Geographic disparities and temporal changes in risk of death from myocardial infarction in Florida, 2000–2014

The rates of deaths from cardiovascular diseases (CVD), such as coronary heart disease (CHD) and myocardial infarction (MI), have decreased in the US in the last five decades [1]. However, CVD remain the leading cause of preventable premature deaths in the US, accounting for one in every four fatalities in the country [2]. MI, or heart attack, contributes significantly to this burden, with approximately 14% of the 790,000 people who experience an MI in the US each year dying from it [3].

Cardiovascular diseases also represent a serious economic burden to the US healthcare system, constituting 17% of national health expenditures in 2014 [3], with MI being the most expensive condition to treat [4]. The burden of MI is particularly high in the southeastern US states, including Florida, where 5.3 and 12% of the adult and elderly (≥65 years) populations, respectively, reported a history of acute MI in 2014 [5]. Moreover, the increase in mean age of the population coupled with an upsurge in risks of obesity and type 2 diabetes [2] are expected to exacerbate the burden of MI and increase its public health and economic costs [6].

Consistent with the trends seen nationally [1], an overall decline in MI/ischemic heart disease mortality risks has been observed in Florida [7, 8]. However, it has been shown that population subgroups defined by geography and other factors may show widening disparities in cardiovascular health, despite reductions in overall CVD mortality risks [9]. Additionally, previous studies showing geographic disparities of MI mortality risks at county- [10] and census tract-levels [11, 12] suggest that geographic hotspots of MI mortality risks may exist in Florida. Therefore, it is strategically advantageous to identify populations with high MI burdens and investigate how the MI burdens change over time to guide control programs geared towards reducing/eliminating disparities and improving population health. Moreover, understanding how MI burdens change over time may reveal the effectiveness of intervention programs and can be used to guide policy decisions and resource allocation. Unfortunately, no rigorous population-level studies have been conducted to determine if the decreases in MI mortality risks have occurred equitably across all communities in the state. Therefore, our objectives were to: (a) investigate the spatial distribution and clusters of MI mortality risk in Florida; and (b) assess temporal changes in geographic disparities in MI mortality risks in Florida from 2000 to 2014.

There were 58,198 MI deaths in Florida between 2000 and 2014. The overall annual age-adjusted MI mortality risks were 55.5 (2000–2002), 43.8 (2003–2005), 33.1 (2006–2008), 29.8 (2009–2011), and 28.1 (2012–2014) deaths/100,000 population over the study period. This represented an overall decrease of 48% in MI mortality risks during the period of interest.

We investigated geographic distribution and spatial clusters of MI mortality risks in Florida over a period of 15 years. We also identified communities with consistently high MI burden over the study period. Study findings will be useful for guiding resource allocation for intervention programs. Florida has a racially and ethnically diverse population with large proportions of minority, immigrant, and elderly populations, hence it foreshadows the demographic structure projected for the US population by the year 2030 [23]. Therefore, Florida’s strategy to address the high MI burden will not only be critical to Florida’s future, but it will be instructive for the rest of the US.

Similar to other studies using county-level data to assess cardiovascular mortality disparities across the US [8, 10], this study found disparities in the burden of MI across Florida, with the north having the highest mortality risks while the south had the lowest risks. This is consistent with the shift in the concentration of counties with high rates of heart disease-related mortality from Northeastern US to socioeconomically disadvantaged areas in the Deep South that was observed by Casper et al. [10] over a 40-year period.

The identification of high-risk clusters mainly in rural north and low-risk clusters almost exclusively in urban south suggests that different segments of Florida’s population have not benefitted equitably from preventive and treatment efforts. Moreover, these findings mirror those of stroke mortality risks in Florida between 1992 and 2012 [24]. Other studies have also reported disparities in MI/heart disease-related mortality risks in southeastern United States based on rurality. For instance, Casper et al. [10] also identified a large persistent low-rate cluster of heart disease mortality in urban counties in southern Florida and 1–2 high-rate clusters in the rural north between 1972 and 2010. Roth et al. [8], also reported clustering of low risks of CVD and ischemic heart disease mortality in South Florida counties and clustering of high risks in North Florida counties in 2014. Odoi and Busigye [12] reported higher MI-mortality risks in rural than in urban neighborhoods in middle Tennessee. Higher mortality rates for CHD, the principal cause for MI, have also been reported for rural/non-metro areas compared to urban/metro areas in southern US [25]. By contrast, Pedigo et al. [11] reported higher odds of urban and suburban neighborhoods being in a high-risk cluster than rural neighborhoods.

We did not investigate the determinants of the identified geographic disparities. However, based on findings from previous studies, the disparities may be associated with disparities in distribution of MI risk factors and access to preventive and treatment services. For instance, rural communities generally have lower prevalence of physical activity [26] and good dietary habits [27] compared to urban populations. Moreover, increased mechanization and automation of farm work has reduced the amount of physically demanding occupations in rural areas [28], making rural lifestyle more sedentary [29]. These contribute to higher risks of obesity, hypertension and diabetes which lead to higher MI-mortality risks in rural than urban areas. By contrast, the prevalence of nonsmoking, normal body weight, and physical activity, etc., are higher in urban than rural counties in US [30].

Most North Florida counties are rural, sparsely populated, medically underserved [31, 32] and have low rates of health insurance coverage [33]. Since health funding is allocated based on population, rural counties tend to have limited resources for adequate prevention and management of CVD and its risk factors [34]. The distribution of health workforce is also geographically skewed, with rural counties having inadequate supply of general practitioners [35] and cardiac specialist [36]. Moreover, cardiac centers tend to be clustered in urban center [37], leading to long travel times and poor MI outcomes.

Socioeconomic status (SES) is one of the most reliable predictors of cardiovascular health disparities, with people of low SES experiencing higher mortality from MI and other cardiovascular health outcome [38]. Clustering of CVD risk factors has been reported among US residents with low SES [39]. Socioeconomic status may also contribute to disparities in MI mortality risks by shaping exposure to unhealthy behaviors during childhood [40]. Since a majority of counties in North Florida have poor socioeconomic conditions [41], it is likely that lower SES for rural residents made them less likely to adopt and, therefore, benefit from improvements in prevention and control programs for MI [42], contributing to higher MI mortality risks in rural areas.

The composition of the populations in the different geographic regions is an important factor that may have also contributed to the disparities in MI mortality risks. North Florida has a higher proportion of African-Americans than the rest of Florida [43]. African-Americans tend to have higher burdens of MI [44] because they are less likely to receive certain cardiovascular interventions than whites [45] and as a result of stressors associated with systematic segregation in socioeconomically deprived neighborhoods during critical life stages [46]. In addition to traditional MI risk factors, environmental exposures such as higher, more variable temperatures in the north than the south [47], may have contributed to higher MI mortality risks in the north [48].

The identification of the lone high-risk cluster in Miami-Dade County was surprising because unlike other persistent high-risk clusters, it occurred in an urban county with a relatively younger population compared to Florida. Additionally, unlike the other persistent high-risk clusters, the Miami-Dade cluster was not identified in earlier county-level studies investigating geographic disparities in heart disease [10] and ischemic heart disease [8] in the US. However, the county has a high prevalence of other major risk factors for MI including hypertension (32.6%), high blood cholesterol (32.2%) overweight/obesity (87.2%), and physical inactivity (56.7%) [49]. Additionally, Miami-Dade County has a high proportion of socioeconomically-disadvantaged immigrant minority uninsured/underinsured population [50]. However, despite the high prevalence of MI risk factors and high under/uninsured rates, utilization rates for low-cost health care programs, such as the Federally Qualified Health Centers, are very low [50]. Therefore, low levels of utilization healthcare services and poorer control of hypertension and other modifiable risk factors for MI may also explain the presence of this cluster.

The reasons for the persistence of some counties in high- or low-risk clusters throughout the 15-year study period are not clear. However, persistence may be reflective of a lack of temporal changes in the geographic patterns for MI risk factors such as prevalence of cigarette smoking [51], hypertension [52], obesity, physical inactivity [53] and socioeconomic factors [54] reported in US counties.

The observed declines in MI mortality risks during the study period imply that population-wide preventive and control efforts to reduce the MI burden have had positive impacts across Florida [55]. These findings are consistent with those of other studies in the US that have shown steady declines in overall MI/CHD-related deaths at the national [56] and regional levels [57]. That a reduction in the prevalence of major risk factors contributed to reduced MI mortality risks in Florida was partly corroborated by a study that reported an 8.8% reduction in MI mortality rates in the state in 2004 following the implementation of the smoke-free ordinance in 2003. Three years prior to the ordinance, the rates declined at only 6.4% per year [58]. However, persistent clustering of MI-mortality risks, coupled with differences in rates of declines among clusters and over time indicate that geographic disparities still exist.

Disparities in geographic patterns and magnitude of rates of declines in MI mortality risks suggest that factors influencing the rates of MI mortality decline are not equitable across the state. According to Phelan et al. [42], the differential rates of decline in MI mortality risks among clusters may be related to disparities in access to social resources that influence adoption and/or the ability to benefit from improvements in MI prevention and control strategies.

The observed decline in MI mortality risks represents remarkable progress in reducing the burden of MI across Florida and is encouraging. However, in light of the fact that elimination of health disparities is one of the goals of the Healthy People 2020 national public health agenda [59], the levelling off of rates of declines from 2009 to 2014 is concerning. Thus, the goal of reducing CVD deaths by 20% by 2020 appears elusive. It is interesting to note that these results mirror the recent temporal trends reported for heart disease deaths in the US. For instance, Ma et al. [60] reported an annual rate of decline of heart disease deaths of 3.9% from 2000 to 2010, and a much slower annual rate of 1.4% from 2010 to 2013. Sidney et al [61]. reported annual rates of decline of CVD mortality of 3.79 and 0.65% between 2000 and 2011 and 2011–2014, respectively. Cardiovascular disease death rates decreased at an average of 3.7% per year between 2000 and 2011 and at less than 1%/year between 2012 and 2014, after which the rates actually increased by 1% in 2015 [4]. A deceleration in decrease in CHD mortality rates in the US was also reported between 2012 and 2015 [62]. These changes in the trajectory of MI and heart disease burden may be due to slowed progression in the favorable trends of MI prevention and/or treatment, coupled with an aging population and dramatic increases in the risks of obesity, hypertension, and diabetes mellitus over the past 25 years [2]. Capewell et al. [63] showed that improvements in survival among CHD patients in the US associated with decreases in the prevalence of CHD risk factors in the wider population were partially offset by increases in the prevalence of obesity and diabetes.

The fact that MI mortality risks for high-risk clusters at the end of the study (2012–2014) were at par with, or higher than the risks in low-risk clusters at the beginning of the study (2000–2002 period) indicates that counties in high-risk clusters lagged behind those in low-risk clusters in the south by at least 1.5 decades in reducing MI-mortality risks. Assuming a continuing downward trend, this implies that high-risk counties would require at least 15 additional years to achieve mortality risks seen in low-risk counties during the 2012–2014 period.

There was substantial progress in reducing the overall MI burden and disparities in MI mortality risks in Florida over time. However, there are persistent geographical disparities, with high-risk clusters occurring primarily in rural northern counties and low-risk clusters occurring exclusively in urban southern counties. Moreover, the reduction in MI death risks in the north lagged behind that in the south by at least 1.5 decades. Since counties within high-risk clusters account for a sizeable proportion of the total population in Florida, prevention and control strategies should be targeted to those counties to maximize efficiency and effectiveness of interventions geared towards reducing health disparities and improving health for all Floridians. Moreover, MI shares similar risk factors with other CVD such as stroke, hence these health conditions tend to have similar geographic distribution. Thus, public efforts targeting those counties we identified as having persistently high MI risks would address not only MI disparities but also stroke and several of their risk factors such as diabetes, high blood pressure, etc. Suffice it to say that it is critical that planning and public health programs need to be guided by empirical evidence such as findings from this study so as to better address issues of health inequity and improve health for all.