Frailty Index

A frailty index quantifies the accumulation of health deficits, symptoms, diseases, and functional impairments across many domains into a single overall score, on the principle that the sheer number of accumulated deficits, more than any single one, predicts vulnerability to adverse outcomes. The claim: a higher frailty index score independently predicts mortality, hospitalization, and loss of independence, often outperforming any individual biomarker used alone.

The Newcastle 85+ Study, a prospective cohort of adults aged 85 at baseline (n=845, mean age 85.5), compared a biomarker-based frailty index (FI-B, combining 40 biomarkers of cellular aging, inflammation, hematology, and immune function) against a clinical-deficits frailty index (FI-CD) and the Fried frailty phenotype for predicting 7-year mortality [1]. Each one percentage-point increase in FI-B raised the mortality hazard ratio by 5.4% (HR 1.05, 95% CI 1.04-1.06), and the FI-B outperformed every individual biomarker tested, with no single biomarker's area-under-curve (AUC) exceeding 0.61 while the combined FI-CD/FI-B model reached an AUC of 0.75 [1]. This directly demonstrates the central premise of frailty-index theory: an aggregate of many small deficits captures risk information that no single measurement contains. A separate analysis from the same Newcastle 85+ cohort, using a clinical Rockwood-style frailty index with a cutoff of 0.25, found baseline frailty was associated with more than double the mortality risk after 7 years compared to non-frail participants, though frailty was not associated with increased hospital time specifically during the last 90 days of life [2].

How to measure it: frailty indices are typically built by counting the proportion of deficits present out of a large checklist, commonly 30-70 items spanning symptoms, diseases, functional limitations, and sometimes biomarkers, divided by the total number of items assessed, yielding a score from 0 to 1 [4]. Common index cutoffs classify scores below roughly 0.12 as low-deficit, 0.12-0.25 as pre-frail, and above 0.25 as frail, though exact thresholds vary by index version [3].

How to intervene on it: no single treatment reverses an elevated frailty index, since it is by design an aggregate of many different underlying conditions; the evidence-based approach is comprehensive geriatric assessment paired with targeted intervention on individual modifiable deficits (nutrition, resistance exercise, medication review, treating undiagnosed conditions), since frailty indices are explicitly built to be dynamic and improvable, unlike a fixed genetic trait.

Critics note that despite strong population-level predictive power, an electronic frailty index has shown low predictive value for mortality at the individual patient level, even near the end of life, limiting its usefulness for individual clinical decision-making [2]. Different frailty index versions, including clinical deficits, biomarker-based, and Fried phenotype, also do not always agree with each other in a given patient, and grip strength alone has been proposed as a simpler frailty-adjacent screening measure with its own separate evidence base [5].

A composite biomarker-based frailty index built from cross-sectional Newcastle 85+ data used many of the same underlying markers, reinforcing that deficit accumulation, not any single biomarker, carries the risk signal [4].

Bryan Johnson's Blueprint biomarker panel does not report a formal Rockwood-style frailty index, tracking many of the same underlying inputs, strength, cardiovascular, and inflammatory markers, as separate individual metrics rather than a single composite deficit-accumulation score [6].

The plain takeaway: frailty indices convert scattered health information into a single meaningful number that predicts mortality more accurately than any individual biomarker at a population level, but the score reflects a compilation of many deficits rather than measuring a single thing, and it works best for population-level risk stratification and care planning rather than pinpointing exactly what to fix in an individual person.