uploaded?
Updated: 2/19/2022
University of Maryland Global Campus
HMGT 400 Research and Data Analysis in Health FINAL EXAM
Data File: HMGTFINALEXAM.csv (Request from Professor after you submit Group Assignment #3)
Software Tool: Analysis ToolPak, RegressItLogistic or XLMiner Analysis ToolPak
Question #1 (15 credits):
The FINAL EXAM data file provides data on hospitals in 2011 and 2012. Analyze the FINAL EXAM dataset. You may calculate the ?Hospital Beds per Population (Per Capita)? variable by dividing ?total_hospital_beds by tot_population. Use the analysis results to complete Table 1 (template) below. Net Hospital benefits = Hospital Revenues minus Hospital Costs.
Table 1. Comparing hospitals in 2011 with hospitals in 2012
2011 
2012 
t Value (Pr<t) 

N 
Mean 
St. Dev 
N 
Mean 
St. Dev 

Hospital Characteristics 

1. Hospital beds 
1078 
229.3831 
206.7204 
921 
1301.885 
32944.54 
0.285168 
2. Number of paid Employee 
945 
1146.951 
1414.9 
354 
653.2254 
875.9757 
1.18516E09 
3. Number of nonpaid Employee 
80 
48.1 
68.10483 
115 
41.87826 
43.81463 
0.438670 
4. Internes and Residents 
279 
79.65591 
138.8227 
44 
4.340909 
3.863624 
0.000377 
5. System Membership 
1079 
0.60797 
0.552588 
922 
0.646421 
0.680688 
0.092113986 
6. Total hospital cost 
1078 
203721472.4 
303617442.7 
919 
183732389.6 
265000930.8 
0.122213 
7. Total hospital revenues 
1078 
477069773.3 
1034436756 
919 
471954365.8 
1093163806 
0.917651 
8. Net Hospital benefits 
1078 
919 

9. Available Medicare days 
1068 
16538.37734 
19225.12023 
919 
16538 
0 
0.99643 
10. Available Medicaid days 
1052 
5311.276616 
9190.481016 
919 
5311 
0 
0.994757 
11. Total Hospital Discharge 
1074 
9312.395717 
10711.43761 
919 
9345 
0 
0.927247 
12. Medicare discharge 
1070 
3206.281308 
3380.252336 
920 
111902.162 
3296796.503 
0.280894 
13. Medicaid discharge 
1064 
1253.279135 
1900.277427 
905 
1171.319337 
1762.961121 
0.34953 
SocioEconomic Variables 

14. Hospital Beds per Population (Per Capita) 
1078 
0.002338 
0.004095 
919 
0.002349 
0.003575 
0.9515 
15. Percent of population in poverty 
1078 
25.96855288 
9.66858643 
919 
25.79129489 
9.528074007 
0.6901 
16. Percent of Female population in poverty 
1078 
15.87625232 
5.562371262 
919 
15.7773667 
5.444037856 
0.6696 
17. Percent of Male population in poverty 
1078 
10.09230056 
4.321048879 
919 
10.01392818 
4.316069396 
0.6933 
18. Median Household Income 
1078 
50137.41373 
13656.34485 
919 
49714.11099 
12863.14896 
0.4460 
Then please answer the following questions.
A) What are the significant differences for each of the hospital characteristics between 2011 and 2012?
The number of paid employees, on the other hand, was larger in 2011 than in 2012, indicating a considerable difference. Between the two years, there is a major difference between Internes and Residents.
B) What are the significant differences for each of the socioeconomic variables between 2011 and 2012?
There are no significant differences for the socio economic variables between 2011 and 2012.
C) Based on your findings did hospitals (in general) have a better performance in 2011 or 2012? Explain.
2011 hospital performance looks better based upon findings. In 2011 there were more hospital beds, more paid employees, more interns and residents, more revenue, more Medicaid and Medicare discharges.
D) How does hospital performance relate to the hospital characteristics and how does it relate to the socioeconomic characteristics?
Question #2 (15 credits):
In the data file, create a new variable called ?hospital net benefits?. Do this by subtracting ?hospital costs? from ?hospital revenues?.
Analyze the dataset and then complete Table 2 (template). In the last column report the Ttest results comparing hospital characteristics between 2 groups of hospitals ?1. Forprofit & Other ownership?, 2.?Nonprofit & public? hospitals. Note in the data file:
If own = 0 then the hospital is nonprofit owned
If own = 1 then the hospital is forprofit owned
If own = 2 then the hospital is public (government) owned
If own = 3 then the hospital is owned by some other organization
Table 2. Comparing ?Forprofit & Other? hospital group to ?Nonprofit & Public? hospital group in two years (2011, 2012)
For Profit & Other 
NonProfit & Public 
t Value (Pr<t) 

N 
Mean 
St. Dev 
N 
Mean 
St. Dev 

Hospital Characteristics 

1. Hospital beds 

2. Number of paid Employee 

3. Number of nonpaid Employee 

4. Internes and Residents 

5. System Membership 

6. Total hospital cost 

7. Total hospital revenues 

8. Net Hospital Benefits 

9. Available Medicare days 

10. Available Medicaid days 

11. Total Hospital Discharge 

12. Medicare discharge 

13. Medicaid discharge 

SocioEconomic Variables 

14. Per Capita Hospital Beds 

15. % population in poverty 

16. % Female population in poverty 

17. % Male population in poverty 

18. Median Household Income 
Then answer the following questions:
1) What are the main significant differences between the ?Forprofit & Other? hospital group and the ?Nonprofit & Public? hospital groups? Which test is the best fit test? Why?
2) Use a column plot or graph to compare ?Net Hospital Benefits? between ?forprofit & other? and ?nonprofit & public? hospitals.
3) Create a scatter plot with the correct variable and axes labels and legend. Use the plot to compare hospital cost (xaxes) and revenue (yaxes) and discuss your findings?
4) If we assume that hospital netbenefit is a good measure of hospital performance, which type of hospitals had the better performance between ?forprofit & other? and ?nonprofit & public? hospitals?
5) Overall, what are the main statistically significant differences between ?forprofit & other? and ?nonprofit & public? hospital groups?
Question #3 (15 credits):
Our data file provides the variable herf_ins called the Herfindahl?Hirschman Index which measures market concentration for the health insurance market. Please note that unlike the class exercise in which you used herf_cat, which measured market concentration for the hospital market, in this assignment you are using herf_ins which measures market concentration for the health insurance market.
Analyze the data to complete Table 3 (template) below:
Table 3. Comparing hospitals in High, Moderate, and Low Insurance Market Concentration in two years (2011, 2012)
High Concentration Insurance Market 
Moderate Conc. Insurance Market 
Low Concentration Insurance Market 
ANOVA 

Count 
Mean 
STDev 
Count 
Mean 
STDev 
Count 
Mean 
STDev 
F Value (Pr<F) 

Hospital Characteristics 

1. Hospital beds 

2. No. of paid Employees 

3. No. of nonpaid Employees 

4. Internes and Residents 

5. System Membership 

6. Total hospital cost 

7. Total hospital revenues 

8. Net Hospital benefits 

9. Available Medicare days 

10. Available Medicaid days 

11. Total Hospital Discharges 

12. Medicare dischargeratio 

13. Medicaid dischargeratio 

SocioEconomic Variables 

14. Per Capita Hospital Beds 

15. Median Household Income 
Then answer the following questions:
1) In a short paragraph describe and explain what the Herfindahl index is. You can use the reference provided in the class exercise or any other citation.
2) Compare the following information between hospitals located in high, moderate, and low concentration health insurance markets?
a) What are the main significant differences between hospitals in the different insurance markets? (use the ANOVA test)
b) What is the impact of being in a highconcentration health insurance market on hospital revenues and costs?
c) Does being in a high concentration market have a positive impact on net hospital benefits?
d) What about the number of Medicare and Medicaid discharges? Are hospitals in high concentration insurance markets more likely to accept Medicare and Medicaid patients?
e) What is the impact of other variables?
(Note: to answer the last question, please compute Medicaredischarge ratios and Medicaiddischarge ratios first and then run two tTests (high concentration vs. moderate concentration, and high vs. low concentration market). Please support your findings with an illustrative graph.
Question #4 (Credits 20)
Analyze the data by running a regression model with ?Net Hospital Benefits? as the dependent variable and present your results using the Table 4 template below.
Table 4 ? Regression Model 1
Coefficient 
ST. ERR 
T Stats 
Pvalues 
Lower 95% 
Upper 95% 

Intercept/Constant 

Total Hospital beds 

Teaching Hospital Dummy 

Count (N) = 

R Square = 
a) Describe and discuss your findings.
b) Do the number of hospital beds or whether a hospital is a teaching hospital or not have a positive or negative impact on hospital netbenefit. (Hospital Performance)? In answering this question, consider statistical significance.
Regression Model 2:
Analyze the data by running a linear regression model and present your results using the Table 5 template below.
Table 5 ? Regression Model 2
Coefficient 
ST. ERR 
T Stat 
Pvalues 
Lower 95% 
Upper 95% 

Intercept/Constant 

Total Hospital beds 

NonTeaching Hospital Dummy 

Count (N) = 

R Square = 
a) Describe and discuss your findings.
b) Do the number of hospital beds or whether a hospital is a nonteaching hospital or not have a positive or negative impact on hospital netbenefits. (hospital performance)? In answering this question, consider statistical significance.
c) Use the results from your Regression model 1 and regression model 2 to comment on any differences or similarities in impact of teaching hospital status or nonteaching hospital status on hospital netbenefits. (hospital performance)?
Regression Model 3:
Analyze the data by running a linear regression model and present your results using the Table 6 template below.
Table 6 ? Regression Model 3
Coefficient 
ST. ERR 
T Stat 
Pvalues 
Lower 95% 
Upper 95% 

Intercept/Constant 

Total Hospital beds 

Teaching Hosp. Dummy 

Medicare discharge ratio 

Medicaid discharge ratio 

Count (N) = 

R Square = 
a) Describe and discuss your findings.
b) Do the number of Medicare or Medicaid patients in a teaching hospital have a positive or negative impact on hospital netbenefits. (hospital performance)? In answering this question, consider statistical significance.
Regression Model 4:
Analyze the data by running a linear regression model and present your results using the Table 7 template below.
Table 7 ? Regression Model 4
Coefficient 
ST. ERR 
T Stat 
Pvalues 
Lower 95% 
Upper 95% 

Intercept/Constant 

Total Hospital beds 

NonTeaching Hosp. Dummy 

Medicare discharge ratio 

Medicaid discharge ratio 

Count (N) = 

R Square = 
a) Describe and discuss your findings.
b) Do the number of Medicare or Medicaid patients in a nonteaching hospital have a positive or negative impact on hospital netbenefit. (hospital performance)? In answering this question, consider statistical significance.
c) Based on your findings please recommend three policies to improve hospital performance. Please make sure to use the final model for your recommendation?
Question #5 (Credits 20)
Logistic Regression Models
Analyze the data by running a linear regression model and present your results using the Table 8 template below.
Use ?being a member of a hospital network? (system_member) as the dependent variable. and the independent variables presented in Table 8 template below
Table 8 ? Logistic Model 1
Coefficient 
ST. ERR 
PValue 
Exp (coeff) 
Exp (z SE) 
Exp (Std. Coeff.) 

Intercept/Constant 

Total Hospital costs 

Count (N) = 

R Square = 
a) Describe and discuss your findings.
Logistic Model 2:
Analyze the data by running a linear regression model and present your results using the Table 9 template below.
Table 9 ? Logistic Model 2
Coefficient 
ST. ERR 
PValue 
Exp (coeff) 
Exp (z SE) 
Exp (Std. Coeff.) 

Intercept/Constant 

Total Hospital Costs 

Total Hospital Revenue 

Count (N) = 

R Square = 
a) Describe and discuss your findings.
Logistic Model 3:
Analyze the data by running a linear regression model and present your results using the Table 10 template below.
Table 10 ? Logistic Regression Model 3
Coefficient 
ST. ERR 
PValue 
Exp (coeff) 
Exp (z SE) 
Exp (Std. Coeff.) 

Intercept/Constant 

Total Hospital Costs 

Total Hospital Revenue 

Medicare discharge ratio 
< 