Validating the Johns Hopkins ACG Case-Mix System for the Elderly in Sweden
Author Information
Author(s): Halling Anders, Fridh Gerd, Ovhed Ingvar
Primary Institution: Blekinge Institute for Research & Development
Hypothesis
Can the Johns Hopkins ACG Case-Mix System's predictive value for polypharmacy in the elderly be improved by adding individual-based data?
Conclusion
The Johns Hopkins ACG Case-Mix System is useful for measuring comorbidity, but additional individual factors can enhance its predictive accuracy for medication use.
Supporting Evidence
- The sensitivity of the prediction increased from 31.9% to 46.3% by adding individual-based data.
- The specificity of the model improved to 90.1% with additional covariates.
- The area under the curve increased by 11.1% when additional factors were included.
Takeaway
This study shows that using a special system can help doctors understand how sick elderly people are, and adding more personal information can make it even better.
Methodology
The study used a cross-sectional design with 1402 elderly participants, applying the Johns Hopkins ACG Case-Mix System to analyze health data and medication use.
Potential Biases
Potential bias due to varying diagnostic practices among different physicians and missing data from some participants.
Limitations
The study's findings may be limited by the lower participation rates in older age groups and missing data in self-administered questionnaires.
Participant Demographics
Participants were aged 60 to 96 years, with 41.7% men and 58.3% women.
Statistical Information
Confidence Interval
95% CI for area under ROC curve: 0.72 to 0.82
Statistical Significance
p < 0.001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website