Skip to main content
Dryad logo

Signage interventions for stair climbing at work: more than 700,000 reasons for caution

Citation

Eves, Frank (2019), Signage interventions for stair climbing at work: more than 700,000 reasons for caution, Dryad, Dataset, https://doi.org/10.5061/dryad.qz612jm91

Abstract

Increased stair climbing reduces cardiovascular disease risk. While signage interventions for workplace stair climbing offer a low‐cost tool to improve population health, inconsistent effects of intervention occur. Pedestrian movement within the built environment has major effects on stair use, independent of any health initiative. This paper used pooled data from UK and Spanish workplaces to test the effects of signage interventions when pedestrian movement was controlled for in analyses. Automated counters measured stair and elevator usage at the ground floor throughout the working day. Signage interventions employed previously successful campaigns. In the UK, minute‐by‐minute stair/elevator choices measured effects of momentary pedestrian traffic at the choice‐point (n = 426,605). In Spain, aggregated pedestrian traffic every 30 min measured effects for ‘busyness’ of the building (n = 293,300). Intervention effects on stair descent (3 of 4 analyses) were more frequent than effects on stair climbing, the behavior with proven health benefits (1 of 4 analyses). Any intervention effects were of small magnitude relative to the influence of pedestrian movement. Failure to control for pedestrian movement compromises any estimate for signage effectiveness. These pooled data provide limited evidence that signage interventions for stair climbing at work will enhance population health.

Methods

Observations from automated counters followed by bootstrapped regressions

Usage Notes

Read me file for UK interventions

Up from ground floor

The coding used to allow analysis of the file with logistic regression is explained below.  The dependent variable is stair and escalator choice.  If you have any queries, please do not hesitate to contact me directly (evesff@bham.ac.uk).

For ascent, file IJERPHMultiSiteUKUp, the coding is as follows below.

INTSTAGE; 1 =baseline, 2=first intervention, 3=second intervention.

BASE0INT1; the coding allows comparison of the baseline (0) with the first intervention (1).

PH2ASE1; the coding allows comparison of the first intervention (0) with the second (1).

ST1LIFT0; 0 for elevator choice, 1 for stair choice measured at the ground floor.

TRAFFUPMC; this is the number of individuals each minute travelling upwards from the ground floor, mean centred separately for each building.

FLOORS; the is the number of floors in the building.

LIFTS; this is the number of lifts in the building.

HOUR; this is the hour of day.

Down to ground floor.

Down to ground floor

The coding used to allow analysis of the file with logistic regression is explained below.  The dependent variable is stair and escalator choice.  If you have any queries, please do not hesitate to contact me directly (evesff@bham.ac.uk).

For descent to the ground floor, file IJERPHMultiSiteUKDown, the coding is as follows below.

INTSTAGE; 1 =baseline, 2=first intervention, 3=second intervention.

BASE0INT1; the coding allows comparison of the baseline (0) with the first intervention (1).

PH2ASE1; the coding allows comparison of the first intervention (0) with the second (1).

ST1LIFT0; 0 for elevator choice, 1 for stair choice measured at the ground floor.

TRAFFDOWNMC; this is the number of individuals each minute travelling down to the ground floor, mean centred separately for each building.

FLOORS; the is the number of floors in the building.

LIFTS; this is the number of lifts in the building.

HOUR; this is the hour of day.

Read me file for Barcelona interventions

The coding used to allow analysis of the file with multiple regression is explained below.  The dependent variables are the number of individuals using the stairs and lifts over a 30-minute period.  The observational data were mean centred for each choice point.  If you have any queries, please do not hesitate to contact me directly (evesff@bham.ac.uk).

For stair and lift usage, file IJERPHBarcelonaMultiSite, the coding is as follows below.

PHASE; the coding allows selection for analysis of the two phases.  Baseline and intervention data for the first intervention are coded as 1.  Baseline and intervention data for the second intervention are coded as 2.

BASE0INT1; the coding allows comparison of the baseline (0) with the intervention (1) within that phase.

STINMC; this is the number of individuals entering the stairs over a 30-minute period at the ground floor.

LIINMC; this is the number of individuals entering the lifts over a 30-minute period at the ground floor.

STOUTMC; this is the number of individuals exiting the stairs over a 30-minute period at the ground floor.

LIOUTMC; this is the number of individuals exiting the lifts over a 30-minute period at the ground floor.

HOUR; this is the hour of day.

 

Funding

Medical Research Council, Award: G0802070/91321

Bupa Foundation, Award: 22096780