Actor sensemaking and its role in implementation of the decentralized drug-resistant tuberculosis policy in South Africa

Abstract

South Africa has a high burden of drug-resistant tuberculosis (DR-TB). A policy to decentralize DR-TB treatment from specialized central hospitals to more accessible district facilities was introduced in 2011, but to date implementation has been suboptimal, with variable pace, coverage, and models of care emerging. This study explored multilevel policy implementation of DR-TB decentralization in two provinces of South Africa, Western Cape and KwaZulu-Natal. Applying interpretive policy analysis, this paper describes how actors across health system levels and geographies made sense of the DR-TB policy and how this shaped implementation. In an embedded qualitative case study, districts of the two provinces were compared, through data collected in 94 in-depth interviews, and analysed using Vickers’ framework of reality, value, and action judgements. Five district cases characterize variation in the pace of implementation and models of DR-TB care that emerged. Individual and collective attitudes for and against the policy were underpinned by different systems of meaning for interpreting policy problems and making decisions. These meaning systems were reflected in actor stances on whether DR-TB care needed to be specialized or generalized, nurse- or doctor-led, and institutionalized or ambulatory. Actors’ stances influenced their actions and implementation strategies adopted. Resistance to decentralized DR-TB care related to perceived threats of budget cuts to and loss of authority of central facilities, and was often justified in fears of increased transmission, poor quality of care, and inadequate resources at lower levels. New advances in diagnosis and treatment to address the growing burden of DR-TB in South Africa will have little impact unless implementation dynamics are better understood, and attention paid to the mindsets, interests, and interpretations of policy by actors tasked with implementation. Deliberative policy implementation processes will enhance the quality of discourse, communication and cross-learning between policy actors, and critical for reaching synthesis of meaning systems.

Key messages

• Sensemaking processes are crucial to policy implementation. Policy documents, by their nature, leave room for interpretation. As a result, actors at different levels of the health system will make sense of policy according to their own experiences and values.

• This study of the decentralization of drug-resistant tuberculosis treatment in South Africa found that actors’ sensemaking significantly affected the pace of policy implementation and the models of decentralized care that emerged.

• This study suggests that deliberative policy implementation processes that facilitate dialogue, address divergent interpretations, and build a shared understanding of policy can help to overcome implementation challenges.

Introduction

Drug-resistant tuberculosis (DR-TB) has emerged as a public health crisis globally due to decades of poorly performing tuberculosis (TB) programmes and inadequate drug development. South Africa is among the countries with the largest incidence of DR-TB cases in the world (World Health Organization 2023). Prior to 2011, a handful of specialized centres of excellence (CoE) were treating DR-TB in South Africa. People with DR-TB (PWDRTB) were hospitalized for the initial 6 months of the ‘intensive phase’ of treatment until they were no longer infectious, returning for monthly outpatient visits for the remaining treatment of 18 months or longer (Loveday et al. 2012). This resulted in CoEs becoming overwhelmed, with long waiting lists and bed shortages, and patients facing significant barriers to accessing care (Wallengren et al. 2011).

A policy of decentralized care for DR-TB was introduced in South Africa in 2011, which proposed transferring responsibility for the treatment of PWDRTB to lower levels of the health care system (decentralization) and reducing the length of hospitalization for those who required admission (deinstitutionalization; National Department of Health South Africa 2011). The policy also made provision for nurse-initiated management of DR-TB care (NIMDR). However, in the years following adoption, districts established decentralized DR-TB sites at different paces and adopted vastly different models of care, based on different interpretations of the policy, district health systems contexts, capacity, and readiness to implement decentralized care (National Department of Health South Africa and the World Health Organization 2016, Cox et al. 2017, National Department of Health South Africa 2019). As a result, despite decentralized care shortening the time to treatment initiation, treatment outcomes remained poor with high loss to follow-up and early mortality (Cox et al. 2014, Loveday et al. 2015).

Policy implementation is a complex and dynamic process that evolves in political, social, and cultural contexts (Brynard 2009) that enable or constrain implementation (Durlak and DuPre 2008). While there are often structural barriers to effective policy implementation, it is also important to consider the role and actions of the agents involved in implementation. As pointed out, ‘People are not passive recipients of innovations. Rather … they seek innovations, experiment with them, evaluate them, find (or fail to find) meaning in them, develop feelings (positive or negative) about them, challenge them, worry about them, complain about them, “work around” them, gain experience with them, modify them to fit particular tasks, and try to improve or redesign them’ (Greenhalgh et al. 2004, p. 598). Actors shape implementation through their individual agency, their lived experiences, and expressions of power (Long 2001). They make choices and can wield power and influence on others ‘with predictable or unpredictable consequences for implementation’ (Damschroder et al. 2009).

Policymakers do not always consider the way policy will be interpreted and shaped by those tasked with implementation. Research has described health system barriers to decentralized DR-TB care, but there has been limited examination of implementation processes, particularly how actors make sense of policy and how their collective sensemaking influences implementation. Adopting an interpretive policy analysis perspective, this paper explores individual and collective sensemaking as a factor-shaping implementation of the decentralized DR-TB policy in 15 districts of two South African provinces, Western Cape (WC) and KwaZulu-Natal (KZN). From this, we draw conclusions on the role of actor sensemaking and associated frames and discourses in explaining the varying pace of establishing decentralized DR-TB sites, and the diverse models of DR-TB care that emerged.

Theoretical positioning

Effective health policy implementation relies on stakeholders creating a shared understanding of the purpose, goals, and intended outcomes of policy. Underscoring the actor-centric nature of implementation is the observation that ‘programs do not work in and of themselves; they work through the reasoning of program subjects’ (Pawson 2010, p. 186). The agendas and frames of reference of implementing actors (as individuals and groups), along with local contexts, shape what is actually enacted (Coleman et al. 2010). Policy actors form ‘appreciations’ of the facts of the problem (‘reality’ judgements), combine this with ‘value’ judgements around what solutions should be implemented, which then inform their judgements of appropriate ‘action’ (Vickers 1965).

Sensemaking is understood as ‘the process through which people work to understand issues or events that are novel, ambiguous, confusing, or in some other way violate expectations, perceiving cues, creating interpretations, and taking action’ (Maitlis and Christianson 2014). Organizational sensemaking is therefore a social process (Weick et al. 2005), using discursive constructions of reality that interpret or explain policy through interactions with others, as ‘meaning is negotiated, contested, and mutually co-constructed’ as a basis for action (Maitlis and Christianson 2014). Sensemaking has been positioned as having a central role in institutional change (Maitlis and Christianson 2014).

Weick’s theory posits that people will engage in sensemaking processes, especially when faced with novel, uncertain, or ambiguous situations that create space for divergent interpretations (Weick et al. 2005). Policy evidence is often ambiguous and contested (Kyratsis et al. 2012). Moreover, given the political nature of policy formulation, goals laid down in policy documents will often be compromises, and therefore susceptible to multiple interpretations (Signé 2017). Ambiguity is most evident where high levels of discretion are exercised (such as in federal systems) (Signé 2017). In a context of multilevel governance, policies are reformulated and adapted by people at various levels that creates further room for contestation (Denis et al. 2009). The sensemaking of implementers inevitably results in evolution of policy as it unfolds at each level (Spillane et al. 2002).

Once adopted, actors through their discourses continue to legitimize, contest and adapt to policy in common or divergent ways (Ciccia and Lombardo 2019, Parashar et al. 2020). Discourses may take the form of arguments in support or against, or justifications for positions (Blanchet Garneau et al. 2019). Divergent framing or prioritizing of evidence may result in power struggles between those that want to implement the policy as originally intended, those that aim to modify its goals, and those that will try to slow down or impede its implementation (Ahrens 2018). Organizational sensemaking makes visible the tensions, power, and political manoeuvrings of policy implementation (Maitlis and Christianson 2014).

Research has described the link between the sensemaking, discourses, and deviations from policy goals by implementing actors (Spillane et al. 2002, Peck 2006). Rather than seeing this as inherently problematic, however, some level of policy adaptation may be unavoidable and necessary, allowing room for innovation and assimilation at local levels (Damschroder et al. 2009). Successful implementation thus requires establishing meanings and framings of policy action that allow for expressions of agency on the part of implementers rather than simply instructing them what to do (Lehmann and Gilson 2012).

Methods

Study design

The research design was a qualitative, interpretive policy analysis of actor sensemaking in DR-TB policy implementation, spanning national, provincial, and district spheres. Interpretive policy analysis is used to examine the influence of sensemaking on health policy implementation (Yanow 2015).

Setting

The districts studied were in the WC and KZN provinces, which account for around half of the DR-TB cases in South Africa (National Department of Health South Africa 2019). These provinces have a long history of decentralized DR-TB services and provided early lessons from the implementation of pilot programmes from 2007 to 2010. Post 2011, these provinces implemented DR-TB decentralization very differently, with WC achieving 100% coverage through primary health care (PHC) and outreach models shortly after the policy was launched, while in KZN, less than half (44%) of 43 sub-districts were initiating DR-TB care by 2019. Treatment success, defined as treatment completion with culture conversion and no evidence of clinical deterioration, for the 2017 MDR-TB treatment cohort, was 55% in WC and 63% in KZN, compared to the national average of 60% (Dlamini et al. 2020).

Site selection

Fifteen embedded district/sub-district cases were selected for study in each province: six sub-districts/districts in the WC and nine districts in KZN, and within these districts, thirteen hospitals and ten PHC facilities were further sampled. Maximum variation sampling took into account issues such as rural/ urban setting, staff mix, patient load, and population groups served by DR-TB units (Glaser and Strauss 1967). Also considered were different models of DR-TB services and time periods of decentralization in each province.

Sampling key informants

Sampling for in-depth interviews was purposive and included DR-TB experts at all levels within the health service, and in non-governmental organizations (NGOs), and research institutions. Sampling was conducted until a point of saturation was achieved. In total, 94 interviews were conducted, 10 at national level, 40 in KZN, and 44 in WC provinces, respectively. Respondents included 6 national policymakers, 5 provincial programme managers, 24 district/sub-district programme managers/coordinators, 21 nurses, 17 doctors, 11 facility managers, and 10 TB researchers/technical experts (Table 1). The sampling strategy thus incorporated multiple perspectives on the DR-TB programme, allowing the researcher to look at differing positionalities and reality judgements.

Table 1.

Participants included in key informant interviews

	Programme		Facility
	Manager	Coordinator	Manager	Doctor	Nurse	Researcher
National	6					4
Western Cape province	3
City of Cape Town metro (six facilities)	4	2	5	2	5	2
Cape Winelands district (five facilities)	2	1	1	5	5
Overberg district (one facility)	1	1			2
West Coast district (two facilities)		1		1	1
Kwazulu-Natal	2					4
eThekwini metro (three facilities)	3	1		2	1
uMzinyathi district (one facility)		1		1	1
uMgungundlovu district (one facility)		1	1		1
uThukela district (one facility)		1	1	1	1
Ugu district (one facility)		1		1	1
iLembe district (one facility)		1		1	1
King Cetshwayo district (one facility)		1	1	1
uMkhanyakude district (one facility)		1	1	1	1
Zululand district (one facility)		1	1	1	1
Total	21	14	11	17	21	10

Data collection

Primary data collected for this study included key informant interviews using a semi-structured interview guide with a core set of themes that allowed for exploration of informants’ insights into implementing the DR-TB policy. Interviews were conducted in person except for three interviews conducted on Skype because the participants were located outside the country or were not available when the fieldwork to those provinces was conducted. Four interviews were conducted with a small group of two to three participants because of limited time available, with the recognition that this may have limited the openness to sharing conflicting opinions. All interviews were audio recorded and then transcribed verbatim.

Data analysis

Interview transcripts, in the first phase of analysis, were read for content familiarity and to get a general idea of the major themes or ideas that emerged. Several district typologies were created that characterized specific models of care. Vickers’ theory (1965) of the appreciative dimensions of policy actors’ responses provided a framework for classifying and thematically organizing explanations of their actions. This approach was applied systematically to the data, using the qualitative data programme Atlas Ti 4.2, to extensively code and organize the data. The interpretative perspective shifted the analytic focus from structural factors to the expression of meaning-making in explaining the action of implementers (Yanow 2015). Coded chunks of data were retrieved, organized into the overall theme, and iteratively written up to develop a narrative thread and ensure that the research questions were addressed.

Ethical considerations

Permissions were obtained from provincial, district, and facility managers to conduct fieldwork. Written informed consent was obtained from all participants to conduct interviews and for audio recording. Identifiers for individuals and health facilities were anonymized, and individuals were categorized using analytically relevant identifiers including their function within the health system (policymaker, health manager, doctor, nurse, or NGO technical expert/researcher), level of health system where they worked (national, provincial, district, or facility), type of DR-TB health care facility (centre of excellence, decentralized hospital site, or satellite site/clinic), and rural or urban setting of the workplace.

Results

Broadly speaking, as the DR-TB burden increased, decentralization in both provinces was inevitable, but the first years of implementation varied because each province had different approaches, with varying pace, standardization, and degree of verticalization/integration. The models that emerged ranged from centralized district hospital units, district hospital units that initiated then down-referred to satellite sites, initiation at sub-district sites, initiation at PHC facilities, decentralization to PHC with outreach support from the DR-TB site, and even districts that had no decentralization. With time, a degree of harmonization was achieved as most districts moved towards implementing a fully decentralized model to satellite sites.

We present five cases that typify the decentralization forms and experiences that emerged across both urban and rural districts in the two provinces. These cases highlight variations in the timing of policy implementation (early versus late), the adoption of different models of care, and prevailing ideas and beliefs that may help explain the diverse nature and pace of decentralized DR-TB care.

District 1 (rural KZN): early district DR-TB site (2008) and early satellites sites (2012)

Hospital A was among the first district decentralized sites established in KZN in 2008, receiving patients from facilities across District 1 for initiation. However, by 2012, they experienced bed shortages and capacity constraints and it became necessary to further decentralize to satellite sites (district hospitals and community health centres) in each sub-district. They adopted a phased approach to capacitate the satellite sites.

District managers and clinicians advocated for establishment of the first district DR-TB hospital in 2008, primarily to reduce waiting times to initiate treatment at the CoE.

Everyone was pushing.. Patients have to wait for a month before they can start treatment.. You keep a patient here and they are deteriorating and we have nothing. You can’t help them. It is not fair. (KZN rural DR-TB unit doctor)

They also believed that decentralization would improve linkage of PWDR-TB as well as continuity of care.

You are a facility that knows the community, you know the geographical distribution, you know the cultures.. So, they are closer, you feel that you can trust somebody you are going to all the time instead of going to a doctor that you’ve only seen one time … when we can’t find a patient, we go to the clinics and we ask the clinics to please trace this person and bring them to us. (KZN rural DR-TB unit doctor)

TB nurses and programme managers, believed that all clinicians should be trained and confident to treat DR-TB, as was done successfully for human immunodeficiency virus (HIV) and drug-sensitive TB. This would ensure patients did not experience delays and incur costs in accessing and initiating treatment early.

I still think that MDR is just like any other condition … We should be able to initiate the patients because they don’t have money to travel around and it costs them so much.(KZN rural TB nurse)

In 2012, an experienced clinician at the district DR-TB site, motivated by the need to reduce pressure on the DR-TB site and facilitate care closer to patients’ homes, began further decentralization to satellite sites. He first down-referred stable patients for the continuation phase, and then supported the satellite sites to become initiating sites through outreach visits, training, mentoring, and telephonic support.

The patient will have eight hours to and from [Hospital A] so they were like no, this is not working. So, the doctor there started to say guys let us start seeing patients where the patients are. He started the decentralization in each sub-district through his own will.. Initially he would just go there and initiate patients there until some people had interest and joined him, mentored by him and then he left them to continue and then he will start another hospital. (KZN provincial TB manager)

However, in time he observed that the district DR-TB site managers became concerned that decentralization would impact on their facility headcount and lead to reduced staffing and budgets. He recounted open sabotage of decentralization, e.g. when facility managers refused to authorize the use of cars for outreach visits.

One of the issues of decentralization is most of the well-known MDR-TB sites once they see that their numbers are going down because of decentralization, it is clear that they are not happy. You have well-known sites with the bed capacity of 100 beds. They have been full of patients, now you come and say ‘send the patient home’. You send the patient home and the bed occupancy drops, it is under threat … there are people who are going to challenge this decentralization … and say “No, our bed capacity has dropped, what is going to happen next?” There’s a lot of speculation around that they may close the hospital, that they may reduce the staff. Now they [facility management] try to make sure that this does not happen … I left because I was seen as a person who wanted to shut down the hospital. They started fighting in all directions. To go and visit a satellite site, for example I needed a vehicle to go, when they were aware of my roster that I was going to this place .. they will tell you that there is no vehicle because they don’t want you to go there. They want decentralization to fail.(KZN urban DR-TB unit doctor)

District 2 (rural KZN): early district DR-TB site centralized model (2008) and late satellite sites (2016)

Following an Extensively Drug-Resistant Tuberculosis (XDR-TB) outbreak in 2005, District 2, supported by NGO partners, established a decentralization pilot project. In 2008, Hospital B was identified as the district DR-TB unit. The site centralized care in the district and delayed further decentralization to satellite sites until 2016.

Despite the 2011 policy advocating for deinstitutionalization and community-based care, District 2 adopted an institutionalized model, admitting patients for initiation and only discharging them once they had achieved culture or smear conversion. The site nurse and doctor believed that in this way, patients were counselled intensively to ensure good adherence, they were monitored for side-effects, and they were discharged when they were no longer infectious to the community.

I know they said that we shouldn’t admit all of our patients … but we would like to admit them just to see that they are adherent, any side-effects, adherence counselling, the social worker can see them, and the audiologist can see them. (KZN rural DR-TB unit nurse)

In other places they just send a patient away while he is on kanamycin, … on bedaquiline, … during the intensive phase. Then in our place we said no let the people first complete the intensive phase. Then we know that their adherence is good. They have gained weight. They have got insight to the programme … Then the patients with comorbidities such as diabetes … we won’t down-refer those, we just keep them because we need to really monitor them. (KZN rural DR-TB unit doctor)

The district DR-TB site employed a centralized model, treating PWDR-TB at the hospital, and requiring them to return for monitoring in the intensive phase. They employed tracer and injection teams rather than have the sub-districts and PHC facilities take responsibility for tracing and community-based care.

It didn’t take place the way it was set in the policy … In actual fact, some of the decentralized sites kind of became mini centralized sites … If you are really decentralizing then from the decentralized sites, the patient will have to go to the sub-district … then the sub-district has to link the patients to the facilities for follow-up and injection, because the injection teams, most of the time are at this level … So what is happening now, you’ll find that the injection teams are here [at the DR-TB site] … then let’s say you’ve got four sub-districts. How are they going to cover all of these areas? (KZN provincial TB manager)

This district delayed further decentralization until 2016, when the district DR-TB unit began to capacitate satellite sites to initiate treatment, trace patients, and provide community-based care. The district DR-TB unit still served as overall support for satellite sites, for registering PWDR-TB, seeing complicated patients, and monitoring the programme.

District 3 (urban KZN): late district DR-TB site (2015) and late sub-district satellite sites (2017)

Decentralization faltered in District 3 because it was served by the provincial CoE located within the district, and was actively resisted by district managers, CoE clinicians, and hospital managers and clinicians until 2015, when a newly appointed district clinical manager established two outpatient DR-TB initiation programmes. By 2017, a DR-TB clinician was providing outreach support to satellite sites to begin initiating treatment.

Concerns were expressed that decentralization would create a risk of transmission in communities. A TB nurse reflected, ‘There was tremendous stigma that developed, [people] were dead scared of DR-TB and did not want to have the patients in their midst’. The first attempted DR-TB units in District 3 were ‘kicked out due to pressure from staff who didn’t want DR-TB’ in their facilities (KZN urban district manager). As a result of these fears, clinicians actively shunned and denied treatment to DR-TB patients and often referred them late to the CoE.

Doctors were resistant to establishing decentralized sites in their hospitals due to inadequate staff, infrastructure, infection control, DR-TB drugs, and monitoring equipment, including audiometry and electrocardiograms (ECG).

We are giving patients kanamycin. This hospital does not have an audiologist and we don’t have a proper audio booth … We don’t have all of the necessary things that we need to manage DR-TB patients. This can be one of the issues around the implementation of decentralization. The policy is good on paper, but you must have on the ground necessary tools to implement it.(KZN urban DR-TB unit doctor)

The provincial CoE was also perceived as being resistant to decentralizing patients to a district DR-TB site due to concerns about the quality of care.

[CoE] did not want to let go of their hold on this programme. In some ways I do understand, you know when you look at statistics and you look at quality … but on the other hand you cannot manage with the numbers, and patients are dying and they are not getting the treatment that they require. There has to be some give and take. (KZN urban DR-TB unit doctor)

The district TB coordinator (a nurse) was not able to navigate resistance from CoE clinicians and secure buy-in for implementation, stating ‘we started 2013, it was very difficult changing people’s mindsets’. Eventually, district management intervened to instruct the CoE to decentralize.

The district management team must support what you are implementing otherwise nobody’s going to listen to you … There was one time where there was a deadlock at [CoE] and we were not moving anywhere and… the district manager said to them, “If you don’t want to implement, please open that door, please leave now … I am not here to ask for your permission, I’m here to say implement.” (KZN urban district TB coordinator)

This led to the establishment of four decentralized DR-TB sites in the district. In 2017, one DR-TB site doctor began visiting satellite sites to capacitate the nurses. They believed that decentralization to all PHC facilities would not be possible because of resource challenges, and rather that one site per sub-district, preferably a hospital or well-equipped community health centre (CHC) with required equipment and doctors, should be capacitated to initiate treatment. They developed initiation sites to cover all 18 PHC areas.

No, we want to keep it at a sub-district level … if you go down to PHC level you will need the ECG. Are we going to be able to have ECGs in all facilities? The answer is no, so it is better if it is decentralized up to the level of the [sub-district] hospital. (KZN provincial TB manager)

They also argued that PHC clinics were underfunded, understaffed, overburdened, and did not have infrastructure and space.

I also don’t think we need to go down that far to that lower level. A lot of the clinics are understaffed and underfunded, you need expertise. They are already overburdened … So, if you unpack it right to the bottom you will need training across the board, you will need support and I don’t think it is going to be feasible to train to that lower level for MDR management. If you keep it at CHC, it is possible. (KZN CoE doctor)

District 4 (urban WC): early district DR-TB site (1998), early PHC satellite sites (2010)

Throughout the 1990s and early 2000s, Hospital D was the only referral centre in the WC. In 2007, they established an outpatient DR-TB unit to allow ambulatory initiation of treatment and monthly follow-up. In 2008, Médecins Sans Frontières and District 4 introduced a pilot project for the provision of decentralized DR-TB care and treatment at PHC clinics. The TB manager began replicating the lessons of the pilot across other sub-districts. Even though the national model recommended decentralization from the CoE to the regional/district hospital then to PHC satellite sites, the district had doctors who could support DR-TB treatment initiation at PHC level, and in 2010, they decentralized directly to PHC clinics in a nurse-led doctor-supported model. Patients could have ambulatory care and did not need to be hospitalized. The district did not adopt community-based treatment and did not employ injection teams.

Decentralization of DR-TB care proceeded early because key managers were committed to patient-centred care, bringing care closer to PWDR-TB and reducing the need for hospitalization.

It was also to, prevent, um the disjuncture within the family unit as well. Bring the services to the patient within the community, and at the end of the day once the patient is better, he can go back to work … You’re not breaking up that family… it was exclusively catered at looking at the patient and doing what’s in the best interest of the patient.. to help that patient get cured, he needs the support of his family. (WC provincial TB manager)

There was strong support for deinstitutionalization. Participants believed that moving care out of hospitals was more cost-effective and allowed hospitals to focus on sicker patients with complications. Furthermore, hospitalization for long periods in an urban hospital that was far-removed from home, meant families could not visit and patients felt isolated and abandoned. Many PWDR-TB refused admission or absconded.

The patients no longer wanted to be admitted. Other people just simply put it to us that they would rather die, they’re not going to go into hospital. We had really extreme and unexpected strong reactions towards being admitted … Their friends and family do not visit them … they feel abandoned, rejected, forgotten and they feel like they’re going to die. (WC urban district manager)

The district focused on what they believed to be the ultimate goal of the policy which was providing DR-TB treatment in all PHC facilities, foregoing the intermediate step of district DR-TB units.

The national model says specialist hospital, down to secondary hospital, down to district hospital, and from the district hospital outreach teams to primary care level, but because local authority had developed the expertise around managing TB anyway … it didn’t make sense to use that national model in our context here and that’s why we went straight from specialist hospital right down … to primary care clinic level.(WC urban district TB manager)

District 5 (rural WC): early district DR-TB site with centralized model (2008), late PHC satellite sites (2017)

By 2008, District 5 established a DR-TB referral centre at Hospital E which provided centralized care. The DR-TB hospital would monitor the patient or provide outreach support to see patients at the PHC clinic, and clinic doctors were not involved. With pressure from the district, in 2015, Hospital E clinicians began to build capacity of satellite sites and patients were down-referred after the intensive phase. Once the satellite site became more confident they started to initiate treatment. Over time, Hospital E reduced the frequency of outreach visits.

Clinicians in District 5 preferred a model of specialized clinical expertise at a central site to initiate treatment, counsel, and optimize treatment, followed by down-referral to PHC clinic to continue treatment, with monitoring at the DR-TB site every 3 months. They argued that managing DR-TB patients was complex, as drug regimens were complicated, patients were often co-infected with HIV and other diseases, and were likely to experience side-effects and drug interactions.

While the national policy framework recommended the approach of NIMDR, DR-TB clinicians and managers argued that the treatment of DR-TB was too complex for nurses to manage. Nursing approaches are protocol-driven, and nurses may not be able to monitor patients adequately and make decisions about regimens and drug substitution. There was also a regular turnover of nurses especially in rural areas, making it difficult to build up sufficient capacity for managing DR-TB.

I think it is a pie in the sky idea .. They won’t understand much of the new regimens. You also cannot expect a NIMDR nurse to make a decision on a dropping Hb [haemoglobin] or deciding whether to stop or to continue with linezolid [DR-TB drug] and all those toxic drugs. I think they are venturing into dangerous grounds. (WC rural DR-TB unit doctor)

Clinicians and managers resisted decentralizing to lower levels, as they believed that decentralized sites would not manage the patients well and that if specialists did not manage DR-TB, it was going to amplify drug resistance.

We didn’t want to decentralize .. We knew what we were doing was the correct way of doing it in terms of treatment and follow-up. We didn’t trust that anybody else outside could do the job … It is one thing to compromise and the patient dies. It is another thing to compromise and the patient is now a treatment failure because now it becomes the health care providers problem again … So if there is a doctor and a hospital system that could manage the numbers of patients and you know you are giving the best clinical care and you’re following up those patients, why would you risk that and give it to someone else that don’t know TB very well or is going to change jobs in the next few months or years. (WC rural DR-TB unit manager)

There was some resistance to decentralization due to inadequate staffing, challenges retaining doctors and nurses in rural areas, doctors who did not feel confident to manage complicated DR-TB patients, and increased workload of clinicians.

The other reason is the doctor [working in PHC] is doing six times more difficult complex work than any GP [general practitioner] in private practice… We want them to treat complex HIV children. We also want them to manage all of the mental health complex clients that should be seen by psychiatrists. We want them to be boffins in chronic disease management. They must do the medical male circumcisions that has been politically put on our system. Now we want them to also do all of the MDR-TB work. And the cost is that the patients in PHC do not get the doctor support service that they need.(WC rural DR-TB unit manager)

While the urban district in WC had achieved early decentralization to PHC since 2010, the impetus to further decentralize in the rural districts came around 2015 from the districts themselves as a result of the growing burden of disease, large catchment areas with no DR-TB sites, insufficient DR-TB clinicians in the district, and the distance and delays patients experienced in accessing care. Rural district TB managers did not want to rely on outreach from the DR-TB site and started to push back against what they perceived as the DR-TB unit being resistant to decentralization and holding on to the patients.

I understand also why he wants to check patients or keep care of the patients. I advocate for releasing the patient so that we look after them.. We don’t want them to keep on telling us that “we are the specialists, we must come and spend hours looking after your patients” and all of that. They may have more reasons why they want to hang on to the patients, but we want to move away from that. (WC rural district HAST manager)

Summary of findings explaining differences in models and pace of implementation

Actors’ sensemaking of the Decentralization Policy Framework significantly influenced both the pace and models of DR-TB site establishment. The stated rationales of the policy were variably prioritized and contested by different actors. Divergent interpretations led to discourses and debates concerning the nature of DR-TB services and models of care, resulting in actions that either supported or resisted the policy. These divergent interpretations of the policy and the perceived needs of the DR-TB programme, manifested in dominant discourses of resistance to implementing decentralized DR-TB care in the way the policy proposed. Resistance was expressed and justified by actors in a variety of ways: fears of increased transmission (District 3), a too rapid process undermining quality (Districts 2, 3, and 5), lack of resources compromising care at proposed decentralized treatment sites (Districts 3 and 5), and concerns about losing ‘headcounts’ and, therefore, funding (District 1). Many of the rationales for resistance appear to stem from reality judgements on the feasibility and likely impact of policy. The net effect of the resistance was that despite the policy and plans being in place to guide decentralization, unwilling actors at sub-national levels could effectively impede its implementation.

Figure 1 maps the value orientations that underpin the meaning systems influencing actors in the implementation of decentralized DR-TB care. The multiple stated rationales of the policy—namely, that decentralized DR-TB treatment is more accessible and socially acceptable to patients, reduces the burden on central hospitals, results in improved clinical outcomes, and enhances the capacity of lower-level health care workers to manage DR-TB—created ambiguity and allowed for divergent interpretations regarding their relative importance. Actors invoked arguments and discourses to support or oppose various aspects of the policy, debating the need for specialized versus generalized care, nurse-led versus doctor-led treatment, institutionalized versus ambulatory care, and treatment at district hospitals versus deinstitutionalized treatment at PHC facilities. These discourses were used to justify actors’ respective approaches to implementation.

Figure 1.

Mapping of the value orientations underpinning meaning systems.

Value judgements differed in whether policy was positioned principally along two axes of interpretation or perspective: first, whether the policy was judged from a patient/family or provider/health service point of view; and secondly, from a clinical or public health approach to DR-TB. In Fig. 1, this is represented as a matrix of perspectives and associated discourses marshalled (e.g. increasing patient access versus safeguarding quality) by policy actors.

In summary, underlying value orientations and reality judgements (represented in Fig. 1) created the frames for supporting or resisting implementation (Table 2), elucidating how collective sensemaking among actors shaped the differential implementation of decentralized DR-TB sites and the diverse models of care that emerged. The conflicting interpretations and practices of the different groups of professionals engaged in the implementation process were compounded by complexities of institutional positions, incentives, norms, and power relationships between groups of actors, affecting their actions in relation to the policy (Table 2). Specifically,

Table 2.

Systems of meaning among categories of implementers

Category of implementer	Value orientation	Reality judgements	Action judgements
Provincial/district managers and TB coordinators	Patients’ access to care	Waiting lists for treatment at COEs	Compliance with policy
Centre of excellence DR-TB clinicians and managers	Duty to treat challenges of complex treatment	Concerns about quality of care	Discretion in implementing
DR-TB sites facility managers	Health service efficiency	Fear of losing headcount and budget as DR-TB site	Contravention of policy
Satellite site facility managers and TB clinicians	Health service efficiency	Lack of resources for decentralization	Discretion in implementing
	Perception of personal risk	Fear of DR-TB transmission in facilities
Satellite site TB nurses	Patients’ access to care	Waiting lists for treatment at COEs	Compliance with policy
	Health service efficiency	Lack of resources for decentralization	Discretion in implementing

1. TB managers and TB coordinators were positioned at the provincial and district levels with oversight responsibilities and authority to ensure compliance with policy directives. They perceived waiting lists at CoEs as a critical issue and thus prioritized compliance with decentralization policies to improve patients’ access to care.

2. CoE DR-TB clinicians and managers were positioned at specialized treatment centres with significant discretion in clinical decision-making due to their specialized knowledge. Their duty to treat and concerns about the challenges of complex treatments drove them to exercise discretion in implementing policies, sometimes prioritizing quality of care over strict policy adherence.

3. DR-TB sites facility managers oversaw operational aspects and resource allocation at local DR-TB treatment sites. Their focus was on health service efficiency and maintaining resources. Fearing a loss of headcount and budget, some may have contravened policy to protect their facility’s interests.

4. Satellite site facility managers and TB clinicians managed decentralized sites and directly interacted with patients. They had less power than central managers but could influence local implementation. Efficiency and resource availability were key concerns, along with safety. The lack of resources and fear of transmission led them to exercise discretion in policy implementation, balancing efficiency with safety concerns.

5. Satellite site TB nurses were frontline health workers, with limited decision-making power but crucial in direct patient care. Their incentives aligned with improving patient access and efficiency. Facing waiting lists and resource shortages, they complied not only with policy where possible but also exercise discretion to manage practical challenges.

Discussion

This study explored how in the early phase of implementation, actors internalized, made sense of, and framed the DR-TB decentralization policy, resulting in competing ideas, contestation and resistance to implementation. Dominant discourses emerged that begin to explain the differences in the way the DR-TB decentralization policy was implemented between districts in the two provinces.

Policy ambiguity provided room for competing ideas and multiple interpretations of the intent and meaning of the DR-TB decentralization policy and its benefits. What emerged at district level depended in part on the way some discourses were able to assert dominance over others. How provinces, districts, and facilities chose to implement was subject to their own interpretation of the need for and the benefit of decentralization, which challenge they perceived was greatest, which intervention offered most benefit, and what they believed about the evidence for the interventions.

These findings echo what has been reported in other studies of DR-TB decentralization in South Africa. One study reported that the introduction of DR-TB services at underprepared facilities created anxiety resulting in resistance to the new service (Vanleeuw et al. 2020). A study of the first four decentralized sites in KZN demonstrated different interpretations and implementation of decentralization guidelines (Loveday et al. 2018). Another KZN study reported durations of hospitalization longer than the policy dictated (380 days versus 2 weeks), suggesting that it was difficult to change the long-standing practices of health care workers (HCWs) who were reluctant to discharge people who were culture-positive (Brust et al. 2012). In rural WC districts, decentralization lagged despite evidence of the effectiveness of community-based DR-TB care in the urban district (Leavitt et al. 2021). A study pointed to the existence of disease-centred champions who placed emphasis on clinical outcomes, and patient-centred champions who focused on improving patient access to care (Le Roux et al. 2022).

Implementation gaps in low- and middle-income countries have traditionally been addressed through central control introducing stronger regulations and laws and strengthening supervision and audit (Sheikh and Porter 2010). These strategies typically do not consider implementing actors’ meaning systems. Our study emphasizes the importance of looking beyond an expectation of frontline practices mirroring intentions of policymakers, towards complex and cascading sensemaking and organizing. These findings compare well to empirical findings in other studies.

Studies have documented various patterns of sensemaking particularly among middle-managers who ‘have the challenge of grasping a change they did not design and negotiating the details with others equally removed from the strategic decision-making’ (Balogun and Johnson, 2004: 543). In South Africa, PHC facility managers and staff engaged in sensemaking of PHC policies and plans, to incorporate them into routine practice (Gilson et al. 2014). When HCWs were not engaged with the development and implementation of a new policy, viz. the introduction of free health care, resistance surfaced, and affected the quality of the services offered (Walker and Gilson 2004). A study in India identified gaps in implementation of HIV policy guidelines resulting from disparities in implementing actors’ systems of meaning, insufficient avenues for dialogue between actors, and power imbalances in the implementation process (Sheikh and Porter 2010). A study of the Nigerian National Health Insurance System implementation, revealed different framings of the policy message and the critical impact of power and politics in sensemaking resulting in ‘silos of self-interest’ impacting policy implementation (Fawehinmi 2018).

Study strengths and limitations

DR-TB decentralization was selected as a case study for understanding policy implementation in the hope of strengthening decentralized DR-TB care, and also to inform other programme interventions. However, there are differences between health programmes in burden of disease, advocacy, funding, and efficiencies of service delivery. The findings from this case study may, therefore, not translate directly to other programmes. The two selected provinces may not be directly representative of the contexts in other provinces, and further limit generalizability. However, it may be argued that the same complexities in implementation exist in all health programmes and across all settings in South Africa, and there are similar service delivery challenges and similarly poor outcomes. Use of qualitative study design and interpretive policy analysis and a systematic research process involving a large number of interviews with actors across all levels of the health system, strengthened the validity and the potential of this study of DR-TB decentralization to be analytically generalizable to other health programmes and similar settings, and provide credible insights into policy implementation in South Africa.

Conclusion

Implementation often fails because it is conceptualized as a simple set of operational steps that need to be taken and does not take into account the way implementing actors attempt to cope with these changes. Health programme managers must recognize that collectives of implementing actors in complex social systems, including provincial and district managers, programme coordinators, facility managers, and clinicians, in the context of their shared working environments, common goals and similar logics of action, will attribute meaning to complex interventions (Sheikh and Porter 2010) and mobilize their skills and resources towards collective action (Greenhalgh and Papoutsi 2019). Discursive framings of the policy problem must therefore be understood and recognized, and their implications for the policy implementation process navigated.

Implementing communities are often insular and lack opportunity to engage in mutual meaningful dialogue (Sheikh and Porter 2010). Deliberative policy implementation processes will allow understanding of commonalities, facilitate dialogue, enhancing the quality of discourse towards reaching synthesis of meaning systems. Engaging stakeholders ensures actors’ perceptions of the intent of the policy are aligned with the policy content, and that the policy content is aligned with evidence, local experiences, and the contexts where they will be implemented.

New advances in diagnosis and treatment to address the burden of DR-TB in South Africa will have little impact unless implementation dynamics are better understood, and attention paid to individual and organizational processes of adopting and integrating complex strategies.

Author contributions

Waasila Jassat (Conceptualization and Designing, Data collection, Analysis and Interpretation, Drafting the manuscript), Helen Schneider (Conceptualization and Designing, Critical revision) , and Mosa Moshabela (Critical revision). All authors provided final approval of the submitted paper.

Reflexivity statement

The authors represent inclusivity and balanced representation in terms of gender, race, regional location, and seniority. The authors include two females and one male; and one person each of Indian descent, Black African, and White race. The authors are all located in South Africa, a middle-income country. The first author is a recent PhD graduate and emerging researcher, and the other two authors are established researchers working in senior positions in South African universities, who have also contributed significantly to strengthening the public health system in the country.

Ethical approval:

The study received ethics approval from the University of the Western Cape Biomedical Research Ethics Committee (BREC) (BM17/7/4).

Conflict of interest

None declared.

Funding

This work was supported by the South African Medical Research Council (SAMRC) through its Division of Research Capacity Development under the Bongani Mayosi National Health Scholars Programme from funding received from the Public Health Enhancement Fund/South African National Department of Health. The content hereof is the sole responsibility of the authors and does not necessarily represent the official views of the SAMRC.

Data availability

The anonymized qualitative interview data are available upon reasonable request to the corresponding author.