Table 1 The independent variables of interest in this study

Ambulation activity

Continuously; 6 times a second on each resident by a RTLS technology; triangulated location and motion data from wristband worn by resident and sensors mounted throughout CLC’s

Path characteristics:

1) Time spent walking in a path (at least 60 s of uninterrupted walking separated by at least 30 s of non-ambulatory intervals before and after the path)

2) Distance covered in a path (distance, in miles, where there is at least 60 s of uninterrupted walking)

3) The number of paths in a week (count)

1) & 2) Spearman correlation with Tinetti Gait subscale (0.32–0.35) and Tinetti Balance subscale 0.37–0.40 (unpublished data)

− 95% concordance in accuracy in ambulatory path, correct location, accurate time with direct observation (observational study, unpublished data)

Tortuosity (random changes in direction during movement measured by deviation from a straight line measured from 0–1)

Spearman correlation with stride-time variability measured by a Gait–Rite mat (0.30)²

Spearman correlation with Mini-Mental State Exam (− 0.47)²

1) Time (minutes) and 2) distance walking (miles)

Spearman correlation with Tinetti Balance subscale 1) (0.11–0.40) Tinetti Gait subscale 2) 0.35 (unpublished data)

Gait speed

Spearman correlation with the Tinetti Performance Oriented Mobility Assessment (0.39) (unpublished data)

Lapping and pacing patterns first identified by hand-coding of 2-D visualizations (e.g., gif files), then using Rubine classifiers to identify patterns

Inter-rater reliability for observations of these patterns in this setting and population (0.89)