實驗安全多重要?能發《Nature Chemistry》!A review of academic lab safety

各位科研人員在剛進實驗室的時候，做的第一件事就是接受安全培訓，因為這太重要了，事關自己和實驗室同事/同學的安全，還關係到了供職單位的形象。但是時間一長，大家往往會忽略安全問題，造成傷亡，因此安全問題這個警鐘一定要長鳴才行。

A Review and Critique of Academic Lab Safety Research

A. Dana Ménar& John F. Trant 

Nature Chemistry (2019)

Abstract

Over the past ten years, there have been several high-profile accidents in academic laboratories around the world, resulting in significant injuries and fatalities. The aftermath of these incidents is often characterized by calls for reflection and re-examination of the academic discipline’s approach to safety research and policy. However, the study of academic lab safety is still underdeveloped and necessary data about changes in safety attitudes and behaviours has not been gathered. This Review article critically examines the state of academic chemical safety research from a multifactorial stance, including research on the occurrence of lab accidents, contributors to lab accidents, the state of safety training research and the cultural barriers to conducting safety research and implementing safer lab practices. The Review concludes by delineating research questions that must be addressed to minimize future serious academic laboratory incidents as well as stressing the need for committed leadership from our research institutions.

On December 29, 2008, Ms Sheharbano Sangji, a research assistant in the lab of Dr Patrick Harran at UCLA, was working with a large quantity of tert-butyllithium when the pyrophoric chemical spilled and ignited her clothing, leading to 2nd and 3rd degree burns over 40% of her body. The 23-year-old died in hospital three weeks later. The California OSHA report1 into the death of Ms. Sangji is provided in the Supplementary Information.

Contributing factors to this accident can be identified at multiple levels: the individual, the laboratory, the department, the institution and the discipline itself1. At the time of the accident, Sangji was not wearing a lab coat1 and was not following the manufacturer’s safety protocols for handling large quantities of a pyrophoric chemical (for example, the reagent bottle was not clamped and a plastic syringe was used instead of glass)2. Despite knowing that she had limited experience working independently in chemistry labs, Harran, her supervisor, stated that he had not trained Sangji in the proper handling of pyrophorics and that the necessary technical guidelines were not readily available in the lab1. A post-doctoral researcher in the Harran group who recalled that he may have offered general guidance to Sangji about the transfer and handling of tert-butyllithium acknowledged that he did not follow the manufacturer’s safety instructions for handling this reagent and did not believe he had ever read them1,3.

It is not clear that a lab coat was actually ordered for Sangji by Harran or anyone else at UCLA1. In fact, use of personal protective equipment (PPE) was not officially mandated by university policy1. Pending completion of renovations to laboratory space on another floor, Harran had been given temporary space that was 30–40% smaller than his requirements and did not have a stockroom for storing chemicals1. In the 14 months prior to Sangji’s death, UCLA had failed to report to the California Division of Occupational Health and Safety two other similar, non-fatal, incidents from other research groups involving burns and facial lacerations to students not wearing appropriate PPE1,4,5. Although experienced researchers will criticize the technique employed for the pyrophoric chemical, an individual who has not been trained in their use cannot be faulted: in virtually all published research where tert-butyllithium is used, the hazards of working with this chemical are rarely spelled out6.

In 2012, Dr Neal Langerman, former chair of the Division of Chemical Health and Safety of the American Chemical Society (ACS), described the UCLA incident as, 「The most serious challenge to the practice of laboratory safety in many years. The lessons learned should result in fundamental cultural changes in the approach to research safety」7. Yet despite the occurrence of this tragedy and other serious high-profile incidents in the intervening years (for example, the death of Michele Dufault at Yale University, and explosions causing significant injuries to Preston Brown at Texas Tech and Dr Thea Ekins-Coward at the University of Hawaii)8,9,10,11, the field of academic lab safety has received little empirical attention, and research efforts in this area have been fragmentary12.

In many cases, questionable research methodologies seriously undermine the reliability, validity and applicability of findings. As a result, policies and procedures around safety are developed in a reactionary, ad hoc patchwork rather than on a solid, comprehensive, empirical foundation13. More than ten years on from Sangji’s death, we can conclude that there is no evidence of sweeping, fundamental changes, nor of major paradigm shifts in how academic lab safety is approached within the discipline. As this was a high-profile case, and was initially expected to be a turning point for academic lab safety, we will return to this example for illustrative purposes. However, we want to emphasize that safety is not a problem unique to the Harran laboratory or to UCLA: the failures that led to the death of Sangji are systemic and could have occurred in many research groups at many institutions. The problem is, and sadly remains, much bigger than any single case14.

This Review aims to critically examine the current state of chemistry laboratory safety research, to discuss the barriers to conducting and implementing results of this research, and to call for a re-examination of, and a commitment to, academic chemistry’s role in accident research and prevention.

Type and frequency of accidents in the academic setting

All practicing chemists in academic institutions are aware of and acknowledge that lab accidents and near-misses (for example, fires, leaks, glassware implosions or explosions, spills, equipment or instrument misuse resulting in equipment failure that do not result in injury) occur regularly15,16. However, to the best of our knowledge, no researcher, university, regulatory body or professional organization has collated the annual incidence of academic laboratory accidents. No comprehensive dataset is currently available on the type or frequency of accidents or near-misses in academic laboratories5,17. Although several researchers have attempted to create accident databases15, participation in these initiatives is voluntary, meaning that these sources are incomplete and inadequate for the purposes of research and comprehensive policy development. This lack of data severely hampers any efforts to understand accidents, to take steps to prevent them, to reduce their frequency and severity or to create evidence-based safety guidelines.

It could be expected that gathering, collating and analysing lab accident data in order to determine prevalence rate might fall under the purview of governmental regulatory agencies. However, to our knowledge, neither the Occupational Safety and Health Administration (OSHA) in the United States nor any of the 13 provincial and territorial safety boards in Canada, despite receiving accident reports and carrying out investigations, has ever compiled or analysed this data as a whole. In addition, OSHA regulations do not apply to all universities or to all lab personnel working in a university, depending on their employment status. Since 2001, the US Chemical Safety and Hazard Investigation Board (CSB) has reported 120 academic research laboratory accidents resulting in 87 evacuations, 96 serious injuries and three deaths (Table 1)18. However, these represent only those accidents that universities have been required to report due to the severity of the consequences. Regulators may thus remain unaware of potentially major incidents or significant near-misses if no one was seriously injured19.

Table 1 A partial list of researchers killed in laboratory accidents at academic institutions (2008–2018)

Full size table

There has been very little academic research into the prevalence and incidence of laboratory accidents. In the only study we could find using a proper multi-institutional epidemiological approach, Hellman, Savage and Keefe, examining 574 accidents occurring at 13 Colorado institutions between 1966 and 1984, found that 81% of accidents occurred in teaching labs, 13% in research labs and 2% in fabrication rooms20. Most accidents occurred in entry-level chemistry lab courses or organic lab courses, and most commonly occurred among younger individuals20.

There have been a few, mostly small, studies focused specifically on the prevalence of research lab-related injuries. In one survey from Nature and UCLA of 2,400 scientists, 30% reported having witnessed a lab injury severe enough to warrant attention from a medical professional21. A small pilot study of 56 lab personnel in Canadian chemistry and biology labs revealed that 15% of those surveyed had sustained at least one injury22. Simmons, Matos and Simpson found that lab accidents, both in teaching and research labs, represented 18.4% of the total incidents reported at Iowa State university from 2001 to 2014, and that student employees were the victims in one third of injury reports23.

Aside from the study by Hellman and colleagues, there have been few studies of injuries sustained in undergraduate teaching laboratories, perhaps because these situations are more carefully controlled and involve less dangerous reagents. However, one study of students enrolled in general chemistry and organic chemistry courses found that 12% sustained an injury, the most common ones being chemical burns, inhalation of irritating or toxic gases and cuts24.

Although this research is incomplete, it certainly paints a troubling picture. One major issue is that research into laboratory injuries tells us nothing about the overall accident prevalence rate. 『Close calls』 involving no injuries are anecdotally far more common than accidents involving injuries, but are rarely even reported unless the property damage is severe. In addition, the true accident prevalence rate is likely worse than these results suggest as there is some evidence to suggest that underreporting is a significant problem in science. Studies conducted in this area have shown that 25–38% of participating lab personnel have been involved in an accident or injury in the lab that was not reported to the supervisor/principal investigator (PI)23,24,25.

Contributing factors in laboratory accidents

Given the lack of research on the prevalence and incidence rates of academic lab accidents, it is perhaps not surprising that there is a similar lack of research on what causes lab accidents. Contributing factors to lab accidents can be conceptualized as occurring at multiple levels: risks associated with the materials or equipment being used, risks related to the skills, knowledge and choices of the research personnel doing the study, characteristics or qualities of the PI and the research lab in which the research is occurring and risk factors arising from the departmental or institutional level.

Risks associated with the materials being used have received the most attention in the laboratory health and safety literature. For example, there have been publications about specific reagents such as diazomethane, organolithiums or dimethyl dioxirane26,27,28,29; these reports are usually presented in the context of why a new methodology is safer or more effective. Discussions on particular reagents are mostly found only in the blogosphere30,31, or in ever-growing compendia of reagents32,33 (https://www.rsc.org/merck-index). However, safety information about reagents is not typically required by journals in the discipline. In 2016, Grabowski and Goode found that only 8% of the 726 chemistry journals they identified required safety factors to be mentioned in the manuscript6. The authors specifically looked at mentions of 11 target compounds that are known to be hazardous (including tert-butyllithium); these compounds were mentioned 107 times but only one article provided cautionary information.

If institutions are not providing comprehensive safety training on the use of reagents, which seems to be the case (see below), it becomes untenable for authors to assume that readers will be aware of the risks associated with particular compounds. This assumption becomes increasingly dangerous as more and more chemistry research is conducted around the world by inexperienced students and trainees. Starting in 2017, the ACS mandated that all experimental publications provide warnings for current or new hazards or risks34. However, this recommendation was tempered by the use of the phrase 『as appropriate』, meaning that the inclusion of safety information is at the discretion of the authors and reviewers. A follow-up to the work of Grabowski and Grave would be informative as to whether the change in policy has changed the content of articles or has been enforced by the journals.

We could not find any studies anywhere that looked at how skills, knowledge, experience or attitudes of the research personnel are associated with the occurrence of lab accidents or other proxy variables (for example, near-misses). Similarly, there have been no studies investigating the occurrence and recurrence of accidents within specific departments or universities, nor has there been research looking at the role of situational factors in causing accidents, such as time of day (for example, late at night). The most complete research to date on the causes of academic lab accidents comes from the previously mentioned epidemiological study of Colorado chemistry departments for incidents occurring between 1966 and 198420. Hellman, Savage and Keefe examined demographic characteristics of victims, details about research activities, type/location of injury, time of day, and time of year for 574 accidents. The value of this data to the contemporary research laboratory is questionable. Most incidents occurred during undergraduate teaching labs and many involved now-obsolete techniques (for example, mouth-pipetting).

There are also a number of historically based factors that limit the current applicability of the results: the majority of accidents happened during afternoons in the academic year, likely because this was when those universities offered undergraduate labs; and most injuries were to men, primarily because a much greater proportion of the undergraduate student population at that time were men. However, the study’s authors highlight the contribution of human factors to lab accidents and call for additional research, saying 「Of all the variables in accident prevention, the human behaviour variable, even with education, was the hardest to control」35,36. The study has never been replicated or updated since it was published and, we note with dismay, has only been cited 7 times according to independent Web of Science, Scifinder and Google Scholar searches conducted on April 5, 2019.

Case studies have been published in response to significant incidents and have typically resulted in the creation of reaction- and equipment-specific guidelines. For example, case studies have been published about Sangji’s death37, the explosion at Texas Tech11, a gas leak at the National University of Singapore38, the mishandling of a drum of radioactive material39 and a sucrose-acid explosion at an unnamed university40. These case studies often take a comprehensive look at the multifactorial contributors to the accident at the individual, laboratory and institutional levels. However, this multi-level approach is not characteristic of safety research in general, nor of the implementation of safety policies in chemistry departments, though it should be. Researchers in the wider field of occupational safety have suggested that accidents are most likely to occur when multiple individual and system failures align (that is, the 『Swiss cheese』 model of accidents)41. These case studies and reagent-specific studies have not led to more comprehensive research, a deeper conversation across the academic discipline or a broader examination of accident causes, despite explicit calls to do so. Fundamentally, case studies represent a collection of anecdotes42, which may be informative and useful in specific situations and with specific materials, but represent an insufficient basis for the creation of wider evidence-based safety policies and procedures. Case studies play a valuable role in building the evidence base, but they are usually intended to be the launching point for broader analysis, not the end-point as they have been to-date in the world of academic lab safety. In fact, the bulk of these publications appeared in the Journal of Chemical Health & Safety, which is not completely indexed in SciFinder, despite being a flagship publication associated with a division of the ACS. Such considerations make this work difficult to find even for those scientists who actively seek it.

Attitudes about safety and behavioural practices

There has been some research on the attitudes and beliefs of lab personnel regarding safety in the lab. For the most part, these studies tend to suggest that researchers have generally positive views towards the concept of lab safety and related concepts. Wu and colleagues in Taiwan assessed perceptions of safety leadership in 465 lab employees; respondents rated levels of 『safety coaching』, 『safety caring』 and 『safety controlling』 at their institution quite highly43. In a study of 85 staff, faculty and grad students, Steward, Wilson and Wang found generally positive attitudes about safety culture in the lab, that is, employees』 perceptions, attitudes, and beliefs about risk and safety44. A large majority of participants in Ayi and Hon’s study (88%) described safety as a high priority in their labs22. Schröder and co-workers found that over 90% of researchers felt that their labs were a safe place to work45. Although these studies are encouraging and potentially relevant to the occurrence of lab accidents, the role of abstract ideas such as safety culture, safety climate, safety leadership, safety coaching and subjective feelings of safety is of limited utility without these constructs being validated against objective measures such as frequency of accidents and injuries, or even of proxy measures such as inspection violations. In other words, it is not clear if individuals who value safety and believe that their workplaces are safe actually make safe choices in their laboratory practices. To date, there has been no research on the correlations between safety attitudes and safety practices.

In contrast to these optimistic findings about safety beliefs, research results regarding behavioural safety practices are concerning. The results from several studies have suggested that researchers are disinclined to conduct safety assessments prior to conducting experiments. In Ayi and Hon’s study, 27% of participants, active experimental researchers, stated that they never conducted any kind of risk assessment before performing laboratory work22. In another study, half of respondents did not search for, or use, safety information in developing experimental procedures, yet 80% considered the existing available information adequate to support risk assessment (suggesting that participants generally thought the information to be sufficient but were disinclined to use it for other, unidentified reasons)46. In Schröder and co-workers』 comparison of researchers in different settings, academic researchers were the least likely to assess risk (only 18% reported doing so) compared to industry (43%) or government (36%)45. To be fair to academia, the low rate of risk assessment identified in this study by researchers in industry and government is also troubling.

There have also been a few studies on the use of PPE, and again, the results are difficult to interpret given that the general positive attitudes towards safety shown in these studies. In a study of undergraduates in teaching labs (arguably, the easiest cohort to observe and control), Sieloff and coauthors found that 94% of students consistently reported wearing eye protection but 65% said they never wore gloves24. These findings are similar to those of Ayi and Hon, who found that only 40% of their participants and academic researchers reported wearing PPE at all times when working22. Schröder and colleagues found that researchers in academia were less likely to wear lab coats (66% consistently wore them) or eye protection (61%) than industry (87% and 83% respectively) or government employees (73% and 76%)45. Again, these numbers are a cause for concern.

From a methodological standpoint, the exclusive use of self-report data in these studies is troubling as results are likely to be inaccurate due to social desirability bias in participants』 responses (that is, the tendency of respondents to answer questions in a manner that will be viewed favourably by others)47, a factor that has not been acknowledged or addressed in any of this research48. For example, researchers know that they should be wearing PPE and may therefore intentionally or unintentionally inflate the numbers they report in studies; a more accurate estimate of PPE usage, which could be gathered through observational studies (but so far has not), may be much worse. However, without proper data, we cannot say with any certainty how the use of PPE relates to accident frequency and/or severity. It may be the case that the semblance of protection can encourage riskier behaviour. For example, research has shown that the use of bicycle helmets is correlated with increased risk of accident49, consumers tend to make higher-calorie choices when provided with calorie counts at restaurants50 and beachgoers often choose to swim outside of designated safe areas on beaches51. The use of PPE might encourage researchers to take more or greater risks and therefore increase the rate or severity of lab accidents. These are the types of questions that should be addressed by researchers of academic lab safety.

Commentators on academic lab safety have noted the role of human factors and remarked on their importance in safety behaviours52. Some have even labelled the cognitive biases at play in safety issues or made safety recommendations based on psychological principles of habit development53. Hendershot expressed concern at researchers』 tendencies to believe that activities must be safe if they are done routinely and nothing has gone wrong, thus ignoring the base rates of accidents36. He cautioned, 「Our personal experience in a few thousand work hours is not statistically relevant when actual performance of the process industries is in the range of a few fatalities in hundreds of millions of exposure hours.」 Human factors have frequently been cited in the write-ups of incident case studies. For example, Schmidt mentioned the bystander effect (that is, the tendency of individuals to offload responsibility for intervening in a critical situation when others are present54) in describing the circumstances that led to a gas leak at a research institute in Singapore38. The results from several studies suggest that many research personnel see the level of risk in their laboratories as low. For example, 59% of participants in Ayi and Hon’s study thought that the level of risk associated with their work was low or very low22. It would also appear that decision-making with regards to PPE is heavily based on respondents』 own assessment of risk: at higher levels of (self-assessed) risk, respondents in Schröder’s study said they were more likely to don the appropriate PPE.

Given the likely impact of individual biases, ensuring perfect access to information and training (for example, Bretherick’s handbook, ACS guidelines, departmental policies and laboratory policies) and making equipment available is not likely to change outcomes without a better understanding of the psychology of safety decision-making46,55. These resources are currently available to many researchers and are not being used. Behavioural data must be collected to inform new practices in training. The relationship between the perception of risk and safety attitudes and behaviour needs to be studied and addressed. However, to date, the champions for safety have been natural scientists and engineers whose research expertise is not in social science methodology and who may be unfamiliar with important and relevant psychological constructs (such as social desirability in responding). The studies examined for this Review article, as noted throughout, suffer from flaws that compromise the validity, reliability and generalizability of their findings where policy is concerned.

More often than not, the consideration of human factors has tended to centre on blaming victims for their behaviours that led to or aggravated an accident, exemplified in the case of Sangji56. Although many write-ups focused on the fact that she did not think to use the lab shower when her shirt caught fire, in fact, neither of the two postdocs who assisted her in the aftermath of the incident thought to do so either1; the tendency of individuals to respond inappropriately in the face of a medical emergency is a well-documented phenomenon, but this has not been accounted for in the development of laboratory safety policies57,58. Or, as Hill and Finster put it, 「It’s easy to blame the individual and not consider why the person acted in this way」59. This attitude seems to be widespread across the profession. In their study of the contributing factors to lab accidents, Hellman and co-workers interviewed chemistry educators, who reported that most accidents happen because students are careless and do not listen to instructions20. The tendency towards victim-blaming has often led to a perception of post-incident investigations as punitive rather than learning experiences39, thus increasing negative attitudes towards safety policies and procedures, poisoning the attitudes of generations of students and increasing the under-reporting rate.

Safety training research

Attitudes and beliefs about lab safety may be shaped as early as during undergraduate study, or even earlier60, and there have been numerous calls for safety training to be incorporated into the undergraduate curriculum in a more meaningful way13,61,62,63. Consequently, there is slightly more research on safety training for undergraduate students enrolled in teaching labs compared to research labs61,64,65,66,67,68,69,70,71,72,73,74.

Several studies have looked at program-wide safety initiatives that incorporated a variety of strategies. Many of these published studies have been done at small primarily undergraduate institutions (PUIs) rather than at large research-focused schools, although there are a few notable exceptions15,73,75,76. Common elements to these programmes include handouts, didactics, the creation of safety databases, self-study programmes, laboratory exams and quizzes, use of safety contracts and the creation of safety-focused courses at some universities63,65,66,67,73,77. Other researchers have looked at the use of more specific strategies for enhancing safety, including safety planning documents78, black lights to demonstrate issues of lab cleanliness71, scavenger hunts as a training strategy64, student safety teams72 and personalized safety videos68. Others have made comparisons between online and in-person safety training programmes61. Shariff and Norazahar had students report on their peers』 safety behaviours and found a reduction in all of the issues researchers identified (for example, use of PPE, keeping work spaces clean, horseplay) over the study time period79, with the exception of cell phone use.

Research on the safety beliefs and practices of undergraduate students is important. Our concern is that many undergraduates quickly learn to see safety training as an institutionally mandated hassle. The risks are often minimal by design, and the information is provided out of context and can appear to be overly restrictive and possibly silly. This negative attitude towards lab safety may be of little consequence in undergraduate teaching labs that are carefully controlled and involve few hazardous reagents. However, should these students continue on to graduate school and further work in academic research labs, a casual disregard towards safety may be a much greater liability when they are working with more dangerous reagents and processes. First impressions are extremely important and can cement attitudes and approaches early.

Unfortunately, much of the existing research on safety training in undergraduates is of questionable validity with regards to evidence-based policy-making on a wider scale or outside of the setting in which the data was gathered. These studies rarely include control groups or randomization to the intervention to ensure that observed changes are due to the programme alone and not to other factors. Studies also typically examine the combined effect of several initiatives simultaneously making it impossible, if there is any measurable change, to disentangle the causative contribution of the interventions. These studies generally do not include pre-and post-measures to assess the efficacy of the program. Because they lack this behavioural follow-up, it is unclear which safety training interventions lead to increased knowledge, better retention, increased compliance with safety rules, a decreased rate of incidents or better results in laboratory safety inspections. The one study that did look at the effects of specific intervention sessions found increases in (self-reported) safety knowledge, safety perception and safety attitude, but crucially, no corresponding increase in safety behaviour70.

The safety programs are usually implemented only at one institution, meaning that it is unclear whether these initiatives would be feasible or applicable elsewhere, and the research is conducted by those who established the safety protocols, making results susceptible to bias. Our intention in critiquing the methodology of these studies is not to throw the metaphorical baby out with the bathwater. The efforts of these researchers who have taken time out of their primary research programme to investigate these issues are laudable, and we would not expect professors whose primary expertise is in the natural sciences to have a detailed knowledge of psychological or pedagogical research methods. Our point is rather that collaboration with social scientists may be the key to building and improving on this area of the research literature and that methodological errors can be readily addressed in the study design, with no more effort on the part of the researchers. At the same time, social scientists, who are not experts in the materials and processes employed in the laboratory setting, need to work with the natural scientists to design the studies and the interventions. This research, by its very nature, requires an interdisciplinary approach.

There have been far fewer studies about implementing safety improvements in academic research labs. Some papers have been published on how to perform safety investigations80, how to learn from close-calls81,82 and how to design hazard-analysis systems83, all indicating that these basic practices are not universally implemented. Others have reported on risk reduction strategies, including systematic approaches to safety management and risk assessment83, the use of facility login software84, the creation of a chemical safety library85 and the creation of a website designed to share safety information and accident findings15. A collaboration between the University of Minnesota and Dow Chemical Company resulted in a safety initiative that included regular lab tours with reports to PIs and Laboratory Safety Officers (LSOs), the use of posters advertising PPE usage and Standard Operating Procedure-compliance, regular communication of safety updates via e-mails and website updates, a 『cleanup』 week and training regimen for LSOs76. Staehle and colleagues studied the implementation of behavioural strategies in a research lab, including twice-daily inspections by lab members, the use of discussions and quizzes during lab meetings and the use of an overnight reaction form75. Huising and Silbey described the 5-year implementation of a comprehensive management system involving lab inspection teams, the registration of PIs』 labs on a database and completion of safety training courses35. Other publications on safety training have looked at how to train staff56, including custodial and maintenance workers who work in laboratories86.

Again, the same methodological issues that plague research on undergraduate safety programmes are also true for academic research labs (for example, lack of control groups and randomization to interventions, inclusion of several interventions at once, no measurement of objective data such as accident frequency or inspection violations) and make interpretation and generalizability of results questionable. We have not been able to locate any studies or articles investigating how PIs train their research personnel, or on how PIs report that they themselves were trained. This lack of research with regards to safety training in academic settings means that in most instances, safety training is a product of institutional memory, anecdotes and, as James Gibson from the Office of Environment, Health, and Safety at UCLA put it, the 『application and misapplication of common sense』 rather than guided by a standard evidence base87. Throughout the investigative report on Sangji’s death at UCLA, it was clear that training was largely conducted through informal interactions and the passing along of knowledge. Although this is an essential component of training and knowledge, it should supplement rather than replace the use of formal training, institutional and laboratory-specific standard operating procedures, protocols and information from manufacturers, professional societies and compendia of reagents. This informal approach to training is particularly troubling because the knowledge being passed down may not conform to best practices, as was clear from the report on Sangji’s death1.

A few studies have been done about perceptions around safety training by academic researchers. In a survey of 2,400 researchers led by Nature and UCLA, 60% of respondents reported having received safety training on specific hazards or reagents21. Schröder and co-authors found that 70% of researchers in academic settings received safety training, but only 26% were trained within 30 days of starting experiments (the average length of the gap between starting work and receiving appropriate training was not reported but is certainly worrying)45. This training was usually conducted by Environmental Health and Safety Officers, with only 35% of participants saying that they had had additional training from their PI. In a smaller study of 85 participants22, 47% of participants did not know how often safety inspections were performed in their labs, 35% did not have access to data or records regarding their lab’s safety and whether or not it complied with legislated requirements and 9% did not know how to handle an emergency such as a fire or a spill. Again, an additional concern here is that these results reflect the self-perception of participants that they could handle a fire or spill, not an objective evaluation of their capacity to do so. Another study found that 25% of researchers had not been trained in the specific hazard with which they worked45. One study showed that only 10% of students, post-doctoral fellows, faculty and staff felt that their safety training had prepared them to assist others and to intervene when others engaged in unsafe behaviours76.

Research on safety training, or lack thereof, stands in stark contrast to findings suggesting that many researchers feel their lab is a safe environment. Although anywhere from 15–30% of researchers report having been involved in an accident or having sustained an injury, and that a large percentage also say that they have not received adequate or timely safety training, most studies have shown that researchers report feeling safe in their labs, perceive the risk level in their laboratories as low and describe their institution as having a good safety culture. What are we to make of this discrepancy between objective injury data and subjective feelings of safety? Our interpretation is that risky practices and a cavalier attitude toward safety are so normalized within academia that the low standards in the field are not troubling or even apparent to those on the inside.

Barriers to safety research

Researchers repeatedly state that laboratory safety is important, but knowledge about laboratory safety has not improved over the past decades22. Faculty attitudes appear to be one of the main barriers to change, dating back to the establishment of the first modern academic labs in 1840s. According to Kekulé, in the 1840s, Liebig welcomed him into his lab as a graduate student by saying: 「If you want to become a chemist…you have to ruin your health. Who does not ruin his health by his studies, nowadays will not get anywhere in Chemistry」88. The belief that injuries, accidents and near-misses are 『just part of the job』 remains common and current across the profession20,21,89,90. Scherz commented on the element of rebel as scientific hero portrayed in the autobiographies of some prominent molecular biologists91. This attitude towards risk has some corroboration in the research. In a series of interviews conducted with chemistry educators, participants reported that minor accidents were of no importance20. The idea of working alone in the lab at all hours of the day, every day of the year is still considered by some academic supervisors to be a positive and desirable attitude in trainees19. (Note that Sangji’s accident occurred on December 29, during the winter holiday break).

Many commentators have remarked that the most important barrier to the initiation and implementation of comprehensive safety programming is the attitudes of PIs16,36,92,93,94,95,96. In the Nature/UCLA survey of 2,400 researchers, the most commonly identified barriers to improving lab safety included 『time and hassle』, apathy, lack of understanding of safety requirements, lack of leadership, and a focus on regulatory compliance21. Other faculty members have cited a lack of knowledge, a lack of funding and disagreements about safety policies as reasons for non-adherence to mandatory policies9. In one study of lab personnel, barriers to safety improvements included time and hassle factors, apathy, inadequate training and competing priorities22. The issue of 『academic freedom』 is often raised as an objection to safety practices. One study found that 15% of researchers believed that safety regulations interfered with productivity, and 23% believed that they impeded the scientific discovery process45.

The community has made definite efforts to improve policy in the past 10 years. This includes the ACS’s adoption of safety as a core value, a mandate that safety information be included in journal articles as appropriate, the appointment of a manager of safety services and an effort to improve safety education across the educational spectrum. This is laudable, but it is unclear whether these efforts are making an objective difference. Many of the studies we have cited throughout this article regarding safety behaviour are all recent and refer to post-UCLA incident practice. Despite efforts from the ACS and other bodies, it is unclear whether this change in attitude is being integrated into the practice of academic chemical research, especially when one considers the high levels of resistance to safety measures observed by Schröder and colleagues in 201645.

Seemingly negative attitudes towards safety practice may be due to a perception by many scientists that workplace inspections are focused more on procedures and regulatory compliance than with a true concern for laboratory safety45,59,78,87. As Kapin pointed out, 「health and safety programs for laboratories are typically oriented around specific regulatory requirements, even though hazards in laboratories seldom respect these boundaries78」. Possibly one of the major issues may be the perception of EHS officers as academic 『interlopers』91, who are not seen by researchers as having the practical experience necessary to elicit compliance with their recommendations. EHS officers working at UCLA at the time of Sangji’s accident reported that they were aware of inconsistent use of PPE in research labs but had no power to impose sanctions or address non-compliance1.

Simply put, the academic discipline does not prioritize safety. Following the UCLA incident, Langerman made a number of recommendations for how safety issues could be addressed by students, faculty, laboratory staff, environmental health and safety officers, funding agencies, professional societies and the ACS97. His growing exasperation over time was obvious to readers7,97,98, and he recommended ensuring the compliance of recalcitrant PIs by taking accident reports, laboratory investigations, and safety policy compliance into account for promotion and tenure and the allocation of departmental resources. He also suggested that grant funding and prizes should be denied to PIs with poor safety records. Although these ideas have been taken up by other commentators91, so far these recommendations have not been implemented. We are not aware of any PIs who have been terminated or denied funding because of a poor safety record, so these policy recommendations have not appeared to have any influence to date.

A call for action

The state of academic safety research is unconscionable and cannot be allowed to continue. Data is required to develop evidence-based policies to address each of the issues we have raised in this article. Currently, there is no central database or organization responsible or funded to collect and analyse the annual number and characteristics of accidents in academic research labs. Data is needed not only on the headline-grabbing accidents that result in fatalities or hospitalizations, but on any close-calls, regardless of the occurrence of injury or significant property damage as the differences between near-misses and major catastrophes may be primarily due to good luck rather than good management. We need to know how big of a problem underreporting of accidents is and what factors are associated with underreporting. Data of this type would enable the identification of variables associated with accident frequency and severity and would help to determine the most appropriate countermeasures.

We need more information about the causes of academic lab accidents. We need to know how, where, when and to whom accidents happen. We need to know what contributes to accidents at the level of the individual, the lab, the department and the institution. We need to know the impact of these accidents on the victim, their friends and family members, their labmates, fellow students, faculty and staff, and the wider academic culture and institutional community. Do students drop out? Do they change their career plans? Are there mental-health repercussions for students, labmates, faculty members, staff members or other members of the university and research community? Is the climate different in universities where there have been multiple incidents? We have none of this information.

We need to delve further into attitudes and beliefs about safety. We need to know how these are correlated with demographic variables, training and lab experiences. We need to know how safety attitudes develop and how and when to intervene such that students view safety as a fundamental priority within science rather than a hassle. We need to know how beliefs and attitudes relate to behavioural practices with regards to PPE usage and risk assessment and how best to address discrepancies to keep personnel safe.

We need more research into safety training. There are two key domains of inquiry here. The first focuses on process: How is training currently done at different institutions? Are there more effective strategies for conveying the content? How should comprehension be evaluated? Under what circumstances is information retained? The second focuses on content: What should the content include? How should new situations be evaluated for safety? Most importantly, both strands of the research need to converge on the most important question: how do training interventions correlate with frequency and severity of accidents in labs?

However, while actuarial data and guidance in training procedures would be of great benefit, it will not necessarily help address the fundamental problem of culture, that is, the 『fiefdoms』 so omnipresent in academic settings99. We need to identify the barriers that prevent the systematic acceptance of the necessity for the learning and application of safety principles among students, faculty and staff. We need to understand what interventions, rewards and sanctions are required to overcome these barriers. We need to better understand the social scripts around scientific identity and culture. We need to understand how best to implement meaningful and impactful safety training starting in first year undergraduate level (or earlier) and how to build upon it continually throughout the degree and into graduate and postdoctoral training and faculty mentoring. We need to use proper methodology to determine the effectiveness of the training methodology and look at quantifiable outcomes. We need to determine how to address inherent challenges to safety research and training in the academic setting, such as high turnover of staff and students56.

Conclusions

Currently, there are 45 universities in Canada offering graduate chemistry programmes with a total of approximately 880 research groups in chemistry departments based on a hand count conducted on December 29, 2018. There are 432 research intensive (R1, R2 and R3) universities in the United States; although a count of research groups would be challenging, we would expect around 10,000 research groups. As graduate education spreads around the world, and with the exponential growth of programmes in China and India, the number of individuals involved in academic chemical research is set to expand. Currently, we are operating completely in the dark with regards to safety policies in both training and practice. We do not even know how many people are hurt every year and how badly, nor how great the damages are to laboratories, buildings and equipment. We simply have no idea about the scale of the issue — on the very day this section of the article was written, three graduate students were killed in a research-lab explosion at Beijing Jiaotong University9, and while the article was under review, another accident occurred at UCLA that involved a brief hospitalization100.

The benefits of establishing academic lab safety research programmes would be substantial. Ultimately, the goal would be to decrease the rate and severity of accidents in academic labs, ensure that lab personnel stay safe and healthy, and that equipment, laboratory and buildings are protected. This is also likely to have a spillover effect into industry — better-trained, more safety-conscious students would make better industrial employees. Undoubtedly, there would be financial savings related to the cost of accidents, insurance rates and lawsuits.

Despite calls for safety studies to form a central part of chemical research including tenure-track positions at major research universities63, and an increased understanding and interest in chemical safety studies by experimental research professors15, we could not identify any scientist whose principle mandate was the study of chemical safety. At present, we know of no tenure-track positions in science safety at any global research university. Despite the need for these positions, we are not optimistic that the university community will address this situation, but sincerely hope to be proven wrong. If action is not taken soon, academic chemical research may come to be seen as too risky for some institutions from a liability perspective: if we as a discipline do not take action, action may well be taken for us.

Scientific research dealing with new methods, new materials and a constant influx of new and inexperienced trainees will always be potentially hazardous, but we must do what we can to manage those risks that can be managed. In 2009, Langerman said, 「I have come to the disheartening conclusion that most academic laboratories are unsafe venues for work or study. I have concluded that only by a major change in the way we practice laboratory safety can we improve the situation」97. In his profile of major incidents in research labs, he noted with depressing repetitiveness that in most cases, virtually identical incidents occurred at the same institutions within 10–15 years, resulting in the destruction or temporary closure of the buildings98. More than ten years after Sangji’s avoidable and tragic death, we have not made nearly enough progress into understanding and addressing academic lab safety issues. We hope a ten-year follow-up to this review will conclude differently.

References

1.

Baudendistel, B. Investigation Report University of California, Los Angeles, Case No. S1110-003-09 (Department of Industrial Relations, Division of Occupational Safety and Health, Los Angeles, 2009).

2.

Technical Bulletin AL-134: Handling air-sensitive reagents (Sigma Aldrich, 2012).

3.

Benderley, B. L. California investigation report explains what went wrong for Sangji. Science http://blogs.sciencemag.org/sciencecareers/2012/01/yesterday-we-pu.html (2012).

4.

Allen, K. A young lab worker, a professor and a deadly accident. Toronto Star (2014); https://www.thestar.com/news/world/2014/03/30/a_young_lab_worker_a_professor_and_a_deadly_accident.html

5.

Benderly, B. L. Danger in school labs. Sci. Am. 303, 18–20 (2010).

Article

PubMed

PubMed Central

Google Scholar

6.

Grabowski, L. E. & Goode, S. R. Review and analysis of safety policies of chemical journals. J. Chem. Health Saf. 23, 30–35 (2016).

Article

Google Scholar

7.

Langerman, N. Warning to all principal investigators. J. Chem. Health Saf. 19, 42–43 (2012).

Google Scholar

8.

Kemsley, J. N. University of Hawaii fined $115,500 for lab explosion. Chem. Eng. News (2016); http://cen.acs.org/articles/94/web/2016/09/University-Hawaii-fined-115500-lab.html

9.

Pinghui, Z. Three students die in blast at Beijing university laboratory. South China Morning Post (2018); http://www.scmp.com/news/china/society/article/2179543/three-students-die-blast-beijing-university-laboratory

10.

Van Noorden, R. A death in the lab. Nature 472, 270–271 (2011).

Article

CAS

PubMed

PubMed Central

Google Scholar

11.

Texas Tech University laboratory explosion (U.S. Chemical Safety and Hazard Investigation Board, 2010).

12.

Kemsley, J. N. 10 years after Sheri Sangji’s death, are academic labs any safer? Chem. Eng. News (2018).

13.

Hunter, K. et al. Guidelines for chemical laboratory safety in academic institutions (American Chemical Society, 2016); https://www.acs.org/content/dam/acsorg/about/governance/committees/chemicalsafety/publications/acs-safety-guidelines-academic.pdf

14.

Kaufman, J. A. Killed in lab accidents: Memorial Wall. Lab Safety https://www.labsafety.org/memorial-wall (2019).

15.

Miller, A. J. M. & Tonks, I. A. Let’s talk about safety: Open communication for safer laboratories. Organometallics 37, 3225–3227 (2018).

Article

CAS

Google Scholar

16.

Young, J. A. How 「safe」 are the students in my lab? Do teachers really care. J. Chem. Educ. 60, 1067–1068 (1983).

Article

Google Scholar

17.

Accidents in waiting. Nature 472, 259 (2011).

18.

Jorgensen, E. F. Development and psychometric evaluation of the Research Laboratory Safe Behavior Survey (RLSBS). J. Chem. Health Saf. 24, 38–43 (2017).

Article

Google Scholar

19.

Peplow, M. & Marris, E. How dangerous is chemistry? Nature 441, 560–561 (2006).

Article

CAS

PubMed

PubMed Central

Google Scholar

20.

Hellman, M. A., Savage, E. P. & Keefe, T. J. Epidemiology of accidents in academic chemistry laboratories. Part 1. Accident data survey. J. Chem. Educ. 63, A267 (1986).

Article

Google Scholar

21.

Van Noorden, R. Safety survey reveals lab risks. Nature 493, 9–10 (2013).

Article

CAS

PubMed

PubMed Central

Google Scholar

22.

Ayi, H.-R. & Hon, C.-Y. Safety culture and safety compliance in academic laboratories: A Canadian perspective. J. Chem. Health Saf. 25, 6–12 (2018).

Article

Google Scholar

23.

Simmons, H. E., Matos, B. & Simpson, S. A. Analysis of injury data to improve safety and training. J. Chem. Health Saf. 24, 21–28 (2017).

Article

Google Scholar

24.

Sieloff, A. C., Shendell, D. G., Marshall, E. G. & Ohman-Strickland, P. An examination of injuries and respiratory irritation symptoms among a sample of undergraduate chemistry students from a Public Northeastern University. J. Chem. Health Saf. 20, 17–26 (2013).

Article

Google Scholar

25.

Probst, T. M., Barbaranelli, C. & Petitta, L. The relationship between job insecurity and accident under-reporting: A test in two countries. Work Stress 27, 383–402 (2013).

Article

Google Scholar

26.

Rathman, T. L. & Schwindeman, J. A. Preparation, properties, and safe handling of commercial organolithiums: Alkyllithiums, lithium sec-organoamides, and lithium alkoxides. Org. Process Res. Dev. 18, 1192–1210 (2014).

Article

CAS

Google Scholar

27.

Mikula, H. et al. Practical and efficient large-scale preparation of dimethyldioxirane. Org. Process Res. Dev. 17, 313–316 (2013).

Article

CAS

Google Scholar

28.

Morandi, B. & Carreira, E. M. Iron-catalyzed cyclopropanation in 6 M KOH with in situ generation of diazomethane. Science 335, 1471–1474 (2012).

Article

CAS

PubMed

PubMed Central

Google Scholar

29.

Busura, S., Khan, F., Hawboldt, K. & Iliyas, A. Quantitative risk-based ranking of chemicals considering hazardous thermal reactions. J. Chem. Health Saf. 21, 27–38 (2014).

Article

Google Scholar

30.

Frontier, A. Laboratory techniques and methods to improve your experimental skills. Not Voodoo http://chem.chem.rochester.edu/~nvd/index.php (2019).

31.

Lowe, D. How not to do it: Tertiary butyllithium. Science Mag https://blogs.sciencemag.org/pipeline/archives/2007/03/01/how_not_to_do_it_tertiary_butyllithium (2007).

32.

Snyder, S. A. Essential Reagents for Organic Synthesis (eds Fuchs, P., Bode, J., Charette, A. & Rovis, T) (Wiley, 2019).

33.

Urben, P. G. Bretherick’s Handbook of Reactive Chemical Hazards 7th edn (Elsevier, 2017).

34.

Bertozzi, C. R. Ingredients for a positive safety culture. ACS Cent. Sci. 2, 764–766 (2016).

Article

CAS

PubMed

PubMed Central

Google Scholar

35.

Huising, R. & Silbey, S. S. Constructing consequences for noncompliance: The case of academic laboratories. Ann. Am. Acad. Pol. Soc. Sci. 649, 157–177 (2013).

Article

Google Scholar

36.

Hendershot, D. C. Process safety: Is safety 「common sense」? J. Chem. Health Saf. 19, 35–36 (2012).

Article

Google Scholar

37.

Kemsley, J. N. Learning from UCLA. Chem. Eng. News 87, 29–34 (2009).

Google Scholar

38.

Schmidt, H. Anatomy of an incident—Multiple failure of safety systems under stress. J. Chem. Health Saf. 25, 6–11 (2018).

Article

Google Scholar

39.

Cournoyer, M. E., Trujillo, S., Lawton, C. M., Land, W. M. & Schreiber, S. B. Anatomy of an incident. J. Chem. Health Saf. 23, 40–48 (2016).

Article

Google Scholar

40.

Phifer, R. Case study – Incident investigation: Laboratory explosion. J. Chem. Health Saf. 21, 2–5 (2014).

Article

Google Scholar

41.

Reason, J. The contribution of latent human failures to the breakdown of complex systems. Philos. Trans. R. Soc., B 327, 475–484 (1990).

Article

CAS

Google Scholar

42.

Young, J. A. How complacency can jeopardize safety. Chem. Health Saf. 6, 5 (1999).

Article

Google Scholar

43.

Wu, T.-C., Liu, C.-W. & Lu, M.-C. Safety climate in university and college laboratories: Impact of organizational and individual factors. J. Saf. Res. 38, 91–102 (2007).

Article

Google Scholar

44.

Steward, J. E., Wilson, V. L. & Wang, W.-H. Evaluation of safety climate at a major public university. J. Chem. Health Saf. 23, 4–12 (2016).

Article

Google Scholar

45.

Schröder, I., Huang, D. Y. Q., Ellis, O., Gibson, J. H. & Wayne, N. L. Laboratory safety attitudes and practices: A comparison of academic, government, and industry researchers. J. Chem. Health Saf. 23, 12–23 (2016).

Article

Google Scholar

46.

McEwen, L., Stuart, R., Sweet, E. & Izzo, R. Baseline survey of academic chemical safety information practices. J. Chem. Health Saf. 25, 6–10 (2018).

Article

Google Scholar

47.

King, M. F. & Bruner, G. C. Social desirability bias: A neglected aspect of validity testing. Psychol. Market. 17, 79–103 (2000).

Article

Google Scholar

48.

Edwards, A. L. The social desirability variable in personality assessment and research. (Dryden Press, 1957).

49.

Wardlaw, M. J. Three lessons for a better cycling future. BMJ 321, 1582–1585 (2000).

Article

CAS

PubMed

PubMed Central

Google Scholar

50.

Finkelstein, E. A., Strombotne, K. L., Chan, N. L. & Krieger, J. Mandatory menu labeling in one fast-food chain in King County, Washington. Am. J. Prev. Med. 40, 122–127 (2011).

Article

PubMed

PubMed Central

Google Scholar

51.

Ménard, A. D., Houser, C., Brander, R. W., Trimble, S. & Scaman, A. The psychology of beach users: Importance of confirmation bias, action, and intention to improving rip current safety. Nat. Hazards 94, 953–973 (2018).

Article

Google Scholar

52.

Bretherick, L. Chemical laboratory safety: The academic anomaly. J. Chem. Educ. 67, A12 (1990).

Article

CAS

Google Scholar

53.

Hill, R. H. Make safety a habit! J. Chem. Health Saf. 25, 12–17 (2018).

Article

Google Scholar

54.

Darley, J. M. & Latane, B. Bystander intervention in emergencies: Diffusion of responsibility. J. Person. Soc. Psychol. 8, 377–383 (1968).

Article

CAS

Google Scholar

55.

Leggett, D. J. Identifying hazards in the chemical research laboratory. Process Saf. Prog. 31, 393–397 (2012).

Article

CAS

Google Scholar

56.

Stuart, R. Emergency response training for laboratory workers. J. Chem. Health Saf. 17, 29–32 (2010).

Article

CAS

Google Scholar

57.

Mogielnicki, R. P., Stevenson, K. A. & Willemain, T. R. Patient and bystander response to medical emergencies. Med Care 13, 753–762 (1975).

Article

CAS

PubMed

PubMed Central

Google Scholar

58.

Shotland, R. L. & Heinold, W. D. Bystander response to arterial bleeding: Helping skills, the decision-making process, and differentiating the helping response. J. Person. Soc. Psychol. 49, 347–356 (1985).

Article

CAS

Google Scholar

59.

Hill, R. H. & Finster, D. C. Academic leaders create strong safety cultures in colleges and universities. J. Chem. Health Saf. 20, 27–34 (2013).

Article

Google Scholar

60.

West, S. S., Westerlund, J. F., Stephenson, A. L., Nelson, N. C. & Nyland, C. K. Safety in science classrooms: What research and best practice say. Educ. For. 67, 174–183 (2003).

Google Scholar

61.

Withers, J. H., Freeman, S. A. & Kim, E. Learning and retention of chemical safety training information: A comparison of classroom versus computer-based formats on a college campus. J. Chem. Health Saf. 19, 47–55 (2012).

Article

Google Scholar

62.

Nelson, D. A. Incorporating chemical health and safety topics into chemistry curricula: Past accomplishments and future needs. Chem. Health Saf. 6, 43–48 (1999).

Article

Google Scholar

63.

Fivizzani, K. P. Where are we with lab safety education: Who, what, when, where, and how? J. Chem. Health Saf. 23, 18–20 (2016).

Article

Google Scholar

64.

Wood-Black, F. Incorporating safety into the general chemistry curriculum. J. Chem. Health Saf. 21, 14–21 (2014).

Article

Google Scholar

65.

Crockett, J. M. Laboratory safety for undergraduates. J. Chem. Health Saf. 18, 16–25 (2011).

Article

CAS

Google Scholar

66.

Bradley, S. Integrating safety into the undergraduate chemistry curriculum. J. Chem. Health Saf. 18, 4–10 (2011).

Article

CAS

Google Scholar

67.

Burchett, S., Pfaff, A., Hayes, J. & Woelk, K. Exploding misconceptions: Developing a culture of safety through learner driven activities. J. Chem. Health Saf. 24, 36–42 (2017).

Article

Google Scholar

68.

Matson, M. L., Fitzgerald, J. P. & Lin, S. Creating customized, relevant, and engaging laboratory safety videos. J. Chem. Educ. 84, 1727–1728 (2007).

Article

CAS

Google Scholar

69.

Karapantsios, T. D., Boutskou, E. I., Touliopoulou, E. & Mavros, P. Evaluation of chemical laboratory safety based on student comprehension of chemicals labelling. Ed. Chem. Eng. 3, e66–e73 (2008).

Article

Google Scholar

70.

Reniers, G. L. L., Ponnet, K. & Kempeneers, A. Higher education chemical lab safety interventions: Efficacious or ineffective? J. Chem. Health Saf. 21, 4–8 (2014).

Article

Google Scholar

71.

Gallion, L. A., Samide, M. J. & Wilson, A. M. Demonstrating the importance of cleanliness and safety in an undergraduate teaching laboratory. J. Chem. Health Saf. 22, 28–31 (2015).

Article

Google Scholar

72.

Alaimo, P. J., Langenhan, J. M., Tanner, M. J. & Ferrenberg, S. M. Safety teams: An approach to engage students in laboratory safety. J. Chem. Educ. 87, 856–861 (2010).

Article

CAS

Google Scholar

73.

Kennedy, S. & Palmer, J. Teaching safety: 1000 students at a time. J. Chem. Health Saf. 18, 26–31 (2011).

Article

CAS

Google Scholar

74.

Makransky, G., Thisgaard, M. W. & Gadegaard, H. Virtual simulations as preparation for lab exercises: Assessing learning of key laboratory skills in microbiology and improvement of essential non-cognitive skills. PLoS ONE 11, e0155895 (2016).

Article

CAS

PubMed

PubMed Central

Google Scholar

75.

Staehle, I. O. et al. An approach to enhance the safety culture of an academic chemistry research laboratory by addressing behavioral factors. J. Chem. Educ. 93, 217–222 (2016).

Article

CAS

Google Scholar

76.

McGarry, K. A. et al. Student involvement in improving the culture of safety in academic laboratories. J. Chem. Educ. 90, 1414–1417 (2013).

Article

CAS

Google Scholar

77.

Ritch, D. & Rank, J. Laboratory safety in the biology lab. Bioscene 27, 17–22 (2001).

Google Scholar

78.

Kapin, J. M. Beyond chemical safety— an integrated approach to laboratory safety management. Chem. Health Saf. 6, 20–22 (1999).

Article

CAS

Google Scholar

79.

Shariff, A. M. & Norazahar, N. At-risk behaviour analysis and improvement study in chemical engineering laboratories. Int. J. Chem. Environ. Eng. 2, 51–55 (2011).

Google Scholar

80.

Wyllie, R., Lee, K., Morris-Benavides, S. & Matos, B. What to expect when you’re inspecting: A summary of academic laboratory inspection programs. J. Chem. Health Saf. 23, 18–24 (2016).

Article

Google Scholar

81.

Ferjencik, M. & Jalovy, Z. What can be learned from incidents in chemistry labs. J. Loss Prev. Process Ind. 23, 630–636 (2010).

Article

CAS

Google Scholar

82.

Young, J. A. The professional practice of chemical safety. Chem. Health Saf. 6, 41–42 (1999).

Article

CAS

Google Scholar

83.

Marendaz, J.-L., Friedrich, K. & Meyer, T. Safety management and risk assessment in chemical laboratories. CHIMIA 65, 734–737 (2011).

Article

CAS

PubMed

PubMed Central

Google Scholar

84.

Camino, F. E. Make safety awareness a priority: Use a login software in your research facility. J. Chem. Health Saf. 24, 22–25 (2017).

Article

Google Scholar

85.

Nitsche, C. I., Whittick, G. & Manfredi, M. Collecting reaction incident information: Engaging the community in sharing safety learnings. J. Chem. Health Saf. 25, 2–5 (2018).

Article

Google Scholar

86.

LaPierre, J. It’s 1:30 a.m.-Do you know who’s in your laboratories? Chem. Health Saf. 6, 31–33 (1999).

Article

Google Scholar

87.

Mulcahy, M. B. et al. College and university sector response to the U.S. Chemical Safety Board Texas Tech incident report and UCLA laboratory fatality. J. Chem. Health Saf. 20, 6–13 (2013).

Article

Google Scholar

88.

National Research Council. Prudent practices in the laboratory: Handling and disposal of chemicals (The National Academies Press, 1995).

89.

Bayer, R. Lab safety as a collateral duty in small colleges. J. Chem. Educ. 61, A259 (1984).

Article

Google Scholar

90.

Kaufman, J. A. Safety in the academic laboratory. J. Chem. Educ. 55, A337 (1978).

Article

CAS

Google Scholar

91.

Scherz, P. Risk, prudence and moral formation in the laboratory. J. Moral Educ. 47, 304–315 (2018).

Article

Google Scholar

92.

Ashbrook, P. Laboratory safety in academia. J. Chem. Health Saf. 20, 62 (2013).

Article

Google Scholar

93.

Ashbrook, P. C. Hazard assessment. J. Chem. Health Saf. 21, 35 (2014).

Article

Google Scholar

94.

Ashbrook, P. C. Accountability. J. Chem. Health Saf. 20, 48 (2013).

Article

Google Scholar

95.

Czornyj, E., Newcomer, D., Schroeder, I., Wayne, N. L. & Merlic, C. A. Proceedings of the 2016 Workshop Safety By Design – Improving safety in research laboratories. J. Chem. Health Saf. 25, 36–49 (2018).

Article

Google Scholar

96.

Backus, B. D. et al. Laboratory safety culture: Summary of the chemical education research and practice – Safety in chemistry education panel discussion at the 46th Midwest and 39th Great Lakes Joint Regional American Chemical Society Meeting, St. Louis, Missouri, on October 21, 2011. J. Chem. Health Saf. 19, 20–24 (2012).

Article

Google Scholar

97.

Langerman, N. Laboratory safety? J. Chem. Health Saf. 16, 49–50 (2009).

Article

Google Scholar

98.

Langerman, N. Reactive chemistry incidents in laboratories. J. Chem. Health Saf. 16, 23–26 (2009).

Article

CAS

Google Scholar

99.

McCroskey, J. C. in Teaching communication: Theory, research, and methods (eds Daly, J.A., Friedrich, G.W. & Vangelisti, A.L.) 471–479 (Erlbaum Associates, 1990).

100.

One injured in lab explosion at UCLA. Los Angeles Daily News https://www.dailynews.com/2019/02/05/one-injured-in-lab-explosion-at-ucla/ (2019).

文章首先提到了2008年12月29日發生在UCLA的事故，一個叫Sangji的實驗助手在處理大量的叔丁基鋰（tert-butyllithium，做有機的同學知道這個藥品的厲害吧！）的時候發生洩漏，並燒著了她的衣服，造成了二級燒傷，40%的身體被燒傷。三周之後，這個剛剛23歲的女生在醫院去世。造成這個事故的原因是多方面的。Sangji在進行試驗的時候沒穿實驗服並且沒有遵循試驗規範；她的導師Harran在明知道她沒有足夠經驗進行獨立試驗也沒有對她進行試驗培訓，並且危險品操作的規範實驗室也沒有。一個在Harran組裡的博後說，他好像有給過Sangji操作手冊，但是他自己也承認從沒有讀過並且沒有按照上面的規範去做。學校的政策上也沒有官方規定說在個人防護裝備的強制執行。事故發生時，Harran的另一層的實驗室正在裝修，使用的實驗室比他要求的要小30-40%，沒有足夠的空間來存放藥品。這個事故的14個月之前，UCLA隱藏了另外兩起非致命的實驗事故。儘管有經驗的研究人員會批評這種易著火化學物質所使用的技術，但沒有受過使用這種化學物質訓練的人是不能被指責的：在幾乎所有已發表的使用叔丁基鋰的研究中，很少說明使用這種化學物質的危險。2012年，美國化學會化學健康與安全部門的前主席Neal Langerman形容UCLA這個事故是多年來實驗室安全最嚴重的挑戰。應該從中學到教訓，從而進行安全研究方法的根本文化變革。本綜述旨在對當前的化學實驗室安全研究現狀進行批判性的審查，討論如何建立和實施本研究的結果，並要求重新審查化學實驗室安全研究的現狀，並對其在事故研究和預防中的作用進行重新審查和承諾。儘管搞化學的人都知道實驗室的事故和小危險經常發生，但是沒有研究者、大學或者專業組織統計每年的事故，沒有資料庫來顯示事故的類型和頻率。收集、整理和分析實驗室事故數據，以確定發生率可能屬於政府監管機構的權限範圍，但是無論是美國的職業安全與保健管理總署(OSHA)還是加拿大的13個專業和地區安全理事會都沒有曾經編譯過或分析過這個數據。雖然這項研究是不完整的，但它確實描繪了一幅令人不安的畫面。一個主要的問題是，對實驗室傷害的研究並沒有告訴我們總的事故發生率。不涉及傷害的事故比涉及受傷的事故更常見，但除非損失嚴重，否則很少有報導。此外，真正的事故發生率可能比這些結果所顯示的還要糟糕，因為有一些證據表明，漏報是科學上的一個重大問題。由於實驗事故報導的缺乏，對實驗事故的因素也因此非常匱乏，但是總的來說，與正在使用的材料或設備有關的風險、與從事研究的研究人員的技能、知識和選擇有關的風險、PI的特點和研究實驗室的質量以及部門或機構層面產生的風險因素。作者對之前的幾個組的實驗室人員進行調查的結果進行了描述，大家都對實驗安全非常重視。儘管這些研究是令人鼓舞的，並且可能與實驗室事故的發生有關，但抽象理念（如安全文化、安全氣候、安全領導、安全指導和主觀安全感）的作用是有限的，那些重視安全並認為工作場所安全的個人是否在實驗室實踐中做出了安全的選擇，這一點尚不清楚。到目前為止，還沒有關於安全態度與安全做法之間的關係的研究。與這些關於安全信念的樂觀結果不同，關於行為安全做法的研究結果令人關注。幾項研究的結果顯示研究人員在進行實驗之前不傾向於進行安全性評估。

實驗室安全管理 12/9~10 上海 Lab Safety 法規標準，安全設計，化學品風險，管理體系

通常情況下，人類因素的考慮往往集中於指責受害者導致或加重事故的行為，如Sangji的例子。雖然她的襯衫著火時，她並沒有想到要用實驗室淋浴，但事實上，在事件發生後幫助她的兩名博士後都沒有想到要這樣做。個人在醫療緊急情況下作出不適當反應的傾向是一個有充分證據的現象，但在制定實驗室安全政策時沒有考慮到這一點。或者，正如希爾和芬斯特所說的，「責怪個人是很容易的，而不考慮為什麼這個人這樣行事」。這種態度似乎在整個行業都很普遍。在他們對實驗室事故的貢獻因素的研究中，赫爾曼和同事訪談了化學教育者，他們報告說，大多數事故發生的原因是學生不小心，不聽從指導。指責受害者的趨勢往往導致對事件後調查的看法是懲罰性的，而不是學習經驗，因此增加了對安全政策和程序的消極態度，毒化了幾代學生的態度，提高了漏報率。有關實驗室安全的態度和信念可能早在本科學習期間，甚至更早塑造，許多人呼籲以更有意義的方式將安全培訓納入本科課程。因此，與研究室相比，對在教學實驗室註冊的本科生進行安全培訓的研究略有增加。研究本科生的安全信念和實踐是重要的。我們關注的是，許多本科生很快就學會了將安全培訓視為一種麻煩。從設計上看，風險往往是最小的，而且信息是在文本之外提供的，而且可能顯得過於限制性，而且可能是愚蠢的。這種對實驗室安全的消極態度，對於那些受到嚴格控制且幾乎不涉及危險試劑的本科教學實驗室來說，可能沒有多大影響。但是如果這些學生繼續讀研究生，並在學術研究實驗室繼續工作，當他們使用更危險的試劑和工藝時，對安全的隨意忽視可能是一個更大的責任。研究人員一再指出，實驗室的安全性很重要，但在過去幾十年裡，關於實驗室安全的知識沒有得到改善。教師的態度似乎是改變的主要障礙之一，可追溯到18世紀40年代建立的第一個現代學術實驗室。根據Kekulin的說法，在1840世紀40年代，Liebig歡迎他作為一名研究生進入實驗室時說：「如果你想成為一個化學家，你就要毀了你的健康。如果他的研究不破壞他的健康，現在將不會在化學領域中得到任何東西」。許多評論人士指出，啟動和實施全面安全方案的最重要的障礙是PI的態度。在2400名研究人員的Nature/UCLA調查中，最常見的改進實驗室安全的障礙包括「時間和麻煩」、冷漠、缺乏對安全要求的理解、缺乏領導能力，以及對法規遵從性的關注。簡單地說，學術規則並沒有把安全放在第一位。UCLA事件之後，Langerman做了關於學生、教員、實驗室工作人員、環境衛生和安全官員、資助機構、專業協會和ACS如何處理安全問題一系列的建議。他建議通過將事故報告、實驗室調查和安全政策與促進和保有權以及部門資源的分配相結合，確保PI遵守規範。他還建議，在安全記錄差的情況下，應拒絕給予PI資助和獎金。雖然這些想法已被其他評論者所採納，但迄今尚未執行這些建議。我們不知道因安全記錄不佳而被終止或拒絕提供資金的任何PI，因此這些政策建議似乎沒有任何影響。學術安全研究的現狀是不合情理的，不能允許繼續下去。數據是需要制定基於證據的政策，以解決我們在本文中提出的每一個問題。目前，沒有中央資料庫或組織負責或資助收集和分析每年發生在學術研究實驗室的事故數量和特點。不僅需要有關導致死亡或住院的重大事故的數據，還需要任何密切聯繫的數據，無論發生了什麼傷害或重大財產損失，因為小事故和重大災難之間的差異可能主要是因為運氣好而不是管理不善。我們需要知道事故報告不足的問題有多大，以及具體是什麼問題。我們需要更多關於學術實驗室事故原因的信息。我們需要知道發生事故的原因、人物、時間和地點。我們需要知道在個人、實驗室、部門和機構級別發生事故的原因。我們需要知道這些事故對受害者、他們的朋友和家庭成員、他們的同事、同學、教師和工作人員以及更廣泛的學術文化和機構社區的影響。學生輟學了嗎?他們改變他們的職業計劃了嗎？對學生、同事、教職員、工作人員或大學和研究界其他成員有精神健康的影響嗎？在有多重事故的大學裡，氣候是不同的嗎？我們沒有這方面的資料。我們需要深入探討有關安全的態度和信念。我們需要知道這些與人口統計變量、培訓和實驗室經驗之間的關係。我們需要知道安全態度是如何發展的，如何以及何時進行幹預，以便學生將安全視為科學中的一個基本優先事項，而不是一個麻煩。我們需要知道信念和態度如何與個人防護設備的使用和風險評估的行為實踐相關，以及如何最好地解決差異，以確保人員的安全。我們需要更多的研究進入安全培訓。這裡有兩個關鍵的調查領域。第一個重點是流程：當前如何在不同的機構進行培訓？是否有更有效的策略來傳達內容？如何評價理解？在什麼情況下保留了信息？第二個重點是內容：內容應包括哪些內容？如何評估新情況的安全性？最重要的問題是，研究的兩個股都需要集中在最重要的問題上：訓練幹預如何與實驗室事故的頻率和嚴重程度相關？然而，雖然精算數據和培訓程序方面的指導將有很大好處，但它不一定有助於解決文化的根本問題，即學術環境中無所不在的「領地」問題。我們需要找出阻止學生、教師和教職員工系統地接受學習和應用安全原則的障礙。我們需要了解克服這些障礙需要什麼樣的幹預、獎勵和制裁。我們需要更好地理解圍繞科學認同和文化的社會腳本。我們需要了解如何最好地實施有意義和有影響的安全培訓，從大學一年級(或更早)開始，以及如何以此為基礎貫穿在整個學位和研究生及博士後培訓和導師中。我們需要使用適當的方法來確定培訓方法的有效性，並看可量化的結果。我們需要決定如何應對固有的挑戰。作者最後提出，在Sangji慘死後的十多年裡，我們在理解和解決學術實驗室安全問題方面還沒有取得足夠的進展。我們希望這次評論的十年後續行動將取得不同的結論。（全文完，文章連結https://www.nature.com/articles/s41557-019-0375-x）****************************************************************近年來，高校的實驗室爆炸傷亡事故報告不斷，每一次都令人無比揪心，扼腕嘆息。那些不經意的細節，如果處理得當，我們本可以避免失誤，付出如此大的代價。2006年3月15日凌晨左右，復旦大學化學西樓一實驗室內突發爆炸，放置室內的試管、容器等相繼發生連鎖爆炸，所幸未釀成人員傷亡。2008年7月11日10時許，雲南大學北院雲南省微生物研究所5樓510實驗室，一名臨近畢業的博士生在做實驗時發生化學爆炸，該名博士生被嚴重炸傷。2009年10月23日下午1點多，北京理工大學新5號樓一實驗室發生爆炸，導致5人受傷，其中有一名實驗室負責老師、兩名學生和兩名設備調試工程師。2011年4月14日15時45分，四川大學江安校區第一實驗樓B座103化工學院一實驗室，3名學生在做常壓流化床包衣實驗，實驗物料意外爆炸，導致3名學生受傷。2011年12月7日上午11點左右，南開大學一名女生在做化學實驗時發生了意外，手部嚴重受傷。2013年4月30日上午9點左右，南京理工大學校內一廢棄實驗室拆遷施工發生意外爆炸，現場施工的4名工人2名重傷，2名輕傷，其中1名重傷人員經醫院搶救無效死亡。2014年12月4日中午11時左右，江蘇省常州工程學院合一樓化工系頂樓實驗室發生爆炸，現場一片狼藉，傷亡不詳。2015年4月5日中午，位於徐州的中國礦業大學化工學院一實驗室發生爆炸事故，致5人受傷，1人搶救無效死亡。2015年4月29日上午，安徽省淮北礦務局朱仙莊礦中學的實驗室突然發生爆炸，事故造成3名教師受傷。2015年12月18日10點，清華大學化學系實驗室發生一起爆炸事故，一名博士研究生在實驗室內使用氫氣做化學實驗時發生爆炸，後被確認身亡。2015年6月17日下午16:30分左右， 蘇州大學物理樓二樓實驗室在處理鋰塊時發生爆炸，蘇州消防調集7輛消防車參與救援，無人員受傷。2016年1月10日中午北京化工大學科技大廈一間實驗室內又突然著起了火。不過幸運的是，現場無人員傷亡。2016年9月21日，位於松江大學園區的東華大學化學化工與生物工程學院一實驗室發生爆炸，兩名學生受重傷，一名學生受輕微擦傷，無教師受傷。2017年3月27日晚，復旦大學邯鄲校區的化學西樓一間實驗室發生爆炸，現場一名20歲男性傷及雙上肢。2018年12月26日，北京交通大學東校區實驗室爆炸，3名學生死亡。面對這些，我想高校管理、實驗室負責人、導師、學生都應該反思，到底是哪個環節出錯了？我們也可以從這篇評論中學到一些東西。面對事故，高校是否設計了預案，實驗室負責人是否強調實驗安全，實驗室消防設備是否齊全，消防演練是否定期進行，逃生通道是否通暢，安全隱患是否定期排查，學生是否會用滅火器，實驗操作是否規範？加強實驗安全宣傳和教育，怎麼重視都不為過，畢竟文章再重要，文憑再重要，也比不上我們的生命寶貴！最後，希望各位在為科研奮鬥的同時，要注意實驗安全！

做實驗之前進行安全性評估，包括藥品安全性和操作規則；

當你的實驗室同學做實驗沒穿實驗服的時候，去溫柔的提醒一下；

當Ta的操作不規範的時候，去幫助一下；

遇到不會的操作時，要虛心向別人請教；

做危險實驗，必要時請別人幫忙；

涉及到實驗放出有毒氣體時，開啟通風櫥，提醒別人迴避；

實驗室是我們共同的工作場所，共同維護實驗安全，利己利人！實驗室安全管理Lab Safety Workshop

中國上海

12/9~12/10/2019

 

課程特點

實驗室事故案例分享；

實驗室的合規的疑惑與方案；

實驗室設計的安全、健康、環保設施設計；

實驗室化學品的全周期管理;

實驗室新實驗的風險管控；

豐富的工具、數據、圖片和影像資料；

合理的練習，確保學以致用

誰應該參加

GM，OM

EHS總監、EHS經理、安全主管和安全工程師

實驗室管理人員、實驗員、試劑師、設施部、維修部經理、主管、工程師

 

第一部分：引言

實驗室的經典案例分析；

實驗室安全管理的特點及挑戰；

第二部分：研發中心&實驗室適用的EHS法律法規

《安全生產法》的實驗室適用性

實驗室新改擴項目『三同時』

實驗室化學品儲存場所法規要求

實驗人員職業健康監護

實驗室的廢水/廢氣/危廢合規處理

練習

第三部分：研發中心&實驗室安全設施設計及配置

實驗室各功能房間平面及縱向布置原則

實驗室通風系統設置

實驗臺的人機工程學設計

實驗室氣體檢測系統設計

實驗室各房間的壓差設計原則

事故排風設置

練習

第四部分：研發中心&實驗室化學品風險及管理實踐

化學品GHS分類及標籤

化學品採購EHS考慮事項

化學品道路運輸安全

化學品現場存儲安全

化學品內部運輸/轉移安全

化學品使用安全

通風櫥的選擇與管理

化學品廢棄物處理

第五部分：研發中心&實驗室安全管理體系

實驗室安全技術委員會

實驗室人員能力矩陣

實驗室行為安全觀察

實驗室新實驗評估&分析

實驗室安全績效指標體系

 

費用為4980元人民幣/人（僅包含培訓與午餐費用，其他交通住宿晚餐自理）報名聯繫[e-mail]或021-69980278

主辦方   : EHSCity Development Co., Ltd

新環康安(深圳)發展有限公司

EHSCity Investment Management Co., Ltd

上海安進投資管理有限公司