Analysis of the 2019 Cities100 Report using this novel approach exposed gaps within the selected interventions: several DRR categories were under-utilized, and some regions were able to integrate more DRR categories compared to other regions. The current analysis provides important insight into DRR strategy gaps, indicating that the novel integrated framework can potentially serve as a useful analytical tool.

Within the 2019 Cities100 Report, one important trend observed was that the “hazard, corrective” category was most frequently used, especially by upper middle-income countries, compared to the other categories (Figs. 1, 2, 3 and 4). A reason for this might be that “hazard, corrective” interventions were cheaper and politically easier to achieve [13] because they place the burden on individuals rather than systemic issues. For instance, numerous countries encouraged residents to bike or recycle as a means of mitigating inner-city pollution (Appendix). Nevertheless, interventions that targeted systemic or organizational issues were in fact sometimes present within the “hazard, corrective” category. For instance, Delhi shut down the most polluting inner-city power stations and Stockholm recycled heat from data centers and used it for heating homes. However, these programs fail to remove dependency on polluting sources of energy because they relocate the pollution rather than removing its presence or preventing its creation, which would require “hazard, prospective” interventions. Doing so would require investment in renewable energy, an intervention that belongs to the “hazard, prospective” category. The novel DRR framework could be a useful analytical tool that uncovers a deficit of specific DRR categories, such as “hazard, prospective”, within the 2019 Cities100 Report.

Another trend observed was that the “exposure, corrective” category was significantly more represented than the other two exposure DRR categories (Fig. 1). This finding is in line with previous research, which points to “challenges of implementation and a lack of investment in preventive action” [36]. The term “preventive” as used by previous research can fit into the “prospective” or “compensatory” categories within the proposed DRR framework. Therefore, it is logical that corrective actions would be overrepresented, as seen in Fig. 1. One reason for the observed trend might be that addressing resilience to potential future exposures (“exposure, prospective” category) requires more planning and political will than reacting to current threats [16]. Additionally, addressing vulnerabilities that cannot be immediately mitigated (”exposure, compensatory” category) through resilience-building requires significant financial investment and the resulting change takes years to realize, making it difficult to justify such funding, especially in the face of immediate threats [19]. As discussed in the introduction, both of these actions are politically unappealing and it would be difficult to mobilize support for them. Moreover, focusing limited financial resources on vulnerable populations could be seen as unfair or wasteful, while resilience-building of the general population might be favored [19]. Continued public education and awareness campaigns on the importance of diverse DRR interventions can help build support for more balanced investments.

Finally, the novel integrated DRR framework revealed regional trends. European cities constituted the largest percentage of overall DRR connections (Fig. 5). Lower middle-income countries had higher percentage of DRR connections compared to upper middle-income countries, despite being in a lower income bracket (Fig. 6). Additionally, lower middle-income countries favored “exposure, corrective” DRR categories over other DRR categories, possibly because financial resources and political coordination are especially limited in lower middle-income countries, leaving mitigation of the exposure as a more viable option [37].

A limitation of the current analysis, which was done using the 2019 Cities100 Report, is that each description of a city intervention takes up exactly two pages within the original 2019 Cities100 report, meaning there was limited information for the authors to extract. Perhaps a longer explanation of each city plan might have contained more relevant information about the interventions or health outcomes, allowing for the authors of this study to deduce more DRR connections. In fact, some interventions utilized DRR categories that were not acknowledged in the explanation. For instance, many of the interventions listed financial gain as a potential benefit, but there was no explanation as to how this financial gain could lead to greater resilience. For instance, Delhi, India provided subsidies for rickshaw drivers to purchase electric rickshaws, mitigating overall city pollution as well as limiting driver exposure and proximity to tailpipe air pollution. However, the writers of the 2019 Cities100 Report did not mention that subsidizing impoverished rickshaw drivers could potentially increase their resilience in the face of an already-polluted city and prevent further vulnerability. Because the link was not explicitly made, the intervention was not counted in our analysis as “vulnerability, corrective” or “vulnerability, prospective.” If such connections were made in the description, or longer timelines were allowed between measurement of intervention and effect, different rate and distribution of DRR utilization may have been observed. Alternatively, the authors of this current paper could have utilized external sources to obtain elements missing from the 2019 Cities100 Report. However, that would alter the results of the DRR analysis because we would introduce data points from outside of the 2019 Cities100 Report. The aim of this current paper was to test the ability of the proposed DRR framework to analyze a single DRR strategy or report, rather than a compilation of DRR reports. This limitation of the 2019 Cities100 Report, however, does not impact the creation, future utilization, or reproducibility of the proposed DRR framework because application of the approach to a different report would offer a different level of detail.

A barrier to implementing the proposed DRR framework is that it requires additional time and resources that might not always be available. First, developing the necessary data sharing and coordination mechanisms needed to compile DRR interventions in a jurisdiction can be difficult in the face of organizational silos. Second, the analysis can be time- and resource-intensive because if the hazard changes, the corresponding actions and potential interventions throughout the framework would change as well. Third, underutilization of a certain DRR category does not necessarily mean that it should be implemented immediately. For instance, encouraging people to bike in heavily polluted cities exposes bikers to higher levels of air pollution, which creates a further vulnerability, so careful examination of consequences to implementation is needed. Finally, the benefit–cost ratio of interventions could vary depending on context [38]. Therefore, contextual analyses might be useful in evaluating whether adding a certain DRR category would be beneficial in certain environments.

Future analysis of a larger or more diverse data set or a variety of hazards could improve statistical significance, increasing the potential generalizability and external validity of the observed trends. Alternatively, application of the proposed DRR framework to a more thorough report or usage of multiple sources could also provide a different level of detail. Additionally, future research could investigate which barriers, such as political feasibility, lack of knowledge, high cost, or slow return on investments, should be addressed to encourage uptake of under-utilized DRR categories. Future focus on low- and middle-income countries might reveal unique gaps in those regions, as this analysis was overly focused on European and high and upper-middle income cities. This information might be particularly useful because limited resources require efficient policy planning to best reduce disaster risk, especially in areas more prone to the effects of climate change or navigating the risks within a fragile state [39]. Finally, it might be useful to analyze the same interventions from various angles/hazards in order to gain full perspective of missing DRR elements.

The current analysis indicates that the proposed DRR framework is useful in categorizing DRR interventions and in elucidating which DRR strategies are under- and over-utilized. For instance, this analysis points to a need for increased and more diverse implementation and/or explanation of DRR categories within African and South American cities (Fig. 5). Additionally, the analysis indicates an over-utilization of “hazard, corrective” DRR categories (Fig. 3), especially in upper middle-income countries (Fig. 4) compared to other DRR categories, as well as an over-utilization of “exposure, compensatory” DRR categories compared to the other compensatory categories, which would target hazard and vulnerability (Fig. 2). The analysis suggests the “hazard, compensatory,” “exposure, prospective,” and “vulnerability, compensatory” categories were under-utilized in this analysis, relative to the other categories from the novel DRR framework (Fig. 2). Further research is necessary to understand whether including such DRR categories would be beneficial to a specific context. Nevertheless, current literature indicates that comprehensive DRR strategies are more sustainable and effective and more balanced implementation may result in better outcomes [16, 24, 26, 30]. Application of the proposed DRR framework elucidates trends and gaps present in the 2019 Cities100 Report, provides insight into where future efforts should focus, and indicates that an integrated framework might be useful in creating more equitable and sustainable DRR.