Frailty is associated with longer length of stay and adverse outcomes in elderly hospitalized patients, including early re-hospitalization and 12-month mortality. Identifying these risk levels on presentation at emergency departments is crucial to deliver family education, plan care, arrange referrals, and anticipate complications. Many instruments are available covering objective and subjective parameters. Our multi-center cohort study of 2,749 elderly patients in Australia, Denmark and Ireland used Fried’s, Rockwood’s, SUHB’s and Rylance’s frailty scores to predict outcomes at hospital discharge and 3-months. The exacerbation of chronic illness, poor recall, different timeframes for the estimates before presentation, absence of an informant for incompetent patients, organizational limitations and stresses of the emergency environment, and hospital policies about stretcher use on transfer impacted on our ability to reliably measure some of the parameters. Analysis showed substantial inconsistencies in the classification into pre-frail, frail or robust by different instruments on admission. Telephone administration of the frailty instruments at follow-up yielded incomplete or inexact scores due to reliance on self-report or proxy-report rather than direct observation. We concluded that doctors would have limited time to accurately assess all objective parameters during routine care. Emergency/aged care nurses and physiotherapists are best placed to conduct these measurements given their familiarity with the frailty components, training in recognizing physical abilities of patients, ongoing opportunity at the bedside, frequent visual assessment, and communication with patients/ caregivers. Fried’s and Rylance’s instruments were affected by many practical limitations. Telephone assessment on follow-up is not recommended to document decline or improvement over time.