016262gr_acc_investigation_pareto_diagram_asse.pdf

8/14/2019 016262gr_acc_investigation_pareto_diagram_asse.pdf

http://slidepdf.com/reader/full/016262graccinvestigationparetodiagramassepdf 1/6

ManagementManagement

I

Incident InvestigationA Problem-Solving

ProcessIdentifying and correcting root causesBy Jerome E. Spear

www.asse.org APRIL 2002 PROFESSIONAL SAFETY 25

INCIDENTS ARE EVENTS that have the potentialto cause injury or damage; they are caused by unsafeacts and/or unsafe conditions. However, unsafe actsand conditions are often symptomatic of an overallorganizational problem. Companies can significant-ly reduce injuries and losses by implementing a con-tinuous incident investigation process that involvesthe reporting of incidents without fear of being

blamed. Such a process uses structured problem-

solving techniques to identify and track root causesand follow up on corrective actions.

Heinrich’s study in 1931 concluded that for everyserious injury, about 29 minor injuries occur with300 other occurrences that produce no injury(Heinrich 90). In another study, Bird and Germainshowed a 500:100:1 relationship among property-damage accidents, minor-injury accidents and dis-abling injury accidents (Bird and Germain 21).Fletcher reported a ratio of 175:19:1 for no-injuryaccidents, minor-injury accidents and serious injuryaccidents (Fletcher 37). Although the precise ratios of these studies differ to some degree, one point isclear—a large proportion of accidents (perhaps

greater than 80 percent) are neglected if safety pro-fessionals respond only to injury-producing events.Therefore, since the only difference between an inci-dent resulting in no injury and a catastrophe mayhave been by chance, every problem should beinvestigated and resolved (Goldberg 33).

Within the context of this article, the author definesan “incident” as a deviation from an acceptable stan-dard or work practice (i.e., an organizational “prob-lem”). The goal of a proactive organization should beto investigate such problems to determine their rootcauses. However, in the author’s opinion, two barri-ers often keep organizations from doing so:

•Employees do not like to report incidents tomanagement, perhaps due to fear of being blamed.

•Even organizations with an advanced safetyculture, where employees openly report incidents,typically do not have enough resources to thorough-ly investigate all incidents to their root causes.

The first barrier may be addressed via trainingand ongoing communication with employees at alllevels acknowledging the importance of reportingand investigating incidents. With few exceptions,management should not blame an employee for

being injured or being involved in an incident. Suchactions are counterproductive and can causeemployees to be fearful of reporting future incidents.Periodic audits and inspections are alternatives toidentify potential problems.

When the safety culture supports incident report-ing without fear of blame, the organization is headedin the right direction. The primary problem, then, isthe limited resources available to investigate theseincidents. To address this situation, incidents should

be evaluated and ranked by risk; those incidents asso-ciated with the highest risk should receive priority.

Setting Priorities

Incident investigation is a problem-solvingprocess. For most technicalprofessionals, problem solv-ing is a primary job function.Often, the most-difficult stepin this process is identifyingthe right problem to solve(Adams 22).

Statistical process controlinvolves the application of sta-tistical techniques to deter-mine whether observedvariations within the safetyprocess are due to “commoncauses” or “assignable caus-

Jerome E. Spear, CSP, CIH, is a

senior consultant for ECS Risk Control Inc.

A professional with more than 11 years’

experience in the safety and health field, he

helps clients evaluate and manage safety- and

health-related issues, including incident

investigations, industrial hygiene, OSHA

compliance, indoor air quality and training.

An OSHA-authorized construction outreach

instructor, Spear holds a B.S. in Industrial

Engineering from Texas A&M University.

He is a member of ASSE’s Gulf Coast Chapter

and a member of AIHA.



26 PROFESSIONAL SAFETY APRIL 2002 www.asse.org

According to Adams, problem solving is notunstructured thinking (22). However, at times, anunstructured approach may be the most-feasibleoption, such as when investigating incidents thathave a low severity potential and a low frequencyrating. If the incident has a rating greater than 0.20on the prioritization matrix (using the severity andfrequency ratings in Table 1), structured problem-solving techniques such as change analysis, hazard-

barrier-target analysis, events and causal factorscharting, tree diagrams and cause-and-effect dia-grams should be used.

Change Analysis

The premise of change analysis is consistent withthe words of the Greek philosopher Heraclitus,“There is nothing permanent except change.” Sincechange occurs constantly, the likelihood of an inci-dent increases if change is not properly anticipatedand/or managed. Through change analysis theinvestigator can identify the cause(s) of a mishap byexploring differences between the problem situationand a problem-free situation.

Change occurs in three ways: 1) mandated; 2) nat-urally (whether noticed or not); and 3) caused—knowingly or not (Wilson, et al 114). For mandatedchange, one must look at how the change was imple-mented. For example:

•What was the magnitude of the proposedchange?•Was the implementation complete?•Were reinforcing factors considered and includ-

ed as part of the implementation?If change occurred naturally or was caused, the

difference between an incident and a similar prob-lem-free situation may not be readily obvious.

Change analysis may be performed in either areactive mode (e.g., incident investigation) or aproactive mode (e.g., identifying potential effects ofchanges before implementing a new procedure). Thetypical analysis includes the following steps:

1) Define the problem.2) Identify a “problem-free” situation.

es.” Common cause variations are inherent to aprocess when it is operating as designed. Assignablecause variations are unnatural variations within a

process (Sower 53). Regardless of whether the safety process is “out-of-con-trol” or “in statistical process control,”many incidents (and problems) likelyoccur each day. As with other businessscenarios, the time spent investigatingincidents must be prioritized.

Early in the incident investigationprocess, a useful tool is a prioritizationmatrix. This matrix allows comparison of

both quantitative and qualitative data

within the same analysis (Sower 44). It is arisk management tool, since it evaluates

both severity potential and frequency of an incident or hazard. Risk is defined asthe product of the frequency and severityof potential losses (Brauer 527). One wayto prioritize incidents by risk is to catego-rize each incident with the appropriateseverity and frequency weightings. Table1 presents weightings (developed by theauthor) that may be used when develop-ing a prioritization matrix.

Root-Cause Analysis

The identification of root causes in anysystem is fundamental to problem solv-ing and continuous improvement(Senecal and Burke 63). These causes can

be identified using 1) unstructured prob-lem-solving techniques, which include intuition,networking (lessons learned) and experience; and 2)structured techniques, which include the systematictools used in root-cause analysis. Through this sys-tematic process, safety professionals can:

•define the problem;•gather and prioritize related data;•analyze solutions;•evaluate the benefits and cost-effectiveness of

available prevention options (Handley 75).

Prioritization Matrix

Severity Potential Severity Index Frequency Frequency Index

Table 1Table 1

Catastrophic: Hazard may cause deathor loss of facility.

Critical: May cause severe injury, severeillness or major property damage.

Marginal: May cause minor injury,minor illness or minor property damage.

Negligible: Probably would not causeinjury or minor illness.

0.9-1.0

0.6-0.8

0.3-0.5

0-0.2

Likely to occurimmediately or in ashort period of time.

Probably will occurin time.

May occur in time.

Unlikely to occur.

0.9-1.0

0.6-0.8

0.3-0.5

0-0.2

Proactive

companies

continue to

emphasize,

througheducation and

actions, that

every incident (or

organizational

problem) must

be reported,

regardless of the outcome.




Events & Causal Factors ChartingAccording to Wilson, et al, events and causal fac-

tors (E&CF) charting is the best root-cause analysistool to employ when investigating an injury or otherdamaging consequence—especially when one must

document the chain of events (42). An E&CF chart(Figure 1) is essentially a flowchart that depicts thesequence of events leading up to the event causingthe injury or property damage and assesses emer-gency response activities. For each event, relevantcausal factors or conditions are identified and notedon the chart. The goal is to uncover the root causesfor the preceding events in order to break the chainof events leading to a negative outcome.

Tree DiagramsA tree diagram is a graphical display of the con-

tributing factors of each event. Common techniquesinclude fault tree analysis (FTA) and the management

oversight risk tree (MORT). Both techniques use stan-dardized symbols to construct the tree diagram.Figure 2 presents a modified approach that cate-

gorizes root causes by management system. The con-cept behind this model is that incidents are eventswhich have the potential to cause injury or damageand are caused by unsafe acts and/or conditions; asnoted, such acts/conditions are merely symptoms of overall organizational problems. Thus, the top event(usually an injury) is actually caused by failures with-in the management system (Stalnaker 37).

Therefore, employees should not be blamed for being injured. Rather than punish workers, manage-ment should strive to examine these failures and cor-rect the underlying faulty management system—a

3) Describe the conditions and compare.4) Determine the direction/amount of the changes.5) Examine the changes for their effect on the

problem.6) Separate the causes from the problems (Ferry

163).

Hazard-Barrier-Target Analysis (Hazard-BTA)A hazard-BTA considers potential hazards, exam-

ines how damage occurs, and assesses the adequacyof installed barriers or other safeguards that shouldeither prevent or mitigate an incident. Like changeanalysis, it may be used as a reactive tool or proac-tive tool.

For a damage-causing incident to occur, theremust be a target, a hazard and a less-than-adequate

barrier. By identifying potential hazards along withpossible targets, barriers can be analyzed. A target issomething of value; it may be a person, part, system,

procedure or process. A barrier separates the targetfrom the hazard and may be physical (such as per-sonal protective equipment, machine guards or safe-ty valves), administrative (procedures or directives)or personal (employee training and supervision)(Wilson, et al 133).

For example, suppose two employees were open-ing a process line at a flange joint in order to unclogthe line. As the line was opened, the clog shifted,causing the product inside the line to splash theworkers. As a result, each suffered multiple second-and third-degree burns. In this case, a hazard-BTA(Table 2) proved to be an efficient assessment toolthat provided valuable information in a relativelyshort period of time.

Hazard-BTA (Line Breaking Incident)

Hazard Target Barrier Analysis

Table 2Table 2

•Thermal temperatures•Corrosive chemical•High pressure

•Contamination

•Employees•Other personnel

•Soil/groundwater

•Piping, valves,flanges, gaskets•Line breakingprocedure•Personal protectiveequipment (e.g., face-shield, protectiveclothing)•PPE written hazardassessment•Training

•Piping, valves,flanges, gaskets•Impervious materi-al (or drip pan)•Procedure•Training

•Process unit periodically becomesclogged and requires employees toopen pipelines to unclog unit.•No formal line breaking procedurehas been developed.•Employees did not wear thermalprotective clothing.•PPE hazard assessment did notaddress this task.•Line breaking training has not beenprovided to employees.

•Process unit periodically becomesclogged and requires employees toopen pipelines to unclog unit.•No formal line breaking procedurehas been developed that includespollution prevention procedures.•Environmental awareness trainingis not currently required.




Lost-time due to disc

herniation from pushing

a commercial dumpster

Machines

PeopleMaterials

Methods

Dumpster LTA

Weight? Size?

Defective wheels?

Selection of worker LTA

No job placement physicals

Supervision LTA

Monitoring/feedback LTA

No formal process

Training LTA

No back injury prevention training

Communication with physicians LTA

Return-to-work program

not communicated

Improper “pushing” technique?Placement of dumpster LTA

Driver required to push

dumpster to truck

Placed where susceptible

to muddy conditions

LTA = Less Than Adequate

Figure 3Figure 3

Cause-and-Effect Diagram

Negative Result(Injury/Property Damage)

Release of

Energy

Unsafe

Act

Unsafe

Condition

EnvironmentMaterialTrainingPersonTask

In Search of the Root Causes

Incident

Management System

Negative Result(Injury/Property Damage)

Release of

Energy

Unsafe

Act

Unsafe

Condition

EnvironmentMaterialTrainingPersonTask

In Search of the Root Causes

Incident

Management System

Figure 2Figure 2

Tree Diagram Model

Using structured

problem-solving

techniques, safety professionals can

define the problem,

gather and prioritize

related data, analyze

solutions, and evaluate

the benefits and

cost-effectiveness

of available

prevention options.

016262gr_acc_investigation_pareto_diagram_asse.pdf

Documents

Transcript of 016262gr_acc_investigation_pareto_diagram_asse.pdf