S.O.1. To estimate the impact of specific interventions modifying the urban physical environment on each selected NCD and/or identify and measure the association between the characteristics of the urban environment and NCD risk behaviours.
Approach
Natural experimentation involves a set of experimental design criteria with great potential for identifying and estimating the ecological association between characteristics of the urban environment and the prevalence of NCD risk behaviours, as well as eventually the aggregate incidence.
Design
In the absence of random assignment, the treatment of territorial units for statistical aggregation (neighbourhoods and/or census tracts) as belonging to the experimental or control group will depend on the object of analysis:
1) To estimate the impact of specific interventions modifying the urban physical environment on the incidence of each NCD, territorial units will be assigned to the experimental or control group according to their belonging to the isochronal area of the intervention analysed, with radius set according to the distance-decay effect identified in the literature (Location: Valencia, Rijeka).
2) To identify and measure the association between the characteristics of the urban environment and prevalence of NCD risk behaviours, territorial units – neighbourhoods – will be clustered according to the physical and functional characteristics of the urban environment. This design would allow for inter-cluster comparisons of the prevalence of associated risk behaviours (Location: Valencia, Rotterdam).
Target population
General population
Methods
In order to 1) estimate the impact of specific interventions modifying the urban physical environment on the incidence of each NCD, Poisson Generalised Linear Models (GLM) with census tract and year fixed effects will be implemented, specifying the annual incidence by census tract as dependent variable. The choice of a Poisson model is relevant to the nature of the dependent variable, based on counts, although another distribution could be assumed if incidence rates were calculated. The model will specify fixed effects, which will allow control for unobserved heterogeneity at census tract and year level.
We assume that unobserved heterogeneity at these levels can serve as a proxy for aggregated risk behaviours that are not directly measurable in an ecological context. The assumption is based on the theoretical and empirical justification suggesting that risk behaviours for NCDs are correlated with specific geographical and temporal characteristics. By controlling for these variables through fixed effects, the model seeks to isolate the net impact of urban interventions on NCD incidence. In addition, we will specify other potential confounding covariates describing the physical-social and functional characteristics of the urban environment, as well as interaction terms to assess how the impact of staggered interventions might vary over time. This methodological approach is particularly useful for capturing the temporal dynamics of interventions and providing a more nuanced understanding of their effectiveness.
To address the task of 2) identifying and measuring the association between urban environmental characteristics and the prevalence of NCD risk behaviours, a regression analysis approach will be employed. Neighbourhoods will have been previously classified using clustering algorithms according to the degree to which the environment is conducive to NCD risk behaviours. This classification will be incorporated into the model as a categorical variable representing the different neighbourhood typologies. This approach will allow an assessment of how the physical, social and functional characteristics of neighbourhoods drive NCD risk behaviours. In the absence of panel data, no fixed effects will be specified at the neighbourhood level. Instead, the multilevel factor will provide the necessary adjustment to measure differences between neighbourhood categories, thus allowing for more robust comparisons in terms of prevalence of risk behaviours. This methodological approach offers an effective way to address the complexity inherent in the relationship between the urban environment and NCD risk behaviours.
S.O.2. To analyse in depth the behavioural causal links driving this association, through the exploration of different casuistries related to the relationship of the neighbours with their urban environment. S.O.3. To understand the enablers and barriers for vulnerable populations to engage with the urban physical environment.
Approach
GIS-supported Qualitative Methods are characterised by the mix of traditional qualitative research methods with GIS tools. In particular, the implementation of techniques such as focus groups or in-depth interviews is supported by the use of GIS for the provision of spatially contextualised information. As recognised by Hassan, the “transformation of people’s opinions into a GIS could be regarded as the valuable measure for a (public participatory) GIS. These dynamics also serve as a framework for obtaining specific information about the territory.
Design
The Citizen Science approach is particularly relevant for addressing the two objectives mentioned above. Specifically, the approach to be implemented in HORUS is based on the notion of collaborative project. As stated by Jacques-Avi et al.,84 collaborative projects “are developed by research teams in which people contribute data and also contribute to the design, interpretation and dissemination of the results”, incorporating the target population affected by a health problem directly or indirectly from its detection. Some design criteria for GIS-supported qualitative research, rooted in the concept of citizen science, include: 1) participants in the activities and stakeholders will have prior knowledge of the main research questions and objectives and will be able to propose questions to be addressed during the development of the activities; 2) participants will play a major role in the process of data collection; 3) the results obtained will be disseminated, discussed and finally validated by the participants and other local stakeholders.
Target population
General population from selected neighbourhoods, vulnerable population, NCD patients and key informants.
Methods
GIS-supported focus groups (6 for each city, with 8 to 10 participants) and semi-structured in-depth interviews (30 for each city) will be implemented with the purpose of presenting the results of the natural experimentation to people with NCDs and other citizens and key informants and further explore the underlying behavioural causal links. The selected locations will depend on the results obtained in the previous phase of the project. These sessions will aim to conduct a qualitative exploration of the identified links between physical-social and functional characteristics of the urban environment and NCDs risk behaviours, as well as the behavioural causal factors driving the public health impacts of urban interventions.
The use of the GIS and the development of a new module of the Healthy Cities Generator – Citizens Module, using direct data from the citizens in an easy-to-use interface, to understand the levels of use, satisfaction and perceived benefits from their direct urban environment, will support the implementation of this task. The online tool will be used to analyse the proposed urban interventions and existing urban environments. This work will provide qualitative insights on what the targeted communities need from their direct environment and understand what barriers and facilitators they perceive when interacting with it. The citizens input will link to the existing evidence from the tool (scientific evidence on urban determinants of health and their linked health impacts). The results of the citizens’ input will support the design of more robust interventions on two levels: hard interventions to the physical environment to maximise its health potential, and soft interventions targeted at promoting the use of these environments to achieve further health impacts (like those related to social interaction and cohesion). The new citizens module will consist of a series of online questionnaires designed to discover the causal effects of behaviour and lifestyle to the urban environment providing a qualitative analysis. The tool will be designed to be easily adopted by other target users (cities, urban planners, policy makers etc.).
S.O.4. To implement interventions focused on empowering citizens, especially among those vulnerable or socially disadvantaged, to improve their relationship with the urban environment in such a way that behaviours decreasing the risk of NCDs are adopted. S.O.5. To promote the use of the urban environment by citizens through behavioural change interventions.
Approach
Pilots studies are characterised by their capacity to test methodological aspects of a larger study, evaluating the adequacy of specific methods and processes and acquiring knowledge or certainty of the operation of the proposed process. The implementation of the pilot study consisting of a cynical study will allow to know the efficacy of the proposed intervention for the prevention and management of non-communicable diseases. Thus, after the implementation of the pilot study we will be able to know to what extent the proposed treatment works in the target population and its differences with conventional treatments in this area. As this is a longitudinal design, it will allow us to compare the results obtained in the initial evaluation of the participants prior to the implementation of the intervention and the changes obtained after the intervention when evaluating the participants at the end of the intervention, as well as to know the safety and efficacy of the treatment administered in the intervention group in comparison with the control group.
Design
The pilot intervention as part of the pilot studies will focus on a longitudinal prospective design. The nature of the intervention will focus on a multilevel approach to behaviour change, framed in the Socioecological Model and supported by the Motivational Interviewing technique in combination with the Transtheoretical Model and the use of the Wakamola APP. The multilevel approach will ensure the care that arises from the interaction between the different systems that affect health, such as the relationships and interactions of the participants with the social and urban environment. The Motivational Interviewing (MI) approach in combination with the Transtheoretical Model of behaviour change is fully enabled by its ability to strengthen intrinsic motivation towards cognitive-behavioural health change in the population that is currently unwilling to change or ambivalent and to promote adherence to medical interventions.
Moreover, far from traditional medical counselling methods of coercive and argumentative character whose effects have not been satisfactory favouring psychological reactance by not recognizing ambivalence, MI is constituted as a supportive therapeutic approach established on an atmosphere of acceptance and empathy that recognizes the patient’s autonomy needs. Fostering and strengthening a relationship of professional patient trust where the professional’s attitude is one of acceptance and empathy towards the patient’s needs, views and experiences, will allow the detection of the stage of change according to the Transtheoretical Model and increase personal motivation and commitment to change by identifying objectives and goals in a process of co-design between the professional and the patient. The overall objective is to increase the intrinsic motivation of the client so that change emerges from within rather than being imposed from outside as it occurs. The Transtheoretical Model as a theoretical model of behaviour change provides the theoretical framework for understanding the processes of behaviour modification by interpreting it as a sequential process involving different stages, with MI being the tool that facilitates the processes for progress and change from one stage to another.
The use of the Wakamola APP will allow HORUS to support the objectives of the intervention, as it will allow monitoring the progress of participants in terms of meeting the objectives set in terms of behavioural change towards healthy lifestyles. The possibility of interacting with the rest of the participants through the APP, as well as the follow-up and rewards granted through the APP after the achievement of the objectives will be key to maintain and increase the motivation of the participants beyond the personal professional-participant interactions within the framework of the Motivational Interviews.
Target population
Vulnerable population, especially low-income population, immigrant population and ethnic minorities, with non-communicable diseases, especially diabetes and cardiovascular diseases or at risk of developing them.
Methods
The longitudinal pre-post controlled design with a baseline-measurement at inclusion and at the end of the implementation of HORUS will involve a sample of 900 participants, 300 participants for the pilot site in Spain, 300 participants for the pilot site in Croatia and 300 participants for the pilot site in the Netherlands. The sample of 300 participants from each pilot site will be split in two groups: intervention group (150 participants) and the control group (150 participants). For more information, see the appendix “Information on clinical studies.
S.O.6. To provide evidence on the effectiveness of interventions empowering citizens, especially those vulnerable or socially disadvantaged, to engage with the urban environment by adopting behaviours that decrease the risk of NCDs. S.O.7. To provide evidence on city planning criteria to prevent the adoption of NCD risk behaviours by the population.
Approach
Policy Recommendations
Design
The policy recommendations preparation process will build on a systematic analysis of how the evidence based interventions highlighted by the project results can be adopted and integrated in other socio-institutional contexts, with particular attention to vulnerable population groups and neighbourhoods. The methodological process of developing policy recommendations is strongly rooted in the collaborative project concept from citizen science, as it is based on participation, collaboration between stakeholders and active engagement with the research itself, providing design and implementation criteria.
Target population
Multi-stakeholder
Methods
As a base step, a systematised contextual analysis will be conducted in order to understand the conditions under which the interventions have been developed and to assess their replicability to other contexts. Workshops and sessions with local stakeholders will be implemented to jointly analyse the project results and prepare a set of policy briefs and recommendations for stakeholders, addressing key issues such as the socio-cultural cross-cutting nature of the results and the criteria for design, replicability and vertical scaling of the different interventions for each category of NCD. Sessions should be differentiated according to stakeholder profile and expertise. A panel of stakeholders will score the policy recommendations according to a set of scoring criteria. The final policy recommendations will be the tangible output of this process.