Examining Macro-Level Barriers and Facilitators to Scaling Up Integrated Care from a Complexity Perspective: A Multi-Case Study of Cambodia, Slovenia, and Belgium

Aim of this paper

Recently published studies show that barriers and facilitators to scaling up IC largely depend on the macro-level (policy/political or structural) context of a country, e.g. cultural inertia, type of health system, laws and regulations [18, 19]. However, studies in health policy and systems research mainly focus on clinical or organisational strategies to improve health outcomes, while macro-level strategies remain underreported [19, 20, 21]. Hence, more knowledge is needed about what hinders and promotes scaling up IC at various levels and importantly, the connection between barriers, beyond the clinical (micro) and organisational (meso) level [22, 23]. This paper addresses the above-mentioned gap and aims to: (1) identify and compare macro-level barriers and facilitators to scaling up IC in different health system contexts being Cambodia, Slovenia, and Belgium; and (2) visualise interactions between macro-level barriers and barriers at other levels. This way, lessons can be drawn on the influence and pathways of different barriers to scaling up IC in different contexts.

Design

We used a multi-case study design, in which each country is a case [24]. We identified and compared barriers and facilitators to scaling up IC in these countries to learn from their context-specific differences and similarities, adopting a reciprocal learning approach [25].

Data collection methods

Primary data was collected on barriers and facilitators to IC scale-up through qualitative methods during the formative phase (2019) of the SCUBY project [15]. Common thematic interview guides (see appendices 3–4) were prepared, focussing on the assessment of the current IC scale-up and barriers and facilitators in each country at three levels: micro, meso, and macro or individual, organisation, and policy, as inspired by the World Health Organisation (WHO) framework for Innovative Care for Chronic Conditions (ICCC) [35]. In-depth, semi-structured interviews (IDI) were conducted with health facility managers, policy makers, civil servants (Ministry of Health (MoH) and health insurance), representatives of professional associations, non-governmental organisations (NGOs), implementers, academics, and patient platforms in the three countries. Additionally, focus group discussions (FGD) were organised with patients, health workers, and community-based actors in Slovenia and Cambodia to enrich the interview data. A full overview of focus group and interview participants per country can be found in appendix 5.

IDIs and FGDs were conducted in person by a minimum of two researchers. They lasted 50–90 minutes and were audio recorded. Written informed consent was obtained. Data was collected until saturation.

We also reviewed relevant literature and documents for each country to supplement the qualitative data. We purposively retrieved internal project reports, official reports, scientific publications, policy documents and grey literature sourced from in-country contacts [11, 33, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47]. Primary and secondary data were triangulated to corroborate findings. Table 1 outlines the details of data collection characteristics per country.

Data analysis

Data analysis was divided into two phases: the per-country analysis phase and the cross-country analysis phase.

The ICCC Framework levels (macro-meso-micro) were used in the per-country analysis, distinguishing (a) patients and health professionals as the micro level, (b) community and health organisations as the meso level and (c) the underlying structures, policies and politics at the national level as the macro level [22, 35, 48, 49, 50]. The IDIs and FGDs were transcribed verbatim. NVivo software (NVivo qualitative data analysis software; QSR International Pty Ltd., https://www.qsrinternational.com/nvivo-qualitative-data-analysis-software/home) was used. At least two independent researchers carried out the analysis in their respective countries. Codes were developed in English and interpretations were discussed within the respective country teams. Aside from the common multi-level framing, three separate qualitative (inductive-deductive) thematic data analyses were conducted which resulted in three separate codebooks (one for each country).

The country analyses [11, 36, 44, 51] offered a starting point for the cross-country analysis. The cross-country analysis phase (2022) contained several steps (see appendix 6). After the set-up of the core analysis team (MM, ČZ, SC, KD) with two supervisors (JVO, GMK) and several initial discussions, we developed a terms of reference guidebook with definitions (see appendix 7) to stimulate the use of a common language. To have a common method of analysing the three codebooks generated in the per-country analysis, a more comprehensive framework was needed. We thus retained the ICCC framework macro level and adapted the WHO’s health system building blocks [52] to include the 7^th building block ‘people’ [53]. Initially, we included wider contextual factors (see appendix 8), but then narrowed it down to the key health system functions and their inner complexities for the sake of scope and focus. Utilising the adapted framework, thematic analysis was conducted in a cyclical process including a deductive and inductive analysis approach. Apart from thematic analysis, visualisations were created to examine the complex interrelationships between identified barriers to IC in each country using Vensim software (Vensim PLE; Ventana Systems Inc.; https://vensim.com/).

Identifying and comparing macro-level barriers and facilitators

Our analysis indicates that all themes of the adapted framework apply: (1) governance, (2) health service delivery, (3) health financing, (4) human resources for health (HRH)/health workforce, (5) medical supply/resources, (6) health information system (HIS), and (7) linkage between health system and community (people or demand side). Sub-themes were added inductively and deductively. Table 2 shows our key findings. In-depth case study information can be found in appendix 9 (with sub-themes) and 10 (with quotes).

Assessing interactions between barriers at macro-level and other levels

Figures 1,2,3 display the interconnections between macro-level barriers and barriers at the meso- and micro-levels in each country case in this study. The arrows indicate a relationship between the elements, with the ‘+’ and ‘–’ symbol indicating a positive or negative polarity respectively.

Figure 1

Figure 2

Figure 3

Most notable in Cambodia (Figure 1) is the lack of (human, financial, material, medicine) resources, which impacts the quality of care and hinders scale-up of IC, in all three dimensions of coverage, content of care and institutionalisation. Since quality of care is especially lacking in the public sector and HCWs are paid better in the private sector, privatisation (through increased demand and supply) and a subsequent rise in OOP are encouraged, negatively impacting access and IC due to increased fragmentation and episodic care. These structural inequities are compounded by weak governance mechanisms, such as limited regulation for the private sector, but also limited implementation guidance and supervision for healthcare managers and providers. Privatisation also has a negative impact on HRH in the public sector: due to general staff shortages, subsequent increased workload, and low pay in the public sector more and more HCWs are enticed to move to the private sector, hence causing brain drain and moonlighting (also called ‘dual practice’ in Cambodia).

Within Cambodia’s governance system, the role of development partners is important, yet can negatively influence the accountability and stewardship by the MoH. Additionally, Cambodia faces data system and e-health challenges; poor internet services and a lack of a centralised web-based system lead to poor data registration and subsequently limited data availability, sharing, and monitoring. This hinders collaboration and coordination of care, and the monitoring of scale-up.

A defining feature in Slovenia (Figure 2) is the existence of a strong PHC system. Its centralised governance (with ‘state capability’ [42, 54]) and public health approach to PHC provision are crucial to this. Well-functioning multi-profile teams have been installed within community health centres (CHCs), which positively impact the provision of public health services and IC. For example, health promotion centres were established and are connected to the CHCs. Nevertheless, there are four important health system challenges affecting the scale-up of IC, namely: GP shortages; an outdated provider payment model; HIS under-use and fragmentation; and the increasing privatisation of care provision. These are all related to each other; due to the outdated costing model, the PHC provider payment is insufficient, which causes both brain drain to the private sector as well as GP shortage. The latter leads to increased workload of GPs and subsequently, the under-use of the HIS.

At governance level, the strong governance—key for IC implementation—is challenged due to a number of structural inefficiencies (including high political turnover, leading to limited leadership; limited accountability mechanisms and stakeholder collaborations), which hinder big reforms related to the four challenges above (HRH, HIS and financial reform). Hence, if these health system and governance-related challenges are not solved, they may threaten the sustainable delivery and quality of healthcare services.

In Belgium (Figure 3), the scale-up of IC is slowed down by the fragmented and segmented decentralisation of healthcare financing and policy making that allows many pilot projects to emerge, but few institutional reforms to come through. This sustains the situation in which formal task shifting and multi-disciplinary team work are not induced. The slow progress on data sharing arrangements limits the possibility to improve continuity of care and care coordination and to monitor progress.

In sum, Belgian’s complex, fragmented, multi-level health governance itself becomes one of the biggest barriers to improving the conditions for IC and to implementation at the meso- and micro-level. The fragmentation within the policy arena leads to limited accountability and policy evaluation use and therefore, policy adaptations.

Commonalities and differences in the barriers identified

We identified commonalities in our three country cases with regards to collaborative governance; health financing; HIS; and HRH. As our cross-case analysis illustrates and is confirmed by other cases, the nature of barriers and facilitators to implementation of IC seems to be remarkably similar and consistent across HIC and low- and middle-income country (LIC/MIC) settings [55], yet the relative importance of barriers might differ. We will elaborate on the key commonalities and the implications for policy and practice.

First, in Cambodia, the quality of healthcare for chronic conditions was often more concerning. This is also true in many LICs/MICs making this a more important barrier to utilisation and effectiveness of care [49]. Stepwise implementation guides such as WHO PEN [28] can be a lever to increase coverage and quality of care for high prevalent chronic NCDs, provided there are sufficient conditions allowing provision of quality of care. Primary care in many LIC/MICs is predominantly staffed with nurses with a shorter training and a less comprehensive skill set than general practitioners.

Second, IC for NCDs are not yet as high a priority in Cambodia, as it is in Slovenia and Belgium. While the health system needs major shifts and funds to adapt to this growing burden, there are still many competing priorities and vested interests that slow down an appropriate policy response. This is the case in many LICs/MICs that are grappling still with competing budgetary needs to put more means into prevention, control and management of NCDs [55].

Third, in HICs, the case studies Slovenia and Belgium, the increasing workload due to increasing number of patients with complex problems clearly identifies the need to change current ways of working towards more collaboration and interdisciplinary care and to involve communities. However, the highly professionalised (and specialist) healthcare context and subsequent vested interests are a constraint to task shifting and community involvement. LMICs can provide valuable lessons and insights on how to decrease professional barriers and boundaries [56].

Interacting barriers

As Pirrone and colleagues noted:

Our complex systems visualisations provided insight into the identified barriers and their interactions specific for each country case. Expectedly, there were differences in these interactions due to the varying contexts of the three country cases.

Previous research on the complexity of insufficient resources [49] may help explain the findings in Cambodia, where it is mainly the inadequacies of the public PHC sector that hamper IC scale-up and quality of care. Consequently, many people go to the private sector, where OOP is higher, negatively affecting access to care and IC itself. Scaling up IC requires investing into quality of care and access to all elements of IC [3, 15]. If this chain is only partly accessible or dysfunctional, then people risk receiving fragmented care. In Cambodia, for instance, this is apparent when people get medication at the private sector, but are not educated about their disease/condition. In Belgium and Slovenia, the care continuum could be further improved as well; e.g. via an appropriate mix of provider payment mechanisms. While access and coverage are not major concerns for the population at large in these HICs given the compulsory health insurance, inequities in access exist with people in vulnerable situations, such as the elderly, people with multimorbidities and people from lower socio-economic backgrounds [15, 57]. The lack of collaboration between the health and social sector is a special concern for vulnerable groups, since many people in these groups have combined healthcare and social care needs. In Slovenia, the sustainability of the PHC system as a strong foundation of IC scale-up is at risk due to its HRH shortages, outdated financial incentive model, and HIS challenges [42, 58]. In Belgium, willingness to change is needed to build a political and organisational culture that fosters IC [44].

There are also interesting similarities in interactions between barriers:

1. In all three countries, there are HRH shortages within PHC and consequently, interviewed stakeholders highlighted the opportunities for task shifting, which could alleviate the burden placed on HCWs. This HRH shortage is not common only to the three country cases, but is rather a universal occurrence globally [59, 60, 61]. In both HICs and LICs/MICs, redistribution of tasks is being advocated and has been demonstrated to be effective [62, 63, 64]. In the three country cases, we found that ‘informal’ or ad hoc task shifting is in place (e.g., peer support groups in Slovenia, CHWs in Cambodia and some primary care nurses in Belgium), while formalisation through accredited training, recognition, and a legal and financial frame (with clear role descriptions and incentive system, respectively) is lagging behind. For instance, CHWs in Cambodia often work as volunteers and are unpaid. In the case of Belgium, medical regulations currently prevent task shifting within PHC practices. In Slovenia, task shifting is quite formalised for registered and community nurses, but not yet for peer supporters. In some cases, task shifting is regarded as a temporary stop-gap rather than a solution [65, 66]. The difficulty in formalising task shifting underlines the importance of clarifying roles, but also illustrates the dynamic and political nature of health systems, because roles and functions keep changing, and require continuous adaptations and negotiations between different actors in the system.

2. The HRH shortage redounds to the HIS: when HCWs are overworked, data registration at practice level is limited and, consequently, data availability, monitoring and even the potential of data sharing is weakened. Hence, the HIS is a major impediment in our country cases, but potential game changer to scale-up IC; electronic health records that allow easy and efficient registration of all relevant parameters related to chronic disease management will facilitate follow-up, clinical decision-making, continuity of care across time and between different HCWs. If extraction of data is possible, these systems can further facilitate monitoring of chronic disease parameters, assessment of quality of care and population management.

3. In both Cambodia and Slovenia, fragmentation of service provision and an increased OOP due to privatisation is a cause for concern. Yet, while privatisation is an increasing trend in Slovenia, the extent to which private services are offered (22% in PHC in Slovenia in 2019 [45]) and are important for NCD delivery is much lower than Cambodia, where almost 60% of people with chronic diseases are diagnosed and treated in private facilities [46]. Moreover, a strict regulation of both quality and remuneration via health insurance in Slovenia partially safeguards against the excesses of private services hampering chronic care.

Scaling up IC remains complex, meaning that actions with a focus on of the three dimensions (coverage, diversification, and institutionalisation), are intrinsically linked with and have ripple effects on the other dimensions [67]. This paper focuses on the macro-level context and processes related to IC scale-up, which are often ‘institutionalised’ health system characteristics. It thereby starts from analysing the institutionalisation dimension, and its linkage with and effect on the dimension coverage of IC and the content and quality of the IC package. The interdependency of the three scale-up dimensions coverage, content and institutionalisation is an important topic for future scale-up research.

Strengths and limitations

Our study has a number of content-related (theoretical) and operational strengths and limitations.

In relation to content and scope, our analysis could be strengthened by further considering contextual (e.g. political, economic, socio-cultural, technological, environmental, and legal) factors beyond the health system [23]. However, this was out of scope for this work.

Several authors have critiqued the use of health system building blocks, arguing that it is not suitable for analysing dynamic, complex and inter-linked systems impacts [58, 68]. For this reason, our thematic analysis was expanded to incorporate the missing ‘demand’ component (theme 7). We also wanted to add scope for interactions between components offering a more overarching, holistic health systems viewpoint. This was done by complementing the thematic analysis with health system visualisations which capture some of the complex relations between barriers to IC. The same authors [58, 68] also emphasised the benefits of the health system building blocks framework, stating it is valuable because of its simplicity and ability to provide a common language for researchers, which were important reasons we opted for this framework.

More emphasis was placed on barriers than facilitators in the analysis. However, facilitators also open up opportunities for reciprocal learning from other contexts. We therefore suggest that more research on facilitating factors could be conducted, e.g.: to understand what constitutes strong public health services within PHC in Slovenia; how and why nurses and/or CHWs work well (or not) in PHC teams and under which circumstances; or examining contextual conditions and stakeholders dynamics of task shifting.

An operational constraint of multi-country research is to reach and maintain a common understanding of the concepts. For instance, the boundaries between micro, meso, and macro levels (of the ICCC framework) were difficult to establish. Even within the health system building blocks framework, such distinctions are not clear-cut. Specifically, the ‘healthcare delivery’ (theme 2) and ‘health system-community linkage’ (theme 7) building blocks act more at organisational (meso) level, but were beneficial to explore when looking at interactions between various levels.

While relationships between barriers are context-specific and our visualisations based on our qualitative evidence, in reality, most of these relationships have multiple causal roots, all of which could not be drawn up. Still, showcasing complex interactions may provide directions to policymakers on how to improve IC implementation and scale-up in a sustainable way [13]. Future in-depth case studies taking a complex systems approach could help to further disentangle important relationships between barriers and generate evidence supporting system-wide interventions tackling the many complex interactions that constrain IC implementation and scale-up.