Are point-of-decision prompts in a sports science and medicine centre effective in changing the prevalence of stair usage? : a preliminary study
Pillay, Julian David
Lambert, Estelle V.
MetadataShow full item record
Objective. To determine the impact of a signed intervention on promoting stair versus lift usage in a health and fitness facility. Design. A 3-week observational study in which a simple timeseries design of collecting data before, during and after the introduction of an intervention was used. Setting. The Sports Science Institute of South Africa (SSISA): a 5-storey building with a centrally located lift lobby and internal stairwell. Method. Observers were placed unobtrusively on the ground floor, with good visibility of lift/stairwell, to observe ascending movement of students, staff, tenants, visitors and patients 4 hours/day (07h00 - 09h00, 16h00 - 18h00), 4 days/week for 3 weeks. During week 2, motivational signs were displayed on the wall next to the lift and stairs and on the floor leading to the stairwell. In week 3, signage was removed. Factors considered in predicting stair use were gender, phase of intervention, and whether persons were staff/students or visitors. Results. A total of 4 256 person-counts were recorded. Prevalence of stair use increased from 43% before the intervention to 53% during the intervention to 50% after the intervention. Odds of using the stairs during the intervention increased by 45% (odds ratio (OR) 1.45, 95% confidence interval (CI) 1.25 - 1.68) (p<0.00001), were 41% higher for staff/students compared with visitors (p<0.00001) and were 55% greater for women (p<0.00001). These effects did not change significantly after the intervention and stair use remained modestly higher than before the intervention. Conclusion. Signed intervention produced significant increases in stair usage during and after the intervention. These findings support the effectiveness of point-of-decision prompts for changing behaviour, and highlight potential factors influencing the impact of such messages.