Does Testing More Frequently Shorten the Time to Detect Disease Progression?

Area of Science:

• Biostatistics
• Health Informatics
• Medical Technology

Background:

• Remote health monitoring using smartphones is increasing.
• Frequent sampling is often assumed to improve early detection of health status changes.

Purpose of the Study:

• To investigate whether increased sampling frequency enhances early detection of health status changes.
• To evaluate the impact of observation correlation on trend detection.

Main Methods:

• A statistical model was developed for linear trends with autocorrelated autoregressive-moving average (ARMA(1,1)) errors.
• The model analyzes the power of detecting linear changes with varying observation frequencies and autocorrelation.

Main Results:

• Increased sampling frequency does not always guarantee faster detection of trend changes.
• The benefit of rapid, successive observations is dependent on the correlation between them.

Conclusions:

• A strategy for earlier disease progression detection via frequent measurements is limited by autocorrelation.
• The proposed statistical model accounts for measurement correlation, impacting the number of observations needed for significant change detection.