UNFINISHED - Methodology

What this project is

This website is an advocacy-oriented public data project that displays modeled counters related to mortality under current U.S. health-system conditions. The counters are not literal real-time observed deaths. They are annual estimates converted into continuously updating per-second clocks.

Current counters

1. Excess mortality

This counter is intended to represent deaths above an expected baseline. In public-health research, excess mortality is usually defined as the difference between observed all-cause deaths and deaths expected from historical patterns. It is commonly used because it captures both direct and indirect mortality shocks. :contentReference[oaicite:0]{index=0}

2. Avoidable / treatable mortality

This counter is intended to represent deaths that would likely not occur under stronger health-system performance. In the literature, avoidable mortality is often divided into preventable deaths and treatable (amenable) deaths, usually for deaths under age 75 and using standardized cause-of-death lists. :contentReference[oaicite:1]{index=1}

3. Deaths attributable to lack of insurance

This counter is intended to represent deaths associated with uninsurance using a population-attributable-risk style framework. In the literature, these estimates are usually derived by combining an uninsured prevalence estimate with a mortality risk ratio comparing uninsured and insured adults. :contentReference[oaicite:2]{index=2}

How the live counter works

Each counter starts from an annual estimate. The site converts that annual value into a per-second rate and then increments the displayed value continuously throughout the year.

Conceptually:

• per_second = annual_estimate / seconds_in_year
• counter_now = per_second × elapsed_seconds_in_year

The website refreshes its baseline values from the backend periodically, then animates the count locally in the browser.

Current implementation status

The current public version uses placeholder annual estimates routed through a data layer in the backend. That means the architecture is set up for real data integration, but the values shown now should be treated as prototype inputs rather than final published estimates.

The underlying project architecture separates:

• data inputs
• model logic
• API response construction
• frontend display

Planned data sources

• Excess mortality: CDC provisional mortality data and related weekly all-cause death series
• Avoidable / treatable mortality: OECD / WHO benchmark data and standardized avoidable-mortality frameworks
• Uninsurance-attributable mortality: Census insurance coverage data combined with published mortality risk estimates

These are the core source categories identified in the project research memo. :contentReference[oaicite:3]{index=3}

Planned update frequency

• Excess mortality: weekly or near-weekly
• Avoidable / treatable mortality: periodic / annual
• Uninsurance-attributable mortality: periodic / annual

Actual update frequency depends on the release cadence and completeness of the underlying public datasets.

Important limitations

• These are model-based estimates, not direct real-time observed counts.
• Different methodologies can produce different totals depending on baseline windows, benchmark countries, and causal assumptions.
• Provisional mortality data are revised over time, so recent estimates can change.
• Avoidable and uninsurance-attributable mortality involve judgment and counterfactual assumptions, not simple administrative tallies.

Transparency goal

The goal of this project is to make the methodology, assumptions, and limitations visible rather than hiding them behind a dramatic visual. As the project matures, this page will expand to include fuller model documentation, source links, update logs, and version history.