Latest / Safe As Podcast / E30: A better way to think about procedures: resources for action

Transcript

E30: A better way to think about procedures: resources for action · Safe As Podcast

• 0:07
  Beyond the checklists and compliance lies the human
• 0:10
  reality of work. A more progressive approach
• 0:13
  argues for a shift in thinking from controlling behaviour to
• 0:16
  supporting it, reminding us that effective procedures are tools
• 0:20
  that empower people, not chains that bind them.
• 0:23
  But what do workers and management think about
• 0:26
  procedures? Let's find out.
• 0:29
  Good day everyone. I'm Ben Hutchinson and this is
• 0:31
  safe as a bulk US dedicated to the thrifty analysis of safety,
• 0:35
  risk and performance research. Visit safetyinsights.org for
• 0:39
  more research. Today's study is from Hendrix
• 0:42
  and Perez 2021 titled Beyond Human Error, An Empirical Study
• 0:48
  of the Safety Model 1 and Model 2 Approaches for Predicting
• 0:53
  Workers Behaviours and Outcomes with Procedures.
• 0:56
  I covered another paper on procedures from these same
• 0:59
  authors in an earlier episode. This study explored the
• 1:02
  relationship between individual and system level variables on
• 1:06
  procedure use and procedural departures by interviewing 174
• 1:11
  industry personnel, and they also measured some other stuff
• 1:15
  that I won't talk about right now.
• 1:17
  Individual factors included things like people's perceptions
• 1:21
  of the utility of procedures, their thoughts around rule
• 1:24
  compliance, and a few other things.
• 1:26
  System variables included the procedural quality, perception
• 1:30
  of the procedure, change process, and safety climate.
• 1:34
  What was really interesting is they framed the analysis by the
• 1:36
  Model 1 and Model 2 concept. Now, just for clarity, this
• 1:40
  isn't Safety One and Safety 2. The concept originally came from
• 1:44
  Sydney Decker, but it was really fleshed out and discussed by
• 1:48
  Hale and Boris in two 2013 papers really quickly.
• 1:53
  So based on Decker's conceptualisation and then Hale
• 1:56
  and Boris expansion of it, it's said that an organisation
• 1:59
  approach to procedural systems can be kind of thought about as
• 2:02
  a Model 1 or Model 2 approach. Model 1 sees procedures as rules
• 2:07
  to be followed closely and strictly.
• 2:09
  Therefore deviations are always a mistake and reflect a problem
• 2:13
  with the worker. Much of the previous research
• 2:16
  has really looked for methods to decrease human error, trying
• 2:19
  ways to fix the worker via rules.
• 2:23
  This Model 1 approach tends to see procedures as things that
• 2:26
  need to be followed rigidly, and workers shouldn't be deviating
• 2:30
  from the procedures. So from the Model 1 perspective,
• 2:34
  rules are very rationalistic and prescriptive, and the rules that
• 2:38
  seem to be the one correct way to do the work, they're to
• 2:41
  protect workers from their own fallibility.
• 2:44
  These rules are seen as essentially static, to be worked
• 2:47
  out as a one off exercise and ended up being modified when the
• 2:51
  activity changes substantially. The Model 1 is essentially top
• 2:55
  down. It gets applied to the workforce
• 2:57
  and then management act as the enforcers.
• 3:01
  In contrast, the Model 2 perspective sees procedures more
• 3:04
  as tools for support work and resources for action.
• 3:08
  So some prior research found that procedures alone weren't
• 3:11
  sufficient to support work, but they do often best serve as
• 3:14
  resources for action. The Model 2 has more of a
• 3:17
  dynamic nature to it, allowing some level of flexibility and
• 3:21
  interpretation of procedural systems, which is then theorised
• 3:25
  to help maximise the utility of a procedure.
• 3:27
  So the Model 2 view of rules sees rules as patterns of
• 3:31
  behaviour, socially constructed, emerging from experience.
• 3:35
  They're characterised as local and situated in the activity.
• 3:39
  So at Model 2 sees rules essentially bottom up and
• 3:42
  dynamic. It recognises that rules can
• 3:44
  never be complete and the written ones are seen as
• 3:46
  essentially underspecified, requiring a process of
• 3:50
  translation and adaptation by workers before application to
• 3:54
  any given specific situation. So rules are seen in Model 2 as
• 3:58
  a support and guidance for the expert, as templates and
• 4:01
  resources for adaptation. Now, importantly, none of this
• 4:06
  suggests that Model 1 is bad and Model 2 is always good, or vice
• 4:11
  versa. We need both Model 1 and Model 2
• 4:14
  thinking with their procedures. What's important is really
• 4:17
  understanding how they interact and how rules are used and
• 4:21
  perceptualized in the field. But one of the key prompts of
• 4:25
  this study is that improving procedural systems might be
• 4:29
  limited if the focus is placed mostly on trying to tightly
• 4:32
  control worker behaviour via the Model 1 approach, rather than
• 4:36
  accounting for broader organisational factors.
• 4:39
  So what did they find? Evidence was found for both
• 4:43
  individual and system variables. A range of things like perceived
• 4:47
  procedure quality, role experience, attitudes towards
• 4:51
  the procedure and procedural utility, and people's own
• 4:54
  compliance attitudes were all related to some types of outcome
• 4:57
  measures. Attitudes towards procedure
• 5:00
  utility was found to be the only predictor of incidents and near
• 5:04
  misses, where better attitudes regarding the utility procedures
• 5:08
  was related to fewer self reported yearly incidents and
• 5:12
  fewer procedural departures. A system level variable,
• 5:15
  perceived procedure quality was overall the best predictor for
• 5:19
  all outcome and behaviour variables.
• 5:22
  Predictably, higher industry experience was related to less
• 5:26
  procedure use and greater rule departures.
• 5:29
  Perception of higher procedure quality was related to fewer
• 5:31
  incidents and near misses in both individual and system
• 5:35
  models. So generally, and maybe not too
• 5:38
  unexpectedly, the more positive people saw the procedures, the
• 5:43
  more likely they were to actually use them.
• 5:45
  The findings also highlight the importance of procedure quality
• 5:49
  as it's the best and unique predictor of both procedural
• 5:52
  deviation and procedure use behaviours.
• 5:55
  So these findings indicate that not only do system level
• 5:58
  variables predict procedure related behaviours and outcomes
• 6:02
  like injuries and incidents, but they uniquely predict above and
• 6:05
  beyond individual level characteristics, suggesting the
• 6:08
  importance of system variables. Further to this point is that a
• 6:12
  Model 1 view of procedures may be necessary but insufficient
• 6:17
  for understanding and predicting procedural system effectiveness.
• 6:21
  Said differently, a view of procedures as strict
• 6:24
  prescriptions of of the idealised way of working isn't
• 6:28
  exclusively valid. Based on these results, when
• 6:31
  workers perceive procedures as tools support them rather than
• 6:34
  as prescriptions to follow or control their behaviour, then
• 6:37
  they're more likely to use them, more likely to adhere to the
• 6:40
  procedures, and the outcomes are more favourable.
• 6:44
  Other system variables like safety, climate and the change
• 6:47
  process evaluation, you know, change management around
• 6:50
  procedural changes, didn't uniquely predict any behaviours
• 6:53
  or outcomes that were studied in this paper.
• 6:56
  A further interest is that while reported procedural deviations
• 6:59
  are corollary to the incidents, they don't provide more
• 7:02
  predictive ability beyond perceived procedure quality.
• 7:06
  That is, procedural departures are often due to the procedure's
• 7:10
  quality rather than any other issues related to people.
• 7:14
  They just don't care enough. They're lazy, they're
• 7:17
  complacent. These things didn't predict the
• 7:19
  procedure used or the procedural departures.
• 7:22
  It was the perception of the quality of the procedures that
• 7:25
  did. So what can we make of the
• 7:27
  findings based on this study and several other studies from the
• 7:30
  same research group suggests that there's value in
• 7:34
  considering rules not as concrete in flexible constructs,
• 7:38
  but as knowledge reservoirs or resources for action.
• 7:42
  They should help sensitise people to how the work needs to
• 7:45
  be performed or should be performed or could be performed,
• 7:48
  what to do, how to do it, when to do it, particular things that
• 7:53
  must be done or must not be done.
• 7:55
  If a rule is important, as in the work has to be done in a
• 7:59
  particular way or sequence, then make it known.
• 8:02
  Ensure people are set up to succeed with the task and the
• 8:05
  procedure. On that point, there's really
• 8:08
  good guidance out there on how to construct effective
• 8:11
  procedures and here's a rule of thumb.
• 8:14
  If these sort of guidelines haven't been used to construct
• 8:17
  your procedures, then there's probably a decent chance that
• 8:21
  your procedures suck. Another idea not discussed in
• 8:24
  this paper is breaking up your rules into different types.
• 8:27
  This is discussed again by Hale and Burroughs and other work,
• 8:31
  but they differentiate between performance goals or performance
• 8:34
  rules, process rules and action rules.
• 8:37
  Performance based rules define only what has to be achieved and
• 8:40
  not how it must be done, so this could include just general
• 8:44
  duties or so far as reasonably practicable.
• 8:47
  Process rules define the process which people have to follow, but
• 8:52
  leave considerable freedom about how they go about it.
• 8:56
  For instance, the requirements to prepare risk assessment or to
• 9:01
  set up an SMS, or requirements to consult with certain people
• 9:05
  when an emergency situation arises.
• 9:08
  Third, action rules. These tend to be the most
• 9:11
  prescriptive and they're usually specified in kind of if and then
• 9:15
  statements. They really prescribe in detail
• 9:17
  exactly how people should behave or how the hardware should be
• 9:20
  designed or tested. Action rules provide the least
• 9:24
  flexibility. They really should only be used
• 9:26
  sparingly for things that are absolutely genuinely critical or
• 9:30
  need to be done in a very specific way in a particular
• 9:33
  sequence. Unfortunately, a lot of the
• 9:36
  rules that we default to in safety tend to be action rules,
• 9:39
  when what we really should be defaulting to are more formants
• 9:42
  or process rules. Also consider the use of meta
• 9:45
  rules. Again, not discussed in this
• 9:47
  paper, but meta rules are rules about how to use rules.
• 9:52
  Or more specifically for this context, when can we work around
• 9:56
  the rule like dispensations for rules or modifications to rules?
• 10:01
  Simple heuristics and rules of thumb can work really well here.
• 10:05
  Dave Snowden, famous from the Kneven, has some really
• 10:08
  interesting ideas on meta rules and heuristics around procedural
• 10:12
  departures and and procedural dispensations.
• 10:15
  But as a final caution, be careful with confusing
• 10:19
  prescribed systems with reality. Procedures are only ever
• 10:23
  incomplete abstractions of work. They're not the work itself.
• 10:28
  Paraphrasing the authors Pentland and Feldman, it's a
• 10:32
  folly of designing artefacts while hoping to change patterns
• 10:35
  of action. Or, as Didirio said, another
• 10:39
  author, documents aren't the routine.
• 10:42
  That is, no matter how beautiful they are, no matter how much you
• 10:47
  work on them, those procedures are never the things that people
• 10:50
  do. So overall, this study suggests
• 10:53
  that Model 1 and Model 2 views of procedures are complementary
• 10:58
  and not conflicting. They both have their role.
• 11:01
  Further, only focusing on procedure quality in itself, you
• 11:05
  know, as in trying to create the perfect procedure without
• 11:07
  accounting for the utility of the procedure as experienced by
• 11:10
  workers may fail to make any significant progress and thus
• 11:15
  highlights the need for workers to create useful procedures as
• 11:19
  defined by themselves. So limitations as you'd expect,
• 11:24
  interviews and cross sectional surveys that can really only get
• 11:27
  you so far. The authors note that it would
• 11:30
  have been useful to have some observational data and probably
• 11:33
  a use of some naturalistic or ethnographic studies of field
• 11:36
  work and use of procedures would be really useful.
• 11:40
  That's it on Safe as I'm Ben Hutchinson.
• 11:43
  Please help share, rate and review and check out
• 11:47
  safetyinsights.org for more research.
• 11:49
  Finally, feel free to support Safe As by sharing a coffee link
• 11:54
  in the show notes.