1. Thank you for agreeing to participate in this experiment focusing on factors that are involved in the decision to discharge a patient from the hospital after complex surgery.
2. In this experiment you will make a series of choices between two options: "Discharge Patient" and "Do NOT Discharge Patient."
3. You will be able to make at most a total of 30 choices of the Discharge Patient option.
4. A choice of Discharge Patient is successful if the patient is not readmitted. A choice of Discharge Patient is unsuccessful if the patient is readmitted.
5. You will be paid $5 for each successful discharge. You will not be paid for unsuccessful discharges.
6. The session will last no more than 2 hours. The session will include no more than 45 experimental days.
7. You will be given a panel of three patient charts during any experimental day. Each patient chart will be accompanied by a brief clinical scenario regarding the surgery that the patient had and the patient's postoperative course up until that particular time.
8. In addition each patient chart will contain relevant clinical data regarding Inpatient Summary (Figure 2), Laboratory Values (Figure 3), Orders (Figure 4) and Vital Signs (Figure 5) for your review.
9. Each chart will also contain a recommended decision of "Discharge Patient" (as in Scenario A), "Do NOT Discharge Patient" (as in Scenario B) or "Physician Judgment" (as in Scenario C) produced by decision support software developed using data from patient records at Emory University Hospital. This recommended decision reflects experience with patients who are similar to your patient in the chart.
10. If you decide to discharge the patient from the hospital based on the clinical information you have reviewed and the recommended decision, you will click on the Discharge Patient button (see Figure 5.A or Figure 5.B). If you decide to keep the patient in the hospital for an additional experimental day, you will click on the Do NOT Discharge Patient button; we will then provide you additional data for the next experimental day and again ask you to make a decision regarding whether or not to discharge this patient.
11. Throughout the experiment, patients whom you have discharged may be readmitted to the hospital. The likelihood of readmission will be determined by the health status of the patient when discharged and historical experience with a sample of discharged patients with similar diagnoses at Emory University Hospital.
12. If a patient whom you have discharged is readmitted to the hospital, that patient will be added back to your panel as soon as the number of patients being served drops below three and you will not be able to make a decision about discharging that patient for three experimental days. During that time the readmitted patient displaces a new patient whom you could have served.
13. We are only interested in your decision to discharge or not to discharge the patient from the hospital. We want you to assume that the patient is being managed at the appropriate standard of care while in the hospital. We don't want you to speculate about tests and/or procedures that you might want to order.
14. When you are ready to consider a patient for discharge, click on the patient's ID in the version of Figure 1 that appears in the decision software screens.
15. Figures 2 - 5 will present information on the patient you have selected for review.

Figure 1: Patient Selection

Figure 2: Inpatient Summary

Figure 3: Laboratory Data

Figure 4: Orders

Figure 5: Vital Signs

Decision support software information display

16. A Figure 6 screen presents decision support software information for a specific patient on a specific experimental day. The left panel displays the estimated readmission probability and the 80% confidence interval around it. It also shows a horizontal line which is the target readmission rate for a patient with this diagnosis code. The other panels present information that the decision support software has identified as being the most relevant to making the discharge decision for this individual patient on this specific day.

17. If the decision support software recommends that the patient should be discharged on the current experimental day, then you will encounter Figure 6.A. The software recommends Discharge when the upper 80% bound on the readmission probability is at or below the target rate, as in the example shown in Figure 6.A.

18. If the decision support software recommends that the patient should NOT be discharged on the current experimental day, then you will encounter Figure 6.B. The software recommends Do NOT Discharge when the estimated readmission probability is at or above target rate, as in the example shown in Figure 6.B.

19. The decision support software will not make a decision about discharge in cases where the target readmission rate is between the estimated readmission probability and the upper 80% bound. In that case, the software will show "Physician Judgment", as in the example shown in Figure 6.C.

Scenario A: Figure 6.A

Scenario B: Figure 6.B

Scenario C: Figure 6.C

Entering and confirming your decision

1. In either scenario if you decide not to discharge the patient from the hospital then please click on the "Do NOT Discharge Patient" button. You will subsequently get new chart information on the health status of this patient to be evaluated for the next experimental day.
2. If you decide to discharge this patient then please click on the "Discharge Patient" button. You will subsequently get another patient to be evaluated for the next experimental day.

You will be paid $5 for each patient whom you evaluate and successfully discharge.