Assessing statistical applications in publications on Alzheimer's disease

Area of Science:

• Biostatistics
• Neuroscience Research Methodology

Background:

• Statistical methods are crucial for the validity of Alzheimer's disease (AD) research.
• Evaluating the quality of statistical applications in AD publications is essential for reliable findings.

Purpose of the Study:

• To develop and validate instruments for assessing statistical reporting in AD publications.
• To determine the feasibility and reliability of these assessment tools.

Main Methods:

• Development of three instruments: Assessment of Statistical Reporting (ASR), Survey of Statistical Designs (SSD), and Survey of Statistical Methods (SSM).
• Pilot testing on 5 AD publications, followed by reliability testing on 30 AD articles and validity testing on 10 AD articles from MEDLINE.

Main Results:

• The ASR demonstrated high test-retest (ICC=0.94) and interrater (ICC=0.84) reliability, with high correlational validity (r=0.74).
• SSD and SSM showed good item-specific test-retest (0.58-1.00) and interrater (0.17-1.00) reliability.

Conclusions:

• The developed instruments are feasible and reliable for assessing statistical applications in Alzheimer's disease publications.
• This evaluation framework can enhance the quality and transparency of statistical reporting in AD research.