With increasingly complex safety-critical systems like healthcare being developed and managed, there is a need for a tool that allows us to understand their complexity, design better strategies and guide effective change. System dynamics (SD) has been widely used in modelling across a range of applications from socio-economic to engineering systems, but its potential has not yet been fully realised as a tool for understanding trade-off dynamics between safety and efficiency in healthcare. SD has the potential to provide balanced and trustworthy insights into strategic decision making. Participatory SD modelling and learning is particularly important in healthcare since problems in healthcare are difficult to comprehend due to complexity, involvement of multiple stakeholders in decision making and fragmented structure of delivery systems. Participatory SD modelling triangulates stakeholder expertise, data and simulation of implementation plans prior to attempting change. It provides decision-makers with an evaluation and learning tool to analyse impacts of changes and determine which input data is most likely to achieve desired outcomes. This thesis aims to examine the feasibility of applying participatory SD modelling approach to safe and efficient staffing level management within hospital pharmacies and to evaluate the utility and usability of participatory SD modelling approach as a learning method. A case study was conducted looking at trade-offs between dispensing backlog (efficiency) and dispensing errors (safety) in a hospital pharmacy dispensary in an English teaching hospital. A participatory modelling approach was employed where the stakeholders from the hospital pharmacy dispensary were engaged in developing an integrated qualitative conceptual model. The model was constructed using focus group sessions with 16 practitioners consisting of labelling and checking practitioners, the literature and hospital pharmacy databases. Based on the conceptual model, a formal quantitative simulation model was then developed using an SD simulation approach, allowing different scenarios and strategies to be identified and tested. Besides the baseline or business as usual scenario, two additional scenarios (hospital winter pressures and various staffing arrangements, interruptions and fatigue) identified by the pharmacist team were simulated and tested using a custom simulation platform (Forio: user-friendly GUI) to enable stakeholders to play out the likely consequences of the intervention scenarios. We carried out focus group-based survey of 21 participants working in the hospital pharmacy dispensaries to evaluate the applicability, utility and usability of how participatory SD enhanced group learning and building of shared vision for problems within the hospital dispensaries. Findings from the simulation illustrate the knock-on impact rework has on dispensing errors, which is often missing from the traditional linear model-based approaches. This potentially downward-spiral knock-on effect makes it more challenging to deal with demand variability, for example, due to hospital winter pressures. The results provide pharmacy management in-depth insights into potential downward-spiral knock-on effects of high workload and potential challenges in dealing with demand variability. Results and simulated scenarios reveal that it is better to have a fixed adequate staff number throughout the day to keep backlog and dispensing errors to a minimum than calling additional staff to combat growing backlog; and that whilst having a significant amount of trainees might be cost efficient, it has a detrimental effect on dispensing errors (safety) as number of rework done to correct the errors increases and contributes to the growing backlog. Finally, capacity depletion initiated by high workload (over 85% of total workload), even in short bursts, has a significant effect on the amount of rework. Evaluative feedback revealed that participatory SD modelling can help support consensus agreement, thus gaining a deeper understanding of the complex interactions in the systems they strive to manage. The model introduced an intervention to pharmacy management by changing their mental models on how hospital winter pressures, various staffing arrangements, interruptions and fatigue affect productivity and safety. Although the outcome of the process is the model as an artefact, we concluded that the main benefit is the significant mental model change on how hospital winter pressures, various staffing arrangements, interruptions and fatigue are interconnected, as derived from participants involvement and their interactions with the GUI scenarios. The research contributes to the advancement of participatory SD modelling approach within healthcare by evaluating its utility and usability as a learning method, which until recently, has been dominated by the linear reductionist approaches. Methodologically, this is one of the few studies to apply participatory SD approach as a modelling tool for understanding trade-offs dynamics between safety and efficiency in healthcare. Practically, this research provides stakeholders and managers, from pharmacists to managers the decision support tools in the form of a GUI-based platform showcasing the integrated conceptual and simulation model for staffing level management in hospital pharmacy.

Indigenous (Métis, First Nation, and Inuit) peoples and communities in Canada, especially in the prairies, continue to experience disproportionate levels of tuberculosis (TB) compared to the rest of the Canadian born population. This inequitable distribution of TB disease burden demands effective policy, program, and practice responses. These have so far failed to materialize, perhaps in part because of limitations in the approaches we have taken to understanding the issue. As well, these responses have largely been grounded in western scientific paradigms. Science is the search and the re-search for knowledge and this varies according to the perspectives and paradigms of the researcher(s) and stakeholders. In this project, the student researcher collaborated with the Métis Nation-Saskatchewan (MN-S) and two volunteer health researchers to adapt and ground a western paradigm and methodology (System Dynamics and Group Model Building) to a Métis research paradigm to understand experiences of tuberculosis (TB) among Métis people. Data collection took place in a 2-day Métis-adapted group model building (GMB) workshop. The outcome is a causal loop diagram with associated stories co-created by the team and the workshop participants. The workshop was evaluated using a storytelling and story listening method that explored the appropriateness of adapting GMB within a Métis research context. The approach was determined to be successful methodologically, and substantively new knowledge was created in our Métis community about the determinants of TB. This research was a journey of diversity, working at the intersection of knowledge systems to produce a new understanding of a health issue as complex as TB.