Human Factors in Medicine

1
Human Factors in Medicine

• Robert Helmreich, FRAeS
• Professor of Psychology
• Human Factors Research Project
• The University of Texas at Austin
• Canadian Anesthesiologists Society
• Quebec City
• June 18, 2004

2
Psychology and Human Factors
Human factors is a sub-discipline of psychology
that deals with humans interacting with
technology This discussion will address the
broader discipline of psychology and its
relevance for medicine
3
Aims of Psychology

• Psychology attempts to understand individual,
  group, organizational, and national behavior
• Psychology includes physiological as well as
  social determinants of behavior
• Psychology studies both normal and pathological
  behavior
• The body of knowledge that forms psychology
  should be applicable to almost all aspects of
  medicine

4
Is psychology useful in the real
world?Opinions
5
Criticism from the left

• Those who study behavior in organizations are the
  softest of the soft - it is not science but mumbo
  jumbo
• There is a dearth of meaningful theory

6
Criticism from the right

• Psychologists look at extremely narrow questions
  that can be studied in the meaningless test tube
  of the college laboratory using two legged white
  rats as subjects

7
Two Extremes

• Academic psychology

• highly theoretical
• rigorous methodology
• findings difficult to translate to real world

• Organizational Psychology and Management

• focus on case study
• intuitive analysis
• findings often fail to replicate
• theories not based on strong empirical data

8
Part of the acrimony comes from a schism between
applied and theoretical researchBut . . .
Nothing is so practical as a good
theory.KURT LEWIN, 1948
9
A psychological theory is a coherent collection
of hypotheses that serve to explain a broad range
of empirical generalizations and facts
10
Rapprochement Psychology is useful if

• You ask meaningful questions
• The questions can be measured empirically
• The answers are based on data
• The limits of generalization are recognized

11
Topics open to systematic study

• Individual characteristics and performance
• Team dynamics and performance
• Organizational culture and climate
• National culture and its sequelae
• Organizational change
• Inter-group dynamics
• Doctors and nurses
• Front line personnel and management

12
Models

• Data can be organized in a model of how factors
  relate
• Useful models are based on empirical data
• Models of flight crew and medical team
  performance have been developed by the University
  of Texas group reflecting our experiences in
  aviation and medicine

13
A Model of Medical Team Performance
14
Input and Process Factors
15
Medical Team Outcomes
16
Filling in the puzzleMethods of gathering data
17
Unsystematic observation

• Walk around and gather information
• Use the brain to integrate and formulate
  explanation
• Subject to human limitations and biases
• Results often fail to replicate

18
Participant observation

• Immerse self in situation
• Distill results from an insiders viewpoint
• Loss of perspective
• Invalid conclusions

19
Survey research

• Develop critical questions and collect from large
  sample or population
• Economical
• Limited by lack of trust or interest
• Bad questions yield bad answers
• Still method of choice to gather baseline and
  change data

20
Interviews Ask and you shall know

• Important source of date but limited by
• Demand characteristics respondents infer
  desired response and answer accordingly
• Lack of trust
• Deception
• Margaret Mead and the Samoans lurid tales of sex
• Bias
• Questions slanted to produce desired response
• Can be intentional or inadvertent

21
Laboratory experiments

• Manipulate conditions in controlled environment
• High certainty about causality
• Limited generality due to lack of reality
• Findings limited because situation
  inconsequential to subjects

22
Field experiments

• Conditions manipulated in real life situation
• Possible influence of extraneous factors
• Ethical and practical limitations
• Use to test important questions
• Example drug testing using random assignment to
  placebo and drug conditions

23
Systematic observation

• Use detailed, reliable coding system to record
  observable behaviors
• Highly reliable data if
• observers well trained
• trust relationship with researcher
• Costly for collection and analysis

24
Case Studies

• Advantage ability to probe deeply into an
  event, its context, and its precursors
• Disadvantage unsure how much findings will
  generalize to other situations and organizations
• Evaluation an important approach that should be
  used to examine both negative and positive events

25
The University of Texas Threat and Error
Management Model

• The model was derived empirically from
  observations of flight crew performance in normal
  line operations
• Line Operations Safety Audit (LOSA)
• UT-TEMM has three parts
• External threats and external errors and their
  management
• Crew errors and their management
• Undesired aircraft states and their management

26
UT-TEMM
27
Use of the Model

• Framework for analysis of data and application of
  empirical taxonomies in
• Line Operations Safety Audits (LOSA)
• Analysis of incidents and accidents
• Training that stresses threat and error
  management as a central focus of Crew Resource
  Management (CRM 6th generation)

28
Threats

• Definition Overt and latent factors external to
  the flight crew that originate outside a flight
  crews influence and must be actively managed to
  avoid becoming consequential to safety
• Threats increase the complexity of the
  operational environment

29
External Threats

• Expected
• Forecast weather
• Terrain
• Unexpected
• Abnormals
• Traffic
• External error
• Air Traffic Control errors
• Dispatch errors

30
External Threat Examples

• Adverse weather
• Terrain
• Traffic
• Airport conditions
• A/C malfunctions
• Automation events
• Communication events

• Operational time pressures
• Non-normal operations
• ATC commands / errors
• Cabin events / errors
• MX events / errors
• Dispatch events / errors
• Ground crew events / errors

31
Errors

• Definition deviations from crew or
  organizational intentions or expectations

32
Global Flight Crew Error Types

• Procedural Followed procedures but wrong
  execution
• example) Wrong altitude setting dialed into the
  MCP
• Communication Missing information or
  misinterpretation within cockpit
• example) Miscommunication by crew with ATC
• Violation - Intentional non-compliance with
  required procedure
• example) Performing a checklist from memory
• Decision Discretionary choice of action that
  unnecessarily increases risk
• example) Unnecessary navigation through adverse
  weather
• Proficiency lack of skill or knowledge needed
  for activity
• example) Inability to program FMS properly

33
Specific Error CategoriesLOSA Handbook - ICAO
Doc 9803

• Aircraft handling
• Flt path deviations, speed, Wx penetration
• A/C systems and radio error
• Automation use
• Checklist
• SOP cross-verification
• Other procedural deviations

• Documentation
• ATIS, Jeppesens, clearances, etc
• External communication
• Crew communication
• Decision
• Violation

34
Latent Threats
35
Latent Threats

• Factors not directly linked to observable threat
  and error that increase risk and the probability
  of error
• Crew management of latent threats is difficult
  because they are not immediately visible
• Latent threat identification is key to accident
  and incident analysis

36
Latent Threat Examples

• Inadequate management oversight
• Inadequate regulatory oversight
• Flawed procedures
• Organizational culture and climate
• Scheduling and rostering practices
• Crew fatigue
• Performance assessment practices
• Inadequate accident and incident investigation

37
Threat and Error Management and Outcomes
38
Threat and Error Management

• In response to a threat or error crews can
• Trap - threat or error is detected and managed
  before it becomes consequential
• Exacerbate - threat or error is detected and the
  crews action makes it consequential
• Fail to Respond no action to deal with a threat
  or error (undetected or ignored)
• Outcome may be inconsequential or consequential

39
Threat and Error Outcomes

• Inconsequential the threat or error has no
  operational effect
• Additional error the response leads to an
  additional error (error chain)
• Undesired Aircraft State - deviation from normal
  flight that compromises safety

40
Accident Examples

• Air Ontario 1363
• Avianca 52

41
Air Ontario Flight 136310 March 1989
42
(No Transcript)
43
Dryden Scenario

• Air Ontario Flight 1363 took off from Winnipeg,
  Manitoba on March 10, 1989 It was a Fokker F-28
  recently bought from a Turkish airline. The
  aircraft had a number of mechanical problems
• Weather was deteriorating and the flight
  experienced delays due to de-icing at Winnipeg

44
Dryden Scenario - 2

• The flight went from Winnipeg to Dryden then on
  to Thunder Bay and back to Dryden
• It experienced additional delays at Dryden, then
  took off, stalled and crashed in woods off the
  end of the runway
• The post-crash fire was so fierce that the voice
  and data recorders were destroyed

45
Investigation

• Despite the relatively minor loss of life and
  identified cause, the largest investigation of an
  air crash to date was launched by a Commission of
  Inquiry, headed by a justice of the Supreme Court
• Robert Helmreich was human factors consultant to
  the commission
• A four volume report investigating all aspects of
  the Canadian aviation system resulted
• This is not to imply that this level of
  investigation is needed to apply the Threat and
  Error Management Model

46
External Threats

• Worsening weather increasing snowfall
• Possibly below minimums
• Increased passenger load
• Need to offload fuel
• Dispatch release with errors
• Small plane lost above airport delaying take off
• Inoperative APU no ground start ability
• CA and FO lack jet experience lt100 hrs
• Pax needs
• Misconnections
• No facilities at Dryden

47
Threat Management

• Inquiry about de-icing capabilities at Dryden

48
Crew Errors

• Fail to inspect wings for ice accumulation
• Fail to de-ice after inquiry about capabilities
• Taxi out in increasingly heavy snowfall
• Fail to re-assess situation after delay for lost
  small plane to land
• Did not verify windshear report

49
Threat and Error SummaryDryden
Threats Mgt Errors
Mgt
50
Undesired Aircraft States

• Initial lift-off and bounce back
• Second lift-off
• Un-recovered stall after second lift-off
• UAS management unsuccessful
• Outcome - accident

51
Latent Threats

• Regulator
• Certification of design that allows cold soaking
  of fuel (and icing of wing)
• Failure to define de-icing requirements
• Failure to audit jet program at Air Ontario
• Organization
• No company manuals or MEL
• Inconsistent training of pilots on F-28 aircraft
• At different airlines with different procedures

52
Latent Threats - 2

• Organizational pressure to fly with ice
• Chief pilot known as The Iceman
• Rostering practices both pilots inexperienced
  in F-28
• Crew
• Lack of experience in type
• Personal pressure to complete flight
• Captain scheduled to marry on arrival

53
Conclusions

• The root cause was the failure of the crew to
  de-ice before take off in heavy snow.
• A number of threats and errors contributed to
  this decision and action

54
Archival Research

• Use of existing records for analysis and
  hypothesis testing
• Can probe important issues
• analysis of crew behavior in accidents from voice
  and data recorder data
• Limited by what is available

55
Which method to use?

• If the question is important, use multiple
  methods and seek convergence
• Be a cautious consumer of research
• Caveat emptor

56

• The University of Texas
• Human Factors Research Project

www.psy.utexas.edu/HumanFactors