Earlier specialist input → better outcomes

Revised 19 October 2026. Re-graded on 19 October against the new frailty foundations, then tested again on timing the same day, by two blind passes each time. The pattern now carries three claims graded apart: specialist care from the start of an admission is moderate; long delays harm is low, at the upper end; moving the same input earlier is very low, because the trials that randomise timing found no gain.

The register recorded this pattern as established. Graded against the written rules, it is provisional. The register moves at its next monthly review.

The claim the evidence supports

Specialist geriatric assessment and care from the start of an acute hospital admission improve outcomes older people value, and long delays in an acute episode are linked to harm. That giving the same input sooner than it is already given helps has not been shown.

• Specialist geriatric assessment and care from the start of an acute admission, compared with usual care, make it more likely an older person is alive and at home, and less likely to lose function or move into a care home. Moderate confidenceTrials are high certainty in their own terms; downgraded once because most are older, from outside the NHS, and say nothing about deprived or dispersed populations.
• Long waits in an acute episode are linked to worse outcomes for older people: many hours or a night in the emergency department, slow transfer to a geriatric ward, hip surgery delayed beyond a day or two. Low confidenceObservational only, so it starts low; held at the upper end of low by a graded dose-response in several countries (in England, one extra death for every 82 patients waiting beyond six to eight hours). Sicker patients waiting longer explains part of it, frailty is rarely measured directly, and one meta-analysis found harm only beyond 24 hours.
• Giving the same specialist input sooner than it is already given improves outcomes: at the front door, before the crisis, or across the other domains. Very low confidenceWhere timing has been randomised the results lean to no gain: hip surgery at six hours rather than 24 did not reduce deaths (HIP ATTACK); very early mobilisation after stroke did harm (AVERT); early against later palliative care in advanced cancer found no difference in quality of life, though survival was better (ENABLE III). Very low means not shown, not wrong.

Starting point

• D01 qualifies foundations: care from admission is moderate to high; earlier still is unsettled
• D02 qualifies time-to-contact is a hypothesis with no study isolating it
• D03 qualifies early palliative care helps in cancer; the trial of timing itself was null on quality of life
• D04 qualifies 'early' is not tested as timing
• D05 silent the OPAT lever is right-sizing IV treatment, not earliness

The evidence

16 test or bear on it 4 cited but do not test it 2 not yet read at source 5 endorsements, not counted Domains earned: D01, D03 of the 5 it spans

What would prove it wrong

For the delay claim: studies that measure frailty and adjust properly for severity finding no excess harm from long waits. For the earlier-is-better claim: further trials of earlier against later input finding no difference in being alive at home, function or quality of life; HIP ATTACK and ENABLE III already point that way.

What the re-grade found

A single low grade blurred three claims. The field's settled evidence supports specialist care from the outset at moderate confidence. Avoiding long delays is better supported than racing to be first: the evidence fits a threshold, where long waits harm but speeding up care that is already reasonably prompt adds little. Its July record reads the HERO trial as a timing boundary; HERO tested extended home exercise after discharge, not timing. On the grade: two passes, reconciled to the more cautious.

A grade measures how certain the evidence we have assessed makes the claim as worded: its own sources and the settled evidence for its domains. Low means unproven, not implausible. What a grade means →