Statistical Methods - Text Mining Techniques for Healthcare Provider Quality Determination - page 95

Table 11. Independent variables

Age Race

Sex Acute myocardial infarction

Congestive heart failure Peripheral vascular disease

Cerebral vascular accident Dementia

Pulmonary disease Connective tissue disorder

Peptic ulcer Liver disease

Diabetes Diabetes with complications

Paraplegia Renal disease

Cancer Metastatic cancer

Severe liver disease HIV

Hospital

Figure 17. Residuals for linear regression

beyond three standard deviations is equal to 20%; the probability assuming normality indicates that only
10% should be beyond that point.
This difference in outliers can be extremely important. For example, the Centers for Medicare
and Medicaid provide information on how hospitals are reimbursed based upon patient procedures:
(Anonymous-PPS, 2008)

The base payment rate is comprised of a standardized amount, which is divided into a labor-related and
nonlabor share. The labor-related share is adjusted by the wage index applicable to the area where the
hospital is located and if the hospital is located in Alaska or Hawaii, the nonlabor share is adjusted by
a cost of living adjustment factor. This base payment rate is multiplied by the DRG relative weight.