Response stability and reliability in longitudinal health evaluations.

Area of Science:

• Gerontology
• Epidemiology
• Biostatistics

Background:

• Longitudinal studies like the Baltimore Longitudinal Study of Aging (BLSA) rely on consistent data collection.
• Assessing the stability and reliability of health questionnaires and medical examination items is crucial for data integrity.

Purpose of the Study:

• To evaluate the long-term stability and intravisit reliability of health questions and clinical medical examination items within the BLSA.
• To identify specific items with potential inconsistencies in responses over time and between assessment methods.

Main Methods:

• Stability analysis involved tracking positive responses to medical history and physical examination items over multiple visits.
• Intravisit reliability was assessed by comparing responses from the general health questionnaire with the Cornell Medical Index during the same visit.
• Statistical measures, including simple and weighted averages, and Kappa statistics, were used to quantify stability and agreement.

Main Results:

• Subject-completed health questions showed a stability of 58% (simple average) to 64% (weighted average).
• Physician-asked questions exhibited lower stability (29% to 36%), and physical examination items showed 34% to 37% stability.
• Intravisit reliability indicated substantial agreement for subject-completed questions (Kappa = 0.74) and moderate agreement for physician-asked questions (Kappa = 0.44).

Conclusions:

• Health questionnaires and medical examination items in the BLSA demonstrate variable stability over time and moderate to substantial intravisit reliability.
• Subject-completed questions generally offer higher reliability and stability compared to physician-administered items.
• These findings highlight areas for potential improvement in data collection instruments for aging studies.