Digital Bioimpedance for Physical Activity Detection in Type-2 Diabetes: Quasi-Experimental Validation Study
Bibliographic record
Abstract
Background: Primary care diabetes management lacks objective, scalable methods for continuous physical activity surveillance. Bioelectrical impedance analysis (BIA), routinely collected in diabetes care, offers untapped potential as an automated digital biomarker but requires validation for behavioral phenotyping. Objective: This study aims to evaluate the feasibility and predictive validity of multifrequency bioimpedance for physical activity detection and its association with glycemic control in type 2 diabetes. Methods: This was a pragmatic quasi-experimental study using temporal allocation across three 4-month periods (January 2021-July 2023) in a Japanese primary care clinic, including comprehensive tracking with BIA-guided counseling (n=65), partial tracking (n=31), and standard care (n=100). Adults with type 2 diabetes (hemoglobin A1c [HbA1c] 7.0%-10.0%) underwent monthly segmental multifrequency BIA. The primary outcome was HbA1c <7% at 4 months. Intervention-outcome associations were examined using chi-square trend tests and multivariable logistic regression adjusted for baseline HbA1c, the Walk Score (0-100), and medication indicators. To assess temporal confounding, we conducted ANCOVA on 4-month HbA1c with baseline adjustment (age and BMI added in sensitivity analyses). Effect modification by built environment was tested via Walk Score×Intervention interaction. Predictive validity of left-arm 50-kHz reactance was assessed using area under receiver operating characteristic curve with 95% CI via 10-fold cross-validation. Results: Among 196 participants, the baseline characteristics (age, BMI, HbA1c, diabetes duration, and medications) did not differ across periods (all P>.05). HbA1c <7% achievement showed a gradient: 80% (52/65) comprehensive, 58% (18/31) partial, and 56% (56/100) standard care (χ²4 for trend=14.23; P<.001). ANCOVA of 4-month HbA1c (baseline-adjusted) showed no linear period trend (P=.25). A significant Walk Score×Intervention interaction was observed (β per 10-point Walk Score=-.55; 95% CI -1.03 to -0.06; P=.028), indicating differential effectiveness by neighborhood walkability. Left-arm 50-kHz reactance predicted target achievement (adjusted odds ratio per 1-SD increase =3.04; 95% CI 1.86-4.97; P<.001; area under receiver operating characteristic curve=0.847, 95% CI 0.784-0.910). Among achievers, reactance change correlated with HbA1c change (r=-0.392; P=.032) but not among nonachievers (r=-0.089; P=.54). After the inverse probability weighting was stabilized, each 1-SD increase in left-arm reactance was associated with a 12.1 percentage-point higher probability of target achievement (95% CI 5.2%-19.0%). Conclusions: This pragmatic implementation study demonstrates that automated BIA is feasible for routine diabetes care and suggests potential as a digital biomarker of activity-related glycemic control. While temporal allocation precludes definitive causal inference, and findings should be interpreted as associational, the observed Walk Score moderation and bioimpedance-HbA1c dose-response patterns are consistent with behavioral mechanisms rather than pure confounding. Left-arm reactance warrants randomized validation as a scalable, passive surveillance tool for precision diabetes management.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.019 | 0.012 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".