Briefing Note

Why supportability depends on when you ask

"Is this supportable?" is an incomplete question. A claim is supportable or not at a particular date, under the evidence record as it stood on that date. Ask it without a date and you will get an answer about today, which is rarely the date you needed.

The claim that was fine last quarter

Consider a market-access team preparing a payer deck in March. The deck includes a claim: "Therapy X reduces hospitalization-related outcomes in high-risk patients with Condition Y." At the time, the evidence supports it — the pivotal trial showed benefit in the high-risk subgroup, the label includes the indication, and the guideline recommends use.

By August, three things have changed: a new trial readout shows no benefit in moderate-risk patients, the guideline narrows its recommendation to patients with prior hospitalization, and the label narrows its indication. The claim is still in the deck. The deck is still in circulation.

Is the claim still defensible? That depends on when you ask. In March, yes. In August, no — at least not as written.

Five timelines, one claim

Every healthcare claim lives on several different clocks simultaneously:

Evidence time

When did the studies, safety data, endpoint analyses, or real-world evidence appear? A trial published after the claim was written can either strengthen or weaken it.

Wording time

When did the claim’s language change? "May reduce hospitalization-related outcomes" became "reduces hospitalization" — two words dropped, supportability changed.

Authority time

When did a guideline, label, payer policy, or regulator adopt, weaken, or retire the claim? A guideline change can undermine a claim even if no new evidence appeared.

Use time

When did the claim move into a higher-stakes context? An internal scientific note tolerates uncertainty that a payer-facing deck cannot.

Review time

When was the claim last checked, within what review window, and what would warrant another look? The longer the gap, the more drift accumulates.

What a dated record has to hold

Answering the as-of question requires a record built for it: what the claim asserted, what the evidence supported at that time, what changed since, and whether the claim still stands as written.

It answers questions that most evidence tools do not:

• Was this claim supportable when the deck was approved?
• Has a guideline softened since the claim was written?
• Did the claim broaden — lose a caveat, drop a subgroup — between versions?
• Has the evidence base shifted so that the claim now requires narrowing?
• Is the claim being used in a higher-stakes context than it was written for?

Why the review date matters

Most evidence tools answer from the present: "What does the current corpus say?" That tells you whether a claim is supportable now. It does not tell you whether a claim was supportable then — at the time of MLR approval, at the time of a payer meeting, at the time of a review memo, at the time of a formulary review.

For audits, disputes, fit checks, and quality review, the as-of question is the one that matters. The system should be able to reconstruct what the world looked like at the decision point — not just what it looks like after the fact.

Why forecasting needs the stricter version

Reconstructing the past is hard. Forecasting demands something harder: a record that can prove no later information leaked backwards into an earlier estimate. What was observable before registration, what arrived after, and which public action resolved the question all have to stay separable — otherwise a forecast can be flattered by hindsight and nobody, including us, could tell.

That is why every forecast is pinned to a dated evidence cutoff and frozen when issued. It is the same discipline this page describes, applied to the harder direction: not what did the record support then, but what does it support about what happens next. The method and limits set out how the cutoff is enforced and how resolution is adjudicated.