People who have multiple social factors of vulnerability are likely to be more exposed to negative post-incident outcomes than those who do not (Morrow, 1999). The mapping of each social factor of the Boston Social Determinants of Vulnerability Framework is most helpful when paired with the results from the citywide correlation analysis used to develop the Framework. The goal is to understand the relationships between the social factors of vulnerability at the neighborhood level. For each neighborhood, emergency planners can begin to consider the social factors people may be facing in those areas in the City of Boston.

The original Social Determinants of Vulnerability Framework indicated that post-incident outcomes (lack of access to post-incident services; displacement; injury, illness, and death; loss of employment; property damage; and domestic violence) were related to three social factors: social isolation, low-to-no income, and people of color. Mattapan, Roxbury, and South Dorchester had multiple census tracts with statistically significant concentrations of people who were socially isolated, low-to-no income, and people of color. Therefore, these three neighborhoods are most likely to have the largest exposure to post-incident impacts. The maps from the hot spot analysis can be found in the Social Vulnerability Hot Spots in Boston section of this website.

Based on the Social Determinants of Vulnerability Framework, the relationships between the seven social conditions and characteristics in the City of Boston become more complicated. Overall, the model remained relatively unchanged. However, some of the relationships between the social conditions and characteristics in the framework are not as interrelated in Boston as the literature suggests. For example, older adults and low-to-no income were not as correlated with other social characteristics. Also, children were not correlated with medical illness or low-to-no income.

The variables added on the basis of my public health and emergency management experience and knowledge of Boston further complicated the model. Social isolation remained a key variable and actually became even more significant in its direct relationship with other social factors. Social isolation was correlated with all attributes directly except people without vehicles.  However, people without vehicles were related to renters, which was directly associated with social isolation.

Based on these results, I was able to modify the original Social Determinants of Vulnerability Framework into a new framework specific to Boston. The original Social Determinants of Vulnerability Framework can be found here, the Boston Social Determinants of Vulnerability Framework is at the top of this page, and the expanded pre-incident Framework for Boston can be found below.

Expanded Boston Social Determinants of Vulnerability Framework

Expanded Boston Social Determinants of Vulnerability Framework

The results of my research suggest that one of the most socially vulnerable populations across the city is women who do not have a vehicle, rent, and have a medical illness. In 12 out of 16 neighborhoods, lack of vehicle, medical illness, renter, and women were strongly correlated with social isolation (r > 0.7 in every case, with most being r > 0.8). Additionally, these four variables were clustered together in eight neighborhoods: Back Bay/Beacon Hill, Downtown, East Boston, Hyde Park, Jamaica Plain, Mattapan, South Dorchester, and South End.  Some of the clusters were correlated with social isolation as well as other social factors, which are described in the summary for each neighborhood. In East Boston, Hyde Park, Mattapan, and Roxbury, social isolation was correlated with all of the social factors in the Social Determinants of Vulnerability Framework. The circumstances of socially vulnerable populations in these neighborhoods were very closely related to one another and, therefore, they have higher vulnerability.

The neighborhood summaries are an overview of the most socially vulnerable populations in each of the neighborhoods. The analysis is based on the correlation analysis showing an existence of a relationship between social isolation and each of the other social factors as described below in Table 5 (r > 0.70, P < 0.05) and the social factors that had a high likelihood of being related to many other social factors. The other consideration was for the geographic concentrations of socially vulnerable people in those neighborhoods based on the hot spot analysis. The breakdown of population for each neighborhood and associated social determinants of vulnerability are listed below from Appendix D from my research.