Root Cause Analysis for Effective Incident Investigation Christy Wolter, CIH Principal Consultant Environmental and Occupational Risk Management (EORM.

Root Cause Analysis Root Cause Analysis for Effective Incident Investigationfor Effective Incident Investigation

Christy Wolter, CIH

Principal Consultant

Environmental and Occupational Risk Management (EORM®)

April 7, 2005

© 2005 EORM, Inc. 2

Outline

Introduction What is Root Cause Analysis (RCA)? How does RCA work? Tips to make your RCA more effective Interviewing techniques

© 2005 EORM, Inc. 3

What is Root Cause Analysis (RCA)? One of several tools suitable for after

the fact investigations Most straightforward method

sufficiently structured to identify, and determine relationships between, various events and issues that may have combined to produce the incident

© 2005 EORM, Inc. 4

How Does RCA Work? Employee fell down Employee was careless Employee under time

pressure Under time pressure

because of overlapping delivery dates

Delivery dates overlap because of poor communication between teams

Poor communication exists because…

Keep going further by asking “why?”

Symptoms vs. Roots

© 2005 EORM, Inc. 5

The Root

The root cause is typically not simply machine failure

The root cause is more typically:– Machine failure due to improper

maintenance, contributed to by both difficulty of maintenance access and unclear procedures, each exacerbated by lack of procedure review because no management of change process….(can we go further?)

© 2005 EORM, Inc. 6

Has a Root Cause Been Identified? Thermocouple probe reading high Wrong manual valve opened Pressure set point incorrect Object lifted was too heavy Procedural step performed out of

order

© 2005 EORM, Inc. 7

How Does RCA Work?

1. Start with a descriptive statement of the incident/near miss

2. Determine what conditions, events, and/or factors might have caused (alone) or contributed to (in combination with other conditions) the incident. These are your primary (1o) factors (i.e., Why?)

3. Determine conditions/events/factors that may have caused or contributed to the primary factors. These are your secondary (2o) factors (i.e., Why?)

© 2005 EORM, Inc. 8

Examples of RCA Documentation

Fishbone (cause and effect) diagram

Simplified logic diagram

© 2005 EORM, Inc. 9

Example Fishbone Diagram

Operator broke wrist on

fall from ladder

MachinesMethods

MaterialPeople

Operator did not heed ladder warning label

New valve access requires ladder

Written checklist did not include warnings re safe

ladder use

Ladder legs uneven

Operator not trainedSteps wet and slippery

Fishbone Diagram Example

Lack of Housekeeping

No hazard analysis of procedure

No discipline for previous safety violations

No Trng Mgt System

Lack of Maintenance

No Mgt of Change analysis

© 2005 EORM, Inc. 10

Example Root Cause Diagram

Low accountability for safety issues

5o

KEY:

Operator suffers chemical burn on hand

PPE did not prevent exposure

Chemical splashed when it was manually added to

tank

PPE not used PPE wrong for this

service

Procedure does not prevent

splashing

Procedure not followed correctly

Operator did not don available PPE

PPE hinders

work progress

No Management of Change process in place

PPE assessment poor/omitted

Change review didn't consider PPE

issues

PPE not available

Deviations not

considered important

Example Root Cause Diagram

"OR"

"AND"

Cause Factor/Contributor

"Root"

2o

3o

4o

"AND"

"OR"

No procedural HAZOP performed

Chemical changed since initial PPE

assessment Procedure unclear

No PPE inventory management system

PPE was not

considered important

PPE assessment poor/omitted

"OR"

© 2005 EORM, Inc. 11

Chronology of an Incident Investigation

1. Event occurs2. Collect information from the scene of

the event3. Gather more information (witnesses,

system information, etc.)4. Conduct detailed RCA5. Write an Action Plan6. Implement the Action Plan7. Review results8. Modify Action Plan as necessary

© 2005 EORM, Inc. 12

Tips For More Effective RCAs Factors concisely written yet

sufficiently descriptive– Will the logic be understandable to

persons not in the session, or to you a few years from now?

Speculation is clearly identified as such Actionable items are clearly defined Conduct analysis as soon as possible

after data have been gathered Disallow blame

© 2005 EORM, Inc. 13

Tips For More Effective RCAs Assemble a knowledgeable team Use the 80/20 Rule Tackle one branch at a time….

this helps keep team’s thoughts organized Use brainstorming techniques Don’t disrupt the brainstorm by trying to

perfect the flow/diagram!– Stay ½-step ahead of your team when

diagramming Prevent skipping levels or jumping to

conclusions

© 2005 EORM, Inc. 14

Focusing the Analysis

Consider the likelihood and magnitude of impact of each potential cause, and assess most deeply (i.e., spend the most time on) those most likely or that may contribute most impact.

Although the team may brainstorm 20+ potential causes, they vary in placement along the continuum…

Defies the laws Happens every day

of physics everywhere

© 2005 EORM, Inc. 15

Conducting RCA Interviews

Create a list of questions to ask in advance

Avoid conducting a Root Blame interview

Ask how injured employees are doing

© 2005 EORM, Inc. 16

Examples Of Questions To Ask Process Equipment Questions

– Were any operating parameters (e.g., temperature, pressure, flow rates) changed just prior to the incident (preceding minutes, hours, or days, depending on length of operation)?

– Were operating conditions leading up to the incident recorded (e.g., strip charts, process control system print outs, instrumentation )?

– Were any reactants changed just prior to the incident (e.g., new chemical used, change in chemical concentration, change in chemical vendor)?

Employee Interaction– Was the employee involved in the incident interacting

with the process equipment at the time (e.g., adjusting valves, performing a manual procedure, servicing, troubleshooting, calibrating)?

– Was the employee involved in the incident using support equipment at the time (e.g., ladder, extension cord, lift devices, portable pumps for maintenance)?

© 2005 EORM, Inc. 17

Examples Of Questions To Ask

Documentation– Do written procedures exist for the

operation/activity performed at the time of the incident?

– Do written maintenance procedures exist for the equipment involved in the incident?

– Was maintenance performed on the equipment involved in the incident?

– Did clearly-written procedures exist for all tasks required for this process/equipment?

– Do written procedures describe the potential consequences of deviations?

– Do written procedures describe the PPE required? Systems Review

– Was the appropriate PPE available and worn?– Have you received training on this process

and equipment?

© 2005 EORM, Inc. 18

Focusing the Analysis

Consider the likelihood and magnitude of impact of each potential cause, and assess most deeply (i.e., spend the most time on) those which are most likely or which may contribute most of the impact. Although the team may brainstorm 20+ potential causes, they vary in their placement along the continuum…

Defies the laws of physics Happens every day everywhere

© 2005 EORM, Inc. 19

Summary

Use Root Cause Analysis for actual or near miss incidents, to prevent recurrence

Maximize effectiveness by gathering the right data and following the approach outlined in this course

Keep the analysis and its documentation as straightforward as possible, to enhance the probability you will continue to use it in the future!