Withings Heart Function Risk Notification Retrospective Study

NCT07654491 | Not Yet Recruiting FDA Device

• 1 other identifier: HFRN
• Study Type: observational
• Target: 9,000
• Locations: 1 country
• Sites: 1

Brief Summary

This is a fully decentralized retrospective, non-interventional study evaluating the diagnostic performance of the Withings Heart Function Risk Notification (HFRN), a software-only medical device (SaMD) that analyzes longitudinal weight / BMI, body-water-percent variability, and activity (step) data collected passively by Withings connected smart scales and the Withings activity-tracking ecosystem. The study estimates Sensitivity (Se) and Specificity (Sp) of HFRN against a US claims-based (HealthVerity) heart-failure reference standard, evaluated on a pre-specified holdout test set. The findings support a 510(k) substantial-equivalence determination to the predicate Viz HCM (DEN230003) within the Cardiovascular machine learning-based notification software generic type (21 CFR 870.2380).

Conditions

Heart Failure

Keywords

mhealth

Outcome Measures

Primary Outcomes (2)

• Sensitivity (Se) and its 95% CI of the HFRN algorithm for heart-failure detection, evaluated on the pre-specified holdout test set.

  13 weeks

• Specificity (Sp) and its 95% CI of the HFRN algorithm for heart-failure detection, evaluated on the pre-specified holdout test set.

  13 weeks

Study Arms (2)

HF Cases

Withings smart-scale users with US HealthVerity claims linkage who meet the claims-based heart-failure case definition (\>= 1 inpatient HF ICD-10 code, or \>= 2 distinct outpatient-visit HF ICD-10 codes; per the first-match-wins claims algorithm). Holdout test set: 545 HF cases.

Device: Withings Heart Function Risk Notification (HFRN); Other: Claims-based heart-failure adjudication (reference standard)

Control

Withings smart-scale users with US HealthVerity claims linkage who meet the claims-based control definition (\>= 2 years of continuous enrollment, no heart-failure ICD-10 evidence). Holdout test set: 8,548 controls. Case/control mix is calibrated to the 2021 NHANES-estimated heart-failure prevalence for US adults \>= 45 years.

Device: Withings Heart Function Risk Notification (HFRN); Other: Claims-based heart-failure adjudication (reference standard)

Interventions

Withings Heart Function Risk Notification (HFRN) DEVICE

Software-only medical device (SaMD): an algorithm that analyzes quarterly-aggregated features computed from data passively captured by Withings connected smart scales and step counts synced into the Withings activity-tracking ecosystem. When the predicted heart-failure risk probability exceeds a calibrated decision threshold, the device emits a notification directing the user to seek confirmatory clinical evaluation. The notification is non-diagnostic.

Control; HF Cases

Claims-based heart-failure adjudication (reference standard) OTHER

Ground-truth heart-failure case / control status derived from US HealthVerity administrative claims by a deterministic, pre-specified rule set. Per-user features (counts of HF ICD-10 codes split by inpatient vs. distinct outpatient visits, continuous-enrollment duration, supporting CPT and laboratory evidence) are passed through an ordered first-match-wins classifier. Each user is additionally assigned a claims-operationalized ACC/AHA heart-failure stage (0 / A / B / C). Adjudication is fully automated; no clinician adjudication is performed.

Control; HF Cases

Eligibility Criteria

• Age: 45 Years - 120 Years
• Sex: all
• Healthy Volunteers: Yes
• Age Groups: Adult (18-64), Older Adult (65+)
• Sampling Method: Non-Probability Sample

Study Population

Adult US consumer users (\>= 45 years) of Withings connected smart scales (Body Cardio, Body Scan, Body Comp) with a source of step counts synced into the Withings ecosystem and successful linkage to US HealthVerity administrative claims, enabling heart-failure ground-truth ascertainment and covariate analyses. The development dataset comprises 54,717 individuals (52,736 controls and 1,981 HF cases). The pre-specified holdout test set comprises 9,093 individuals (8,548 controls and 545 HF cases), with case/control mix calibrated to the 2021 NHANES-estimated HF prevalence for US adults \>= 45 years.

You may qualify if:

• Users of commercially purchased, consumer-grade Withings smart scales (e.g. Body Cardio \[WBS04\], Body Scan \[WBS08\], Body Comp \[WBS12 / WBS12C\]) used in a standard home environment
• A source of step counts synced into the Withings ecosystem (a Withings activity tracker, a smartphone, or a third-party activity tracker)
• US HealthVerity claims linkage available, enabling heart-failure ground-truth ascertainment
• Age \>= 45 years
• At least one 13-week quarter in which at least one of the three quarterly features (mean BMI, standard deviation of body-water percent, 90th-percentile daily steps) can be computed

You may not qualify if:

• Users for whom heart-failure status cannot be determined from claims (unknown label; e.g. controls with \< 2 years of continuous enrollment)
• Users with no usable measurement after per-reading plausibility screening (weight outside 20-400 kg, BMI outside 10-60 kg/m2, body-water percent outside 40-80%, daily steps outside 0-100,000, height outside 100-250 cm, or age outside 45-120 years)
• Users without an individual Withings account (shared accounts excluded to prevent data contamination)

Sponsors & Collaborators

• Withings: lead

Study Sites (1)

Withings (Sponsor) - Decentralized Retrospective Study, No Physical Sites

Issy-les-Moulineaux, Île-de-France Region, 92130, France

Location

MeSH Terms

Conditions

Heart Failure

Interventions

Reference Standards

Condition Hierarchy (Ancestors)

Heart Diseases; Cardiovascular Diseases

Intervention Hierarchy (Ancestors)

Weights and Measures; Investigative Techniques

Study Officials

• Pierre Escourrou, MD, PhD

  Centre Interdisciplinaire du Sommeil, Paris, France

  PRINCIPAL INVESTIGATOR

Central Study Contacts

Aline Criton, PhD

CONTACT

+33 1.41.46.04.60; clinicaltrials-us@withings.com

Study Design

• Study Type: observational
• Observational Model: COHORT
• Time Perspective: RETROSPECTIVE
• Sponsor Type: INDUSTRY
• Responsible Party: SPONSOR

Study Record Dates

• First Submitted: June 12, 2026
• First Posted: June 17, 2026
• Study Start: June 1, 2026
• Primary Completion (Estimated): September 1, 2026
• Study Completion (Estimated): September 1, 2026
• Last Updated: June 17, 2026

Record last verified: 2026-06

Data Sharing

• IPD Sharing: Will not share

Locations

FR (1)