Authors: Lillemor LundinOlsson Jane Jensen Lars Nyberg Yngve Gustafson
Publish Date: 2013/07/25
Volume: 15, Issue: 1, Pages: 51-59
Abstract
Background and aims It is of great importance to consider whether a tool’s predictive value is generalizable to similar samples in other locations Numerous fall prediction systems have been developed but very few are evaluated over a different time period in a different location The purpose of this study was to validate the predictive accuracy of the Mobility Interaction Fall MIF chart and to compare it to staff judgement of fall risk and history of falls Methods The MIF chart staff judgement and fall history were used to classify the risk of falling in 208 residents mean age 832±68 years living in four residential care facilities in northern Sweden The MIF chart includes an observation of the ability to walk and simultaneously interact with a person or an object a vision test and a concentration rating Staff rated each resident’s risk as high or low and reported the resident’s history of falls during the past 6 months Falls were followed up for 6 months Results During the followup period 104 residents 50 fell at least once indoors Many of the factors commonly associated with falls did not differ significantly between residents who fell at least once and residents who did not fall In this validating sample the predictive accuracy of the MIF chart was notably lower than in the developmental sample A combination of any two of the MIF chart staff judgement and history of falls was more accurate than any approach alone more than half of the residents classified as ‘high risk’ by two approaches sustained a fall within 3 months Conclusions Residents classified as ‘high risk’ by any two of the MIF chart staff judgement and history of falls should be regarded as particularly prone to falling and in urgent need of preventive measures
Keywords: