Development and psychometric properties of the Balance in Daily Life (BDL) scale in a population of frail older people

Abstract

Background

Balance disorders in older people cause falls, which can have serious functional and economic consequences. No existing scale relates fall risk to daily life situations. This study describes the development, psychometric properties and construct validity of the Balance in Daily Life (BDL) scale, comprising seven routine tasks including answering a phone, carrying a heavy bag, and sitting down and getting up from a chair.

Methods

Frail patients aged 65 years or more were prospectively recruited from the geriatric rehabilitation department of a French university hospital. Inclusion criteria included autonomous walking over 20 m and modified Short Emergency Geriatric Assessment score 8–11. Patients with motor skills disorders and comprehension or major memory difficulties were excluded. Patients were assessed on Day 3 and Day 30 with the Balance in Daily Life scale, Timed Up and Go, one-leg stance time, sternal nudge and walking-while-talking tests. The scale was assessed for acceptability, quality, unidimensionality, internal consistency, reliability, temporal stability, responsiveness and construct validity.

Results

140 patients (83 ± 6 years) were recruited, of whom 139 were assessed at Day 0 and 133 at Day 30. Acceptability was satisfactory (134/139 patients completed the test), quality assessment showed a slight floor effect (6 % of patients with minimal score) and evaluation of item redundancy found no strong correlation (Spearman <0.7). Unidimensionality was verified (Loevinger H coefficient > 0.5 for all items except item 6 = 0.4728). Internal consistency was good (Cronbach alpha = 0.86). Reliability and temporal stability were excellent (ICC = 0.97 and ICC = 0.92). Responsiveness was verified by significant score change p < 0.0001 between Day 0 and Day 30 (decreased by 1 [0; 2] point), in line with other score changes. Construct validity revealed that the Balance in Daily Life scale was convergent with results of the timed up-and-go and one-leg stance time (p < 0.0001 for both) and tended to be higher for participants who had not fallen in the previous 6 months (p = 0.0528). The new questionnaire was divergent to sternal nudge tests (p = 0.0002) and not related to the walking-while-talking test (p = 0.5969).

Conclusion

The Balance in Daily Life scale has good psychometric properties for this population. Its simplicity and innovative nature mean that it can be applied in institutions while being easily modifiable to domestic settings.

Study registration on clinicaltrials.gov: NCT0334382.