ESTIMATING INDIVIDUAL-LEVEL MEDICAL COST SAVINGS FROM DIGITAL CHRONIC DISEASE MANAGEMENT: A BIOMETRIC-TO-PMPM PROXY METHODOLOGY
Author(s)
Yifat Fundoiano-Hershcovitz, PhD1, Rafi Herzfeld, FSA2, Omar Manejwala, MD3.
1VP of Clinical and Scientific Affairs, Dario Health, Caesarea, Israel, 2Dario Health, Caesarea, Israel, 3Dario Health, New york, NY, USA.
1VP of Clinical and Scientific Affairs, Dario Health, Caesarea, Israel, 2Dario Health, Caesarea, Israel, 3Dario Health, New york, NY, USA.
OBJECTIVES: To describe a replicable methodology for estimating individual-level medical cost savings attributable to a digital chronic disease management platform by translating observed biometric improvements into projected per-member-per-month (PMPM) cost reductions derived from published clinical evidence.
METHODS: For each member, three biometric endpoints were tracked: systolic blood pressure (mmHg), estimated A1c (units), and weight (BMI). Baseline (first 30-days) measurements were established at enrollment; follow-up values were assessed at standardized cut points (6-24 months) using subsequent 30-day rolling averages. When exact readings were unavailable, three proximity substitution rules were applied (nearest preceding, nearest subsequent, or nearest in either direction). Observed biometric changes were annualized via linear duration scaling (factor: 12/observation month, for example, measurements at 6, 12, and 24 months were scaled, 2x, 1x, 0.5x respectively). Annualized changes were multiplied by literature-derived PMPM rates: $11.25/mmHg SBP reduction; $114.42/unit eA1c reduction; and BMI-based rates $80.50 (reduction) versus $21.08 (increase) per unit. Individual projected annual savings were computed as: Annual Savings ($)=(Observed Biometric Change×[12 / Observation Month])×Literature PMPM Rate×12 Members with worsening values retained negative estimates to avoid upward bias.
RESULTS: The framework produces individual- and population-level savings estimates across one or more program tracks (blood pressure, blood glucose, weight management). Multiple program savings are additively and conservatively aggregated despite literature supporting higher multiplicative savings for comorbid improvements. The three proximity substitutions support sensitivity analysis. Program costs are captured via PEMPM billings enabling a group-level ROI ratio of projected savings relative to program expenditure.
CONCLUSIONS: This proxy model approach produces ROI estimates consistent with comparable published benchmarks using only biometric data and no medical claims. Limitations include population generalizability of PMPM rates and linear scaling assumptions. Validation against actual claims data is recommended where feasible.
METHODS: For each member, three biometric endpoints were tracked: systolic blood pressure (mmHg), estimated A1c (units), and weight (BMI). Baseline (first 30-days) measurements were established at enrollment; follow-up values were assessed at standardized cut points (6-24 months) using subsequent 30-day rolling averages. When exact readings were unavailable, three proximity substitution rules were applied (nearest preceding, nearest subsequent, or nearest in either direction). Observed biometric changes were annualized via linear duration scaling (factor: 12/observation month, for example, measurements at 6, 12, and 24 months were scaled, 2x, 1x, 0.5x respectively). Annualized changes were multiplied by literature-derived PMPM rates: $11.25/mmHg SBP reduction; $114.42/unit eA1c reduction; and BMI-based rates $80.50 (reduction) versus $21.08 (increase) per unit. Individual projected annual savings were computed as: Annual Savings ($)=(Observed Biometric Change×[12 / Observation Month])×Literature PMPM Rate×12 Members with worsening values retained negative estimates to avoid upward bias.
RESULTS: The framework produces individual- and population-level savings estimates across one or more program tracks (blood pressure, blood glucose, weight management). Multiple program savings are additively and conservatively aggregated despite literature supporting higher multiplicative savings for comorbid improvements. The three proximity substitutions support sensitivity analysis. Program costs are captured via PEMPM billings enabling a group-level ROI ratio of projected savings relative to program expenditure.
CONCLUSIONS: This proxy model approach produces ROI estimates consistent with comparable published benchmarks using only biometric data and no medical claims. Limitations include population generalizability of PMPM rates and linear scaling assumptions. Validation against actual claims data is recommended where feasible.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
EE180
Topic
Economic Evaluation, Patient-Centered Research, Real World Data & Information Systems
Disease
Cardiovascular Disorders (including MI, Stroke, Circulatory), Diabetes/Endocrine/Metabolic Disorders (including obesity)