Obesity and Healthcare Cost

Introduction

The intent of this paper is to explore the correlation between obesity and increased cost to companies for their healthcare coverage of all employees. Our hypothesis is that obesity is increasing the cost of healthcare company wide due to obesity and the diseased associated with this epidemic. Our null hypothesis is that the effects and treatment of obesity is not having a direct impact on the total healthcare cost of organizations.

Obesity has both direct and indirect effects. Direct effects we will explore are lost time due to increased use of sick days and restrictive work duties. A few of the indirect costs are the cost of medication for diseases directly related to obesity, such as but not limited to, Diabetes, High Blood Pressure, High Cholesterol, heart related issues and cost of medication to treat and control these diseases.

In an effort to accomplish this task we will review recent studies conducted on this subject by nationally accredited institutions such as the Center for Disease Control (CDC), American Diabetes Association and the American Heart Association along with other government and public agencies. We have included the results of one of the most in-depth recent studies completer by the Medical Expenditure Panel Survey and published in the International Journal of Obesity in January 2005. We will also reference other data from other sources as well in an effort to prove the hypothesis.

Sample Size

Selecting a sample size is necessary when testing theories because there would never be enough time to test an entire population and it might be very costly. The sample size should be a fair and unbiased representation of the population. In order for the sample to be unbiased it should be selected randomly. Random selection ensures that everyone has the same amount of information and opportunity.

When researching a population the sample size depends on three factors, "the level of confidence desired, the margin of error the researcher will tolerate, and the variability in the population being studied" (Lind, 2002). When utilizing the level of confidence, the value can be anywhere between 0 and 100 percent. If a higher level of confidence is selected then the larger the sample size.

The next factor is allowable error. This is computed by adding and subtracting the allowable error to the sample mean. This calculation helps determine the end points of the confidence level. "A small allowable error will require a large sample. A large allowable error will permit a smaller sample"(Lind, 2002).

The last factor in order to determine a sample size is the population standard deviation. With the research data gathered, since the population is widely dispersed, a large sample is required. It would also be essential to use an estimate for the population standard deviation. Three suggestions for finding the estimate are to use a comparable study, use a range-based approach and to conduct a pilot study. The comparable study is used when there is an estimate available from another study. The range-based approach is utilized when the estimate of the largest and smallest values in the population has to be known. The last suggestion is to conduct a pilot study which is the most common method. This is when the validity is tested then the standard deviation is calculated and then the value is used to determine the appropriate sample size

The criteria we would need to select our sample size would be height, weight, corporate (work force) status, health issues and insurance carrier. From this information we would be able to see how much people are spending on health insurance as well as how much money companies are spending.

Methodology

There are numerous methodologies for the study of obesity and the related healthcare costs for organization associated with this epidemic. One of the most detailed studies comparing cost and obesity was the Medical Expenditure Panel Survey, completed in late 2004 and published in January 2005 in the International Journal of Obesity. The study consisted of a cross-sectional analysis of 16, 262 adults from a nationally representative survey of non-institutionalized civilian population in the United States. The researchers used the Body Mass Index (BMI) categories to classify participants in the study. The study used a two part, multi-variable model, adjusted for age, gender, race, income, education level, and type of health insurance, marital and smoking status (Arterburn, Maciejewski & Tsevat). The results of the survey are as follows:

The two most common research techniques are to randomly select a nationally representative sample and conduct either a phone or