# Harvard University Hospital Readmission Rate Case Study

Case Study Scenario

The hospital readmission rate is often considered an indicator of an undesirable patient outcome. You quality improvement team is located at a metropolitan hospital that serves a large population. The quality improvement team is conducting a study pertaining to readmissions, smoking and number of drugs administered to admitted patients with congestive heart failure (CHF). The team is interested in reducing the number of readmissions among patients discharged with a principal diagnosis of CHF.
The team believes that the high admission rate is due to the difficulty that these patients have in controlling the number of drugs that they typically take. The team believes that by improving patient/family education regarding drug administration, the readmission rate could be reduced. Thus, they have developed the screen “CHF patients taking three or more drugs” to identify these patients before discharge. To evaluate the effectiveness of the measure, the team conducts a study on all CHF patients discharged the previous year. The results are shown below:

 CHF patients Number of Drugs Administered Readmitted Not readmitted Total ≥3 drugs 200 40 240 <3 drugs 100 900 1000 Total 300 940 1240
• Your team needs to decide how to measure the effectiveness of this measure. You think the best way would be to compute the values of the sensitivity, specificity and predictive value. Show your work and provide an interpretation of these measures.
• On the basis of your results, is this an effective measure? Why or why not?
• Once you assess the effectiveness of this measure, and your team feels confident that this is an appropriate measure to use, the Health Administrator of the Unit asks your team to examine the odds of patients being readmitted for those who were administered at least 3 drugs. Calculate the odds ratio and explain this finding to your administrator.

Next, your team is examining the number of cigarettes smoked per day in this sample and sex-specific death rate per 1000 per year among CHF patients admitted to the hospital. The following table shows the results that you obtained.

• Your team would like to calculate the attributable risk or attributable proportion of CHF death rates due to smoking 1-14 and 15-24 cigarettes per day separately for males and females. Do you notice any differences within and between these groups in death rates due to CHF based on the different number of cigarettes smoked per day?
• What can the team conclude about the death rate due to CHF for males and females separately when zero cigarettes per day are smoked? How do they compare?
• Obtain a regression equation for the data shown above that predicts readmissions based on the number of drugs administered. How is the number of drugs administered associated with readmissions for these patients?
 Cigarettes/Day Death Rate/1000/Year due to CHF Males Death Rate/1000/Year due to CHF Females 0 0.01 0.008 1-14 0.27 0.11 15-24 1.23 1.15 25+ 2.00 1.50

Your quality improvement Team is revisiting the following patient readmissions and would like to predict the number of readmissions based on the number of drugs administered.

 CHF patients Number of Drugs Administered Readmissions 1 100 2 100 3 200 4 400 5 500 6 550 7 600 8 630 9 660