Obesity and health factors as predictors of county life expectancy

¶ … predictors of average life expectancy at the county level by examining data collected nationwide on each county with respect to a number of variables. The dependent variable of average life expectancy for a county is determined by the independent variables of rates of obesity, access to a primary care physician, smoking habits and exercise activities. The most important independent variable in this relationship is hypothesized to be rates of obesity because obesity has been consistently shown to be a predictor of a number of related health conditions that can lead to reduced life expectancy. For example, obesity is linked with diabetes and heart disease, two of the leading causes of death in the U.S.A.

Identify the primary independent variable and defend why it is important.

The primary independent variable in this study is obesity. This variable represents the percentage of adults in the county who have a BMI of 30.0 or more, which is the cut off point for what is considered obese. BMI stands for Body Mass Index. Obesity has been linked to reduced life expectancy through a number of studies. It has been suggested that the average life expectancy overall in the U.S.A. can be expected to drop by 5 years as a result of the increasing rates of obesity (Dollemore, 2005). Obesity has also been linked to heart failure and heart disease, which are leading causes of death in the U.S.A., and thus promote the hypothesis that obesity is a strong predictor of average life expectancy (Contalado, Passanisi, Finelli & de Simone, 2002).

Write the general form of the model.

Predicted Average Life Expectancy = Rate of Obesity + Access to Primary Care Physician + Smoking Habits + Exercise Habits + Constant

This model is written irrespective of coefficient value (positive or negative), but simply reflects the variables in question that will be used to predict the average life expectancy within a given county.

Definition of Variables: Define & Defend All Variables

The independent variable of obesity has been described above. The remaining variables in the model are explained below.

All variables are calculated at the county level, or in other words, the subject is the county. Thus, the average life expectancy variable refers to the average life expectancy in a particular county.

Average Life Expectancy (Dependent Variable)

Average Life Expectancy refers to the average number of years an individual can be expected to live given the year that they were born. The calculation of this variable is based on the assumption that current mortality trends continue as they are now.

Access to Primary Care Physician

This variable represents the total number of non-federal physicians per 100,000 in population in the year 2007. Thus, this variable represents the level of access individuals in a particular county have to a primary care physician. Access to a primary care physician can make a large difference in the health of an individual. If a population does not have adequate access to physicians, then their health care needs might go unmet, which could severely diminish an individuals health. In particular, not having access to a primary care physician may delay the detection of life threatening diseases such as cancer, leading to a higher mortality rate and thus impacting average life expectancy.

Smoking Habits

This variable represents the percentage of adults in the county who responded "yes" to the question of whether or not they are currently a smoker. Smoking has been linked directly to a number of life threatening cancers, especially lung cancer, as well as to other life threatening illnesses such as heart disease and high blood pressure (Rosamond et al., 2008). As such, it is expected that smoking habits of a county would influence the average life expectancy within that same county.

Why should each independent variable predict the dependent variable?

Obesity, exercise and smoking are all variables that have been linked to health related outcomes in the past. Thus it is expected that they will be important predictors of average life expectancy at the county level. In addition, access to a primary care physician can greatly impact the health of individuals, with respect to their ability to receive care in the proper amount of time and their avoidance of serious illnesses.

What sign is expected for the independent variable's coefficient (positive, negative)?

Obesity will have a negative coefficient.

Smoking will have a negative coefficient.

Exercise variable will have a negative coefficient.

Access to a Primary Care Physician will have a positive coefficient.

Data Description

The source of the data for this analysis comes from the Community Health Status Indicators 2009 Project, or the CHSI 2009 Study. The data was collected through a variety of methods and represents data on all 3,141 counties in the United States. Data was collected through direct survey methods as well as through access to other nationwide databases, including Census data. The entire data set includes more than 200 variables, but just a section of these variables were used in the current analysis. The project originally began in 2000, but ran out of funding and was archived in 2004. In 2009 the study regained funding and has since updated the dataset. The variables used in this sample represent data from 2006 to 2007.

When was the data collected (Years)?

The data was predominantly collected in 2006 and 2007.

Are there any data limitations?

The data was part of a large nationally representative sample, and consists of verified data at the county level for each county within the U.S.A.

Present and Interpret the results of a Multiple Linear Regression Analysis?

A multiple linear regression analysis was conducted, entering all independent variables together predicting the dependent variable of average life expectancy. The regression equation with all 4 predictor variables was significantly related to average life expectancy at the county level, R2 = .628, adjusted R2 = .627, F (4, 1894) = 798.75, p = .000. The results of the analysis indicate that, together, the four variables account for nearly 63% of the variance in average life expectancy, thereby making these variables significant contributors to the determination of average life expectancy. The value of the coefficients for each variable were as expected, Access to a Primary Care Physician being the only coefficient with a positive value, and all of the others having a negative value. All were significant predictors at the p = .000 level.

Present the regression output & Evaluate

Write the regression equation

The regression equation is as follows:

Predicted Average Life Expectancy = -.090 Obesity -.116 Exercise - .097 Smoking + .000 Access to Primary Care Physician + 84.063

More useful for comparing the relative importance of predictors are the standardized weights, labeled as Beta on the output. The prediction equation for the standardized variables is as follows:

ZPredicted Average Life Expectancy= -.191 ZObesity -.342 ZExercise -.237 ZSmoking + .487 ZAccess to Primary Care Physician

Identify and interpret the adjusted R2

The adjusted R2 I this particular analysis differed from the non-adjusted R2 by only .001. This measure indicates that the variables, together, account for nearly 63% of the variance in average life expectancy. For a model to be able to predict 64% of a dependent variable is to suggest that the model is quite strong and that the variables within the model are very important predictors of the outcome measure. In this case, the R2 suggests that taken together, obesity, lack of exercise, smoking, and reduced access to a primary care physician can contribute significantly to a reduced life expectancy. The variability left unaccounted for in this model is only 37%.

In this particular sample, looking at the adjusted R2 is somewhat pointless, as the data is based on a population, not a sample, and as such, the adjusted R2 does not add a significant contribution to our interpretation of the data.

Discussion of the empirical results from the regression analysis

The results of the regression analysis support the predictions that each of the identified independent variables is an important and significant predictor of average life expectancy at the county level. Higher percentages of obese individuals, smokers, low rates of exercise and small numbers of active primary care physicians predicted lower average life expectances at the county level. These results echo past research that has linked these variables to specific health outcomes and highlight the importance of focusing on these issues as important areas for intervention with respect to improving health, reducing health care costs, and general health promotion.