Validation of a Record-based Frailty Assessment According to the Multidimensional Prognostic Index

Background Various phenotype and cumulative frailty assessment tools have been developed and compared in research. For use in an in-hospital setting, a cumulative and graded frailty assessment method is preferred in order to identify subgroups of patients at risk of adverse events during discharge and transition to primary care. Comprehensive Geriatric Assessment (CGA) is the gold standard to assess frailty. Most medical wards have limited access to specialized geriatric team support capable of performing CGA. Thus in these wards there is a need for a CGA based instrument to identify the frail patients and to quantify the level of frailty. The Multidimensional Prognostic Index (MPI) is based on CGA and is a comprehensive cumulative deficit frailty assessment tool validated in a Danish geriatric department. MPI is fully applicable in the everyday clinical work and supplies useful information to clinicians. It can predict readmission and death, and it is well-suited to assess the degree of frailty. Enabling identification of patients at risk of adverse events facilitates targeting of the interventions in order to improve patient outcomes. The MPI is a bedside assessment. However, in observational record-based research the patient is not accessible for the researcher. To assess and identify hospitalized frail patients retrospectively for clinical research, a valid record-based frailty assessment method is needed. The aim of this study is to compare the accuracy of a record-based MPI assessment with a bedside performed MPI assessment in order to use the record-based MPI when access to bedside MPI is impossible.