For several years medical research has linked low level of medical care with an increased risk of cancer.This study uses statistical regression methods to analyze the relationships of cancer death rate among other variables; some of these are the rate of nurses per 100,000 civilian population, the percent of 65 years old and over resident population, the percent of people education with high school graduate or higher and the percent of social security recipients. There are 51 cases (50 states and Dist. Columbia) and 63 variables in the data set were censused by the US Census Bureau. The basic technique used in the analyses is the stepwise regression procedure.All results of testing hypotheses are obtained by the F-statistic numerical values for determining the statistical significance of regression models at each step under consideration.