Journal of Loss Prevention in the Process Industries 64 (2020) 104027

Contents lists available at ScienceDirect

Journal of Loss Prevention in the Process Industries

journal homepage: [Link]

Analysis of laboratory incident database

Nirupama Gopalaswami *, Zhe Han
Mary Kay O’Connor Process Safety Center, Artie McFerrin Department of Chemical Engineering Department, Texas A&M University System, College Station, TX 77843-
3122, USA

A R T I C L E I N F O A B S T R A C T

Keywords: With the growth of academic institutions, the number of labs handling hazardous chemicals has increased.
Laboratory Although the chemicals used in academic institutions are well-known, the dosage, usage, and management of
Incidents these chemicals pose significant threat to researchers at all levels. As a step towards incident prevention, a
Database
laboratory incident database was developed to record incidents at universities and secondary schools. The data
Accidents
Data analytics
consisted of 128 entries occurring during the years 2012–2015. Incidents were classified by institution type,
hazard type, consequences, substance type, body parts injured and direct causes. Of 128 incidents, 65% of the
incidents were taking place in universities. Chemical spills were observed to be the most recurrent hazard type
contributing to 45% of the incidents following by explosions (23%) and fires (21%). The consequence which
most frequently occurred in incidents was personal injury and hazmat response contributing to 22% each. It was
observed that in 41% of the incidents, the body parts injured in the incident was not known or reported. Of the
total 128 incidents, about 50 incidents occurred due to improper storage and handling.

chemicals used in academic institutions are well-known, the dosage,

usage and management of these chemicals pose a significant threat to
1. Introduction researchers at all levels. Documentation of lab incidents can facilitate
the identification of direct causes of said events.
Laboratory researchers, governmental agencies, industries, and
universities have increasingly focused on laboratory safety in recent 2. Laboratory safety and safety climate
years (National Research Council, 2014). The importance of such efforts
is obvious considering lab personnel are exposed to hazards which can Laboratory safety and training is often relegated to those experts who
cause serious injury and death (National Research Council, 2014). study and work as Principal Investigators in a given field (e.g., chemists,
Laboratories in the U.S. are strictly regulated by Occupational Safety biologists) (National Research Council, 2014). A certain level of
and Health Administration (OSHA) standards and university- and knowledge is necessary to appropriately develop safety regulations.
lab-specific policies (OSHA, 2008). (see Table 1) These domain experts, however, are not the only individuals involved in
Negative outcomes ranging from minor incidents to fatalities laboratory safety.
continue to occur despite these efforts. In a recent survey, 46% of 2374 Psychologists have long been interested in safety and have effectively
laboratory personnel indicated that they sustained at least one injury applied psychological principles to explain workplace safety and
while working in a lab (Harper and Watt, 2012). Highly publicized in­ behavior (Jex et al., 2012). Often the underlying cause of a given inci­
cidents in recent years have spurred the reevaluation of university pol­ dent is psychological in nature (Jex et al., 2012). In a lab, for example, a
icies concerning laboratory safety (US Chemical Safety Board, 2010). chemical spill may occur because a student tries to complete a task too
Incidents are not solely relegated to university labs. As an example, four quickly.
students were recently hospitalized when a fire broke out in their lab at a According to psychological models and theories of workplace safety,
high school (Nicholson, 2014). Incidents such as these occur regularly the most proximal determinant of accidents and injuries is safety
and clearly highlight the need for improving accident prevention in labs. behavior (Christian et al., 2009). Researchers generally accept the
Additionally, with the growth of academic institutions, the number distinction between safety compliance and participation as aspects of
of labs handling hazardous chemicals has increased. Most laboratories safe behavior (Griffin and Neal, 2000). Safety compliance refers to
utilize some type of hazardous chemicals (Furr, 2000). Although the

* Corresponding author.
E-mail address: nirupama89@[Link] (N. Gopalaswami).

[Link]
Received 10 March 2019; Received in revised form 3 December 2019; Accepted 3 December 2019
Available online 19 December 2019
0950-4230/© 2019 Elsevier Ltd. All rights reserved.
N. Gopalaswami and Z. Han Journal of Loss Prevention in the Process Industries 64 (2020) 104027

Definitions Hazard An inherent chemical or physical characteristic that has the

potential for causing damage to people, property, or the
Accident An accident is defined as an unplanned event that results in environment
personal injury or property damage Hospitalization Person(s) involved in the incident were taken to the
Corrosive Containing Acid or Base. Acid is a substance having a pH hospital
of less than 7. E.g. Hydrochloric acid. A base is a substance Improper Storage and Handling Improper storage and handling can
that reacts with an acid in an acid-base reaction having a refer scenarios where hazardous substances are not stored,
pH of more than 7. E.g. Sodium Hydroxide. handled or disposed properly
Chemical Explosion A chemical or compound that causes a sudden, Incident An incident is defined as an unplanned event that does not
almost instantaneous release of pressure, gas, heat and result in personal injury but may result in property damage
light when subjected to sudden shock, pressure, high or is worthy of recording
temperature or applied potential Lab incident Actual and potential adverse events related to
Chemical Fire Chemical Fire is the rapid oxidation of a chemical in laboratory
the exothermic chemical process of combustion, releasing Laboratory A room or building in school or university equipped for
heat, light, and various reaction products scientific experiments, research, or teaching, or for the
Flammable and Combustible Flammable material has flash point manufacture of drugs or chemicals
below 100 � F (37.8 � C). Combustible material has a closed- Lack of PPE Person(s) involved do not wear appropriate personal
cup flash point at or above 100 deg. F (37.8 � C) protective equipment
Community Disruption Incident that involved disruption to the Lack of training Lack of training/poor training regarding the
community where people were evacuated, and the incident chemicals, equipment or other relevant information
site was temporarily closed. Evacuation involves the action required to conduct a lab experiment
of evacuating people from a place of incident. Temporary Personal Injury A personal injury is an abnormal condition or
site closure refers to the temporary shutdown of a single disorder. Injuries include cases such as, but not limited to,
site of employment, or one or more facilities or operating a cut, fracture, sprain, or amputation
units within a single site of employment Property Damage Harm or damage to the building/property in
Consequence The undesirable result of a loss event, usually which the incident occurred
measured in health and safety effects, environmental Procedure Violation A procedure violation occurs when a lab
impacts, loss of property, and business interruption costs member is aware of the procedures, but chooses not to
Equipment Failure Failure of the equipment due to shortage of fuel follow them
in burner, power shortage, breaking of equipment parts Reactive Chemical Having a tendency to react chemically
Explosion Explosion is a release of energy sufficient to cause a Runaway Reaction A reaction that is out of control because the rate
pressure wave of heat generation by an exothermic chemical reaction
Fires Fire is a chemical reaction in which a substance combines exceeds the rate of cooling available
with oxygen and heat is released. Spill/Leakage An instance of a liquid spilling or being spilled from
Hazmat/Firefighter Response Incidents which were responded by its storage containment
hazmat team or fire fighters

behavior in line with stated safety procedures, whereas safety partici­

Table 1 pation refers to proactive behavior at improving safety (Griffin and Neal,
Example of news article coding in lab incident database. 2000). A laboratory member may, for example, follow all stated safety
Database Categories Description
policies, but do little more to improve safety, whereas another member
may actively seek out ways to improve it.
Incident ID 32
Psychologists have also identified and studied a variety of person and
Date 18-Dec-2012
Year 2012 situational variables that contribute to safety behavior and outcomes
Link [Link] (Christian et al., 2009). Situational factors include safety climate and
-20773740 leadership (Christian et al., 2009). Person-related determinants of safety
Short Description A student has been taken to hospital after an explosion at a behavior include safety motivation and knowledge, personality char­
school laboratory. Emergency crews were called to reports of
an explosion. The 18-year-old pupil was taken to the Hospital
acteristics, and job attitudes. In the following sections, we focus on the
suffering from extensive burns to his hands and face. His safety climate of a laboratory while acknowledging that there are
injuries are not thought to be life changing and no-one else several predictors of safe behavior.
was hurt in the incident. The incident had involved the
student working on an individual chemistry experiment.
Type of Institution School 2.1. Safety climate
Chemical Involved Unknown
Substance Type Unknown Research on safety climate began with (D. Zohar, 1980). Safety
Number Injured/ 1
climate is a multidimensional, group-level construct defined as shared
Affected
Consequence Personal Injury perceptions of safety policies, practices, and procedures (D. Zohar,
Body Parts Injured Face and Hands 1980). Safety climate, however, goes beyond the stated safety policies,
Spill/Leakage/ No encompassing the relative priority of safety compared to other organi­
Fumes zationally relevant behaviors (e.g., speed, efficiency) (Dov Zohar, 2010).
Explosion Yes
Fire No
As a group-level construct, safety climate reflects perceptions of multiple
Direct Cause Improper Storage and Handling individuals; it is most often examined by collecting individual percep­
tions and aggregating them to the workgroup, department, or organi­
zational level.

2
N. Gopalaswami and Z. Han Journal of Loss Prevention in the Process Industries 64 (2020) 104027

Within the safety literature the similar phrase “safety culture” is also � reducing the frequency and risk of incidents by providing public
used [e.g.,(National Research Council, 2014)]. However, psychologists information on causes and consequences of incidents;
differentiate between organizational culture and organizational climate. � identifying trends and patterns in the database for better under­
Culture refers to those values, assumptions, and beliefs of an organiza­ standing of incidents;
tion passed down from its founders (Pettigrew, 1979). Organizational � developing training materials and procedures to promote laboratory
climate is based on perceptions of what happens in an organization, safety and incident prevention in labs;
whereas culture helps to explain why things happen in a certain way � communicating lessons learned from incidents to researchers and
(Ostroff et al., 2013). Safety should be an integral aspect of organiza­ laboratory safety personnel.
tional culture, wherein an organization is founded on and indoctrinates
employees concerning safe practices above all else. Organizational The database will help facilitate the acquisition of knowledge about
values and beliefs are most often measured through qualitative methods and recognition of commonly occurring hazards, and assist in the
such as interviews. On the other hand, safety climate is a group-level (e. evaluation of risks as a means of preventing incidents in the future. In-
g., laboratory-level) construct assessed using self-reports, which are then depth knowledge of the procedures, equipment, chemicals, and biolog­
aggregated to measure shared perceptions. Consequently, we use the ical materials is required to understand the class of hazards that one can
phrase safety climate in this paper. be exposed to. The database may provide information on leading in­
Safety climate is a multidimensional construct consisting of the dicators of laboratory incidents in university and secondary school
following dimensions: management commitment to safety, safety environments.
communication, co-worker safety practices, safety training, safety The importance of such an undertaking is somewhat obvious given
involvement, safety rewards, and safety equipment & housekeeping that there is not currently such a readily accessible database of labora­
(Beus et al., 2010). These seven dimensions incorporate perceptions of tory incidents. Most universities develop and maintain their own labo­
co-worker and supervisor dedication to and involvement in safety, as ratory safety manuals and database of incidents that occurred at their
well as the sufficiency of safety communication, training, rewards, and institution. For instance, several institutions focus on fire safety and
equipment. Empirical evidence indicates that management commitment maintain records of fire-related incidents (The Center for Campus Fire
is fundamental to the safety climate construct (Flin et al., 2000). Un­ Safety, 2015). Unfortunately, this information is often not readily
derstandably, leadership practices are very influential in how an indi­ available to the public. There are also significant institution-specific
vidual and group perceive the importance of safety. resources available concerning laboratory safety in the form of bro­
Safety climate is one of the most extensively studied safety-related chures, posters, handouts, videos, e-books and toolkits (Michigan State
constructs (Christian et al., 2009; Jex et al., 2012). Multiple previous University, 2019; Ohio State University, 2014).
meta-analyses support safety climate as a robust predictor of safety A description of the development of the database is provided in the
behavior and outcomes (Clarke, 2006). Meta-analytic estimates of the following section. This is followed by an overview of the categories
relationship between safety climate and behavior (i.e., compliance and included in the database, summary of the incidents, and important
participation) are 0.48 and 0.59, respectively (Christian et al., 2009). trends.
Estimates of the relationship between safety climate and accidents and
injury rates range from 0.22 to 0.39 (Christian et al., 2009; Clarke, 3. Methodology
2006).
A database was populated with information from the internet. Inci­
dent details were collected by creating email alerts on newspaper arti­
2.2. Safety climate in labs cles. The mail alerts were created by using key words such as “chemical
fire”, “chemical explosion” and “lab incident”. When there is match
Based on earlier definitions, safety climate in a laboratory refers to between keyword and the title of the incident, email alerts are triggered.
shared perceptions among lab members of safety policies, practices, and Incidents are then filtered based on relevancy to this study. Incidents
procedures. Lab safety (particularly in a chemical lab) involves (1) were recorded in Microsoft Excel. The database is continually being
awareness of the physical and chemical properties of materials; (2) updated to include recent incidents. Incident recording was initiated in
availability and use of equipment; (3) knowledge of additional practices 2012. This paper focuses solely on those incidents recorded between
to reduce risk; (4) ability to complete emergency procedures; (5) a well- 2012 and 2015. Because not all data are accessible, and all incidents are
designed and organized workspace; and (6) use of PPE (National not likely reported, the database should be interpreted as a sample of
Research Council, 2014). laboratory incidents that occurred within the specified timeframe. There
Safety climate has also been defined as the relative priority of safety. are undoubtedly incidents that took place that are not accessible and
There are a variety of competing priorities to safety in the lab environ­ therefore not included in the database. It should also be noted that the
ment (National Research Council, 2014). The pursuit of new knowledge incidents that are recorded in the database are acute in nature. Chronic
is fundamental in any research lab (National Research Council, 2014). incidents were not considered as it requires periodic inspection of
Principal Investigators prioritize the acquisition of new knowledge, personnel and labs for a long duration of time. Institutions rarely pub­
which may engender resistance to safety policies and oversight from licize chronic incident data even if they record them internally.
others (National Research Council, 2014). Additionally, the quality of Each incident was checked for duplication and assigned a unique
the work and the knowledge that can be obtained is often at least identification number. To date, the database includes 128 incidents. The
partially contingent on how much external funding a researcher can get. database is provided as Appendix A in this manuscript. The database
Safety procedures and regulations can be viewed as hindrances to get­ includes incidents that occurred in the U.S. as well as other parts of the
ting work published and external funding. Teaching labs are somewhat world. Details about each incident such as the date of the incident,
different. The purpose of teaching labs is to impart knowledge and skill. location, news source and incident description were entered into the
For students, the competing priorities to safety may be getting a good database. The following information was compiled for each incident:
grade on a project or completing an experiment in the allotted time (i.e., institution type, hazard type, consequences, body part(s) injured, sub­
speed). stance type, and causes.
The purpose of this paper is to describe the development of a uni­ Fig. 1 shows the number of incidents that occurred on the yearly
versity and secondary school laboratory incident database and to sum­ basis. There are about 30–35 lab incidents occurring on a yearly basis
marize the information contained therein to date. The objective of the (see Fig. 2–4).
database is to archive laboratory incidents which will facilitate: The coding of incident was based on the news alert received from the

3
N. Gopalaswami and Z. Han Journal of Loss Prevention in the Process Industries 64 (2020) 104027

Fig. 1. Incidents on a yearly basis.

Fig. 3. Incident classification based on hazard type.

whereas the incidents in universities primarily involved independent

researchers (J. N. Kemsley, 2009). It is important to note that all aca­
demic institutions in the U.S. are regulated by OSHA standards. How­
ever, there is no uniform policy on how to manage chemistry
laboratories.

3.2. Hazard type

Incidents were also classified based on the following types:

Fig. 2. Incidents classification based on Institution type. � Fires – Fire is a chemical reaction in which a substance combines
with oxygen and heat is released.
internet. For example, a news article in the alert is recorded in the � Explosions – Explosion is a release of energy sufficient to cause a
database as follows pressure wave. Explosions can be caused by reactive chemicals,
Please note that the weblinks might have expired over the year. To flammable gases, or combustible liquids.
retrieve original data regarding the incident, one can search the internet � Spill– An instance of a liquid spilling or being spilled from its storage
using the short description. containment
� Spill and Explosion – incidents involving both spill and explosion
� Spill and Fire-incidents involving both spill and fire
3.1. Institution type � Fire and Explosion – incidents involving both fire and explosion
� No Hazard– incident did not involve any hazard
This category refers to the type of educational institution. The in­
cidents were classified as “university” and “secondary schools”. In A majority (45%) of the incidents were primarily due to spill or
addition to high schools, private academies and tutoring schools were leakage of chemicals. Most of the spills could be avoided by evaluating
also classified as secondary schools. Fig. 1 shows a pie chart of incidents the chemicals used in the lab. This will help to identify the highly
classified based on institution. A majority of the incidents occurred at dangerous chemicals and the occurrence of secondary hazards like fire
universities (65%) when compared to secondary schools (35%). Ac­ or explosion. A spill kit should also be readily available in all labs to
cording to National Center of Educational Statistics, the number of properly clean up after a spill.
secondary schools in the U.S. is 10 times greater than the number of The next most frequent (23%) major incident type was explosions.
universities (Injury et al., 2010). However, the number of labs in uni­ One of the major contributors to explosions is a failure to recognize the
versities surpass the number of schools in the U.S. A university contains hazards posed by chemicals especially when information is not provided
one or more departments with varied set of labs, rather than school regarding potential reactions between chemicals. For example, a
which contains only science labs (Physics, Chemistry and Biology). Due researcher in an university lab was injured when trying to make tri­
to this reason, incidents at universities involved different kinds of methysilyl azide with sodium azide (J. Kemsley, 2014). The synthesis
agents, chemicals and a wide variety of equipment. In conclusion, uni­ was based on previously published methods with some alterations. One
versities tend to have proportionally more labs with greater risk. Many of the alterations involved the use of Poly Ethylene Glycol (PEG) to
of the secondary school incidents took place during a group demon­ remove the clumping of chemicals. However, the chemical behaved
stration of an experiment or a phenomenon to a class (Nicholson, 2014), counter actively and exploded causing damage to both lab personnel and

4
N. Gopalaswami and Z. Han Journal of Loss Prevention in the Process Industries 64 (2020) 104027

Fig. 4. Incident classification based on consequence.

the lab. One way to avoid similar incidents is to perform a thorough � Hazmat/firefighter Response and Community Disruption – incidents
hazard assessment. While performing hazard assessment, one can use that involved both Hazmat/Firefighter Response and Community
the chemical compatibility matrix to understand the potential reaction Disruption
between chemicals (Fisher Scientific, 2003). � Personal Injury and Hazmat/Firefighter Response – incidents that
Laboratory incidents also included fires (21%) resulting from involved both Personal injury and Hazmat/Firefighter Response
burners, flammable gases and combustible liquids. Other incident � Property Damage and Community Disruption – incidents that
involved more than one hazard. 3% of the incidents involved both spill involved both Property Damage and Community Disruption.
and explosion, 1% of the incidents involved spill and fire and 5% of the � None – no major consequence was observed in the incident.
incidents involved both fire and explosion. 2% of the incidents did not
involve any hazards. These incidents did not result in any spill, fire or A majority of the incidents (22%) led to Personal Injury. Among the
explosion. incidents that were recorded in the database, about 47% could be
classified under OSHA recordable injuries. OSHA recordable injuries are
3.3. Consequences death, days away from work, restricted work or transfer to another job,
medical treatment beyond first aid, or loss of consciousness (OSHA,
The third variable that was recorded for each incident was the 2011). OSHA has an extensive database recording incidents from several
consequence to the people involved or the environment they work in: educational institutions. However, OSHA does not differentiate between
incidents which occurred in labs and those which happened elsewhere.
� Personal Injury – A personal injury is an abnormal condition or Additionally, OSHA records ergonomic injuries (e.g., repetitive motion
disorder. Injuries include cases such as, but not limited to, a cut, injuries) which are not included in our database. The incidents recorded
fracture, sprain, or amputation. in the current database were restricted to lab incidents.
� Property Damage – harm or damage to the building/property in About 22% of the incidents required response from Hazmat teams or
which the incident occurred. firefighters. Hazmat teams are specially trained to handle hazardous
� Hazmat/Firefighter Response – incidents which were responded by materials or dangerous goods. There are three response levels in Hazmat
hazmat team or fire fighters. or three type of incidents (NFPA, 2002).
� Community Disruption – incident that involved disruption to the
community where people were evacuated, and the incident site was � Level 1 - An incident involving hazardous materials that can be
temporarily closed. Evacuation involves the action of evacuating contained, extinguished, and/or abated using resources immediately
people from a place of incident. Temporary site closure refers to the available to the public sector responders having jurisdiction. Level 1
temporary shutdown of a single site of employment, or one or more incidents present little risk to the environment and/or to public
facilities or operating units within a single site of employment. health with containment and cleanup.
� Hospitalization – person(s) involved in the incident were taken to the � Level 2- An incident involving hazardous materials that is beyond
hospital. the capabilities of the first responders on the scene and could be

5
N. Gopalaswami and Z. Han Journal of Loss Prevention in the Process Industries 64 (2020) 104027

beyond the capabilities of the public sector responders having � Eyes

jurisdiction. Level 2 incidents might require the services of a state or � Face
regional response team or other state or federal assistance. This level � Face and Hand
can pose immediate and long-term risk to the environment and � Multiple Parts
public health. � Hands
� Level 3- An incident involving hazardous materials that is beyond � Lungs
the capabilities of a single state or regional response team and re­ � Skin
quires additional assistance. Level 3 incidents can require resources � None
from state and federal agencies and private industry. These incidents � Unknown
generally pose extreme, immediate, and/or long-term risk to the
environment and public health. About half of the incidents did not result in an injury as depicted in
Fig. 5 These incidents might involve only site closure, property damage,
From the incident database, no information could be extracted or temporary evacuation. The most common injuries were to the skin,
regarding the response levels of Hazmat. However, it will be useful to lungs or face. Burns were classified as a skin injury and can be distin­
record this information if it is available in any source. guished based on the degree of exposure. Skin injuries account for 7% of
20% of the incidents in the database lead to community disruption. the total number of incidents. If multiple body parts were injured, they
Community disruption was in the form evacuation or temporary closure were classified as Multiple Parts. This amounts to around 1% of the
of the building in which incident took place. incidents. If only a single body part was affected, it was classified
There are many incidents which involved more than one conse­ separately based on the parts which were injured. Among the single
quence. For instance, one incident (Nicholson, 2014) led to Personal body parts, the face and hands appear to be the most affected body parts
injury and Hazmat/firefighter response. Such incidents were classified (see Figs. 6 and 7).
under “Personal Injury and Hazmat/FireFighter Response”. About 8% of Among the 128 incidents, 2% of the injuries could have been avoided
the incidents resulted in both personal injury and Hazmat/Firefighter by wearing personal protective equipment (PPE). In 2008, a fatal acci­
Response. Similarly, Hazmat/Firefighter Response and community dent happened in an university where a researcher suffered severe burns
Disruption contributed to 9% of the incidents and Personal Injury and from a fire that occurred in the chemistry lab. The researcher was using a
Community Disruption contributed to 10%. plastic syringe to transfer the pyrophoric reagent tertbutyl lithium from
5% of the incidents in the database involved hospitalization where one sealed container to another. The syringe broke spilling the chemical
person injured in the incident were taken to hospital. There were also and ignited the fire. It is to be noted that the researcher was not wearing
incidents where students were admitted in the hospital due to insistence a protective lab coat and her clothing caught fire which resulted in se­
from parents. Finally, there were about 1–2 incidents where property/ vere burns and death (J. N. Kemsley, 2009). The importance of PPE
building was damaged during the incident. needs to be emphasized to all lab personnel. While wearing PPE, it is also
important to make sure that it is appropriate for the task being con­
3.4. Body parts injured ducted in the lab as well as the hazards and experimental conditions. For
instance, a chemist in the U.S. was exposed to a chemical called dimethyl
For all incidents involving an injury, the body part that was injured mercury despite wearing gloves. The gloves were penetrable to mercury
was recorded as one of the following: and later the researcher died due to mercury poisoning (Edicott, 1998).
Incidents continue to occur due to lack of PPE or improper PPE. PPE
� Arm should not be considered as a primary hazard control, but rather as

Fig. 5. Incident classification based on body parts injured.

6
N. Gopalaswami and Z. Han Journal of Loss Prevention in the Process Industries 64 (2020) 104027

Fig. 6. Incident classification based on substance type.

Fig. 7. Incident classification based on root causes.

additional protection. It is important to eliminate or minimize hazards 3.5. Substance type

prior to starting an experiment.
The incidents were further categorized based on the following sub­
stance types:

7
N. Gopalaswami and Z. Han Journal of Loss Prevention in the Process Industries 64 (2020) 104027

� Compressed Gases ○ Placing material in an incompatible container

� Corrosive � Procedure violation - A procedure violation occurs when a lab

� Flammable and Combustible member is aware of the procedures, but chooses not to follow them.
� Oxidizer � Runaway reactions – reactive process when an exothermic reaction
� Pyrophoric generates more heat than expected and goes out of control, often
� Radioactive resulting in an explosion.
� Reactive Chemical � Lack of training –lack of training/poor training regarding the
� Toxics chemicals, equipment or other relevant information required to
� Water Reactive conduct a lab experiment.
� Unknown � Unknown –the cause of the incident is unknown

A majority of the substances involved in the incidents were corrosive Improper storage and handling comprised on the most frequent
(22%). Among the corrosives nitric acid, Sulphuric acid and hydro­ direct cause of incidents contributing to 40% of the incident. In an
chloric acid based incidents occurred in both universities and schools, university laboratory, solvent bottles fell from the top shelf of a four-
accounting for 22% of the incidents. However, there have were also shelf flammable-solvent storage cabinet because the clips used to
many incidents where the substance type was not known, accounting for secure the shelf failed. The clips had been modified improperly
49% of the incidents. (Borchardt, 2006). Similar to our analysis, the Board of Chemical Sci­
Flammable chemicals were involved in 12% of the incidents. Flam­ ence and Technology has also identified improper storage and handling
mable gases need to be handled carefully in the lab. Among the flam­ as a major contributor to accidents. Their analysis was based on accident
mable gases, hydrogen, needs to be handled with utmost care in investigations and OSHA incident database (National Research Council,
laboratories. An incident involving hydrogen gas occurred in 2006. A 1995). The National Research Council put forward a detailed document
fire began in an autoclave when the rupture disc collapsed, which focusing on prudent practices in the laboratory for handling and disposal
released, oil, distillates and hydrogen out of the autoclave. The oil and of chemicals to avoid incidents due to improper storage and handling
gas ignited and the fire eventually led to an explosion (Hydrogen tools, (National Research Council, 1995). Proper chemical storage involves
1983). One person received minor burns on his arms and the lab organizing the lab through good house-keeping procedures. Such
incurred a huge monetary loss due to the physical damage. Given the housekeeping controls include proper segregation and labeling and
hazardous nature of hydrogen, there are separate lab accident databases regular inspection of different areas in the lab to ensure proper storage
maintained for hydrogen alone (Hydrogen tools, 1983). of chemicals. This procedure helps to avoid slips, trips and falls in the
Toxic materials can be identified as liquids and gases which are toxic lab.
to humans on exposure. Toxic substances amount to 6% of the incidents. Runaway reactions contributed to 12% of the incidents. Many of the
An incident occurred in Australia in 2014 when dozens of school chil­ chemical reactions conducted in the ab tend to have an industrial basis.
dren were exposed to sodium hydroxide fumes during a lab demon­ Many industrial chemical processes involve exothermic or heat gener­
stration (Philip, 2014). To avoid such exposure during demonstration, ating reactions. While conducting these exothermic reactions, runaway
one can protect themselves from toxic substances using proper respira­ reactions can occur due to failure of reaction’s cooling system, mis­
tory equipment. charged reactants, contaminants and other reasons (EPA, 1999). It is
Reactive chemicals often contain different types of organic perox­ important to understand the reaction mechanisms and properly follow
ides, shock-sensitive materials involving functional groups such as the precautions to avoid thermal runaways. The Health and Safety Ex­
acetylide, azides, diazo or nitroso compounds. These shock sensitive ecutive in UK provides a leaflet outlining the hazards of thermal run­
materials are susceptible to heat or shock. The water reactive react aways and detailed guidance on preventing such runaway reactions
violently to create heat, oxidizing gas, fumes or corrosive acids. Pyro­ (Health and Safety Executive, 2014).
phorics can ignite spontaneously on exposure to air. Examples of such A procedure violation occurs when a lab member is aware of the
pyrophoric chemicals include alkali metals like sodium and potassium. procedures, but chooses not to follow them. This amounted to 3% of the
These categories contributed to 1% of the incident respectively. In 3% of incidents.
the incidents, explosives were involved. Another important category contributing to the incidents was
equipment failure contributing to 14%. At an university, organic
3.6. Direct causes chemistry students were storing samples in an unlabeled refrigerator
that was not explosion proof. The refrigerator exploded and the doors to
The incidents were coded to obtain the direct causes. Direct cause is a the main lab were blown to the other side of the room where they hit an
primary cause leading to an incident. It is an immediate basis without apparatus used to purify solvents (The Center for Campus Fire Safety,
which an injury would not have happened. The direct causes of the in­ 2015). Major causes of equipment failure are due to utility failure, which
cidents were recorded based on the following types: disrupts electrical operations.
An incident was categorized as lack of training when there was no
� Lack of PPE –person(s) involved do not wear appropriate personal appropriate procedure or training in place for laboratory. Many stu­
protective equipment dents, especially those with little experience, are not aware of the risks
� Equipment Failure – equipment failure or factors like power shortage associated with the materials and equipment in their lab. This can lead
leading to equipment failure. to accidents. Such incidents contributed to 6% of the incidents. It is the
� Improper Storage and Handling - Improper storage and handling can role of the institution to inculcate lab safety to persons in lab. In the
refer scenarios where hazardous substances are not stored, handled absence of institution-based lab safety, one can resort to various other
or disposed properly. Following causes can be considered as external sources of lab training available currently. These lab training
improper storage and handling. are offered in many ways such as courses, books, pamphlets and videos.
○ Incidents which lead to spill due to dropping of chemical con­ There were 2% of the incidents were researcher was lacking PPE in the
tainers like beakers or flasks. lab.
○ Spilling of chemicals when transferring materials from one Despite proper planning and implementation of lab prevention
container to another. strategies, lab accidents continue to occur. It is important to report the
○ Holes or cracks in containers or transfer equipment such as pipes, incident immediately to responsible authorities. Incident reporting is
hose and valves. Placing key to gain information about the incident; however, this is not strictly

8
N. Gopalaswami and Z. Han Journal of Loss Prevention in the Process Industries 64 (2020) 104027

followed in all educational institutions. institution type, hazard type, body parts injured, substance type, causes
and consequences. It was observed that 65% of the incidents take place
4. Recommendations for improving lab safety in the universities when compared to secondary schools. The most
common hazard observable in incidents is spill of chemicals contrib­
Currently there is no incident database that is comprehensive. uting to 45% following by explosions (23%) and then fires (21%). The
Various databases have different recording trends and there is unclear major consequence observed in the incident is personal injury (22%)
definition for incident categories as well as information that needs to be and Hazmat response (22%). In most of the incidents, the body parts
recorded, which makes consolidating data challenging and difficult to injured were unknown. About 22% of the incidents involved the use of
draw conclusions and provide reasonable recommendations. The au­ corrosive substances. The main direct cause of the incident was attrib­
thors would recommend that when recording incidents, more details uted to improper storage and handling.
should be provided. The categories reported in this research can serve as The paper also provides an overview of safety climate as an influ­
an example. ential predictor of safe behavior and accidents in the laboratory envi­
In terms of consequences, the severity should be recorded for each ronment. However, it was difficult to obtain data on safety climate in
consequence. To be specific: laboratories. This involves interviews with institutions and detailed
mining of incidents. The database is an ongoing activity and future work
� Personal injuries - the body injury part and level of injury will include further refinements and more detailed classification of the
� Property damage - the material loss incidents including safety climate factors.
� Community impact - length of evacuation/site closure; The paper highlights the importance of having an incident database
� Hazmat/Firefighter response - if there is Hazmat or firefighting because information obtained from press reports stay in the internet
response only for a limited amount of time. Collection of such incident informa­
� Hazmat response level - level 1, 2, or 3 tion need to coded and analyzed to avoid future incidents. While there
� Hospitalization - length of hospitalization and treatment received are numerous codes and standards regulating academic institutions,
� Environmental impact - the chemical released and location/impact recent incidents indicate that enforcement can play an important role in
of release (e.g., to air, land, or public) reducing incidents in labs.

Additionally, one can record the type of lab (science, chemistry, Acknowledgement
biology, mechanical, electrical and so on). To be able to do this, suitable
key words needs to be increased in the google alert list. One can also This paper is dedicated to our PhD advisor Dr. Sam Mannan who
classify the incident as cognitive or in-cognitive. This will aid in passed away on September 11, 2018. The authors would like to thank
analyzing the database from safety culture perspective. Michael O’Connor from Mary Kay O’Connor Process Safety Center,
Based on the analysis performed in the current study, recommen­ Texas A&M University, College Station for his support and guidance.
dations are proposed for improving lab safety in schools and The authors would also like to thank Dr. Stephanie Payne and Nate
universities. Keiser from Department of Psychology at Texas A&M University, Col­
lege Station, USA for their help in this paper.
1. Conduct a comprehensive hazard analysis before performing any
demonstration or experiment. The American Chemical Society (ACS) Appendix A. Supplementary data
has published guidelines on identifying and evaluating hazards in
research laboratories (American Chemical Society, 2013). It provides Supplementary data to this article can be found online at [Link]
general information and multiple tools to identify and assess hazards org/10.1016/[Link].2019.104027.
in research laboratories for the laboratory researcher. To implement
safety culture in labs one can resort to guidance put forward by APLU
References
in their book published “A Guide to Implementing a Safety Culture in
Our Universities”(APLU Council on Research Task Force on Labo­ American Chemical Society, 2013. Identifying and Evaluating Hazards in Research
ratory Safety, 2016). Laboratories.
2. Understand the risk involved in experiment or demonstration thor­ APLU Council on Research Task Force on Laboratory Safety, 2016. A guide to
implementing a safety culture in our universities. Washington DC. Retrieved from.
oughly before initiating work in the lab. [Link]
3. Follow proper procedures and techniques that need to be used in the Beus, J.M., Payne, S.C., Bergman, M.E., Arthur Jr., W., 2010. Safety climate and injuries:
experiment. an examination of theoretical and empirical relationships. J. Appl. Psychol. https://
[Link]/10.1037/a0019164. Beus, Jeremy M.: Department of Psychology, Texas
4. Take necessary training like written procedures, appropriate PPE and A&M University, 4235 TAMU, College Station, TX, US, 77843-4235, jeremybeus@
equipment before handling hazardous chemicals [Link]: American Psychological Association.
5. If risk is involved, use appropriate measures to enhance the safe­ Borchardt, J.K., 2006. Lab Safety Requires Training and Commitment. Retrieved May 12,
2015, from. [Link]
guards. While performing demonstrations, implement a safety bar­ training-and-commitment.
rier between audience and demonstrator. Christian, M.S., Bradley, J.C., Wallace, J.C., Burke, M.J., 2009. Workplace safety: a meta-
6. Establish regular inspection schedules and procedures for checking analysis of the roles of person and situation factors. J. Appl. Psychol. [Link]
org/10.1037/a0016172. Christian, Michael S.: Eller College of Management,
the lab safety and appropriate PPE University of Arizona, Department of Management and Organizations, McClelland
7. Follow all regulations and guidelines governing lab safety including Hall, P.O. Box 210108, Tucson, AZ, US, 85721-0108, msc@[Link]:
labeling, storage, handling and disposal of chemicals. American Psychological Association.
Clarke, S., 2006. The relationship between safety climate and safety performance: a
8. Use safety check list to before conducting experiments.
meta-analytic review. J. Occup. Health Psychol. [Link]
8998.11.4.315. Clarke, Sharon: Manchester Business School, The University of
5. Conclusion Manchester, Booth Street West, Manchester, United Kingdom, M15 6PB, sharon.
clarke@[Link]: Educational Publishing Foundation.
Edicott, K., 1998, April. The Trembling Edge of Science, vols. 1–10. Dartmouth Alumni
This paper provides an overview of lab incidents occurring in uni­ Magazine. Retrieved from. [Link]
versities and schools over a three-year period based on a database that ty/trembling_edge_science.pdf.
was developed to record said events. Documentation of incidents is EPA, 1999. How to Prevent Runaway Reactions, vol. 73. EPA.
Fisher Scientific, 2003. Chemical Compatibility Chart. Fisher Scientific. Retrieved from.
essential as a means of identifying leading indicators and preventing [Link]
accidents in the future. To that end, incidents were analyzed based on [Link].

9
N. Gopalaswami and Z. Han Journal of Loss Prevention in the Process Industries 64 (2020) 104027

Flin, R., Mearns, K., O’Connor, P., Bryden, R., 2000. Measuring Safety Climate: NFPA, 2002. NFPA 471 - Recommended Practice for Responding to Hazardous Materials
Identifying the Common Features. Incidents (Quincy, Massachusetts).
Furr, A.K., 2000. CRC Handbook of Laboratory Safety. CRC Press, Baco RAton, FL. Nicholson, K., 2014, September. Chemistry Lab Fire Burns Students at Denver School.
Griffin, M.A., Neal, A., 2000. Perceptions of Safety at Work: A Framework for Linking Denver Post. Retrieved from. [Link]
Safety Climate to Safety Performance, Knowledge, and Motivation. Journal of ree-denver-middle-school-students-injured-lab-mishap.
Occupational Health Psychology. US: Educational Publishing Foundation. https:// Ohio State University, 2014. The Institutional Biosafety Manual. Cinncinnati: Ohio State
[Link]/10.1037/1076-8998.5.3.347. University. Retrieved from. [Link]
Harper, L., Watt, F., 2012. Laboratory Safety Culture Survey 2012 – Draft Report. f.
Retrieved from. [Link] OSHA, 2008. Occupational Safety and Health Administration, Recording and Reporting
lture-draft_report.pdf. Occupational Injuries and Illness (29 CFR Part 1904). Retrieved from. [Link]
Health, Safety Executive, 2014. Chemical Reaction Hazards and the Risk of Thermal [Link].
Runaway, vol. 254. Retrieved from. [Link] OSHA, 2011. 29 CFR Part 1904, Pub. L. No. C. e-CFR, USA. Retrieved from.
Hydrogen tools, 1983. Fire and explosion in autoclave cell. Retrieved November 17, [Link]
2010, from. [Link] ¼STANDARDS&p_id¼9638.
Injury, O., Classification, I., Version, M., Statistics, L., 2010. Bureau of Labor Statistics Ostroff, C., Kinicki, A.J., Muhammad, R.S., 2013. Organizational culture and climate. In:
Occupational Injury and Illness Classifications Major Characteristics for Handbook of Psychology: Industrial and Organizational Psychology, second ed., vol.
Occupational Injuries and Illnesses, vol. 2010 (September). 12. John Wiley & Sons Inc, Hoboken, NJ, US, pp. 643–676.
Jex, S.M., Swanson, N., Grubb, P., 2012. Healthy Workplaces. John Wiley & Sons, Inc, Pettigrew, A.M., 1979. On studying organizational cultures. Adm. Sci. Q. 24 (4),
Hoboken, New Jersey. 570–581. [Link]
Kemsley, J., 2014. More details on the University of Minnesota explosion and response. Philip, J., 2014, August 28. 57 students treated after school science accident. The
Retrieved September 11, 2015, from. [Link] Western Star, p. 2. Retrieved from. [Link]
7/more-details-on-the-university-of-minnesota-explosion-and-response/. udents-taken-to-hospital/2366127/.
Kemsley, J.N., 2009, August. Learning from UCLA. Chem. Eng. News 29–31. Retrieved The Center for Campus Fire Safety, 2015. Campus fire safety. Retrieved May 12, 2015,
from. [Link] from. [Link]
Michigan State University, 2019. Biosafety Manual. Michigan State University, Michigan. US Chemical Safety Board, 2010. Texas Tech University Laboratory Explosion. https://
Retrieved from. [Link] [Link]/10.1037/e528452011-001 (No. 2010- 05- I- TX). October. Washington ,D.C.
National Research Council, 1995. Prudent practices in the laboratory. In: Handling and Zohar, D., 1980. Safety climate in industrial organizations: theoretical and applied
Disposal of Chemicals, first ed. NATIONAL ACADEMY PRESS, Washington, D.C. implications. J. Appl. Psychol. 65 (1), 96–102.
National Research Council, 2014. Safe Science: Promoting a Culture of Safety in Zohar, Dov, 2010. Thirty years of safety climate research: reflections and future
Academic Chemical Research. The National Academies Press. Retrieved from. http directions. Accid. Anal. Prev. 42 (5), 1517–1522. [Link]
://[Link]/catalog/18706/safe-science-promoting-a-culture-of-safety-i aap.2009.12.019.
n-academic-chemical.

10