. Epidemiological evidence suggest that genetics, structural and chemical changes in the brain, childhood trauma, and previous and active substance use may be possible risk factors [5,6]. Therefore, this study examines the relationship between SMI and common substance use such as alcohol, khat, tobacco, and cannabis, by comparing individuals with SMI to age-and sex-matched controls.

Substance use disorders (SUDs) are common and often comorbid with SMI, with the comorbidity reported to be between 30% and 50%, more than two times higher than the rate of SUDs in the general population [7]. Available evidence, largely from high-income countries, demonstrates substance use disorders as risk factors that worsen symptom prognosis, causing morbidity and premature mortality among people with SMI [6,8]. In addition, SUDs significantly impact the quality of life, economy, and social relationships of affected individuals and their family members [9]. The dual burden often leads to adverse health outcomes, including poor physical well-being, inadequate self-care practices, homelessness, and criminal justice system involvement, thereby increasing the burden on healthcare systems [10,11].

Mental health problems and substance use disorders have been on the rise in low-and middle-income countries (LMICs) [11]. A 2018 meta-analysis showed that the overall prevalence of substance use was around 40 percent in sub-Saharan African countries [12]. The predominant substances used in Eastern African countries, such as Ethiopia, which is the second most populous sub-Saharan African country, are alcohol, inhalants, and khat [12]. Alcohol and khat, in particular, have seen significant increases in usage, particularly among younger populations in recent years [13]. A few cross-sectional studies conducted in Ethiopia explored the magnitude of substance use comorbidity among patients with SMI. They found that the current prevalence of use varies from 25% to 40% [14–17], and dependence was reported in 10–25% [16,18,19]. The cross-sectional nature of the previous studies and the higher prevalence of both conditions in males and younger age groups make it difficult to understand the magnitude difference in substance use between people with SMI and the general population [20]. Thus, this study will examine the association of common substance use among people with SMI compared to their age and sex-matched controls in Ethiopia.

Methods

Results

Descriptive characteristics of participants

The characteristics of the study participants are summarized in Table 1. A total of 13,000 participants were included in this study with 6,500 cases and 6,500 controls. The case group included 4,276 male participants (65.8%), with a mean age of 36.9 years with a standard deviation (SD) of 10.3 years, while the control group comprised of 4,204 males (64.7%) with a mean age of 36.2 (SD = 10.9) years. More than half of the cases (56.3%) were never married/single, while 28.2% were married and 15.4% were divorced or separated. Among controls, 37.9% were never married/single, while 50.9% were married and 11.2% were divorced or separated. More than one-third of the cases had secondary education (36.7%), while 33% had primary education and 24.2% had a college degree or more. In controls 43.9% of participants reported having a college degree or more and 30.2% had a secondary level education and 23.5% had a primary level education. In addition, more than two-thirds of both cases and controls lived with parental or spousal families (72.8% vs 81.5%), respectively.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Demographic characteristics of participants.

https://doi.org/10.1371/journal.pone.0332107.t001

The primary diagnosis of the cases was schizophrenia spectrum disorder (75.5%), while the remaining were diagnosed with bipolar disorder (24.5%). Of these, only 12.2% were recruited from inpatients, while the remaining were outpatients.

Pattern of substance use

Substance use patterns are summarized in Fig 1 and Table 2. Alcohol and khat were the most common substances consumed among both cases and controls, with a proportion of 44.7% (95% CI: 43.5–45.9%) and 44.9% (95% CI:43.7–46.2%) for cases and 64.1% (95% CI: 62.9–65.2%) and 31.5% (95% CI: 30.4–32.7%) for controls respectively. Tobacco and cannabis use were 33.2% (95% CI: 32.0, 34.4%) and 19.9% (95% CI: 18.9–20.8%) for cases and 5.5% (95% CI: 4.9–6.1%) and 5.1% (95% CI: 4.6–5.6%) for controls, respectively. Current use of alcohol was higher among controls than cases: 41.5% (95% CI: 40.3–42.7%) vs. 17.2% (95% CI: 16.3–18.1%) respectively; whereas current use of other substances like khat and tobacco were higher among cases than controls (khat: 20.4% (95% CI: 19.4–21.4%) vs. 12.5% (95% CI:11.8–13.4%); and tobacco: 15.1% (95% CI: 14.2–15.9%) vs. 6.0% (95% CI: 5.5–6.6%)). However, cannabis use was similar across cases and controls: 1.0% for cases (95% CI: 0.8–1.3%) vs 0.8% for controls (95% CI: 0.6–1.1%) (Fig 1).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Unadjusted and adjusted odds ratio (OR) and 95% confidence intervals (CI) for substance use in cases compared to controls.

https://doi.org/10.1371/journal.pone.0332107.t002

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Patterns of substance use among cases and controls.

https://doi.org/10.1371/journal.pone.0332107.g001

Similarly, khat was the most common regularly used substance reported by cases with 16.2% (95% CI: 15.3–17.1%), while alcohol was the commonly used substance reported by controls with 17.5% (95% CI:16.6–18.5%). Regular tobacco use was higher in cases than controls: 13.2% (95% CI: 12.4–14.1%) vs. 4.8% (95% CI: 4.3–5.3%) with a similar trend in daily use: 11.6% (95% CI: 10.8–12.4%) vs. 3.5% (95% CI:3.1–3.9%). Similarly, regular cannabis use was reported as 0.5% (95% CI: 0.4–0.7%) by cases and 0.4% (95% CI: 0.2–0.5%) by controls.

Discussion

This study examines the association between SMI and the use of common substances—alcohol, khat, tobacco, and cannabis— by comparing individuals with SMI to age- and sex-matched controls. The study has three important findings. First, substance use was frequently reported among individuals with severe mental illness compared to controls. Of note, tobacco smoking and chewing khat were more than two times higher when compared to the controls. Second, current alcohol use was 70% lower among cases compared to controls. Lastly, no significant difference in regular cannabis use was found between cases and controls. However, irregular cannabis use was reported to be 70% lower among cases. This study highlights the association between commonly reported substances and individuals with and without SMI in Ethiopia. Such understanding can inform the design and implementation of effective intervention strategies in the country. Current tobacco use was three times higher among cases than controls, with 15.1% (95% CI; 14.2%, 15.9%) reported by cases and 6.0% (95% CI: 5.5%, 6.6%) by controls. Among female participants, the odds of current use was more than three times higher among cases, although the prevalence was notably low in both groups, with 1.5% of female cases and 0.4% of female controls. These findings align with previous studies conducted in Ethiopia and Nigeria [18,29]. Recent meta-analyses report a much higher pooled prevalence of tobacco among people with SMI [30,31]. However, these studies are primarily from high- and middle-income countries with higher national tobacco-use prevalence, such as China (26%), Japan (18%), the United States (13%), and Canada (12%) [32]. The national prevalence of tobacco in Ethiopia is approximatley 6%, which is similar to the prevalence among controls in this study [32,33]. The lower national prevalence might be explained by sociocultural and religious factors in this setting. The odds of tobacco smoking among cases is four times higher than controls. This finding is consistent with a well-established body of epidemiologic literature largely from the global north, showing that individuals with psychiatric disorders have a prevalence of tobacco use approximately two to four times higher than that of the general population [30,31,34].

Khat (Catha edulis) is a plant whose leaves and young buds are chewed to get a state of euphoria as a stimulant. The second most commonly used substance in Ethiopia, khat is usually chewed at social gatherings and is highly intertwined with cultural and religious beliefs. In the present study, the odds of khat chewing was almost two times higher in cases than in controls, with a lifetime prevalence of 45% among cases, consistent with other studies in similar settings [35,36]. This higher prevalence of khat use among cases might be explained by patients’ need to manage their symptoms, i.e., self-medication or tackle the side effects of the medications [37]. This is not surprising as the khat plant parts contain cathinone, an amphetamine-like substance, and small magnitudes of pharmacological compounds like meru cathine, phenylpentenylamines and pseudomerucathine [38]. Similar to other substance types and other study reports, khat use was more common among males, with only 3.3% of female cases and 2.4% of female controls reporting current use [35,39]. Cannabis was the least commonly reported substance, with a lifetime and current prevalence of 5.5% and 1.0% for cases and 5.0% and 0.8% for controls. There was no significant difference in regular cannabis use between cases and controls after adjusting for other variables; however, irregular cannabis use was 0.7 times lower among cases compared to controls. The overall reported magnitude of cannabis use was much lower compared to other similar studies [40,41], which might be attributable to the possible underreporting due to its illegal status in Ethiopia.

Regular and irregular use of alcohol were reported by 8% and 37% of cases and 17% and 46% of controls, respectively. In other words, the odds of drinking alcohol were 70% higher in controls than in cases. This finding is inconsistent with most [42–44], but not all, prior studies [42–45]. For instance, a large-scale study using the Genomic Psychiatry Cohort (GPC) reported that people with SMI reported higher alcohol use with an odds of 4.0 (95% CI 3.6, 4.4) compared to the general population [46]. Similarly, a systematic review and meta-analysis by Di Florio et al. also reported higher odds of alcohol use among people with bipolar disorder, along with significant heterogeneity influenced by factors such as geographical location and male-to-female ratio in the studies [42]. However, in agreement with this study, a study from South India by Kumar et al. reported the odds of alcohol use as three times higher in controls than cases [45]. Similarly, studies from Denmark and the United Kingdom found a higher prevalence of drinking in controls than cases [47,48]. This report of lower prevalence of alcohol consumption among cases might be explained by different reasons beyond geographical location. First, the reason for use of these substances may differ; patients with negative symptoms tend to prefer stimulants over substances with sedative effects, such as alcohol [49]. Even though further studies are required to validate this inconsistent findings across studies, the reported lower frequency of alcohol use might also be due to the difference in symptom level and symptom types of participants in the different studies. Second, another possible reason might be attributed to the financial demand that alcohol use requires compared to other common types of substances. Patients with SMI may already be financially disadvantaged because of their mental illness, potentially resulting in them not being able to spend as much when other common substances are available for a cheaper price [50]. Third, they may be likely to receive advice from health professionals and their caregivers against the use of alcohol, due to its interaction with psychiatric medications [51]. This may also increase the possibility of social desirability biases, resulting in underreporting by the patients.

Although this study examined the patterns of substance use among individuals with and without SMI in a largely representative sample across age groups and genders, certain study limitations should be noted. First, participants with acute or severe symptoms at the time of data collection were excluded, which may have led to an underestimation of the observed associations; however, due to the extended duration of the data collection, those with acute were more likely to be recontacted once they were in remission. Second, the self-reported nature of data collection may introduce recall or reporting bias, potentially resulting in underreporting. Third, the controls used for the study were not ideal controls, as they included hospital staff and students who were present at the hospital; moreover these controls were not entirely healthy controls.as such, the majority of the controls were individuals visiting the hospital for health conditions potentially influencing the reported substance use patterns, given the high comorbidity of substance use and chronic conditions [52]. Fourth, the primary aim of the study differed from the current objectives, which limited the assessment of certain potential confounders; however, this limitation may be partially mitigated by the large sample size included. Lastly, although the study grouped substance use into regular and irregular use, it was still measured based on solely on frequency, it did not assess the severity or apply diagnostic criteria for substance use disorder.

Conclusion

In conclusion, the study found that tobacco and khat use were significantly higher among psychosis cases compared to controls, while no statistically significant difference was observed for regular cannabis use. However, alcohol and irregular cannabis use were more prevalent among controls than cases. Khat and cannabis are known to exacerbate existing severe mental illnesses, which adversely affect the course and outcome of the illness. Tobacco also contributes to higher physical morbidity and premature mortality in patients with severe mental illness. Therefore, patients need to be routinely screened for substance use, and early detection and integration of substance use treatment within the mental health care services are recommended. Further studies are needed to replicate these findings in similar research settings and confirm their generalizability. Additionally future researh should also examine the underlying motivations for substance use among these groups to support the development of targeted interventions that address the issue at its source.

Supporting information

Acknowledgments

We are deeply grateful to all participants in the NeuroGAP-Psychosis Ethiopia study for their helpful and valuable contributions to the study. Our sincere appreciation also goes to the lead institutions and the staff members of Addis Ababa University for their unwavering dedication and support.