Prediction of Hospital Inpatient Charge Levels at Admission for Individual Cases

• In Event: Breakfast with Research Roundtable Forum Session

Abstract

"Abstract of:
Prediction of Hospital Inpatient Charge Levels at Admission for Individual Cases

Any form of business would be interested in determining prior to purchase which individual clients would be likely to spend the most money upon a mix of products and services. A hospital is no different. In this paper, the problem considered is that of predicting whether or not total charges accrued to a given hospitalization episode of an individual will be high or low. Hospitals could use such data for cost monitoring and possibly cost control. We develop our model from information specific to each case that is available to the facility at the time of patient admission. Proper categorization of episodes of care in regard to charge levels at the time of admission would enable the facility to respond appropriately to the individual case prior to the actual accrual of charges
The resulting model shows which of the available variables most influence the charge level of hospital inpatient cases. The model is shown to be stable over time by a number of measures. The model is also shown to be a significant improvement over random expectation. This method could also be used successfully in other fields."

Author

• Gerald S. Silberstein, Sage College of Albany