Predictive Validity of the Life-Space Assessment in the Canadian Longitudinal Study on Aging (CLSA)

OBJECTIVES: The Life-Space Assessment (LSA) is a measure that assesses community mobility; however, its predictive validity has not been assessed in Canadians. This study assessed the predictive validity of the LSA among community-dwelling Canadians for adverse outcomes after 3 years. DESIGN: A longitudinal analysis of the Canadian Longitudinal Study on Aging (CLSA) comprehensive data set. SETTING AND PARTICIPANTS: Data from the CLSA were used (n = 22,695). The average age of the sample at baseline was 62.9 years (SD, 10.1), with a median LSA score of 90 (interquartile range, 74-100). METHODS: Receiver operating characteristic curves were conducted to determine whether the LSA scores at baseline could distinguish between those who experienced adverse outcomes at 3-year follow-up. Analyses were stratified by age and sex. We expected an area under the curve (AUC) of ≥0.7. RESULTS: AUC values ranged from 0.50 to 0.58 for predicting emergency department visits, nursing/convalescent home visits, hospitalizations, serious injuries, and injuries because of a fall. When stratified by age and sex, AUC values remained low. CONCLUSIONS AND IMPLICATIONS: The LSA did not show evidence of predictive validity for adverse outcomes among middle-aged and older Canadian adults after a 3-year follow-up. This may be because of the CLSA sample being relatively highly functioning at baseline, the 3-year-long gap in assessment, and the low prevalence of adverse outcomes in this sample. Further research is needed to determine the usefulness of the LSA for predicting these adverse outcomes at the population level.