A retrospective investigation of spatial clusters and determinants of diabetes prevalence: scan statistics and geographically weighted regression modeling approaches

Abstract

Background Diabetes and its complications represent a significant public health burden in the United States. Some communities have disproportionately high risks of the disease. Identification of these disparities is critical for guiding policy and control efforts to reduce/eliminate the inequities and improve population health. Thus, the objectives of this study were to investigate geographic high-prevalence clusters, temporal changes, and predictors of diabetes prevalence in Florida. Methods Behavioral Risk Factor Surveillance System data for 2013 and 2016 were provided by the Florida Department of Health. Tests for equality of proportions were used to identify counties with significant changes in the prevalence of diabetes between 2013 and 2016. The Simes method was used to adjust for multiple comparisons. Significant spatial clusters of counties with high diabetes prevalence were identified using Tango’s flexible spatial scan statistic. A global multivariable regression model was fit to identify predictors of diabetes prevalence. A geographically weighted regression model was fit to assess for spatial non-stationarity of the regression coefficients and fit a local model. Results There was a small but significant increase in the prevalence of diabetes in Florida (10.1% in 2013 to 10.4% in 2016), and statistically significant increases in prevalence occurred in 61% (41/67) of counties in the state. Significant, high-prevalence clusters of diabetes were identified. Counties with a high burden of the condition tended to have high proportions of the population that were non-Hispanic Black, had limited access to healthy foods, were unemployed, physically inactive, and had arthritis. Significant non-stationarity of regression coefficients was observed for the following variables: proportion of the population physically inactive, proportion with limited access to healthy foods, proportion unemployed, and proportion with arthritis. However, density of fitness and recreational facilities had a confounding effect on the association between diabetes prevalence and levels of unemployment, physical inactivity, and arthritis. Inclusion of this variable decreased the strength of these relationships in the global model, and reduced the number of counties with statistically significant associations in the local model. Conclusions The persistent geographic disparities of diabetes prevalence and temporal increases identified in this study are concerning. There is evidence that the impacts of the determinants on diabetes risk vary by geographical location. This implies that a one-size-fits-all approach to disease control/prevention would be inadequate to curb the problem. Therefore, health programs will need to use evidence-based approaches to guide health programs and resource allocation to reduce disparities and improve population health.

Determinants of disparities of diabetes-related hospitalization rates in Florida: a retrospective ecological study using a multiscale geographically weighted regression approach

“…Population density, which was used as a measure of rurality in this study, was significantly associated with DRH rates, with more densely populated areas tending to have lower rates of hospitalizations. In the present study, the association between population density and DRH rates was strongest in northern Florida, where spatial clusters of pre-diabetes and diabetes prevalence [ 20 , 22 , 88 ], as well as stroke prevalence [ 89 ] and myocardial infarction mortality [ 90 ] have been identified. Previous research suggests that rural residents are more likely to report delaying health care due to cost than urban residents [ 91 ].…”

“…In Florida, the most populous state in the Southeastern US, over 2 million adults have been diagnosed with diabetes [ 18 ], and it is estimated that an additional 546,000 Floridians have diabetes that is yet to be diagnosed [ 19 ]. Within the state, there is significant geographic variation in the distribution of diabetes prevalence [ 20 – 22 ]. In addition, recent research has identified local geographic hotspots of diabetes-related hospitalization (DRH) rates [ 23 ], indicating that some communities in the state bear a disproportionately high burden of potentially preventable diabetes complications.…”

Investigation of geographic disparities of diabetes-related hospitalizations in Florida using flexible spatial scan statistics: An ecological study

“…Previous research in Florida identified county-level geographic disparities in pre-diabetes and diabetes prevalence [ 17 – 20 ]. The locations of those identified high-prevalence diabetes clusters and the high-rate DRH clusters identified in this study exhibited some overlap, particularly in south-central Florida and near Tallahassee [ 18 – 20 ].…”

“…This is consistent with findings of the aforementioned Duval County study [ 22 ]. The proportion of non-Hispanic Black residents has also been reported to be a predictor of county-level pre-diabetes [ 17 ] and diabetes prevalence in statewide investigations in Florida [ 20 ]. Racial disparities with respect to diabetes-related quality of care and management have been reported in numerous studies in the United States, and may contribute to disparities in adverse outcomes [ 55 , 64 , 66 – 68 ].…”

“…However, there is a lack of information regarding the distribution of diabetes-related outcomes in Florida. Furthermore, while county-level geographic disparities in pre-diabetes prevalence, diabetes prevalence, and participation in validated self-management education and support programs have been identified within the state [17][18][19][20], information at the sub-county level is limited to model-based estimates [21] or focused, single-county investigations [22]. Filling this knowledge gap by characterizing the local distribution of diabetes-related hospitalization rates across the state will be essential to reduce disparities of adverse diabetes outcomes.…”