Related Experiment Videos
Beyond Fixed Thresholds: Optimizing Summaries of Wearable Device Data via Piecewise Linearization of Quantile
Junyoung Park1, Neo Kok1, Irina Gaynanova1
1Department of Biostatistics, University of Michigan, Ann Arbor, Michigan, USA.
Statistics in Medicine
|June 30, 2026
Summary
Data-driven thresholds for wearable devices like continuous glucose monitors (CGMs) optimize data analysis. These personalized thresholds improve insights compared to fixed ranges across diverse patient populations.
Area of Science:
- Biomedical Engineering
- Data Science
- Endocrinology
Background:
- Wearable devices (actigraphy, continuous glucose monitors [CGMs]) generate high-frequency data.
- Current analysis relies on fixed thresholds (e.g., activity levels, glucose ranges), potentially suboptimal.
- Optimality and population-specific applicability of fixed thresholds remain unclear.
Purpose of the Study:
- To define and optimize thresholds for wearable device data using loss functions.
- To develop data-driven thresholds that accurately represent full data distributions.
- To assess the impact of population-specific thresholds on data interpretation and clinical relevance.
Main Methods:
- Introduced two loss functions based on Wasserstein distance to quantify distribution discrepancies.
- Reformulated threshold optimization as piecewise linearization of quantile functions.
- Employed stepwise algorithms and differential evolution for optimization; developed semi-supervised approaches.
Main Results:
- Data-driven thresholds derived from continuous glucose monitor (CGM) data varied significantly across populations (type 1 diabetes, type 2 diabetes, normal glycemic control).
- Optimized thresholds demonstrated improved discriminative power compared to fixed thresholds.
- Population-specific thresholds showed stronger associations with clinical variables.
Conclusions:
- Fixed thresholds for wearable device data may not be optimal or universally applicable.
- Data-driven, population-specific thresholds enhance the analytical power of wearable sensor data.
- This approach offers a more precise method for analyzing CGM and actigraphy data in clinical practice.
Related Concept Videos
01:23Linear Approximations
For a differentiable function of two variables, linear approximation estimates values near a known point by replacing the curved surface with its tangent plane. Consider the function\begin{equation*}f(x,y)=x^2+3y^2\end{equation*}near the point (2, 1). The exact value at this point is f(2, 1) = 22 + 3(1)2 = 4 + 3 = 7.The linear approximation of f(x, y)) near (a, b) is\begin{equation*}L(x,y)=f(a,b)+f_x(a,b)(x-a)+f_y(a,b)(y-b)\end{equation*}First, compute the partial derivatives: fx(x, y) = 2x and...
01:045-Number Summary
In a dataset, the 5-number summary includes the minimum data value, the data value of the first quartile, the median data value or data value of the second quartile, the data value of the third quartile, and the maximum data value. These 5 data values can be visualized as a box and whisker plot.
In a box plot, the minimum and maximum data values represent the lower and upper whiskers in the graph, and the median is designated as the center of the box in the chart. The first quartile and third...
In a box plot, the minimum and maximum data values represent the lower and upper whiskers in the graph, and the median is designated as the center of the box in the chart. The first quartile and third...
01:29Mechanistic Models: Compartment Models in Algorithms for Numerical Problem Solving
Mechanistic models play a crucial role in algorithms for numerical problem-solving, particularly in nonlinear mixed effects modeling (NMEM). These models aim to minimize specific objective functions by evaluating various parameter estimates, leading to the development of systematic algorithms. In some cases, linearization techniques approximate the model using linear equations.
In individual population analyses, different algorithms are employed, such as Cauchy's method, which uses a...
In individual population analyses, different algorithms are employed, such as Cauchy's method, which uses a...
01:26Linearization and Approximation
Linearization is a mathematical technique used to approximate complex, nonlinear functions with simpler linear models in the vicinity of a chosen reference point. The method is based on the idea that, although a function may be difficult to evaluate exactly, its behavior near a specific input value can often be closely approximated by the tangent line at that point. This approach is particularly useful when small deviations from a known value are involved.Consider the square root function, for...
01:28Piecewise-Defined Functions
Piecewise defined functions are mathematical models where different expressions define a function over distinct intervals of the domain. These functions are useful for representing systems with varying behaviors depending on input values.For example, the function: uses a linear rule for inputs less than or equal to –1 and a quadratic rule for values greater than –1. Although it has two formulas, it still defines a single function.Another common type is the absolute value function, given...
01:29Application of Linearization and Approximation
A drone flying through complex terrain often relies on more than one sensing method to estimate small changes in altitude. Along with direct measurements, air pressure provides a useful indirect indicator of vertical movement. Atmospheric pressure decreases as altitude increases, and this relationship is commonly described using an exponential model. Although accurate, converting pressure measurements into altitude values requires calculations that are too complex to perform repeatedly during...