Initiatives and partnerships in an Australian metropolitan obesity prevention system: a social network analysis

To address study objective 1 (audit the physical activity, nutrition and obesity prevention initiatives taking place in the area), descriptive statistics were used to summarise the Systems Inventory data. To address study objective 2 (identify the most influential stakeholders and collaborations) the barriers and contributions identified by each organisation during the Organisational Network Survey were analysed using descriptive statistics, along with analysis of the Organisational Network Survey data using SNA. Descriptive statistics were analysed using SPSS Statistics Package 25 [45], and SNA data were mapped and analysed using UCINET 6 for Windows [46] and Netdraw software [47]. Each of the relationship domains (share information or knowledge; share resources; engage in joint planning or run joint programs; receive funding; provide funding; share funding or apply for joint funding) were treated as individual networks. The individual networks were also combined/flattened [48] to create a composite network. Networks were mapped visually and the analysis examined the role of organisations in the network relative to others by using whole network and cohesion scores. Structural properties that were investigated included: average degree (the number of ties an organisation has); density (the proportion of possible ties between organisations); diameter (the shortest path between the two organisations furthest from each other); and degree centralization (tendency of the network to focus connections on a single organisation) [23, 26, 49‐51]. Organisation characteristics were used to determine the basis of clustering to enable the identification of potential organisational collaborations and gaps in the service delivery system across the study area.

Core-periphery analysis was conducted using a composite of the network across the six relationship domains to identify densely connected core-nodes and sparsely connected periphery-nodes [29, 52]. Core organisations were characterised by dense connections with other members of the core, and peripheral organisations were characterised by relatively less dense connections with other members in this group [29]. Non-respondents were excluded from the core-periphery analysis. Nodes were de-identified for presentation.

The Systems Inventory results address study objective 1: audit the physical activity, nutrition and obesity prevention initiatives taking place in the area. Fifty-one staff from government and non-government organisations participated in the Systems Inventory interviews. A total of 189 nutrition, physical activity and obesity prevention initiatives were identified. A summary of the objectives, strategies, target groups, setting and funding sources of the initiatives is presented in Additional File 3.

The majority of initiatives were physical activity focused (50%, n = 95), 35% were nutrition focused (n = 66) and 15% obesity prevention focused (n = 28). Overall, the most common objectives were: influence attitudes (87%, n = 164); behaviour change (86%, n = 162); increase knowledge (85%, n = 160); build skills (82%, n = 156); and raise awareness (80%, n = 152). The least common objectives were: influence changes to the built environment (32%, n = 60); advocate for change (32%, n = 17); and to develop regulation (9%, n = 18).

The most common strategies to implement initiatives were communication or education based: online communications (79%, n = 150); printed resources (76%, n = 144); social media (70%, n = 132); group education sessions (67%, n = 127); and unpaid media (61%, n = 116). Partnership development was a strategy in over 70% of initiatives. Implementing policy and guidelines (24%, n = 45); paid media (20%, n = 38); and school curriculum (10%, n = 18) were the least common strategies.

The primary target groups were varied, from everyone (18%, n = 34), adults aged 50 years or more (10%, n = 19), to employees (9%, n = 17). Community centres (39%, n = 73), leisure centres (30%, n = 57) and workplaces were the main settings initiatives occurred in. Early childhood settings were the least commonly used, (e.g. play groups (4%, n = 7) and day care centres (3%, n = 5)). Thirty initiatives (16%, n = 30) operated state-wide.

Government was the most commonly reported source of funding, and included state (14%, n = 26), local (12%, n = 23), or federal government (2%, n = 4). State government funded more obesity prevention initiatives (21%, n = 6) compared with nutrition or physical activity.

The Organisational Network Survey results address study objective 2: identify the most influential stakeholders and collaborations.

Most prevention initiatives in this study focused on behaviour change strategies at the individual level. The 189 initiatives were predominantly aimed at raising awareness, and changing attitudes, knowledge, behaviour and skills using media and education strategies. Few initiatives focused on influencing changes to the built environment, advocating for change, or developing policies and regulations. An individual’s ability to meet dietary and physical activity guidelines to maintain a healthy body weight is influenced by their knowledge and skills as well as the built environment (e.g., access to recreation, available food options) [30]. Research suggests that a comprehensive approach is needed to address the determinants of obesity [53, 54], and individual-level single solution behaviour programs are likely to be difficult to sustain and may have limited success in the long term [30, 55‐57]. Consideration of approaches beyond individual-level behaviour change will better meet the diverse needs of a population, and create an environment supportive of positive, long-term health outcomes [28, 58, 59].

Environmental change to enhance health behaviours is challenging, especially in the complex area of obesity. Achieving long-term change requires overarching governance, shared vision, clear leadership, communication, trust, capacity and performance indicators [24, 44, 60]. Where uncoordinated action occurs, local communities may receive inefficient and ineffective initiatives that are unevenly distributed, with organisations competing for limited resources [44]. However, strategic priority setting and funding can encourage consideration of the social, environmental and behavioural determinants of health [2], with interventions encompassing a broader range of objectives and strategies to strengthen the effectiveness of prevention efforts [14]. This process requires a strategic and coordinated approach to achieve changes across a network, ideally with the support of central organisations.

Organisations in the present study were well connected and interacted with one another across information and resource sharing domains; however, joint funding and joint planning were not well established across the network. This is consistent with previous research [29] that has found informal networks (e.g., information and resource sharing) are densely connected, while formal networks such as joint funding applications and joint planning have weaker connections; and suggests that some of the network connections may not be reaching their full potential. Informal networks are likely to move to more formal structures over time [61]; however, this can be challenging for self-organising networks operating in systems with limited strategic oversight and diffuse governance [12].

Although formal collaborations and partnerships are recommended as a way of sharing limited resources and reducing duplication of initiatives, barriers can exist at the funding and organisation levels [29]. The present study found that limited funding, staffing, and in-kind resources all present barriers to prevention initiatives. Collaborations, particularly formal collaborations, come with time and human resources costs, and these costs need to be offset by the benefits stemming from collaborations [12]. Some funding bodies will mandate collaborations between organisations which can lead to “partnerships on paper” in order to meet funding requirements, even though funded collaborations require a high level of commitment and do not always result in better collaborations [62]. Organisations may find it difficult to align their priorities when delivering joint initiatives [50, 63], and funding and political cycles can create short-term relationships between organisations making it difficult to maintain collaborations long term [50]. Organisations applying for the same funding may also see each other as competitors which can create tensions [29]. A range of strategies may address some of these barriers, such as: rewarding long-term investment in collaborative practice by providing funding to existing relationships [26]; local organisations building on existing infrastructure of informal networks when demonstrating capacity to meet funding requirements and deliver services; and funding bodies recognising local initiatives rather than imposing top-down models of effective partnerships [29].

The core-periphery analysis provided additional information on the study’s network structure. SNA can provide researchers with insight into whether organisations are working together, and the core-periphery analysis goes one step further by answering questions such as: “Who is in the best position to benefit, to lead or to bring the network together? Could opportunities be distributed more equitably? Who should be connecting with whom, considering their role and resources?” [60]. Universally, core organisations have been found to provide strategic direction and expertise across the network having rich connections that can be used to access resources and information compared to periphery organisations. Core organisations are active in overseeing or knowing about the initiatives of others and supporting effective functioning of the system, while peripheral organisation are focussed on delivering initiatives [60].

Core-periphery structure of a network can be further explored in the context of the complementary contributions made by government and community organisations. A previous study investigating the contributions to prevention by different types of organisations [64] found key differences between LGAs and community health organisations. The perceived contributions to prevention by LGAs in the previous study were: cross-sectoral relationships; influence on the structural, cultural, social and physical determinants of health; and a population-wide reach. LGA functions are diverse (urban planning, community wellbeing, leadership, lobbying to state government) and they can mandate structural change within community and have influence over businesses, residents and workplaces. Conversely, community health organisations and NGOs were seen to contribute to local prevention through: community development approach and philosophy; community connections; and knowledge and expertise. Practitioners becoming more aware of the key contributions made by other types of organisations in the network may be able to encourage collaborations by highlighting and building on the complementary nature of these contributions.

The terms “core” and “periphery” are not intended to imply that any type of organisation is more valuable or important than another, but to assist in understanding of the role and density of each type of organisation in a network. There are limitations when the density of relationships at the core is high, such as the risk of organisations becoming “closed-off” to new information and operating under the usual way of working [60]. Provision of additional resources and opportunities for those organisations in the periphery and increasing practitioner awareness of the benefits and limitations of the core-periphery structure in the local system have been shown to distribute resources and collaborations more evenly [60].

This study highlights the complementary nature of the contributions made by different organisation types operating in a system. It provides an opportunity to consider these findings and identify opportunities to continue current practices or modify these practices to enhance prevention strategies and intended outcomes. Opportunities include: consideration of prevention initiatives that emphasise environmental supports [30]; encouragement of formal and informal information sharing and joint planning across the system so that limited resources are best utilised [61]; advocacy for funding bodies to reward long-term collaborative practice by providing funding to existing collaborations and recognition of local initiatives [29]; increasing practitioner awareness of the contributions made by all organisations in the network to encourage collaborations [60]; recognition that core organisations play a vital role in transmitting information, with the ability to oversee and understand what activities are occurring [60]; and more engagement with sectors outside of health to enable prevention work to occur outside of the usual bureaucratic and political silos [31].

The use of Health Promotion Officers from the Health Service to recruit and collect data ensured a certain level of trust with organisations in the community. Data were collected quickly without relying on significant financial resources. An additional strength is use of the Systems Inventory as a baseline measure that can be used for planning and assessing the prevention response. It should be noted that the study did not assess the quality and impact of initiatives and this is recommended for future research endeavours. Additionally, the study does not comment on changing organisational positions in the network across time [60]; however, there is currently a dearth of evidence regarding the methods and tools to capture this longitudinal type of data [17].

The study recruited via “word-of-mouth” and it is possible that some organisations and their initiatives may have been omitted. Furthermore, there may have been a lack of shared understanding of prevention and the systems thinking concepts and language between practitioners and researchers. Previous research has shown that even those in senior positions who consider themselves to be systems thinkers are often considering the health system specifically, rather than the whole system that would be considered by systems scientists [65]. In addition, the identification of the core network organisations were based on defined factors (e.g. share information or knowledge; share resources; engage in joint planning or run joint programs; receive funding; provide funding; share funding or apply for joint funding), therefore should other factors or organisations be examined it would potentially reveal a different core-periphery structure [60].