Assignment - Introduction

This portion of the practice asks you to apply your quantitative analysis skills to addressseveral questions about spending for end-of-life care for Medicare decedents with chronic illness during their last two years of life (deaths occurring 2001-2005). The Assignment 4 Excel file includes data related to spending, general information about hospitals, resource inputs, and patient experience during end-of-life care.  A list of the variables included in the data file and their definitions are included on the third page of this document.

Question 1: Review the data in the Excel file as well as the information about the dataset on the next page. 

1. Divide the sample into two halves based upon one of the variables within the data set (e.g., you might want to divide the sample based on outpatient spending low vs. high - you'd simply sort the data on the outpatient spending variable and then split the group [somewhat arbitrarily] into roughly two halves based on this single variable).

2. Choose a second variable of interest.  For example, you might want to see if there is a difference in the reimbursement per patient day between the facilities with high outpatient spending and those with low outpatient spending. 

3. Complete some basic descriptive statistics related to your variables so that you can describe the variables and interpret some descriptive differences between the two groups.  Include at least one frequency table and one chart in your analysis. 

4. Based upon the two groups - high and low - complete a one-sample t-test to see if there is a significant difference between the two groups based on the second variable that you choose.  Use the hypothesis testing steps that we reviewed in class and utilized in Assignment.

5. Show your work for Question 1 in a new tab in the Excel file, leaving the "raw data" in the first tab, the "Data" tab.

6. In a separate Word file, provide basic descriptive statistics, including your frequency table and chart, regarding your data in a single paragraph.  Your write-up should be similar to how you would see the descriptive statistics narrative in a typical journal article.

7. In this same Word file, write out your hypothesis testing steps (from #4 above) and provide an inferential analysis for your t-test.

Question 2: Use this data set to address the following question:

How much of the variation in total Medicare spending (Spending_TotalMedicare) can be explained by a combination of a) type of physician involvement (medical specialists (EoL_MS), vs. primary care physicians (EoL_PCP), b)total ICU days per decedent (EoL_ICU_Days), c) total hospice days per decedent (EoL_Hospice_Days), d) number of home health visits (EoL_HHVisits); and e)CMS quality score (CMS_CompositeQualityScore.

Based on this question, you should:

For Question 2, complete your Excel work in a new tab in the same Excel file you used for Question 1, but in a new tab that you will label as Question 2.  For your narrative work below, complete this work in the same Word file that you used for Question 1.  (So, you will submit one single Excel file and one single Word file for Assignment.)

1. Conduct univariate statistics on the variables of interest.  Convey your results in a few paragraphs and appropriate tables as would be found in a Results section of a journal article.

2. State an explicit overall hypothesis to test bivariate relationships.

3. Conduct appropriate analyses to test your hypotheses. 

4. Indicate the specific questions that are addressed with your analyses(in addition to the overall hypothesis).

5. Concisely interpret your results.

6. Convey your results in a few paragraphs and appropriate tables as would be found in a Results section of a journal article.

7. Choose two of the outcome variables from the regression model and generate a correlation analysis of these two variables, including a scatterplot and Pearson's correlation value.  Interpret your results in just a couple of sentences.

8. Run a multiple regression to address the above question.  Conduct bivariate statistics on the variables of interest.

9. Note any limitations or concerns that you might have (if any) in conducting your analyses and describe what could be done to address these limitations.     

Data Set Information

This data set, End of life care for Medicare decedents with chronic illness 2001 to 2005.sav, was derived from CMS data. Data reflect Medicare spending and resource utilization for decedents with chronic illness during their last two years of life (deaths occurring 2001-2005). The data set includes information from 1,737 acute care hospitals across the country. Specific variables are Medicare spending, Medicare Part B spending; reimbursement for hospitals (facilities), reimbursement for physician visits, resource inputs per 1000 decedents, and patient experience of end of life. Variables reflecting CMS hospital compare technical process-of-care quality measures include: composite quality score, acute myocardial infarction (AMI) scores,  congestive heart failure (CHF) scores, and pneumonia scores (this set of variables includes data for all patients, 2005).

Note: Variables ending with "_1K" reflect "Resource inputs per 1,000 decedents". Variables were labeled with "per decedent" toreflect values of X per decedent.

Attachment:- Data.rar