Intra- and interrater reliability and agreement of the Danish version of the Dynamic Gait Index in older people with balance impairments

Abstract

Objectives: To examine the intrarater and interrater reliability and agreement of the Danish version of the Dynamic Gait Index (DGI) in hospitalized and community-dwelling older people with balance impairments.

Design: Reliability study.

Setting: University hospital and outpatient rehabilitation.

Participants: A convenience sample of older people (≥65y); 24 subjects from a hospital and 24 from an outpatient rehabilitation center. All subjects had either 1 or more falls within the last year or balance impairments evaluated by a physical therapist.

Interventions: Not applicable.

Main outcome measures: All subjects carried out the DGI twice with a 1.5-hour interval. Each subject was rated by 3 physical therapists in the first attempt (1 for intrarater and 2 for interrater comparison) and by the intrarater in the second attempt, in both settings. The reliability was calculated using the intraclass correlation coefficient (ICC, 2.1), while agreement was calculated as the smallest real difference (SRD).

Results: The ICC for intrarater and interrater reliability of the total DGI was .90 and .92 at the hospital, while the SRD was 2.72 and 2.58 points, respectively. Correspondingly, the ICC for intrarater and interrater reliability of the total DGI at the rehabilitation center was .89 and .82, while the SRD was 3.49 and 3.99 points, respectively.

Conclusions: The intrarater and interrater reliability of the total DGI ranged from good to excellent in hospitalized and community-dwelling older people. Improvements of 3 and 4 DGI points for hospitalized and community-dwelling older people, respectively, should be regarded as a real change (with a 95% certainty).