Routine labs estimate patient vulnerability

A laboratory frailty index from routine blood tests strongly identified adults at higher risk of death and hospital use.
*Retrospective multicohort prognostic study; Level 2b.

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. doi:10.1371/journal.pmed.1005004

Background

Frailty and illness severity are hard to summarize quickly in acute care. This study tested whether routinely available blood tests could provide an automated measure of patient vulnerability across ages and care settings.

Patients

Adults from emergency departments in Boston and London, plus a United Kingdom community cohort: 902,582 total records or participants. Exclusions included absent linkable same-day laboratory data and invalid records.

Intervention

Automated laboratory frailty index: proportion of abnormal results among commonly ordered blood tests; reference model used 25 tests with at least 15 required.

Control

Age, sex, and the National Early Warning Score 2 where available.

Outcome

Primary outcome: all-cause death within 1 year. Secondary outcomes included hospital admission, in-hospital death, readmission, and length of stay.

Follow-up Period

1 year primary follow-up; up to 10 years in the community cohort.

Results

Cohort Predictor 1-year mortality result
Boston emergency department Laboratory frailty index HR 2.04 (95% CI 1.99 to 2.09)
London emergency department Laboratory frailty index HR 2.55 (95% CI 2.42 to 2.69)
United Kingdom community cohort Laboratory frailty index HR 1.87 (95% CI 1.80 to 1.94)
All cohorts Age HR range 1.72 to 2.70 (95% CI reported in article)

HR: hazard ratio; CI: confidence interval.

Results were observational, not based on intention-to-treat assignment. Performance was broadly consistent across age, sex, ethnicity, and settings. About 20 to 40 common tests captured most of the signal.

Limitations

Retrospective design cannot prove that using the score improves care. Data came from high-income settings with strong electronic records. Some patients without blood tests were excluded, limiting generalizability. The index may reflect both chronic vulnerability and acute illness.

Funding

Public and academic funders; no funder role reported.

Clinical Application

Clinicians may consider automated laboratory frailty indexing as a risk-context tool, not a replacement for bedside judgment, pending prospective testing.