Validating a Method to Identify Nursing Home Admissions
Author Information
Author(s): Ilene H. Zuckerman, Masayo Sato, Van Doren Hsu, Jose J. Hernandez
Primary Institution: University of Maryland School of Pharmacy
Hypothesis
Can an algorithm be developed to accurately identify nursing home admissions using administrative claims data?
Conclusion
A reliable algorithm for identifying nursing home admissions was developed and validated using administrative claims data.
Supporting Evidence
- The algorithm showed 99.9% agreement with expert review for identifying nursing home residence.
- Sensitivity was 96.7% and specificity was 100%.
- Algorithm 2 had a higher Kappa statistic indicating better agreement with the expert review.
Takeaway
Researchers created a computer program to help find out if people are living in nursing homes based on their medical bills, and it worked really well.
Methodology
The study used algorithms based on provider type, place of service, and procedure codes from a claims database, validated against expert reviews.
Potential Biases
The algorithms may not be applicable to other claims databases due to unique characteristics of the MarketScan database.
Limitations
The 'gold standard' for validation was also based on claims data, which may not reflect actual nursing home residence.
Participant Demographics
Mean age of the sample population was 74 years old; 55 percent were female.
Statistical Information
P-Value
<0.0001
Confidence Interval
95% CI for Kappa: 0.93 (0.85, 1.00)
Statistical Significance
p<0.0001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website