The public mental health conceptual framework brings together academic literature, grey literature, and consultations with practitioners, policy makers, and members of the public. This paper provides a detailed description of the development process of a public health tool grounded in academic research, lived experiences, and practitioner perspectives in order to create a comprehensive view of public mental health, produced in an accessible format. The iterative consultations helped us shape the framework to ensure it is informative and meaningful to our stakeholders.
Identifying determinants using multiple sources: This comprehensive conceptual framework was made possible by bringing together information from multiple sources. One of the interesting aspects of this research was to observe what determinants were identified across all three data source and which ones were unique to a particular perspective. While there were many determinants that appeared in all sources, the public mind maps identified many determinants which had not been discussed in the academic or grey literature. This additional richness achieved through extensive public consultation speaks to value that meaningful public involvement can have in the development of conceptual frameworks, which has not been described in the development of other frameworks. Further, this project has demonstrated that mind maps may be an accessible and useful tool for soliciting public input into future research projects.
Defining determinants: This is the first time many of these determinants have been defined in relation to mental health, representing an important contribution to our understanding of the drivers of public mental health. A further challenge was to determine the appropriate language for the tool, which will communicate the scientific and technical aspects into accessible language that meets the needs of our stakeholders.
Prioritising determinants: During our stakeholder workshop, we asked stakeholders review the potential determinants list, identify any missing determinants, and prioritise potential determinants based on their importance to public mental health and their amenability to change through public health action. When sorting determinants by importance to public mental health, many stakeholders asked if we had estimates of the strength of the relationship between each potential determinant and mental health outcomes. A limitation of our approach was that we did not have systematic estimates of these effects, which meant participants had to rely on their pre-existing knowledge of the literature or personal experiences to provide their view on the importance. Ranking the determinants based on amenability to change through public health action was similarly difficult. First, public health interventions often lack thorough, long-term evaluation, so the evidence for effectiveness of interventions was limited. Further, many factors may be amenable to change, but constraints around resources cannot be ignored, as resource allocation and political will are critical factors to consider.
Overall, this was a challenging task, as participants were asked to comment on 72 potential determinants within a single session. Other prioritisation exercises the authors have participated in have divided potential determinant lists into subsections to be reviewed separately by small groups. While this has an advantage of allowing for more conversation for each potential determinant, it does not provide each participant with an overall view of what is being considered for inclusion in the framework. One of the goals of this exercise was to identify completeness and gaps, which would not have been possible without reviewing the full potential determinant list. Further, there was rich discussion about which level several of the determinants should be included at, which was informative for our later decisions.
The online prioritisation exercise also had challenges. While each participant was able to take as much time as they needed to complete the prioritisation, there was little context provided to inform their decision on how to rank each determinant. The online survey was helpful in that it permitted a higher number of stakeholders to participate, however, there was limited ability to discuss broader conceptual issues or provide context and clarification to participants in the online format.
Overall, asking stakeholders to prioritise an extensive list of determinants during an in-person or online consultation risked simplifying a complex process. Each determinant is intricately linked with other determinants in the framework, so using a prioritisation exercise to justify investing in some areas while dismissing others may ignore the complex system within which determinants arise. In order to address this challenge, we have attempted to embed systems thinking to the way we visualised the connections between determinants and worked to ensure the conceptual framework was able to illustrate the intersectionality of determinants for public mental health.
Sorting determinants into levels and groups: At the various development stages, we had between 48 and 72 determinants, which became difficult to review and discuss in detail. As mentioned above, within each of the determinants could be multiple additional aspects to consider. To simplify the overall view, we sorted determinants into four levels: individual, family, community, and structural. These four levels had face validity and were similar to existing frameworks for mental health, well-being, and the social determinants (12). We further divided determinants within each level into 15 groups, which allowed us to create small groups of closely related determinants. However, the process of sorting and assigning determinants to specific levels and groups had some limitations. In particular, there were several determinants which reasonably fit in multiple levels. Each determinant is undeniably linked to factors across the framework and several could be comfortably placed in several levels. We chose to categorise each determinant at a single level for simplicity and accessibility of the tool. We used the connected determinants link to communicate that each determinant is linked to other determinants within the tool.
We also grouped similar determinants within each level. For example, we created a ‘sociodemographic’ group which includes education, employment, housing, and income, which were thought to be closely related to each other. When defining the final list of determinants, we combined several of the determinants in an effort to balance nuanced specificity and the need for simplicity and accessibility. Following the online consultation, the research team met to discuss if any individual determinants could be reasonably combined with another without losing its meaning. Some examples of combined determinants include social support and networks, genetic and biological factors (e.g. age, hormones, brain chemistry), and health behaviours (including physical activity, nutrition, sleep, and substance use). The results are therefore a parsimonious list of determinants which still capture the breadth of factors that affect public mental health.
When observed as a whole, it was apparent that the distribution of determinants was not equal between the levels. While the individual level boasted 21 determinants, the family, community, and structural levels have far fewer: 9, 12, and 13 respectively. While the literature on public mental health widely recognises the importance of higher-level determinants, the preponderance of determinants at the individual level may reflect a bias in measurement and research. Much public mental health research is based on population surveys, which have provided rich information on many individual factors, like socioeconomic status, life experiences, and health behaviours. Few large-scale population studies have included scales which measure social norms, system performance, or political factors. This impacts the amount of evidence we have on the relationship between each determinants and mental health. The paucity of evidence, particularly at the structural level, may be due to gaps in measurement, rather than lack of an important association. Rather than continuing to replicate known associations at the individual level, public health researchers should consider which higher-level constructs may be most relevant to public mental health and how these could be measured. There may be opportunities to develop new measurement approaches or work in multidisciplinary teams to access alternate data sources.
Multiple sources of knowledge informed the framework: A key strength of the development process was that we used an inclusive, collaborative process which brought together voices beyond the academic literature to create a co-produced and comprehensive picture of public mental health. Frameworks based solely on academic literature can miss determinants that are important to the lived experience of public mental health, particularly if these determinants are difficult to measure. We brought together well-established determinants from the academic literature and grey literature with expertise from lived experience and public health practice.
The substantial overlap between the determinants identified in academic research, reflected in policy documents, and supported by members of the public demonstrate a wide acceptance of several key determinants of public mental health. This was further shown in the importance rankings from the online survey, where most determinants received an average score of moderately important, reflecting broad consensus of the determinants of public mental health.
The unique determinants identified in each of the sources highlighted the utility of bringing together multiple sources to capture a comprehensive view of public mental health, as a framework based on only one form of evidence would miss important aspects. Notably, the emergence of personal traits and attributes within the public mind maps, including sense of self, aspirations, self-regulation, and sense of contribution were not discussed in the academic or grey literatures. This suggests that there was an evidence gap for the relationship between of personal traits and attributes on mental health outcomes which may require further investigation.
The nuanced detail that was included in the public mind maps was also distinct from the academic and grey literatures in many cases. For example, the mind maps emphasised that myriad aspects of education, including accessibility, inclusion, quality, and completion, were all determinants of public mental health. The predominant measure of education explored in the academic and grey literature was level of education, which might reflect the measure that is most commonly included in surveys and studies but that does not capture the complexity of the relationship between education and mental health.
In the area of trauma and adversity, the academic search had identified literature related to specific types of abuse and trauma, including economic abuse, sexual abuse, emotional abuse, physical abuse, intimate partner violence, rape, stalking, and more. These were captured more broadly by life stage in the other two sources – adversity experienced in childhood and adult trauma and adversity. The level of detail included in the measures from the academic literature may reflect research exploring hypotheses estimating the specificity of trauma type on mental health outcomes, while the focus on overall life course exposures may reflect a broader perspective on how adverse experiences during childhood and adulthood may have different effects on mental health.
The lack of commercial determinants represented in the public mind maps might indicate a relatively low level of public knowledge of how corporations and private businesses can impact mental health. This might represent an opportunity for further public discourse about the role that private corporations have on mental health outcomes and motivate further research to better understand the mechanism linking commercial determinants to mental health.
Iterative consultations: The highly consultative process we followed allowed us to ground the framework in research as well as lived experience, public health practice, and policy. This enabled us to capture a broad range of perspectives, which has not previously been done in public mental health. The iterative approach enabled us to revise the determinant list at each stage and encourage conversations across the stakeholder groups to further adapt the framework.
Format: The public mental health conceptual framework (www.publicmentalhealth.co.uk) has been created to be a highly visual representation of the drivers of public mental health. These visuals, in combination with the evidence-based research, resources, and lived experience perspectives, has created an innovative and engaging tool. We have developed the framework as an interactive online tool, which includes a simple overview as well as detail within each determinant. These different views allowed us to create a tool which may be useful to a variety of audiences. The extensive user testing and consultation allowed us to adapt the online tool to meet the needs of our stakeholders. We have also created a printable version, which captures the full conceptual framework at this time and is a snapshot of our current knowledge of the determinants of public mental health (Appendix B).
While we made an effort to create a comprehensive framework which summarised the current knowledge of the determinants of public mental health, there are some limitations to note.
Strength of association between determinants and mental health: We used literature reviews and consultations to identify determinants for this framework, but we did not have the capacity to explore the strength of associations between each determinant and mental health. Thus, our framework does not provide information on which determinant was most strongly related to mental health, which might be useful information to inform the likely impact of interventions to modify the drivers of mental health problems.
Difficulties representing intersectionality in a meaningful way: From the beginning of this research, we identified that it was critical to represent the intersectionality of determinants in the framework. In the final framework, we included links to connected determinants, but this does not fully capture the interplay between multiple factors across all levels. While the inclusion of the connected determinant links highlights our desire to consider intersectionality, this solution falls short of what could be achieved in a dynamic system map which identifies complex and changing relationships between multiple factors.
Limitations around effective interventions: We ran scoping searches to identify key resources and interventions designed to address each determinant. However, the list of included resources is not comprehensive and further exploration of the evidence around effective interventions would strengthen this tool.
Investment in development and sustainability: This tool has represented a significant investment of time and resources. We initiated this project in Spring 2019, and it has taken more than two years to complete the research, consultations, and design. This has been a resource-intensive project, requiring skills from numerous collaborators. Early feedback from our stakeholders has indicated that they find the tool interesting, engaging, and potentially useful for their work by highlighting resources, interventions, and lived experience perspectives. However, this iterative consultative process and the interactive visual product may not be suitable for future research projects which have shorter timelines and fewer resources.
Sustainability: This conceptual framework is more complete than others that are currently available, representing a major step forward in bringing the disparate evidence on public mental health together. However, continued investment would be needed to update the tool as new evidence emerges.