Objective: Vital signs are often used in triage, but some may be difficult to assess in low-resource settings. A patient's ability to walk is a simple and rapid sign that requires no equipment or expertise. This study aimed to determine the predictive performance for death of an inability to walk among hospitalized Malawian adults and to compare its predictive value with the vital signs-based National Early Warning Score (NEWS).

Methods: It is a prospective cohort study of adult in-patients on selected days in two hospitals in Malawi. Patients were asked to walk five steps with close observation and their vital signs were assessed. Sensitivities, specificities, and predictive values for in-patient death of an inability to walk were calculated and an inability to walk was compared with NEWS.

Results: Four-hundred and forty-three of the 1094 participants (40.5%) were unable to walk independently. In this group, 70 (15.8 %) died in-hospital compared to 16 (2.5%) among those who could walk: OR 7.4 (95% CI 4.3-13.0 p<0.001). Inability to walk had a sensitivity for death of 81.4%, specificity of 63.0%, positive predictive value (PPV) of 15.8%, and negative predictive value (NPV) of 97.5%. NEWS>6 had sensitivity 70.9%, specificity 70.6%, PPV 17.1%, and NPV 96.6%. An inability to walk had a fair concordance with NEWS>6 (kappa 0.21).

Conclusion: Inability to walk predicted mortality as well as NEWS among hospitalized adults in Malawi. Patients who were able to walk had a low risk of death. Walking ability could be considered an additional vital sign and may be useful for triage.