Normalization of deviance in diagnosis .... primary problems

Woods, David woods.2 at OSU.EDU
Mon Jan 12 14:39:15 UTC 2015

The article referenced makes a common error when discussing Challenger, deviations from procedure, and normalization of deviance.

The article presumes that operators deviations from procedures, practices and policies is normalization of deviance.  This is not the original finding.

The original observation and subsequent observations of normalization of deviance was as a safety management problem.  In these cases, evidence was coming in that the underlying engineered/physical processes was operating outside of the boundaries of well understood disturbances and system behavior — and this evidence was discounted by management. This is simplest to see in Columbia accident where, among other factors, the energy of foam strikes was 100x greater than expected/previously analyzed in terms of risk, yet rationalized away by managers under production pressure.  The observations refer to the difficulties in conceptual change and how this applies to safety management.  These processes could occur in operators but the findings occurred with management / supervisory roles under production pressure who then discounted incoming evidence that operations were ‘drifting' outside of the well understood risk envelope.  The best and shortest description is in Feynman’s dissent from the Challenger report (It’s Appendix F).  Conceptual change is hard and there are findings about how to enhance this, for example how diverse perspectives help.  A good example and success story in health care is the work of George Blike on pediatric sedation funded a number of years ago by NPSF.

How do the findings related to normalization of deviance apply to operator behavior when operators appear in hindsight to deviate from procedures, practices and policies?  When operators normal behavior is observed to deviate from a standard, the first hypothesis is that the distant parties’  model of the work system is off and needs updating and revision.  This is not to say the operator behavior is correct but rather that the behavior is the result of systemic factors and operator attempts to accommodate multiple pressures over time. The common phrase the specialists use today is that there is a difference between “work-as-imagined and work-as-practiced”.  The difference arise from workarounds that are adaptations to try to deal with crunches and  bottlenecks that occur at the sharp end of work.  A good example is that distant views of a work system often underestimate or miss how workload bottlenecks accumulate as tasks congregate in time or how small problems can quickly cascade into more demanding situations.  The role of supervision is to recognize this gap, investigate the work system, and learn what are the factors, difficulties, and demands that have been underestimated or missed and then to revise their model of how safety is created.  The law of fluency captures the difficulty in seeing the gap:  " ‘Well'-adapted cognitive work occurs with a facility that belies the difficulty of the demands resolved and the dilemmas balanced.”  Redesigning the work system and managing the work system begins with understanding what factors create the gap, and effective organizations continually spend effort to do this.

When people are doing something different than you expect them to do,
• the first response is that your model of the work system is off as people are accommodating factors that you have missed;
• the second response is that your model of the work system is off as people are accommodating factors that you have missed;
• the third response is  that your model of the work system is off as people are accommodating factors that you have missed.
The hypothesis that the observed behaviors are due to unappreciated systemic factors is fundamental to proactive safety management which is now a requirement in aviation organizations.
Maybe after looking hard, the model of the work system is on track and the operator behavior is so off that some individuals should be moved to other less safety critical roles, but usually (and I mean usually) it turns out that system factors actually create the behavior patterns.  As is basic in systems safety:  “a system does what it was designed to do, its just not what the designers intended.”

There are a variety of things people could look at on safety management, conceptual change in safety management, the Columbia accident as a good case study, cognitive work systems, and resilience engineering.

 Hope this helps as background as you try to focus in on critical issues in improving diagnosis.


David Woods, PhD  木材
Releasing the Adaptive Power of Human Systems

• Lead, Initiative on Complexity in Natural, Social & Engineered Systems
• Co-Director, C/S/E/L Cognitive Systems Engineering Laboratory

[cid:0a56a0af-9e73-4a26-8d58-a8da49b506ac at]
Department of Integrated Systems Engineering
The Ohio State University
290 Baker Systems | 1971 Neil Ave Columbus, OH 43210
woods.2 at<mailto:woods.2 at>

• Past-President and Executive Committee, Resilience Engineering Association

On Jan 11, 2015, at 10:14 AM, Lorri Zipperer <Lorri at ZPM1.COM<mailto:Lorri at ZPM1.COM>> wrote:

This short piece discusses the normalization of ineffective behaviors and their effect on medication safety.

What poor behaviors are normalized that degrade the diagnostic process? What about not looking at evidence but relying on only one resource or only one colleague’s assessment to inform decision making?


Lorri Zipperer, editor
Patient Safety: Perspectives on Evidence, Information and Knowledge Transfer. London, UK.
Gower. June 2014. ISBN: 978-1-4094-3857-1
free chapter
LinkedIn Group / Patient Safety Partners:

"everyone has to be his own leader, on his own sled."
Joseph Beuys



To unsubscribe from IMPROVEDX: click the following link:


Visit the searchable archives or adjust your subscription at:

Moderator:David Meyers, Board Member, Society for Improving Diagnosis in Medicine

To learn more about SIDM visit:

To unsubscribe from the IMPROVEDX:
or click the following link: IMPROVEDX at LIST.IMPROVEDIAGNOSIS.ORG


For additional information and subscription commands, visit:

http://LIST.IMPROVEDIAGNOSIS.ORG/ (with your password)

Visit the searchable archives or adjust your subscription at:

Moderator: David Meyers, Board Member, Society to Improve Diagnosis in Medicine

To unsubscribe from the IMPROVEDX list, click the following link:<br>
<a href="" target="_blank"></a>

HTML Version:
URL: <../attachments/20150112/e975e5a6/attachment.html> ATTACHMENT:
Name: osu-emailsig.png Type: image/png Size: 5963 bytes Desc: osu-emailsig.png URL: <../attachments/20150112/e975e5a6/attachment.png> ATTACHMENT:
Name: pastedGraphic.pdf Type: application/pdf Size: 51130 bytes Desc: pastedGraphic.pdf URL: <../attachments/20150112/e975e5a6/attachment.pdf>

More information about the Test mailing list