Our Software Saves Lives by Predicting Workplace Injuries


Using Your Inspection Data to Improve Safety Performance through a Leading Indicator Safety Incentives Program

What comes to mind when you hear the term “safety incentives”?  Pizza parties held after 30 days worked without an injury?  Prizes raffled off after six weeks without a lost-time incident?  Maybe you’ve even run an incentives program in the past and have had direct experiences with the benefits and challenges of implementing such programs.

These days, support for incentive programs is mixed.  Some argue that incentive programs, especially those based on lagging indicators of safety (e.g. injury rates or counts), lead to a decrease in injury reporting – not actual injuries.  Others are more optimistic about the programs, noting that they can help to improve workplace safety and that rewards can serve to increase morale, even if just for a short period of time.

As an occupational health and safety researcher, I want to help answer the question of whether safety incentive programs are effective in improving working conditions and injury rates. I’m interested, however, in safety incentive programs based on leading indicators of workplace safety, which are recorded before an incident occurs. My research team and I developed a novel safety incentive program (called B-SAFE: Building Safety for Everyone) that rewards based on measures of workplace safety recorded before an incident occurs, not lagging, post-incident measures.  Our program relies on leading indicators of safety such as safety inspections, like those entered into SafetyNet, to evaluate the safety on-site and determine whether or not workers should receive a reward.

Yet before we could implement the program on worksites, we needed to determine what level of safety should be rewarded.  Ideally, a program would reward the site only when all of the observations are safe (i.e. 100% safe).  However, this is an unrealistic goal and would likely result in a program where workers are never rewarded and as such, have less motivation to change unsafe work practices.  Conversely, a threshold that is too low might result in workers easily achieving the reward each month, causing them to see no reason to continuously improve safety-related behavior. A reward threshold should feel attainable to all workers on-site, but it should also be competitive enough to encourage improvements in safety-behavior.

What is Considered “Safe” enough to Reward in a Leading Indicator Incentive Program?

To answer this question, our team evaluated multiple approaches to determining the numerical value of a safety inspection score that should be used in an employee safety incentive program to active a reward.  We examined the effect of different reward thresholds by using safety inspection data collected over a 19-month period at 65 Harvard University-owned construction projects.  We grouped this data together in five different ways: owner, general contractor, project, trade, or subcontractor.  We calculated the monthly cumulative safety inspection score for each of the groupings and selected the median value as the reward threshold.

We then applied the five thresholds to data from a completed Harvard-owned construction project in order to calculate what the frequency and distribution of rewards would look like in a monthly safety incentive program.  Each approach was then evaluated qualitatively for consistency, competitiveness, attainability, and fairness.

Categorizing inspection data by owner resulted in a threshold score of 96.3% and met all of our qualitative evaluation goals.  In this simulation, the reward frequency would have been approximately 33%, making it a competitive, yet attainable.  Additionally, the site was evaluated as a whole to ensure that all workers were treated equally, and the threshold value remained the same throughout the duration of the project.

Using this Method at Your Own Site

As a site owner or general contractor, you can adapt this process of selecting a reward threshold when establishing a leading indicator based safety incentive program on your worksite.  Here’s how:

  1. Collect safety inspection data from a program such as SafetyNet that has been generated from your previous projects within the last year and a half.
  2. Generate cumulative monthly safety performance scores (percent of safe observations from the total observations) based on the inspections for all of projects.
  3. List the monthly safety performance scores from lowest to highest, and select the middle value.  That’s your threshold.
  4. Continue your normal safety inspection process and generate a cumulative worksite safety performance score at the end of each month.
  5. If this value is at or above the threshold, recognize the worker’s achievements with a reward.  Reset the monthly cycle and re-evaluate the site at the end of the following month.

What is the Effect of Leading Indicator Incentive Programs on  Injuries?

This study is part of an effort by our research group to evaluate the  true impact of safety incentive programs on work-related injury and  safety culture.  We are currently collecting data on safety inspections,  injury rates, and worker perceptions of safety culture on worksites in  the Boston area with and without a leading indicator-based safety  incentive program.  This is the first study to evaluate the effectiveness  of such a program on multiple sites.  Anecdotally, the results from our first year of data collection indicate a positive impact of the safety incentive program.  Safety managers on sites with the program indicated higher levels of teamwork and employees noted the increased awareness around the importance of safe work practices.

Additional Information

More information on the threshold determination process can be found here: http://www.sciencedirect.com/science/article/pii/S0925753512001531.  This project is funded by CPWR – The Center for Construction Research and Training.  A project description and abstract can be found on their website at: http://www.cpwr.com/abstracts/2009-2014_jdennerlein.html.