Part One

Model Overview

Reducing healthcare costs is a crucial step toward improving life expectancy and overall health outcomes in the United States.

Consider these striking statistics:

“The U.S. allocates nearly 18% of its GDP to healthcare, yet Americans have shorter life expectancies and poorer health outcomes than residents of other high-income nations.”

“In 2021, U.S. healthcare spending per capita was almost double that of comparable countries—$12,197 vs. $6,514 on average.”

“Over half (54%) of individuals with employer-sponsored insurance who delayed or skipped care due to high costs reported that their health conditions worsened as a result.”

To explore potential solutions, this model estimates the cost savings of pre-hospital telehealth programs by applying findings from the 2016 study: “Cost Benefit Analysis of Telehealth in Pre-Hospital Care.”

Conducted in Houston, Texas, the study assessed the financial impact of the ETHAN telehealth system, which redirects eligible patients from expensive emergency services (ambulances and ERs) to more cost-effective alternatives (taxis and urgent care centers). [1]

Key Findings

Model projections for 2024 suggest that state-level cost savings from implementing similar programs could range from 2.48% to 16.22% of annual health and hospital expenditures, highlighting the transformative financial potential of pre-hospital telehealth adoption across the U.S. (Figure 1)

Figure 1

Methodology

A. Model Inputs

To simulate 2024 cost savings across 31 U.S. states, the model incorporates:

2016 ETHAN study cost savings per patient and provider

Annual health inflation rates (2016–2021)

2021 regional price parity (RPP) indices per state

Total 2021 treat-and-release emergency department (ED) visits⁽¹⁾

2021 primary care-related emergency department (PCRED) costs per region

B. Patient Savings Calculation

Total patient savings were determined by combining 2021 PCRED costs with state-specific ED visit data. Since Nevada’s ED visit data was unavailable, it was estimated using multivariate forecasting in Python.

Bi. Forecasting Emergency Department Visits

To estimate missing ED data, the model applies multi-series recursive forecasting, which accounts for underlying trends impacting state emergency department visits.

Dataset: 11 years of historical ED visit data

Approach: Splitting data into "training" and "test" sets for evaluation (Figure 2)

Tool Used: skforecast.org package

Figure 2

Short-Term Trend Analysis

The model leverages the LightGBM recursive forecasting method with a lag of 1, ensuring each predicted year’s value informs the next. This technique captures short-term fluctuations, such as:

Changes in insurance policies

Federal healthcare spending adjustments

The impact of public health emergencies (e.g., COVID-19)

By using a shorter rolling window, the model preserves year-over-year variations instead of smoothing them out. (Figure 3)

Figure 3

Prediction for Nevada (2021 ED Visits):

Model Output: 985,809 visits

Confidence Interval (95%): 849,600 – 1,022,703 visits

C. Provider Savings Calculation

Provider savings were estimated using the ETHAN study’s $103 per-visit cost reduction.

1. Adjusting for State Differences: The real price parity (RPP) index was used to scale Texas-based cost savings across 31 states.
2. Forecasting Future Savings: The 2021 provider savings estimate was projected forward to 2024 using compounded annual healthcare inflation rates.

D. Combined Cost Savings

The final total cost savings per state was calculated by combining patient and provider savings (Figure 4).

Figure 4

Key Takeaways

Telehealth can significantly lower emergency department costs by diverting patients to lower-cost care settings.

Cost savings vary by state, with greater benefits in regions with lower healthcare spending per capita.

Data-driven forecasting ensures accurate projections and highlights trends that impact healthcare expenses.

By expanding telehealth solutions nationwide, the U.S. can reduce healthcare spending, enhance access to care, and improve long-term health outcomes.

Footnotes & Data Sources

⁽¹⁾Total emergency department visit data was sourced from the State Emergency Department Database (SEDD), sponsored by the Agency for Healthcare Research and Quality (AHRQ).

Graphic: Deepai.org