The Proof › Outcomes

What the monitored population actually did.

Headline outcomes

−14.1 mmHg

mean systolic, 30,098 monitored patients

−8.0 mmHg

mean diastolic reduction, 30,098 monitored patients

70.8%

below 140/90 at last reading, 30,098 monitored patients

−$10,932

median annual total cost of care per patient, 597 high-acuity Medicare beneficiaries

Blood-pressure figures are drawn from the full monitored population. Cost figures are from 597 high-acuity Medicare beneficiaries over one year. Observational and single-arm.

Utilization

−40.9%

inpatient hospitalizations, 597 high-acuity Medicare beneficiaries

−19.9%

emergency department visits, 597 high-acuity Medicare beneficiaries

52.8% → 25.8%

share with at least one hospitalization, 597 high-acuity Medicare beneficiaries

All three from the same 597-beneficiary cohort. Do not read these as causal: the design does not support it.

Peer-reviewed publications

  • Published

    Healthcare (MDPI) 2024;12(16):1583

    Remote Patient Monitoring Is Associated with Improved Outcomes in Hypertension: A Large, Retrospective, Cohort Analysis

    n = 6,595

    DOI 10.3390/healthcare12141451

    In 6,595 patients monitored a mean of 289 days, uncontrolled hypertension fell from 66.3% to 40.2% and Stage 2 from 37.5% to 19.1% (both p<0.01). Systolic improved 7.3 mmHg overall and 16.7 mmHg in Stage 2. Co-authored with investigators at Prisma Health Upstate and VCS. Editor’s Choice, open access.

  • Published

    Telemedicine and e-Health 2026

    Short-Term Impact of Remote Patient Monitoring on Total Cost of Care Among High-Acuity Medicare Beneficiaries

    n = 597

    DOI pending

    Median total cost of care fell $10,932 per patient in one year across 597 beneficiaries; hospitalizations fell 40.9% and ED visits 19.9% (all p<0.001), with reductions significant in all four baseline cost quartiles.

Conference research

Five peer-reviewed presentations to American Heart Association Scientific Sessions.

  • Engagement dose-response

    Presented 2025

    n = 20,353

    Every 10% increase in transmission frequency associated with roughly 1.2 mmHg additional systolic lowering.

  • Home versus office concordance

    Presented 2025

    n = 3,399 matched per group

    Home and office systolic reductions agreed within 0.4 mmHg (p=0.44).

  • Interruption and reinitiation

    Presented 2025

    n = 9,942

    Stopping monitoring was followed by measurable rebound; re-enrolling restored prior gains within five months.

  • Readings in target range

    n = 2,987

    Among patients whose 12-month average was controlled, only 55.3% had at least 80% of readings in range.

  • Longitudinal alert dynamics

    211 health systems

    Alert-day rates fell 75–83% across every alert type once normalized to transmission volume.

≤3 d−11.5 mmHg
4 d−13.2 mmHg
5 d−15.0 mmHg
6–7 d−16.9 mmHg

More engagement, larger reductions — inside every severity band.

Patients transmitting 6–7 days a week averaged −16.9 mmHg with 79.8% reaching target, against −11.5 mmHg and 60.4% for those transmitting three days or fewer. The gradient held in all four baseline severity bands.

Mean systolic reduction by transmission frequency. And it is not cumulative time on the program: the highest-transmitting patients averaged 181 fewer days enrolled than the lowest.

The care model is the mechanism

The readings do not manage themselves.

Across 68,431 patients on one platform, identical devices, portal, alerting, EHR integration and billing, the only variable that differed was who provided the care navigation. Patients supported by HealthSnap navigators transmitted on 46.1% of days against 26.1% for other navigation, fell 18.5 mmHg systolic against 14.9, and reached control 66.5% against 56.7%.

The counter-intuitive part: navigated programs generated 36% fewer medium-severity and 44% fewer high-severity alerts per patient despite nearly twice the transmission volume, and median response was 4.2 hours against 6.3. Because technology is held constant and only staffing varies, this is closer to a natural experiment than a vendor comparison.

The care model, compared against itself

18.5 vs 14.9 mmHg

systolic reduction, navigated vs not, 68,431-patient within-platform comparison

66.5% vs 56.7%

reached control, 68,431-patient within-platform comparison

4.2 hours

median response to a high-severity alert, navigated patients, 68,431-patient within-platform comparison

A 68,431-patient within-platform comparison: identical devices, portal, alerting and billing, with only the staffing model differing. Navigated patients transmitted 77% more often and generated fewer medium- and high-severity alerts despite nearly twice the transmission volume.

What the calls actually carry

1,354,099

care-navigation notes, 14,503 patients

60.5%

of notes document medication-adherence support, 1,354,099 notes across 14,503 patients

24.2% of notes carry physical activity counseling, over 327,000 notes, and 22.7% nutrition counseling, over 307,000 notes. Across 1,354,099 care-navigation notes spanning 14,503 patients, sodium reduction appears in 14.8% and guideline-named DASH or Mediterranean patterns in 7.9%. These are not adjacent wellness topics: each domain maps to a quantified blood-pressure effect in the ACC/AHA guideline, DASH to roughly 11 mmHg, sodium reduction to 5–6, aerobic activity to 5–8. First-line therapy, delivered at a frequency the office visit cannot match.

Surveillance & safety

Patients stabilize — and the alert burden falls with them.

Among 22,102 patients followed across nine 30-day periods from their first alert, mean alert days per patient fell 48% and the share of the cohort alerting fell from 53.3% to 32.4%. 77% of that decline reflects patients leaving the alerting state entirely rather than alerting less often. The curve plateaus rather than decaying to zero — about a third of the cohort is still surfacing events at nine months, which is what a chronic surveillance need looks like in data.

The alert nobody counts: more patients alerted for blood pressure that ran too low (10,082) than too high (8,459), and at least 1,610 triggered both. Continuous measurement is therefore also the routine safeguard against over-treatment and fall risk in older adults — a trade-off that is real, common, and observable the day it appears.

Cardiovascular risk

−15.5%

relative reduction in major cardiovascular events, modeled projection from a sustained 8 mmHg reduction

0.4 mmHg

home-versus-office agreement, 3,399 matched pairs

Modeled projection from randomized-trial hazard ratios — not an outcome measured in the HealthSnap population. An individual participant-level meta-analysis of 48 randomized trials and 344,716 participants associates each 5 mmHg of systolic reduction with roughly a 10% lower risk of major cardiovascular events. Applied conservatively to a sustained 8 mmHg reduction, that corresponds to an estimated 15.5% relative reduction in major adverse cardiovascular events.

Methods & limitations

Our platform analyses are observational and single-arm, without a concurrent control group or adjustment for medication changes. We report associations, not causation. Severity bands come from transmitted device readings rather than diagnosis codes. Where regression toward the mean could inflate a result, we say so and stratify to test it.

Our published studies are co-authored with independent health-system investigators; HealthSnap authors are employees and equity holders, and disclose it.

Two specific limits worth stating plainly. The 30-day readmission result was −16.0% and did not reach significance (p = 0.131), so it is not presented as a finding anywhere on this site. And the cardiovascular-risk projection is a model, not a measurement.

Reports

Read the full evidence binder, as submitted to CMS.

Patient clinical outcomes, cardiology clinical outcomes, the in-office blood-pressure validation findings, and the cost-savings and utilization findings.

Download the binder

Read it with our Chief Scientific Officer.

Bring the methodological questions. That is the conversation we want.

Book a Demo