Shared decision-making (SDM) has become the preferred model for patient care as it supports patients’ fundamental rights of autonomy and bodily integrity by actively involving patients in their health decisions.² Shared decision-making can be described as a collaborative process where clinicians and patients jointly make healthcare decisions, incorporating the best available evidence with the patient’s values, preferences and circumstances.³ Shared decision-making thus assists the clinician to appreciate the subjective aspect of the patient’s illness.⁴ The United Kingdom (UK) and United States (US) have seen progress in the implementation of SDM, which can be attributed to strong policy mandates, legal precedents, development of support tools and patient autonomy.⁵ Despite the benefits, there remains a massive gap between aspiration and practical application of SDM, specifically in developing countries such as South Africa.⁶ The four styles of decision-making that have been widely adopted in clinical care are the paternalistic model, the SDM model, the consumerist model and the interpretive model.⁷ The SDM model was developed in 1997.⁸ The model is comprised of four key elements, which include a clinician and a patient, both parties sharing information, and both parties trying to reach a consensus regarding the treatment plan.⁹ Shared decision-making honours a ‘patient-led consultation’ and ‘choice’, thus placing the patient at the core of patient-centred care.^10,11 The informative model or consumerist model allows the patient to make the final decision about their healthcare based on all the information and choices presented to them by the healthcare worker (HCW). The paternalistic model allows the HCW to act as the patient’s ‘guardian’ and make decisions on the patient’s behalf without patient interaction. The interpretive model involves the patient presenting many different options to the HCW, and the HCW then advising them which option is the best based on the patient’s clinical condition and their expertise.¹² The primary health care (PHC) setting is considered the starting point of SDM in a patient’s healthcare journey.¹³ Shared decision-making has been shown to increase patient satisfaction, which in turn improves treatment adherence and overall quality of life.^13,14 Shared decision-making is also linked to decreased healthcare costs, prevention of unnecessary referrals to specialist care and promotion of patient autonomy.^15,16 This reinforces the critical role of SDM in resource-limited healthcare facilities across South Africa.^4,17 A study conducted in South Africa revealed that involving patients in their treatment plan increased their cooperation and understanding, thus highlighting SDM as a crucial component in the South African context.¹⁸ In Europe, a large population-based study revealed that most respondents (96%) preferred to be offered choices and be asked for their opinions.¹⁹ This contrasted with a study in West Africa, which found that the paternalistic model was still the preferred model and that not enough was being done in healthcare to advocate for the uptake of SDM or to promote healthcare training surrounding the SDM approach.²⁰ A South African study found that almost half of the patients interviewed at a district hospital (46%) wanted to be involved in SDM. Despite these findings, doctors assumed that patients did not want to engage in this type of collaborative consultation due to lower educational and socioeconomic status.²¹ South African doctors do not readily provide information to patients who do not show interest in their treatment plans and assume that all patients want a paternalistic approach in their treatment decisions.¹⁸ A predominant factor influencing the uptake of SDM is gender, with women taking a more active role than men.²² Age was also found to be a significant influence, with younger patients assuming an active role.^22,23 Furthermore, people with higher education wanted to be more involved in their treatment choices.¹⁷ The nature of the illness also determines the wish to engage in SDM, that is, the more life-threatening the illness, the less inclined a patient is to participate in SDM.²⁴ Asian and African patients preferred family involvement in treatment decision-making.^23,25,26 Family engagement in South African healthcare is often observed in primary health care, primarily because African cultural beliefs are rooted in the notion that a person’s identity is deeply embedded in their collective existence.²⁶ A qualitative study conducted in London found that the major barriers to patients becoming involved in SDM included difficulties in communication, noisy environments and difficulty involving a carer or escort.²⁷ Healthcare practitioners also cite time constraints as reasons for not engaging with patients and not providing concise patient education regarding treatment and procedures.¹⁸ The main facilitators that prompted the uptake of SDM included early involvement in decision-making, involvement of the whole clinical team, positive relationships with staff members, good administration and a friendly doctor.²⁷ In another subset of patients, including patients with lower literacy levels, it has been found that the use of decision aids could assist with language and cultural barriers.^8,18

Africa is lagging in the clinical application of the SDM model.¹⁵ Thus far, there has been no research conducted in South Africa that focused on the implementation of SDM in primary health care or the barriers and facilitators that affect its uptake. In the African healthcare system, a massive socioeconomic and literacy divide exists, which poses a major challenge to implementing SDM. In an ideal situation, SDM would form a compulsory component of a consultation, as it forms the basis of active participation, which reinforces the patient’s importance in their treatment decision.²⁸ Shared decision-making significantly improves medication adherence among individuals with chronic conditions, including diabetes and hypertension, human immunodeficiency virus (HIV), and even patients undergoing treatment for cancer.²⁹ This research aimed to determine the proportion of patients reporting SDM and highlight the barriers, facilitators and sociodemographic determinants of patient engagement in SDM.

The study was an analytical cross-sectional survey conducted at two community healthcare clinics in the Johannesburg Metropolitan District, which explored the patient perspective.

The Alexandra Community Healthcare Centre (CHC) and Chiawelo Community Healthcare Clinic are public healthcare facilities that serve the communities of Alexandra and Chiawelo, respectively. Hereafter, the clinics will be referred to by number to ensure anonymity of the individuals working there. The study was conducted in the chronic care sections of both clinics. The non-communicable diseases seen at these clinics are comparable to the national non-communicable disease profile.³⁰ These facilities offer an array of services, including acute care, chronic care, child and maternity care, HIV management, prevention services, radiology and rehabilitation services. The doctor–patient ratio, as per Gauteng provincial averages, was estimated at 1:4024. According to the 2017 General Household Survey, 71.2% of the population use public providers as a first choice of consultation when ill, with PHC serving as the first point of entry into the health system.³¹

The total number of patients registered at Site 1 in the general chronic outpatient department was 12 000, and the number of registered patients at Site 2 in the general chronic outpatient department was 33 585 annually. Inclusion criteria were all patients older than 18 years of age who consulted a HCW in the morning. Patients who had poorly controlled psychiatric conditions and patients who needed acute emergency assistance were excluded from the study.

The sample size was calculated using a two-step sample calculation.³² The first formula was used to calculate the total sample size with a p-value of < 0.05, a confidence interval (CI) of 95% and a 0.05 margin of error. The required sample size was calculated at 384. At the end of the study, our sample size was 391, resulting from 103 participants at Site 1 and 288 at Site 2.

Non-probability consecutive sampling was utilised for this study.¹² After concluding their consultation with a HCW, all participants who fell within the inclusion criteria were approached in order of their position in the pharmacy queue to be part of the study.

The English version of the 9-item Shared Decision-Making Questionnaire (SDM Q-9) is a validated tool used to assess whether SDM was done during the last consultation.³³ The tool has an excellent internal consistency (Cronbach’s alpha > 0.9) and construct validity. The tool has a 6-point Likert scale with a final score of 45 points; a higher score implies that SDM was done.³³ It comes in a doctor’s and a patient’s version and has been translated into many different languages.

The pilot study conducted on a smaller sample in a different CHC revealed that South African patients found the original 6-point Likert scale challenging to interpret. In our study, we had opted to use the SDM Q-9 with 4 options on the Likert scale (strongly disagree, disagree, agree, strongly agree) rather than the 6-option SDM Q-9 (completely disagree, totally disagree, somewhat disagree, somewhat agree, totally agree, completely agree). The 4-option SDM Q-9 scored a Cronbach’s alpha of 0.92, which shows excellent internal consistency, validity and reliability.^33,34 Each option on the Likert scale was scored 1, 2, 3 or 4, giving a total score at the end of each questionnaire (36 points).³⁴ We then followed Rodenburg-Vandenbussche et al.³⁵ and categorised all scores falling below the midpoint (< 50%) as ‘no’, meaning no SDM took place. All scores that fell on or above the midpoint (≥ 50%) were labelled as ‘yes’, meaning that SDM took place. The preferred decision-making style questionnaire required the patient to express which was their preferred decision-making model. Patients were shown pictures depicting a scenario of each decision-making model. The options presented to patients included the SDM model, paternalistic model, informative model and interpretive model. The final questionnaire looked at the commonly reported barriers and facilitators to SDM in South African PHC settings and asked participants to choose three barriers and three facilitators from pictures based on their experience. They were then asked to put these pictures in order of the influence they had on SDM in their most recent consultation. (worst to least bad and best to least good, respectively). Patients were provided with options so that data could be statistically analysed in a consistent manner.

Patients were sampled between February 2024 and April 2024 every Monday, Wednesday and Thursday between 9:00 and 11:00. Patients meeting the inclusion criteria were recruited from the pharmacy waiting area. Patients attending the outpatient department were asked if they would like to participate in the study and were assured that their place in the pharmacy queue would be kept. They were approached in a consecutive manner based on their position in the pharmacy queue. This ensured that everyone had an equal opportunity to be included in the study.³⁶ Once they agreed, they were taken to a private area in the clinic so that confidentiality and privacy could be maintained. Potential participants were requested to read the information sheet before signing the electronic informed consent form. If they consented, they proceeded to complete the questionnaires on a tablet linked to the electronic questionnaires on the RedCap online survey platform.³⁷ The researcher facilitated the process by managing the visual aids to ensure that all patients could understand the concepts referred to in the same manner. The researcher received assistance from two non-clinical staff members from each respective site to translate sentences into IsiZulu, the most widely spoken language in Gauteng.³⁸ The assistants were asked to use a neutral tone and follow the questionnaires as they appeared on the online system so as not to influence the participants’ options. The researcher had to assist those who had trouble navigating the RedCap online system (submitting questionnaires or moving on to the next form). The patients were thanked and taken back to their place in the pharmacy queue (Figure 1).

Data captured on RedCap were exported to Microsoft Excel before being uploaded to the STATA version 18 application for analysis.³⁹ Data distribution was determined using skewness and scatter plots, as data were not normally distributed. Descriptive statistics were used for categorical variables, and the median and quartile ranges for continuous variables. The Pearson Chi² test was utilised to determine the differences and statistical correlations between groups. Univariate logistic regression was used to determine the crude odds ratio (COR). The multivariate analysis provided us with the adjusted odds ratio (AOR). Variables with a p < 0.05 in the multivariate logistic regression were classified as statistically significant.

The tablet used to collect data was password-protected, and data were backed up online on the Cloud. Every completed questionnaire was saved in the database, and a new, clean questionnaire was opened for the next patient. No identifiers were used to ensure that the study complied with the Protection of Personal Information Act (POPI) (Act No. 4 of 2013).⁴⁰

Ethical clearance to conduct this study was obtained from the University of the Witwatersrand, Human Research Ethics Committee (Medical) (No. M220311).

A strength in this study was that recall bias was minimised. Questionnaires were in English and verbally translated for people who struggled to understand. The use of visual aids was helpful in this regard, as everyone had the same image to decide on; however, in the translation, meaning could be lost. The study only addressed SDM from a patient perspective and not the HCW perspective. Response bias and social desirability bias must be considered, as these patients do not have the option of other health care facilities. The study’s sample showed a similar distribution of patient demographics compared to probability and non-probability studies done in the South African PHC setting.^21,52 A comparison of patient demographics in this study to demographics of the Chiawelo and Alexandra communities shows that age, gender, most widely used language and employment were similar.^53,54 Selection bias was minimised by utilising consecutive sampling, sampling from more than one site, and consistent and validated data collection tools.³⁶ Although these factors do not eliminate the selection bias associated with convenience sampling, they do minimise it. The barriers and facilitators tool only allowed patients to choose from a predetermined list of options. The reason why researchers chose to include predetermined options was to avoid coding inconsistencies and difficulties; however, including open-ended questions in this section would have increased the study’s depth.⁵⁵

Qualitative studies in the future may render good insights into patient and HCW perspectives on SDM. Leaders in PHC and stakeholders must focus on system issues that prevent patients from engaging in SDM. Undergraduate and postgraduate training in family medicine should not only advocate for SDM but also focus on how to implement SDM in busy work environments. Patient education via social media, radio stations and newspapers regarding SDM should be advocated for.