Dementia Prediction in Primary Care
Author Information
Author(s): Frank Jessen, Birgitt Wiese, Horst Bickel, Sandra Eiffländer-Gorfer, Angela Fuchs, Hanna Kaduszkiewicz, Mirjam Köhler, Tobias Luck, Edelgard Mösch, Michael Pentzek, Steffi G. Riedel-Heller, Michael Wagner, Siegfried Weyerer, Wolfgang Maier, Hendrik van den Bussche
Primary Institution: Department of Psychiatry, University of Bonn, Bonn, Germany
Hypothesis
Can a prediction score based on easily obtainable information in primary care effectively identify individuals at risk for dementia?
Conclusion
The prediction score has useful accuracy and can help identify individuals at risk for dementia for appropriate interventions.
Supporting Evidence
- The prediction score achieved a prediction accuracy of 0.84 for Alzheimer's disease.
- At a cut-off of >=9 points, the score achieved a sensitivity of 85.5% in the first cohort.
- The negative predictive value (NPV) was 97.8% in the test cohort.
- The study included 3055 individuals, with 193 developing Alzheimer's disease during follow-up.
- The score was validated in a second cohort with an AUC of 0.79.
- Significant predictors included age, subjective memory impairment, and cognitive test performance.
- The study's findings suggest that the score can guide clinical decisions in primary care.
Takeaway
This study created a simple score to help doctors figure out who might get dementia based on easy-to-get information, like memory tests.
Methodology
A longitudinal cohort study with 3055 non-demented individuals over 75 years, using baseline variables to create a prediction score.
Potential Biases
The study may have selection bias due to the exclusion of individuals with severe illness or those unable to provide informed consent.
Limitations
The study's prediction score is only applicable to individuals aged 75 and older and may not be valid for younger populations.
Participant Demographics
Participants were elderly individuals (mean age 80.1 years), with 65.5% women and 34.5% men.
Statistical Information
P-Value
0.0002
Confidence Interval
95% CI: 0.80–0.88
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website