Does the get up and go test improve predictive accuracy of the Triage Risk Screening Tool or Rowland questionnaire in older patients admitted to the emergency department?

OBJECTIVES: To assess the diagnostic characteristics of the get up and go test (GUGT) as a stand-alone test and in combination with the Flemish Triage Risk Screening Tool (fTRST) and Rowland questionnaire. One aim was to determine whether the diagnostic accuracy of these instruments could be improved for predicting unplanned emergency department (ED) readmission following ED discharge. METHODS: We carried out a prospective cohort study at the ED of the University Hospitals Leuven, Belgium. All patients aged at least 75 years (n=380) completed fTRST, Rowland, and GUGT testing at the index ED admission. Diagnostic characteristics for unplanned ED readmission were determined for hospitalized and discharged patients 1 and 3 months after the index ED visit. RESULTS: In both hospitalized and discharged patients, fTRST and Rowland (cut-off ≥2) had good to excellent sensitivity and negative predictive value (NPV) but low to moderate specificity and accuracy; GUGT had low sensitivity and good to excellent NPV and specificity. The combined fTRST/GUGT or Rowland/GUGT had moderate to excellent NPV (56.3-94.3%). The combined fTRST (cut-off ≥2)/GUGT had low sensitivity and moderate to excellent specificity. Sensitivity of the combined Rowland (cut-off ≥4)/GUGT was good at the 1-month follow-up and moderate at the 3-month follow-up for hospitalized patients; it was low for discharged patients. Specificity was low for hospitalized patients and good for discharged patients. CONCLUSION: Neither the objective measure of mobility (GUGT) nor the combined fTRST/GUGT or Rowland/GUGT improved the results. Our analysis shows that the predictive accuracy of the stand-alone, self-reported screening instruments fTRST and Rowland (cut-off=2) is still good. This study also confirmed their previously known limitations.