On Tue, 1 Aug 1995, MR GEOFFREY F FOUNTAIN wrote:
> I work at a very large site (14,000 workers) for a primary contractor
> that works for the federal government. The primary contractor took over
> from the original contractor in 1989, which built in operated the site
> from the early 50s until '89. The original contractor had a very
> strong safety culture. It was obvious to everyone (and to no one's
> surprise) that the managers of the new contractor were less than
> prepared to reinforce this safety culture when they came on board.
> headquarters guidance. A real recipe for organizational disaster.
[...snip...]
>
> In the last ten months, our division (one of about five large facility-
> based organizations within the site) was on track to set a new record
> of 10 million work hours without a Lost Workday Case (LWC). We were
> being hailed as an organization who put safety at the top, etc, etc.
> Then we had our first LWC. Site E-mails came out from our site
> management noting that even though we had just missed the record, we
> should be proud of our achievement. Then over the next four to six
> weeks we had four more LWCs.
>
> Recently our department manager had a safety "stand down" meeting with
> all employees. He talked for fifteen, emotionally filled minutes,
> emphasizing with a pointed finger at least fifty times how concerned he
> was about us having another injury and how we must turn this situation
> around.
[...snip... including discussion of measuring precursors in addition to
accidents...]
I like your idea of measuring pre-cursors. Accidents are infrequent,
random, and a poor measure. If we could measure the incidence of unsafe
behaviors, this would give us a lot more sensitivity, more meaningful
data, and probably earlier indications of where action is required.
The FAA monitors near misses and incidents, and has an elaborate system
to get reports of unsafe situations. These are monitored as closely as
accidents themselves (although a real loss-of-life accident is studied
more carefully, of course).
Balancing Loop?? Or Reinforcing?
--------------------------------
One of my favorite systems examples is safety: If safety is driven from a
balancing loop (accidents --> accident reduction programs) then what
happens when safety improves? Well, the pressure comes off, there is less
attention to safety, and things go to hell.
Here is the balancing loop diagram for what I'm trying to describe:
-> Accidents Target
/ \ Level of
/ \ Accidents
| \ /
| v v
| B Gap
| /
\ /
\ v
-- Safety
Programs
(Ugh! This is a lot easier to draw in PowerPoint!)
Caption for Diagram: If safety is driven from a balancing loop, then if
Accidents increase, this causes an increase in safety program which will
reduce accidents. But when accidents decrease, the pressure comes off, and
Safety Programs wither. No one ever meets and sets an explicit target
level, "OK, if accidents fall below level X, that's good enough!", but the
effect is just the same. It's as though there is an implicit target.
With accidents, especially those that injure, main and disable, we want
to drive safety from a reinforcing loop which will keep up the pressure
even when there are no accidents. One that will drive accidents and
dangerous situations lower and lower.
Richard Karash ("Rick") | <http://world.std.com/~rkarash>
Innovation Associates, Inc. | email: rkarash@world.std.com
3 Speen St, Framingham MA 01701 | Host for Learning-Org Mailing List
(508) 879-8301 and fax 626-2205 | <http://world.std.com/~lo>