Conduct a test of hypothesis
to determine if there is a difference on average for total costs between
males and females?
Create a multiple regression
that predicts total costs from gender and length of stay. When a
nominal variable with two levels is included in a multiple regression
two parallel lines are estimated, one for each level. The regression
coefficient for the nominal variable represents ½ of the distance
between the two lines or one half of the effect of gender on total
costs.
Compare the difference
in the gender mean total costs found in part a with the gender effect
(twice the gender regression coefficient) in part b. Why do they
differ?
The Alice Hyde Medical
Center is a 76-bed hospital located in Malone, New York, and is also
a part of the University of Vermont Health Network.
Create a JMP data table
from the file ADK_Newborns_2014.jmp that contains only the data for
infants born at the Alice Hyde Medical Center.
Conduct two simple regression
analyses, one that predicts total costs from length of stay and one
that predicts total costs from birthweight (lbs). Compare these regressions
to those from the CVPH found in the case “Building a Simple
Predictive Model for Health Care Costs for Newborns in Adirondack
Hospitals.”
Perform a multiple regression
for the Alice Hyde Medical Center newborn total costs with length
of stay and birthweight (lbs) as predictors. Compare your results
to those from CVPH found in this case.
Select another New York
State hospital outside of the Capital/Adirondack health services area
and download the data for newborns from SPARCS for 2014. Repeat the
analysis of this case and compare your results to those of the Champlain
Valley Physicians Hospital.
Explore JMP’s
Partition platform (Analyze > Predictive Modeling > Partition)
with the Decision Tree option to predict total costs from the available
data in CVPH_Newborns_2014.jmp.
How does this method
compare to multiple linear regression?
How do the results compare
between the two methods?