Data Mining in Home Healthcare
Author Information
Author(s): Elizabeth A Madigan, Olivier Louis Curet
Primary Institution: Frances Payne Bolton School of Nursing, Case Western Reserve University
Hypothesis
The study aims to understand the performance of home healthcare practices in the US using a data mining approach.
Conclusion
Patients over 85 years of age experience different outcomes based on their condition and agency type, with lower lengths of stay generally for hospital-based agencies.
Supporting Evidence
- Age was a significant factor in discharge outcomes for patients over 85.
- Different agency types affected the length of stay for patients.
- Data mining techniques like CART can effectively classify patient outcomes.
Takeaway
This study looks at how older patients do with home healthcare and finds that those over 85 have different experiences based on their health conditions and the type of agency providing care.
Methodology
Data were analyzed using CART (Classification and Regression Trees) on data from the 2000 National Home and Hospice Care Survey.
Potential Biases
Potential biases include the lack of comorbid condition data and the structure of questions regarding functional status.
Limitations
The study has limitations including small sample sizes for certain conditions and the inability to track patient outcomes post-discharge.
Participant Demographics
The study focused on home healthcare patients, particularly those with chronic obstructive pulmonary disease, heart failure, and hip replacement, with a significant portion being over 85 years old.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website