Routine labs estimate patient vulnerability

A simple frailty score from routine blood tests was strongly associated with death and hospital use.
*Retrospective multicohort prognostic study; Level 2b (OCEBM).

Citation

Logan Ellis H, Hanlon P, Dunnell L, Whyte M, Davis DHJ, Bates J, et al. Evaluation of the FI-Lab, a laboratory-based, automated frailty index for acute care: A multicohort study. PLoS Medicine. 2026;23(9):e1005004.

Background

Clinicians need quick ways to estimate how vulnerable a patient is, especially in acute care. This study tested whether abnormal routine blood results can provide an automated measure of vulnerability across adult ages.

Patients

Adults from emergency department cohorts in Boston and London, plus United Kingdom Biobank participants. Patients without linkable same-day blood tests and implausible records were excluded.

Intervention

Laboratory frailty index: proportion of abnormal results from commonly ordered tests; main version used 25 tests with at least 15 required.

Control

Age, sex, and, in emergency cohorts, National Early Warning Score 2.

Outcome

Primary: 1-year all-cause mortality. Secondary: hospital admission, in-hospital death, readmission, and length of stay.

Follow-up Period

Up to 1 year; United Kingdom Biobank also assessed 10-year mortality.

Results

OutcomeCohortLaboratory frailty index result
1-year mortality (primary)Boston emergency departmentHR 2.04 (95% CI 1.99 to 2.09)
1-year mortality (primary)London emergency departmentHR 2.55 (95% CI 2.42 to 2.69)
1-year mortality (primary)United Kingdom BiobankHR 1.87 (95% CI 1.80 to 1.94)
30-day readmissionBoston; LondonHR 1.32 (95% CI 1.30 to 1.33); HR 1.35 (95% CI 1.33 to 1.36)

HR = hazard ratio; CI = confidence interval.

The score’s association with death approached age and exceeded the early warning score. Results were based on observational data, not a treatment trial.

Limitations

Retrospective design cannot prove improved care. Data came from high-income settings with mature electronic records. Test ordering may reflect clinician judgment and illness severity. Real-time clinical benefit remains untested.

Funding

Academic/public funders; no funder role. Frailty-scale licensing conflicts noted.

Clinical Application

Consider FI-Lab as an automated risk-context tool, not a replacement for judgment; prospective testing is needed before practice standards change.