The aim of our study was to establish the profile of adult admissions to the acute wards at LGH. Most patients were men (67%), had a secondary level of education (63%), were unemployed (88%), had a history of concurrent substance abuse (62%) and/or concurrent violence (37%), and presented with psychotic symptoms (90%).

A previous study found that patients admitted to LGH between 01 August 2012 and 31 January 2013 were also mostly men (65.6%) and aged younger than 35 years (58%).¹² In the current study, admissions were predominantly male, with an overall younger mean age of 29 years. In comparison, a study conducted in the UK found a majority of female admissions with an older mean age.¹³

The SACENDU data indicate high levels of substance use, especially illicit drug use, among the general population within the LGH drainage regions.⁶ Our findings mirror this pattern as 61.6% (n = 353) of the patients were found to have misused substances, particularly cannabis and methamphetamines. Younger age, first-episode psychosis, violence, male gender, and being unmarried and unemployed were features significantly associated with acute admissions who abused substances. Those patients who were married, educated to a tertiary level, presented with a mood disorder (mania or depression), and received a disability grant were significantly less likely to abuse substances. Comorbid substance use among psychiatric inpatients is associated with several negative outcomes, including aggression and violence, medication non-compliance, rehospitalisation, and poor social reintegration after discharge.^14,15,16,17

Psychosis was the predominant finding in more than 90% of the cases, with a slightly higher prevalence among men. First-episode psychosis constituted only 28% of all admissions. This might suggest that the current community mental health care system is failing to maintain remission. This could be due to ill-equipped primary health care services and a lack of psychosocial rehabilitation facilities.¹⁸ The repeated admission of patients to specialised psychiatric hospitals due to relapse demands attention.

Notably, 37% of the overall admissions in the current study were found to have displayed violent behaviour during the period leading up to, or upon, admission. In contrast, Chaput et al. reported that only 7.4% of visits to participating psychiatric emergency services in Quebec, Canada were marked by current aggression.¹⁹ Furthermore, our study found the violent cohort to be significantly associated with the male gender and substance use.

Rothärmel et al. found substance use and the severity of psychotic symptoms to be significant factors associated with violent behaviour in psychotic patients.²⁰ Luckhoff et al. found acutely psychotic and violent patients to be at an increased risk of requiring treatment in specialised seclusion rooms (rather than in dormitory rooms).²¹ This requires some consideration in the South African context where the management of psychotic and violent patients inside standard dormitories poses considerable risk to fellow patients, as well as to staff members. However, a reliance on specialised seclusion rooms has been associated with an increased morbidity and mortality within tertiary psychiatric centres.²¹

In terms of the 72-h observation requirements of the Mental Health Care Act,¹¹ a higher proportion of intoxicated and violent patients may be prioritised for admission to tertiary psychiatric centres (Level 2 hospitals such as LGH). Such patients may not be able to be contained by the limited infrastructure and the smaller number of mental health clinicians (medical, nursing, security personnel) within the district centres, such as Mitchells Plain and Khayelitsha District Hospitals. The result may be a system where specialised mental health centres are disproportionately managing intoxication, whereas patients with enduring psychiatric illnesses are managed with an early down-referral to primary health clinics, which are unable to provide the full range of services. Certainly, this merits consideration and possible investigation; it is critical that available resources are utilised in a manner that ensures the delivery of optimal and cost-effective mental health services.

Combined unemployment statistics for Khayelitsha and Mitchells Plain showed an unemployment rate of 32% in 2014.²² The current study shows the psychiatric population to be particularly vulnerable in this regard, as only 12% of the admissions were employed. More than a third of admissions were in receipt of a disability grant, and almost a third of grant recipients misused substances. This is an obvious ethical and systemic concern because it appears government may be funding substance use among persons with mental illness. Only a small minority – who were more likely to be female – were married. This leaves the majority of psychiatric admissions to rely on parents or extended family, where available and willing, for caregiver support. The burden on those taking care of patients with a chronic psychiatric illness negatively impacts their quality of life in various areas.²³ These factors highlight the need for greater cooperation, planning, and possibly policy generation in the Department of Social Services.

The LGH catchment area is characterised by severe socio-economic challenges. Violence and other contact crimes abound within its communities.^3,4 Khayelitsha’s proportion of violent crimes, captured as crimes against a person (contact crimes), are considerably higher than the national average, and crime is also prevalent in Mitchells Plain.²⁴ Previous studies have shown violence to be a considerable problem within South African psychiatric centres.^25,26 Forte et al. report that hyper-vigilance and fear among staff as a result of exposure to violent behaviour negatively impact care.²⁷ The issue of violence in the LGH setting could increase the medico-legal risk. Furthermore, the risk of burn-out among staff also warrants urgent consideration in this setting.^25,28

Comorbid medical illness also requires further consideration in our psychiatric population. Although uniform testing for HIV is not included in standard operating procedures at LGH, this is advised where there is a high index of suspicion based on clinical grounds. HIV infection brings with it other concerns, such as associated medical complications and related cost of care.²⁹ The prevalence of HIV infection in our population was found to be 13% and significantly higher among women admitted to LGH. The considerations around HIV testing probably deserve attention as it is unclear why significantly fewer females, compared to males, were tested during the psychiatric admission process. Syphilis is still a major consideration among psychiatric patients as 9% tested positive in our study. The disease was significantly associated with substance misuse in this cohort. The overlap of these two epidemics of sexually transmitted diseases that have the potential for neuropsychiatric sequelae is a concerning feature of the local mental health landscape.

A limitation of this retrospective study is that our data depended on the accuracy of the original source, as reliance was almost solely placed on self-reported data and collateral information. This is particularly relevant to our substance use data, as very few patients had toxicology done at referral centres. The sample period of 6 months may also limit the accuracy of data, especially regarding seasonal variance, but we believe a large cohort allows for sufficient generalisation of data. Lastly, the retrospective nature of our study limits the inference of causality.