Participation Exercise for Module 10: Relationships and Prediction You work on a
ID: 3289406 • Letter: P
Question
Participation Exercise for Module 10: Relationships and Prediction
You work on a post-surgical floor at IU Health and your floor has been assigned a task to find out why patients perceived they did not get adequate pain relief during the first hour post-op. This came about because the clinical nurse specialist found that patient satisfaction surveys showed about 75% of patients that submitted surveys stated they had uncontrolled pain during the first day of admission. However, the protocol on the floor is to have pain medication administered within 5 minutes of the patient’s request.
The floor will be collecting data to determine if there is a relationship between the pain scores (mean scores) and the amount of time the patient states it took to receive the pain medication from the nurse after the request was made (in minutes). The goal is to see if better communication with patients about the timing of medications will help improve pain control and satisfaction survey results.
Data will be collected over 1 month. Pain will be collected using a pain scale using 1-10. Amount of time will be measured in minutes per the patient’s documentation.
1. What is the null hypothesis?
2. What is the alternate hypothesis?
3. How could this data change practice on the floor?
4. What statistical tests would be considered and what will determine which one is selected?
After 1 month data was collected on 55 patients. An anger questionnaire was also added to determine if patients were angry with the time it took to get pain meds (1 not at all angry, to 10 extremely angry).
1. What statistical test can you use to determine the relationships?
2. What would be your added null and alternate hypothesis for the relationship between anger and time to get pain meds?
3. What rule have you broken since the sample was not randomly selected?
4. How will the results possibly change the protocol for the floor in terms of pain management?
Now let’s look at the data together!!
1. Was the data normally distributed for all variables?
2. What was the relationship between pain scores and perceived minutes to get the pain medication?
a. Is a weak, moderate, or weak relationship?
b. Is it positive or negative?
c. Is it statistically significant?
d. What is your inference?
e. What type of error could have been occurred?
3. What was the relationship between perceived minutes to get pain medication and anger scores?
a. Is a weak, moderate, or weak relationship?
b. Is it positive or negative?
c. Is it statistically significant?
d. What is your inference?
e. What type of error could have been occurred?
4. How would you change the protocol for your floor?
Explanation / Answer
1. The null hypothesis is the hypothesis of no difference. Therefore, it states that all population mean pain scores are same. That is there is no relation between pain scores and the amount of time patient states it took to receive pain medication from nurse after request.
H0:mu1=mu2=...=mu5
2. The alternative hypothesis wants to see if there is relation between pain scores and amount of time patient states it took to receive pain medication from nurse after request.
H1: at least one pair of population mean is different.
3. If the test is conducted and null hypothesis is rejected, then one can conclude that there is significant difference between two timining of medication. Henceforth for which two timings, the paitients are not getting adequate medication should be determined.
4. Assume that each pain score is independent of other, data come from a normally distributed population, and all population have common variance, perform one-way ANOVA test. In case of overall significance of the test, perform Bonferroni procedure to find which pair differs significantly.