Data Mining Techniques for Patient Satisfaction Data in Home Care Settings

Abstract

The purpose of this article is to describe tools (top box analysis, bottom box analysis, and segmentation analysis) that allow more thorough use and interpretation of patient satisfaction data. Techniques are demonstrated using data gathered from 22,937 home care patients at 174 agencies during the first quarter of 1999. Results provide unique insights into home care performance. Top box analysis revealed that nursing issues are most likely to be given the highest ratings. Bottom box analysis showed that very low ratings were most prevalent in explanations of billing and cost, resolution of problems, and involvement in decision making. Simple segmentation analysis showed significant differences in satisfaction based on patient characteristics. Complex segmentation analysis identified the most and least satisfied groups of patients using combinations of patient characteristics.