Homework answers / question archive / Analysis #4: Multiple linear and logistic regression, sensitivity and specificity, and statistical control charts (10 points)   Background and Objectives   We asked young adults at our emergency clinic to complete a short survey at the end of their visit

Analysis #4: Multiple linear and logistic regression, sensitivity and specificity, and statistical control charts (10 points)   Background and Objectives   We asked young adults at our emergency clinic to complete a short survey at the end of their visit

Statistics

Share With

---

Analysis #4: Multiple linear and logistic regression, sensitivity and specificity,

and statistical control charts (10 points)

 

Background and Objectives

 

We asked young adults at our emergency clinic to complete a short survey at the end of their visit. The survey collected information on a few of their personal characteristics and perceptions: their age, gender, insurance status and also contained a few short screening tools that assessed their fear of doctors, confidence in their provider’s experience, and a diabetes screen. It also inquired about two different outcomes: their satisfaction with their visit and if they would recommend the clinic to a friend. We would like to see if we can find a relationship (predict) between the patient characteristics and their perceptions and if they were satisfied with their visit as well as what variables might be associated (predict an increase likelihood) with them wanting to recommend our clinic to others.

 

Data from the two sources were compiled into datafile (emersat.ANAL4_F20.xlsx). 

 

In this exercise you will demonstrate your ability to:

1)         Conduct multiple linear regression, interpret the output and summarize the findings

2)         Interpret logistic regression output and summarize the findings

3)         Describe the specificity and sensitivity of a screening tool

4)         Prepare a run chart and summarize findings

 

Assignment

 

Part A.  Multiple Linear Regression

Task: Evaluate if there is a relationship (predict) between the personal characteristics and perceptions and their satisfaction score. Prepare a short description of what was done and what you found. IV=independent variable, DV= dependent variable

 

Conduct a multiple linear regression to predict satisfaction using all of the personal characteristics and perceptions variables as predictors (if appropriate).

Follow the guide in Module 9 of how to conduct this analysis and include in your description what you did such as the following:

1. 1. Define the hypothesis
   2. Describe each variable using appropriate descriptive statistics; no need to recode anything but make sure dummy coding is correct; create a ‘table 1-remember analysis exercise 1’for this step
   3. Run bivariate associations (why? need IV by each IV to check for _________)
   4. Run the full model (DV and multiple IVs ) –check assumptions, etc (for this exercise it is ok to enter the selected IVs all at once in one ‘block’)
   5. Summarize the above (a-d) and the results in your OWN words.
   6. Include IS raw output view or the Excel output.
       

 

 

Part B. Multiple logistic regression

 

Task: Now we would like to see if we can find a relationship (predict) between satisfaction and some of the personal characteristics and the chance of recommending the clinic to others.   Prepare a short description including the following information:

 

1. Is running a multiple logistic regression appropriate for this task? Explain why it is or is not appropriate.
2. Define the hypotheses
3. How many and what percent of patients indicated they would recommend the clinic?
4. You do not need to run logistic regression in EXCEL or IS. Use the output below to write a summary of what we find.

 

Recommend 0=no/1=yes		B	S.E.	Sig.	OR	95% C.I.
						Lower	Upper
	age	-.182	.115	.112	.834	.666	1.044
	female	1.439	.604	.017	4.218	1.292	13.767
	insurance	.293	.612	.632	1.341	.404	4.447
	confid	-.113	.056	.046	.894	.800	.998
	sugar	.092	.074	.215	1.096	.948	1.267
	Constant	1.188	4.504	.792	3.282

 

 

Part C. Sensitivity & Specificity

Recall that our survey included a 7 item diabetes screen. We want to assess if the results of these screening questions are valid and accurate by comparing our screening result with a gold standard  (blood glucose reading). We identify 10 true positives out of the 23 our screener identified as being positive and 35 true negatives.

 

1. Fill in the following table

	Gold standard positive	Gold standard negative	Total
Screened positive	10	13	23
Screened negative	10	35	45
Total	20	48	68

 

2. Calculate the sensitivity of our screening test. 
3. Calculate the specificity of our screening test. 
4. What does this mean (also think about false positive and false negatives)—was our diabetes screening tool detecting those at risk for having a high glucose result?  Would you want to use this screening tool?                                                          

 

Part D. Run Chart   

Use the data in HPVvaccine.xls to create a Run Chart to help track an effort to improve the rate of initiation of the HPV vaccination series in an inner city pediatric service. The following briefly describes the context:

 

Human papilloma virus (HPV) is a sexually transmitted infection with a national prevalence of greater than 70 million. Most infections are among persons 15-24 years of age. The HPV vaccine has been shown to have 100% efficacy for protection against the carcinogenic strains when practitioners administer 2–3 doses before natural exposure. Currently, the Advisory Committee on Immunization Practices recommends routine HPV vaccination at age 11 or 12 years, and vaccinations can be given starting at 9 years old. Despite these recommendations, national vaccination rates remain less than 50%.

 

The quality improvement plan incorporated several plan-do-study-act interventions: 1) an educational staff in service to improve knowledge about the safety and efficacy of the vaccine, 2) two cycles targeting providers ability to recommend the vaccine strongly and at the ability of the residents to address parental concerns , 3) implementing a standardized script to use when communicating with patients and families about the HPV vaccine, 4) providing parents information about vaccine during triage, and 5) lowered the EMR prompt to remind providers for the need from age 11 to 9. These occurred at staggered time periods. The data file indicates when each intervention was started. 

 

Baseline HPV vaccination rates was determined via chart reviews of random samples of 25 patients per month for April through October 2017. Progress toward their goal was evaluated monthly. Eligible patients were a) children between 9 and 13 years of age presenting for all acute and well child visits, and b) those who had not initiated the HPV vaccine series. The number of eligible patients was highly variable from month to month, ranging from 11 to 65 eligible patients within each month abstracted. They set a goal to increase the percentage of HPV vaccination series initiated in children 9 through 13 years of age to 65% over 7 months of the Plan Do Study Act cycles.

 

In addition to creating a figure that illustrates the run chart, provide a summary of the context of the analysis and your run chart findings that include answers to the following questions:

1. Did the proportion of children initiating HPV vaccination change—what was the mean before and after the program change?
2. Did all the changes that were made lead to improvements?
3. What data would you want to start collecting to determine other steps for quality improvement in these patients?