Following written informed consent, the patients were interviewed and the socio-demographic and clinical information, in accordance with a questionnaire developed by the researcher (A.A.), was obtained. The socio-demographic details obtained included gender, race, relationship status, highest level of education, employment status, whether receiving a disability grant, religion and handedness.

This was then followed by the administration of both the alcohol use disorders identification test (AUDIT) and drug use disorders identification test (DUDIT) by the principal investigator. The hospital’s clinical records were used to establish the date of admission and discharge, admission status in accordance with the Mental Health Care Act, who had brought them to hospital, presenting symptoms, the discharge diagnosis (including medical, substance-related and psychiatric diagnoses), treatment prescribed, number of psychiatric admissions and disposition following the current admission.

The AUDIT and DUDIT questionnaires have both been confirmed as valid and reliable tools amongst acute psychiatric inpatients internationally²⁵ and have also been validated in South Africa.²⁶ The questions are based on Diagnostic and Statistical Manual of Mental Disorders, Fourth edition (DSM-IV) and International Classification of Diseases, tenth revision (ICD-10) criteria for substance abuse and dependence. The AUDIT was developed by the WHO as a screening tool to identify people with hazardous, harmful or dependent alcohol use and has a sensitivity of 83% and specificity of 90% when tested against DSM diagnostic criteria.²⁷ As recommended by the AUDIT manual, cut-off scores of 8 for men and 6 for women were used to identify hazardous or harmful alcohol use and a score of 20 or more for both sexes as an indication of alcohol dependence.²⁷ The DUDIT was developed as a parallel instrument to the AUDIT for identification of drug use other than alcohol, with a sensitivity of 90% and specificity of 78%.²⁸ As recommended by the DUDIT manual, cut-off scores of 6 for men and 2 for women were used to identify hazardous or harmful drug use, and a score of 25 or more for both sexes as an indication of drug dependence.²⁸ The tools have been validated in English, with the investigator translating to Afrikaans for those needed.

In this article, participants were considered to have a SUD if they screened positive for hazardous, harmful or dependent alcohol on the AUDIT, or screened positive for hazardous, harmful or dependent drug use on the DUDIT, or clinically met the DSM-5 criteria for SUD by their treating doctor.

Categorical variables were summarised by frequency and percentage tabulation, and continuous variables were described by the mean, standard deviation, median and interquartile range. These study variables were compared between the groups of patients with and without a SUD. The χ² test was used to assess the relationships between categorical variables and SUD group. Fisher’s exact test was used for 2 × 2 tables or where the requirements for the Χ² test could not be met. The relationship between continuous variables and the SUD group was assessed by the t-test. Where the data did not meet the assumptions of the test, a non-parametric alternative, the Wilcoxon rank sum test, was used. Data analysis was carried out using Statistical Analysis System (SAS) version 9.4 for Windows. The 5% significance level was used, meaning a p-value < 0.05 implies statistical significance.

Ethical clearance was obtained from the Human Research Ethics Committee of the University of the Witwatersrand (HREC M151013).

Of the 150 participants, 62.7% (n = 94) were men, 76.7% (n = 115) were under the age of 40 years and 67.3% (n = 101) were black. Two-thirds (n = 99) of the participants were brought to hospital by a relative, parent or spouse; 16.0% (n = 24) by the police; and 7.3% (n = 11) by the ambulance (Table 1). For 40% (n = 60) of the participants this was their first psychiatric presentation. The remaining 60% (n = 90) were repeat presentations, with 44.0% (n = 66) of the study sample having two or more admissions in the last 2 year (Table 1).

Almost 80% (n = 115) of the participants presented with psychotic symptoms on admission and over 50% (n = 77) were aggressive (Figure 1). Rapid tranquillisation was necessary in 67% (n = 101) of the participants, of which approximately 15% (n = 22) required combinations of three or four tranquillising agents. According to the Mental Health Care Act, the admissions were predominately involuntary (58.7%, n = 88) or assisted (36.0%, n = 54), and the mean duration of hospital stay was 14 days (median 11). The most frequently occurring medical condition was HIV and AIDS, diagnosed in 20 participants, followed by hypertension (n = 6), epilepsy (n = 5), head injury (n = 4) and diabetes mellitus (n = 3).

From the clinical notes by the treating doctor, it was observed that SUD was diagnosed in 60.0% (n = 90) of the participants. Substance-related conditions were the most frequent discharge diagnoses, with 30.7% (n = 46) of participants diagnosed with substance-induced psychotic disorder, 12.0% (n = 18) with intoxication or withdrawal, and 8.0% (n = 12) with substance-induced mood disorder (figure 2).

Administration of the AUDIT showed that 44.6% (n = 67) of participants had an alcohol use disorder, and the DUDIT showed that 47.3% (n = 71) had a drug use disorder (Figure 3). A total of 25.2% (n = 38) of participants had both alcohol and drug use disorders. Overall, the AUDIT and DUDIT tools identified two-thirds of the participants (67.0%, n = 100) as having a current SUD (Table 2).

The findings of the two methods did not correspond in 23% (n = 36) of the participants: 15.3% (n = 23) of participants were identified with SUD by the questionnaire, but not clinically diagnosed with SUD. A total of 8.7% (n = 13) participants were clinically diagnosed with SUD, but did not have evidence of SUD on administration of the questionnaire (Table 3). Had the clinical diagnosis of SUD been augmented with the questionnaires, the overall yield would have been 75.3% (n = 113) SUD.

The demographic and clinical characteristics of those who screened positive for SUD by the AUDIT and DUDIT tools were compared with the group who did not (see Table 4).

There was a statistically significant association between SUD and younger age (p = 0.0010), male gender (p = 0.012) and being more likely to arrive at hospital with police or ambulance escort, or patients by themselves rather than by a family member (p = 0.025). Although unemployment was similar amongst those with SUD (56%) and those without SUD (50%), those with a SUD were significantly less likely to receive a disability grant (10% vs. 26%) (p = 0.015). There was no statistical difference in mental healthcare act status between the two groups (p = 0.073).

Those with a SUD on AUDIT or DUDIT were more likely to have a clinical diagnosis of SUD or a substance-related condition (p < 0.0001). Although those with SUD were less frequently diagnosed with schizophrenia or bipolar I disorder than those without SUD (p = 0.002 and p = 0.0059, respectively), 41.4% of participants with schizophrenia and 47.1% of with bipolar disorder admitted to have comorbid SUD.

A total of 72.7% participants (n = 72 SUD, n = 37 non-SUD) were discharged to home, 10.7% (n = 14 SUD, n = 2 non-SUD) were transferred to Sterkfontein Hospital (a restrictive psychiatric institution) for further involuntary care, 7.3% were transferred for further inpatient care at a less restrictive psychiatric hospital (n = 6 SUD, n = 5 non-SUD) and 8.0% (n = 7 SUD, n = 5 non-SUD) went to a long-stay medium-care facility. Two participants were referred to an inpatient substance rehabilitation centre upon discharge from hospital, which was arranged privately by the family members of the participants.

We found that those with SUD were more likely to be male and single person, which is consistent with the findings of local and international studies.^21,37 Although 54% of the sample was unemployed, which is higher than the 2011 official 25% unemployment rate amongst the general population of Johannesburg,³⁸ there was no difference between those with and without SUD. This may be related to a complicated relationship where mental illness increases the risk of unemployment,³⁹ and unemployment may increase the risk of mental illness.⁴⁰ The difference in disability grants may reflect differences in attitude to substance users by doctors, social workers and the officials approving grant applications or even the substance users’ inability to apply for disability grants.

Whilst the SUD group was more likely to be brought to hospital by police or ambulance, there was no increased rate of aggression, psychosis or mania, which are conditions that could result in a need for police or ambulance escort. Estrangement from family members as a result of the negative effects of SUD may be a possibility. Even though we did not confirm higher rates of psychosis and aggressive symptoms amongst those with SUD, 14 of the 16 participants referred for further involuntary care (usually related to aggressive, disruptive behaviour) screened positive for SUD. It is also not known if any of the 10 participants who were ineligible for study inclusion because of immediate transfer to Sterkfontein Hospital had used substances.

Despite high rates of substance use, only two participants were referred to inpatient rehabilitation centres upon discharge, which was arranged privately by their families This finding is consistent with the low referral rate described by the South African Community Epidemiology Network on Drug Use (SACENDU).⁴¹ In their data on help-seeking substance users from rehabilitation centres, Gauteng province statistics recorded 2734 admissions to 15 treatment centres in the first half of 2018, of which 59% were referred to the centres by themselves, families or friends; 14% by social workers; and 2% by a healthcare professional. Further exploration regarding accessibility of rehabilitation centres is required.

Psychiatric services generally do not offer substance rehabilitation. In South Africa, substance use management falls under the National Department of Social Development, which is obliged to implement inter-sectoral strategies for early detection and treatment under the Prevention of and Treatment of Substance Use Act No. 70 of 2008.⁴² In collaboration with the Department of Social Development and other government departments, the Department of Health is responsible for meeting the medical and psychiatric needs of substance users. Notably, the National Drug Master Plan,⁴³ within the Department of Social Development, which should provide a blueprint for inter-sectoral engagement, expired at the end of 2017, and a new plan is in the process of development. Therefore, collaborative networks and referral pathways between the two departments are awaiting revision.

Clients with both psychiatric and SUD have reported greater satisfaction with an integrated dual diagnosis treatment approach.⁴⁴ In Gauteng province, one government inpatient dual diagnosis unit is available at the remote, stand-alone Sterkfontein Psychiatric Hospital; however, this is not easily accessible, not all are candidates for such intensive programmes and no referrals were made to it from Helen Joseph Hospital during the study period. Outpatient government dual diagnosis services are available only at one central hospital, Chris Hani Baragwanath Hospital.

In 76% (n = 114) of the participants, the questionnaire and clinical diagnosis were consistent; 24.7% (n = 37) of the participants were non-SUD and 51.3% (n = 77) were SUD. However, 15.3% (n = 23) of patients screened positive for SUD upon administration of the questionnaire but were not identified clinically, and 8.7% (n = 13) of patients were not screened positive by the questionnaire but identified so on clinical assessment. Augmentation of clinical screening with a questionnaire would have improved the detection of SUD, a finding that is consistent with the recommendations made by Morejele et al. and WHO.^7,27,28