Вы находитесь на странице: 1из 44

www.monash.edu.

au/muarc

Accident Causation and Analysis:


Human Factors Theory and Methods

Dr Paul Salmon
Ms Amy Williamson
Ms Eve Mitsopoulos-Rubens
Dr Christina Rudin-Brown
Dr Michael Lenne

Human Factors Group,


Monash University Accident Research Centre
1
www.monash.edu.au/muarc

Content

Introduction to Human Factors


Accidents and accident causation models
Accident analysis methods
Case study analyses
Future directions
www.monash.edu.au/muarc

What is Human Factors?


www.monash.edu.au/muarc

Human Factors
the scientific study of the relationship between man and
his working environment (Murell, 1965)

the study of how humans accomplish work-related tasks


in the context of human-machine systems (Meister, 1989)

applied information about human behaviour, abilities,


limitations and other characteristics to the design of tools,
machines, tasks, jobs and environments (Sanders &
McCormick, 1993)

Studying and enhancing the performance of sociotechnical


systems
www.monash.edu.au/muarc

Accidents and accident causation


www.monash.edu.au/muarc

Accidents

Latin verb accidere = to happen


ad + cadere = to fall
a short, sudden, and unexpected event or
occurrence that results in an unwanted or
indesirable outcome (Hollnagel, 2004)
Event or occurrence directly or indirectly the
result of human activity
www.monash.edu.au/muarc

Human error

a generic term to encompass all those


occasions in which a planned sequence of
mental or physical activities fails to achieve its
intended outcome, and when these failures
cannot be attributed to the intervention of some
chance agency (Reason, 1990, p. 9).
www.monash.edu.au/muarc

Theoretical perspectives

Person approach (or old view)


- Errors at the sharp end
- Errors result from psychological/physical factors within
individuals
- Individual focussed strategies and countermeasures
Systems approach (or new view)
- Error as a systems failure
- Human error as a consequence of latent conditions residing
throughout the system
- Systems-based strategies and countermeasures
www.monash.edu.au/muarc

Systems versus person perspectives


www.monash.edu.au/muarc

System and cultural issues

Unsafe acts (Errors and Violations )

Equipment failures (Hardware Software )

1955 2005

1960 s 1970 s 1980 s 1990 s 2000 s


Aberfan Flixborough Chernobyl Paddington Linate
Ibrox Seveso Zeebrugge Long Island Uberlingen
Tenerife Bhopal Alabama Columbia
TMI Piper Alpha Estonia
Mt Erebus
www.monash.edu.au/muarc

Complex systems cannot be understood by


studying parts in isolation. The very essence
of the system lies in the interactions between
parts and the overall behaviour that emerges
from the interactions. The system must be
analysed as a whole (Ottino, 2003)
www.monash.edu.au/muarc

Swiss Cheese model of human error (Reason, 1990)

Reason, J. (1990). Human Error. New York, Cambridge University Press.


www.monash.edu.au/muarc

Accidents happen because (Reason, 2008)

Universals ever present hazards, create

Conditions latent factors that collectively


produce defensive weaknesses that

Causes permit the chance conjunctions of


local triggers and active failures to breach all
the barriers and safeguards
www.monash.edu.au/muarc

Rasmussens risk management framework


Changing political climate
Public opinion Government
and public awareness

Law s Regulators ,
Associations

R egulations
Changing market
Company conditions and financial
pressure

C ompany
Policy
Management

Changing competency
levels and education

Plans
Staff

Fast pace of
Action
Work technological change

Hazardous process
www.monash.edu.au/muarc

Herald of Free Enterprise Zeebrugge disaster

March 6th 1987


Ferry capsized
150 passengers & 38
crew killed
Ferry set sail with
inner bow doors
open
www.monash.edu.au/muarc

Herald of Free Enterprise Zeebrugge disaster

Assistant bosuns
failure to shut
bow doors
Fatigue
Captains leaves
Negative port with bow
Choppy
reporting culture Doors open
sea

Inherent unsafe Poor rostering Pressure to


top heavy ferry design depart early

Not my job
Culture
Failure to install bow
Door indicator
www.monash.edu.au/muarc

Systems approach predictions (Rasmussen, 1997)

Safety is an emergent property impacted by decisions of all actors, not


just front line workers alone
Threats to safety are caused by multiple contributing factors, not just a
single catastrophic decision or action
Threats to safety can result from a lack of vertical integration across
levels of a complex sociotechnical system, not just from deficiencies at
one level alone
Lack of vertical integration is caused, in part, by lack of feedback
across levels of a complex sociotechnical system
Work practices are not static, the migrate over time and under the
influence of financial and psychological pressures
Migration occurs at multiple levels of complex sociotechnical systems
Migration of work practices cause system defences to degrade and
erode gradually over time, not all at once. Accidents are caused by a
combination of this migration and a triggering event(s)
www.monash.edu.au/muarc

Other complex safety critical domains?

Aviation Safety Reporting System


NTSB accident database
Theoretically underpinned, systems-based
accident analysis e.g. HFACS
Systems-based countermeasures
Boeing Safety Management System
www.monash.edu.au/muarc

Figure showing the paths between the categories across the four HFACS levels ; Bold lines denote associations that
are highly statistically significant (p<.005), while dashed lines represent associations at the conventional level of
significance (p<.05). All associations represent positive correlations , aside from that between physical environment
and violations , which is negative
www.monash.edu.au/muarc

Exploratory case study analyses


www.monash.edu.au/muarc

Aim

To ascertain whether systems approaches


are applicable in the led outdoor activity
domain
To determine which of three approaches is
the most suitable for analysing, and learning
from, led outdoor activity accidents and
incidents
www.monash.edu.au/muarc

HFACS (Shappell & Wiegmann, 2003)

Aviation accident investigation approach


Based on Reasons Swiss Cheese model
Applications in a range of domains
- General aviation
- Maritime
- Mining
- Rail
- Road transport
- Construction
www.monash.edu.au/muarc

UNSAFE SUPERVISION

Unsafe
supervision
PRE -CONDITIONS FOR UNSAFE ACTS

Inadequate Planned F aiel d to Supervisory PRECONDITIONS


supervisoi n ni appropriate correct a vioal tions FOR UNSAFE
operations problem ACTS

ORGANISATIONAL INFLUENCES Environmental Condition of Personnel


factors operators factors

Organisational
influences
Physical T echnological Crew Personnel
environm ent environm ent resource readiness
m anagem ent
Resource Organisational Organisational
m anagem ent clm
i ate process
Adverse Adverse Physical &
m ental physiool gical m ental
states states m
il tiations

UNSAFE ACTS

UNSAFE
ACTS

Errors Violations

Skil - based Decision Perceptual


Routine Exceptional
errors errors errors
www.monash.edu.au/muarc
www.monash.edu.au/muarc
www.monash.edu.au/muarc

AcciMaps (Rasmussen, 1997)


Graphically represent accident causation trajectories
Six organisational levels
- Govt policy and budgeting;
- Regulatory bodies and associations;
- Local area Govt planning & budgeting;
- Technical operation and management;
- Physical processes and actor activities;
- Equipment and surroundings.
Applied to range of accidents including gas plant explosions
(Hopkins, 2000), loss of space vehicles (Johnson & de Almeida,
2008), aviation accidents (RAAF, 2001), public health (Vicente &
Christoffersen; 2006), and
Applicable to catastrophic road
failures anddomain
in any rail accidents (Svendung &
Rasmussen, 2002; Hopkins, 2005).
www.monash.edu.au/muarc

Example output

Jenkins, D. P., Salmon, P. M., Stanton, N. A., & Walker, G. H. (In Press). A systemic approach to accident analysis: a
case study of the Stockwell shooting. Ergonomics
www.monash.edu.au/muarc

Root cause model (Davidson, 2007)


www.monash.edu.au/muarc

Case study accidents

Lyme Bay Kayaking tragedy

Ripswing incident

Entrapment and near drowning incident


www.monash.edu.au/muarc

Case study accidents

Lyme Bay Kayaking tragedy

Ripswing incident

Entrapment and near drowning incident


www.monash.edu.au/muarc

The Lyme Bay Kayaking Tragedy

Introductory open sea kayak activity


Initial capsizes
Group becomes separated and swept out to
sea
High wind and wave conditions lead to further
capsizes and abandoning of kayaks
Delayed rescue
4 students drowned
www.monash.edu.au/muarc

Lyme Bay Root cause model analysis

Instructor root causes


Mismatch of skills/Experience
Equipment traps
Environment traps
Poor judgement
Inaccurate assessment of risk
Managerial root causes
Poor safety management systems
www.monash.edu.au/muarc

Lyme Bay Accimap


Go ve rnm en t p ol icy a nd
No gov ernment
legislat ion to Lyme Bay incident Accimap
control ac tiv it y
b ud ge tin g centres

Abs ence of an Ac creditati on body


outdoor centre failed to ins pec t
Re gu la tory b od ie s & regulat or or s ea act ivi ties at
lic ensing body centre
ass oci ati on s

Company I nadequate Com pany Inadequate Fail ure t o s upervis e F ai lure of Devon S outhway s taff LE A check lis t not
management s afety employed emergenc y cent re manager to Count y Council to failure to follow s taff us ed by Southway
Lo ca l ar ea Go ve rnm en t procedures ins truct ors without
ignored lett er proc edures ens ure canoeing ins pec t cent re handbook f or in planning of t rip
p la nn in g & b ud ge tin g, regarding poor rec om mended was being s afely rec reat ional
C omp a ny ma na g eme n t s afety procedures BCU qualific ations t aught by centre ac tiv iti es

I ns truct ors Ex erc is e


Fai lure to Required rat io of Coas tguard given
inadequately not s ui table as certain c anoe competent P oor
wrong inf ormat ion
Centre manager quali fied/ f or nov ic es inst ruc tors to communic ation
experi enc e about miss ing
Te chn ica l & o pe ra ti on al ignored oral ex perienc ed for
st udents not met s tudents
warnings ex erc is e
ma na ge me nt regarding poor S tudent s given
i nadequat e No advic e giv en S tudent s Poor Coas tguard
s afety procedures ins truct ed Ov ers ight by
Failure t o t raining e. g. no on group unders tanding of not i nf orm ed
not to inflate coas tguard
c heck weat her caps ize drill organis ation of fs hore winds & until 3.07pm
forecast lif ejack ets affec t on canoes

Teacher Delayed,
given c anoe Coc kpit s Abandonment
Teacher Inabilit y to right Groups s wept I nadequate
unsuitable fi lled with of c anoes searc h and
c apsiz es c aps ized out t o s ea
f or nov ices water
Phy sica l p roc esse s & a cto r c ont inual ly k ayak resc ue operation
acti vitie s
Students /t eac hers Los s of contac t Capsiz e of Fai lure to
Raft ing
inexperienc e of between fl ooded init iat e in-wat er Drownings
S tudent t oget her of
operating c anoes groups c anoes procedures
c aps iz ed group

Failure t o supply appropri at e equipment f or act ivit y e.g. Failure t o


- Failure t o prov ide radi os ; inflat e
- Failure t o prov ide s pray deck s; lifejac kets
E qu ip men t & surr ou nd in gs
- Failure t o prov ide dis tres s fl ares;
- Failure t o prov ide c anoes with s uppl em ent ary buoyanc y;
- Failure t o prov ide towlines;
- Failure t o prov ide adequate clot hi ng for s tudent s;
- Failure t o prov ide s urv ival bag;
- Failure t o prov ide whist les for every st udent ; A dverse
- Failure t o prov ide c ol oured helmets ; c ondi tions
- Failure t o prov ide c anoes fit ted wit h deck lines;
- Failure t o prov ide esc ort boat.
www.monash.edu.au/muarc

Lyme Bay HFACS output


Procedures Oversight
Resource Resource Organisational Culture
Proc edures / Es tablis hed s afety /
Organisational influences m anagem ent m anagem ent Clim ate Values , beliefs and
ins truc tions about ris k management
Staffing Training Communic ation attitudes
proc edures programs

Inadequate Inadequate Inadequate


Planned inappropriate Supervisory violations
supervision supervision supervision
Unsafesupervision Failed to prov ide Failed to prov ide Failed to prov ide
operations Authoris ed unqualified
Inadequate c rew pairing c rew for ac tiv ity
proper training adequate proc edures profes s ional guidanc e

Physical /Mental Technological


Physical Physical /Mental CrewResource
Lim itations Personal Readiness Environment
Preconditions for unsafe environm ent Lim itations Management
Inadequate ex perienc e for Inadequate training Lack of appropriate
acts Weather Lack of aptitude Poor communication
c omplex ity of s ituation equipment

Decision error Routine violation


Skill based error Decision error Skill -based error
Inadequate Students giv en
Unsafeacts or operations Inability to right
k nowledge of
Wrong res pons e to
Inadequate training
Failed to c hec k
c aps iz ed k ay ak emergenc y weather
s y s tems , proc edures for ac tiv ity

Group s wept out to


s ea , c aps iz e and
drown
www.monash.edu.au/muarc

Systems-based accident analysis?


www.monash.edu.au/muarc

Validation of Rasmussens framework


Safety is an emergent property impacted by decisions of all actors, not
just front line workers alone
Threats to safety are caused by multiple contributing factors, not just a
single catastrophic decision or action
Threats to safety can result from a lack of vertical integration across
levels of a complex sociotechnical system, not just from deficiencies at
one level alone
Lack of vertical integration is caused, in part, by lack of feedback
across levels of a complex sociotechnical system
Work practices are not static, the migrate over time and under the
influence of financial and psychological pressures
Migration occurs at multiple levels of complex sociotechnical systems
Migration of work practices cause system defences to degrade and
erode gradually over time, not all at once. Accidents are caused by a
combination of this migration and a triggering event(s)
www.monash.edu.au/muarc

Lyme Bay Accimap


Go ve rnm en t p ol icy a nd
No gov ernment
legislat ion to Lyme Bay incident Accimap
control ac tiv it y
b ud ge tin g centres

Abs ence of an Ac creditati on body


outdoor centre failed to ins pec t
Re gu la tory b od ie s & regulat or or s ea act ivi ties at
lic ensing body centre
ass oci ati on s

Company I nadequate Com pany Inadequate Fail ure t o s upervis e F ai lure of Devon S outhway s taff LE A check lis t not
management s afety employed emergenc y cent re manager to Count y Council to failure to follow s taff us ed by Southway
Lo ca l ar ea Go ve rnm en t procedures ins truct ors without
ignored lett er proc edures ens ure canoeing ins pec t cent re handbook f or in planning of t rip
p la nn in g & b ud ge tin g, regarding poor rec om mended was being s afely rec reat ional
C omp a ny ma na g eme n t s afety procedures BCU qualific ations t aught by centre ac tiv iti es

I ns truct ors Ex erc is e


Fai lure to Required rat io of Coas tguard given
inadequately not s ui table as certain c anoe competent P oor
wrong inf ormat ion
Centre manager quali fied/ f or nov ic es inst ruc tors to communic ation
experi enc e about miss ing
Te chn ica l & o pe ra ti on al ignored oral ex perienc ed for
st udents not met s tudents
warnings ex erc is e
ma na ge me nt regarding poor S tudent s given
i nadequat e No advic e giv en S tudent s Poor Coas tguard
s afety procedures ins truct ed Ov ers ight by
Failure t o t raining e. g. no on group unders tanding of not i nf orm ed
not to inflate coas tguard
c heck weat her caps ize drill organis ation of fs hore winds & until 3.07pm
forecast lif ejack ets affec t on canoes

Teacher Delayed,
given c anoe Coc kpit s Abandonment
Teacher Inabilit y to right Groups s wept I nadequate
unsuitable fi lled with of c anoes searc h and
c apsiz es c aps ized out t o s ea
f or nov ices water
Phy sica l p roc esse s & a cto r c ont inual ly k ayak resc ue operation
acti vitie s
Students /t eac hers Los s of contac t Capsiz e of Fai lure to
Raft ing
inexperienc e of between fl ooded init iat e in-wat er Drownings
S tudent t oget her of
operating c anoes groups c anoes procedures
c aps iz ed group

Failure t o supply appropri at e equipment f or act ivit y e.g. Failure t o


- Failure t o prov ide radi os ; inflat e
- Failure t o prov ide s pray deck s; lifejac kets
E qu ip men t & surr ou nd in gs
- Failure t o prov ide dis tres s fl ares;
- Failure t o prov ide c anoes with s uppl em ent ary buoyanc y;
- Failure t o prov ide towlines;
- Failure t o prov ide adequate clot hi ng for s tudent s;
- Failure t o prov ide s urv ival bag;
- Failure t o prov ide whist les for every st udent ; A dverse
- Failure t o prov ide c ol oured helmets ; c ondi tions
- Failure t o prov ide c anoes fit ted wit h deck lines;
- Failure t o prov ide esc ort boat.
www.monash.edu.au/muarc

Generic failure types


Government policy and LACK OF
budgeting LEGISLATION

ACTIVITY/DEVICE FAILURE TO
LACK OF FAILURE TO IDENTIFY UNSAFE VOLUNTARY
Regulatory bodies & OVERLOOKED
associations REGULATORY BODY INSPECT CENTRE PRACTICE SCHEME ONLY
IN STANDARDS

FAILURE TO INADEQUATE
CULTURE
Local area Government ASSESS RISKS SUPERVISION
planning & budgeting , INADEQUATE INAPPROPRIATE FAILURE TO COMPLY
Company management
PROCEDURES RECRUITMENT WITH LEGISLATION FAILURE TO RESPOND
FAILURE TO RESPOND
TO REGULATORY BODY
TO SAFETY CONCERNS
REQUESTS
INADEQUATELY INAPPROPRIATE FAILURE TO POOR SITUATION INADEQUATE POOR
Technical & operational QUALIFIED ACTIVIITY CHECK WEATHER AWARENESS TRAINING COMMS
management

FAILURE TO COMPLY
INEXPERIENCE VIOLATIONS
WITH LAW, PROCEDURES

FAILURE TO
SKILL-BASED DECISION VIOLATIONS WRONG POOR ACTION MISTAKE
Physical processes & actor FOLLOW
activities ERRORS ERRORS ACTION COMMS TOO LATE
PROCEDURES
LAPSE INEXPERIENCE

UNSAFE EQUIPMENT
INAPPROPRIATE DESIGN FAILURE
Equipment & surroundings
INAPPROPRIATE EQUIPMENT
USE OF NO FAIL SAFE
EQUIPMENT NOT PROVIDED
EQUIPMENT UNSAFE ENVIRONMENTAL
EQUIPMENT CONDITIONS
www.monash.edu.au/muarc

The way forward?


www.monash.edu.au/muarc

Data challenges

In complex systems, event analyses are


constrained by the quality of the data
gathered, the maturity of the associated
reporting system, and the training and
background of the investigator and reporter.
Such constraints place limits on the
adequacy and strength of analyses
conducted with the data.
(Grabowski et al, 2009)
www.monash.edu.au/muarc

Data challenges

In complex systems, event analyses are


constrained by the quality of the data
gathered, the maturity of the associated
reporting system, and the training and
background of the investigator and reporter.
Such constraints place limits on the
adequacy and strength of analyses
conducted with the data.
(Grabowski et al, 2009)
www.monash.edu.au/muarc

Data and analysis challenges

Standardised incident reporting system


Standardised accident and incident database
Systems-based accident analysis method
Investigator training in HF/systems-based models
Outdoor activity contributory factor taxonomies
Instructor/student error taxonomies
In-depth analysis

Feasibility study?
www.monash.edu.au/muarc

Benefits

Study of accidents widely accepted as a way


of acquiring knowledge that can improve
safety (e.g. Cassano-Piche et al, 2009)
Development and application of standardised,
universally accepted, theoretically
underpinned data collection, storage and
analysis approaches will lead to an increase
in knowledge, the appropriate application of
which is likely to lead to a reduction in safety
compromising accidents and incidents
www.monash.edu.au/muarc

Questions?

paul.salmon@muarc.monash.edu.au