Knowledge, confidence, and practices of clinical associates in the management of mental illness

Background Additional human resources are needed to provide mental health services in underserved areas in South Africa (SA). Clinical associates, the mid-level medical worker cadre in SA, could potentially be used to deliver these services. Aim The study explored the self-reported knowledge, confidence, and current practices of clinical associates related to mental health assessment and management. Setting South Africa. Methods A cross-sectional study was conducted. The link to the electronic questionnaire was distributed to clinical associates via databases and social media. Data were analysed with Stata v17. Results Of the 209 participants, 205 (98.1%) indicated they had training on management of patients with mental illness during their undergraduate degree and 192 (91.9%) had a mental health rotation. Few (10.7%) had any additional mental health training. Most participants rated their knowledge of priority mental disorders as ‘good’ or ‘excellent’. Only 43.2% of the participants felt quite or very confident to perform a mental health examination. Participants who felt quite or very confident to manage patients presenting with suicide risk, aggression, and confusion were 44.9%, 46.9% and 53.1%, respectively. Factors associated with a confidence score of 75% and higher were male gendered, working in Gauteng or Northern Cape provinces, and in a rural area. The majority of participants were already involved in mental health assessment and management in their current work. Conclusion Clinical associates have a contribution to make in mental health service provision, but this may need to be supplemented by additional practical training. Contribution Potential gaps in training have been identified.

Mpumalanga, North West, and Northern Cape) have less than 1.0 psychiatrist per 100 000 population compared to Gauteng and the Western Cape which has 2.6 psychiatrists and 5.0 psychiatrists per 100 000 population.respectively. 15ccess to specialist mental health nurses in rural primary health care (PHC) facilities is also a challenge with one study reporting a ratio of 0.68 mental health nurses per 100 000 population. 13 Given that non-specialist medical doctors and nurses are also in short supply in many of these areas, the use of clinical associates in mental health task sharing needs to be considered.
The scope of practice of clinical associates includes taking mental health history, performing a mental health examination, and mental health counselling. 16Participants in a recent study conducted among university-and facilitybased staff involved in clinical associate training as well as clinical associate student representatives expressed support for a role for this cadre in providing mental health services. 17otential roles for them were the immediate care of mental health patients in emergency settings, provision of basic inpatient care, follow up of patients in outpatient settings, and home visits. 17Similar cadres have been used to varying degrees to deliver mental services in several countries including Kenya, Malawi, the United Kingdom, and the United States of America (USA). 18,19,20,21,22,23study involving all three universities offering the degree confirmed that mental health forms part of the training in all three programmes. 24The training ranges from two to four weeks in the final year of study with a mix of theoretical and practical components varying between the programmes. 24he authors concluded that there may be limitations in the roles clinical associates could play in mental health based on the training gaps identified. 24What is not known is the readiness of clinical associates to take on mental health task sharing and their level of current involvement in providing mental health services.In this study, we explored clinical associates' knowledge, confidence, and current practices related to mental health assessment and management.In addition, we determined the demographic, training, and employment factors associated with participants' confidence to perform a mental health assessment and manage patients with mental illness.

Study design
A cross-sectional study utilising an electronic questionnaire was conducted between 01 December 2021 and 31 July 2022.

Study population and sampling
The study population were clinical associates based in South Africa.In order to be included in the study, they needed to have been qualified for a minimum of 6 months at the time of the survey.Clinical associates who had emigrated from South Africa, had subsequently qualified with a medical degree or who were pursuing a medical degree at the time of the survey were excluded.The survey did not involve sampling as all clinical associates who could be reached were requested to participate.

Measurement tools
An electronic questionnaire was developed using Qualtrics software.The questionnaire included sections on selfassessed mental health knowledge, confidence to carry out various aspects of a mental health assessment, and their current involvement in mental health work.The knowledge, confidence, practice, and interest items that used scales underwent expert validation for representativeness, clarity, and relevance. 25This process was followed by individual cognitive interviews with five clinical associates to ensure respondents interpret items as intended by the researcher. 25etail on the expert validation and cognitive interview processes followed has been detailed by Moodley et al 26 previously.

Data collection
The Professional Association of Clinical Associates in South Africa (PACASA) assisted with distributing the survey information and link to the Qualtrics questionnaire using their available platforms including social media.In addition, two of the three universities distributed the survey information and the questionnaire link to their alumni.Potential participants who clicked on the link had to complete five questions related to the inclusion criteria.If they met the criteria, they were then provided with a participant information leaflet and needed to provide electronic informed consent prior to the first question.In order to improve the response rate, incentives were offered in the form of five gift vouchers to the value of R1000.00 each awarded to five randomly-selected respondents.The available functionality in Qualtrics was utilised to prevent individuals from completing the survey more than once.As the survey was anonymous, an end-of-survey redirect was used for completion of contact information for the incentives ensuring that the contact information was not linked to the questionnaire. 27

Data management and analysis
The data were downloaded from Qualtrics and stored on a password-protected computer with a backup stored in a password-secured Cloud account.The data were imported into Stata version 17 (Statacorp; http://www.stata.com)for analysis.Proportions were calculated for the demographic variables and for each of the knowledge, confidence, and practice items that used scales.Each knowledge item was scored from 0 to 4, as was each confidence item.A knowledge score for each participant was calculated by adding the scores of the 12 knowledge items and a confidence score was calculated by adding the scores of the 27 confidence items.Categories were created for the knowledge and confidence scores based on the scoring equivalent of 75% or more, 50% -74%, 25% -49% and 0% -24%.Scoring 81 or more out of a possible 108 was, therefore, regarded as a high confidence score.Participants were categorised as either having a high confidence score or not, which was our main outcome measure.A bivariate analysis of high confidence score and sociodemographic, employment, and training variables was conducted.Variables with p < 0.25 were included in our initial multivariate logistic regression model.A manual backwards stepwise elimination process was used to arrive at a final model.

Ethical considerations
The study had ethical approval from the University of Pretoria Faculty of Health Sciences Research Ethics Committee (778/2020).Informed consent was obtained from all participants.

Results
A total of 216 individuals who met the inclusion criteria consented to participate in the study.Of these, two individuals provided no responses and five individuals provided only some demographic information and no other responses.These seven individuals were dropped from the analysis.A total of 209 participants were, therefore, included in the analysis.Using the HPCSA register of clinical associates as a proxy, this represents approximately 11.6% (209/1797) of clinical associates in the country. 10

Demographic and employment characteristics
Just under half of participants (49.8%) were aged between 25 and 29 years (Table 1).The majority of participants (72.1%) were female.Almost half of participants (49.5%) indicated that they currently worked in Gauteng province.Most participants (57.7%) indicated that they worked in an urban area.Roughly one-third of participants (34.6%) were employed by a provincial department of health, while 28.8% were employed by a private health facility or private medical practice.The most common work settings of the participants were district hospitals (26.0%), private general practice (17.8%), and academic institutions (13.0%).

Training characteristics
Fewer participants had completed their Bachelor of Medicine in Clinical Practice (BMCP) degree at Walter Sisulu University (WSU) (South Africa) (13.9%) than Bachelor of Clinical Medical Practice (BCMP) degrees at University of Pretoria (43.1%) or University of the Witwatersrand (43.1%) (Table 2).Only two participants (1.0%) indicated they had not received training on assessing patients with mental illness during their BCMP or BMCP degree and four participants (1.9%) indicated they had not received training on managing patients with mental illness.Most participants (91.9%) confirmed they had a mental health rotation during their degree.Just over one-tenth of the participants (10.7%, 22/206) indicated they had received additional mental health-related training after completing their degree.

Mental health knowledge
The participants' self-assessment of their knowledge of various mental disorders that are considered important in the South African context are shown in Table 3.Approximately 70% of the participants rated their knowledge of substance use disorders (69.7%) and depressive disorders (70.2%) as good or excellent, while only 50.5% rated their knowledge of dementia as good or excellent, and even fewer rated their knowledge of attention-deficit hyperactivity disorder (ADHD) (29.5%) as good or excellent.With respect to their knowledge of the management of common mental health presentations, most participants rated their knowledge of the confused patient (66.3%), suicide risk (59.0%), and the aggressive patient (55.6%) as good or excellent.

Confidence in mental health assessment and management
With respect to conducting a mental health assessment on individuals presenting with mental health symptoms, there was some variation in confidence in conducting the different components of the assessment (Table 4).
Approximately half of participants (50.3%) felt quite or very confident taking a mental health history, but only 43.2% of the participants felt quite or very confident in carrying out a mental health examination.A greater proportion of participants were quite or very confident doing a physical examination (59.1%) and ordering relevant investigations (61.2%).
More participants felt quite or very confident managing patients presenting with confusion (53.1%) than patients presenting with suicide risk (44.9%) or aggression (46.9%).
The participants were also asked to rate their confidence in carrying out mental health management tasks for specific conditions.Less than half of participants were quite or very confident to prescribe pharmacological treatment for all of the conditions listed.More than half of participants (51.5%) were quite or very confident to provide counselling to a patient with depressive disorder.
While 47.4% of the participants were quite or very confident sedating an aggressive or violent patient, only 27.7% of the participants were quite or very confident managing serious adverse events from emergency psychiatric medication.Only 41.6% of the participants were quite or very confident completing the relevant forms for 72 h observation of a mental health patient.
A total of 169 participants answered all 27 confidence questions.The median score was 64 (range 0-108).The mean score was 62.18 with a s.d. of 21.89.Just over one-fifth of participants (n = 34, 20.1%) scored 81 and higher with just under half of participants (n = 84, 49.7%) scoring between 54 and 80. Fewer participant (n = 39, 23.1%) scored 27 to 53 with only 12 participants (7.1%) scoring between 0 and 26.There was high correlation between confidence score and knowledge score (Figure 1) with a correlation co-efficient of 0.76.
With respect to demographic, employment, and training characteristics associated with a high confidence score (defined as a minimum of 81/108 i.e., 75% and higher), Following logistic regression, the variables found to be significantly associated with a high confidence score were gender, province of current work, and area of current work (Table 5).Being male, working in Gauteng province and Northern Cape province, and working in a rural area significantly increased the odds of a high confidence score.

Current practices
The current work of most participants (Table 6) included taking a mental health history and doing a mental health examination with 35.6% of the participants indicating that their current work 'often' involved taking a mental health history from patients suspected of a substance use disorder and 34.1% indicating that their current work 'often' involved doing a mental health examination on patients suspected of a substance use disorder.The majority of participants provide counselling to patients with mental illness as part of their

Discussion
In this study, we explored the readiness of clinical associates to take on mental health task sharing by determining clinical associates' knowledge, confidence, and current practices related to mental health assessment and management.The study confirmed previous findings that mental health forms part of the curriculum of all three undergraduate programmes 24 with virtually all of the participants in our study indicating they had received some training in mental health assessment and management as part of their undergraduate training.
Seventeen participants indicated that they did not have a mental health rotation even though all three universities include this in their curricula. 24It is possible that these students were based at facilities where there was no mental health training even though the rotation existed on paper.
Most clinical associates indicated 'good' or 'excellent' knowledge of all the listed mental illnesses with the exception of ADHD.Deficiencies in ADHD knowledge in this study confirms previous findings of gaps with respect to training in childhood disorders that had been highlighted at two of the universities. 24The high levels of knowledge with respect to substance use disorder may in part be related to one of the universities having a focus on this and including an option to spend a week of their mental health rotation at a communitybased substance use programme. 24The findings with regard to the high levels of knowledge of depression contrast with studies among non-specialist health professionals in Africa that report limited knowledge of depression. 28,29A study amongst PHC workers in South Africa and Zambia found moderate mental health literacy, 30 while a study amongst nonspecialist medical practitioners in the public and private sector in South Africa found adequate knowledge of mental illness. 31he use of a self-rating tool in the study may have meant that participants could have overestimated their knowledge.
While there was a strong correlation between the confidence score and knowledge scores, self-rated knowledge scores did not necessarily translate into confidence with mental health assessment and management.Almost 88% of the participants had a self-rated knowledge score of 50% or more, while approximately 70% of the participants had a confidence score of 50% or more.This finding concurs with a previous finding that most of the undergraduate practical training in mental health at two of the universities is by chance, thereby limiting their opportunity to apply their knowledge. 24nsurprisingly, participants were more confident doing a physical examination and ordering relevant investigations which they would usually do in the normal course of their work rather than the mental health specific aspects of the assessment.They also appeared to be more confident managing confused patients which is more likely to be a medical diagnosis than aggression or suicide risk.Suicide risk assessment has previously been noted as a possible gap in their training. 24e number of participants with high levels of confidence in prescribing pharmacological treatment was relatively low.This could be potentially related to scope of practice as clinical associates are only permitted to prescribe medicines (under supervision of a medical practitioner) up to Schedule 4. 16 Psychoactive medicines such as anti-depressants and sedatives are Schedule 5. 32 Although all three universities include pharmacological management in their mental health curriculum, it is not clear how much of attention is given to the topic as there is only evidence of one university having a dedicated lecture. 24Managing side effects and adverse events from psychotropic medication also appeared to be a gap for many participants.With respect to depressive, anxiety and substance use disorders, it appeared that a greater number of participants had higher levels of confidence to provide counselling rather than prescribing pharmacological treatment.
Given that clinical associates are likely to encounter aggressive or violent patients in emergency settings, it is worrying that less than half of participants had high levels of confidence to sedate such patients.
The authors found that a high confidence score was associated with gender.Males had significantly higher odds of a score of 75% (81/108) and higher in our multivariate model.A systematic review that explored the confidence gap between male and female medical students, residents, and faculty found that male self-reported scores in a number of areas including knowledge, skills, and procedural confidence were higher than females in 24 of 31 studies. 33In the remaining studies, there were no gender differences in five studies with only two studies finding higher scores among females than males. 33Vajapey et al. 33 concluded that female health professionals perceive deficiencies with respect to their abilities more frequently than male counterparts, even though there is no difference in clinical performance.Gavinski et al. 34 note that self-reported confidence does not necessarily mean a higher level of competence.Our findings suggest that the gender differences in confidence seen in medical and other health professionals may also be applicable to clinical associates.
High confidence scores were also associated with province of work.It is not clear why working in Gauteng province or the Northern Cape province would increase the odds of a high confidence score particularly as these provinces are very different.Gauteng province is South Africa's economic hub with a high population density and the Northern Cape province is South Africa's largest province by surface area with a low population density. 35The authors also found that working in a rural area significantly increased the odds of a high confidence score.This finding appears contradictory as Gauteng province, which was also associated with high confidence scores, is predominately an urban province.It should be noted that the finding with respect to Gauteng province was borderline significant with a p-value of 0.047 and a lower limit of the 95% confidence interval of 1.02.Lower confidence scores in urban areas other than Gauteng province could have resulted in the apparently contradictory findings.A potential explanation for higher confidence scores among clinical associates working in rural areas could be a greater opportunity for them to be involved in mental health assessment and management because of health workforce shortages in these areas. 36ndings indicate that a majority of participants are already doing work related to mental health assessment and management.This finding is not surprising taking into account the work settings of the participants and the high prevalence of mental illness in South Africa. 37,38A study amongst a similar cadre viz., physician assistants in the USA found they see and evaluate patients with mental illness on a regular basis with 62% of physician assistants doing so at least weekly. 23A large number of participants in our study indicated they were prescribing pharmacological treatment for mental illness.This reported practice is despite their scope of practice limiting them to prescribing no higher than Schedule 4 medication. 16e overall findings suggest that clinical associates generally have good self-assessed knowledge of mental health conditions and have opportunities to assess and manage mental health patients; but there appears to be a confidence 'deficit' in performing some key mental health tasks.As an example, 70.2% participants indicated their knowledge of depressive disorders was good or excellent and more than 70.0% of the participants indicated their work 'sometimes' or 'often' involved taking a mental history and doing a mental health examination in patients suspected of having a depressive disorder.However, the proportion of participants that felt quite or very confident in taking a mental health history or carrying out a mental health examination were only 50.3% and 43.2%, respectively.While 73.8% of the participants indicated their work sometimes or often involved providing counselling to patients with a depressive disorder, the proportion of participants who indicated they were quite or very confident performing this task was only 51.5%.While 59.8% of participant indicated that their current work sometimes or often involved prescribed pharmacological treatment to patients with a depressive disorder, only 46.1% of the participants indicated they were quite or very confident to pharmacological treatment to a patient with a depressive disorder.These confidence 'deficits' were also present to varying extents for anxiety disorders, substance use disorders, schizophrenia and suicide risk and suggest specific areas that need to be addressed in training.
In addition to strengthening training, changes in national policy may be required to use clinical associates more effectively in providing mental health services.The Mental Health Care Act, 2002 predates the development of this cadre and, therefore, does not mention clinical associates in their definition of 'mental health practitioner' which currently includes psychiatrists, medical doctors, nurses, psychologists, occupational therapists, and social workers. 39An amendment to the Act to include clinical associates should be considered particularly as one of the two mental health practitioners who needs to examine mental health care users during the application for involuntary care needs to be qualified to perform a physical examination and one of the two mental health practitioners responsible for 72 h observation of involuntary patients must be a medical practitioner. 39The formalisation of clinical associates as mental health practitioners could, therefore, reduce the clinical and administrative workload of medical doctors with respect to mental health patients.A further change in policy to be considered relates to the prescription of Schedule 5 medication by clinical associates.The current limitation could be eased based on additional training for example, a prescribing course for psychiatric medication.

Limitations
We were reliant on PACASA and university alumni databases to reach potential participants.These may have been incomplete and/or lacked up to date contact information.We did not have access to a WSU alumni database and as a result clinical associates from WSU were underrepresented in our study.It is not clear how many clinical associates in total were reached and it is, therefore, not possible to calculate a response rate.We did, however, offer incentives, send out reminders and extend the closing date of the survey to encourage as many clinical associates to participate as possible.It is possible that the clinical associates who participated were those interested in mental health so generalisation to all clinical associates in South Africa should be done with caution.As our study relied on self-assessment of knowledge, this may have been overestimated by participants and the possibility of the Dunning-Kruger effect cannot be ruled out. 40

Conclusion
Clinical associates are currently involved in mental health assessment and management.Their self-assessed knowledge of mental illness was generally good.Based on their confidence scores, there appeared to be some gaps with respect to practically performing aspects of mental health assessment and management.There is an urgent need to provide additional training to close the gaps that exist among those clinical associates already involved in mental health care as well as strengthen undergraduate training in mental health.Consideration needs to be given to amending the Mental Health Care Act, 2002 to include clinical associates in the definition of 'mental health practitioner' and formally allowing clinical associates with appropriate training to prescribe psychiatric medication.

TABLE 1 :
Demographic and employment characteristics of participants.

TABLE 2 :
Characteristics of participants' undergraduate clinical associate training.

TABLE 3 :
Participants' self-assessment of their knowledge of mental health conditions and presentations.

TABLE 4 :
Participants' confidence in mental health assessment and management.

TABLE 6 :
Mental health asssessment and management as part of current work of participants.

TABLE 5 :
Multivariate model of characteristics associated with a high confidence score. http://www.sajpsychiatry.orgOpenAccess