BMED $4702$

Medical Information systems

Midterm Assignment

The data set we'll be using in the midterm assignment, ClaimsData.csv$,$ is structured to

represent a sample of patients in the Medicare program, which provides health insurance to

Americans aged $65$ and older, as well as some younger people with certain medical conditions.

The observations represent a $1\%$ random sample of Medicare beneficiaries, limited to those

still alive at the end of $2008.$

Our independent variables are from $2008,$ and we will be predicting cost in $2009.$

Our independent variables are the patient's age in years at the end of $2008,$ and then several

binary variables indicating whether or not the patient had diagnosis codes for a particular

disease or related disorder in $2008$: alzheimers, arthritis, cancer, chronic obstructive pulmonary

disease, or copd, depression, diabetes, heart.failure, ischemic heart disease, or ihd, kidney

disease, osteoporosis, and stroke.

Each of these variables will take value $1$ if the patient had a diagnosis code for the particular

disease and value $0$ otherwise.

Reimbursement$2008$ is the total amount of Medicare reimbursements for this patient in $2008.$

And reimbursement$2009$ is the total value of all Medicare reimbursements for the patient in

$2009.$

Bucket$2008$ is the cost bucket the patient fell into in $2008,$ and bucket$2009$ is the cost bucket

the patient fell into in $2009.$

These cost buckets are defined using the thresholds determined by data supplier.

So the first cost bucket contains patients with costs less than $$3,000,$ the second cost bucket

contains patients with costs between $$3,000$ and $$8,000,$ the third cost bucket contains

patients with costs between $$8,000$ and $$19,000,$ and the fourth cost bucket contains patients

with costs between $$19,000$ and $$55,000,$ and fifth cost bucket contains patients greater than

$$55,000.$

Q$1$: Calculate the patient number percentages of each bucket by creating a table of the

variable bucket$2009$ and divide by the number of rows in Claims.

Our goal will be to predict the cost bucket the patient fell into in $2009$ using a CART model.

But before we build our model, we need to split our data into a training set $($ClaimsTrain$)$ and a

testing set $($ClaimsTest$).$ Therefore, load the package caTools, and then set our random seed to

$88$ so that we all get the same split. And set SplitRatio to be $0\mathrm{.}6.$

Q$2$: What is the average age of patients in the training set, ClaimsTrain?