Advances in Safety Management and Human Performance

Table of Contents

1. Frontmatter

2. Resilience and Human Performance

   1. Frontmatter

   2. Examination of Design and Human Factors Supporting Sensemaking, Resilience and Performance in the Ship Accident Helge Ingstad in Norway

      Stig O. Johnsen, Brit-Eli Danielsen

      Abstract

      We examine design and Human Factors supporting sensemaking, resilience and performance on the bridge of the frigate HNoMS Helge Ingstad that collided with the tanker Sola TS in 2018. We are presenting a framework to evaluate the design based on a system perspective. The frigate was a modern ship, with seven seamen on the bridge, performing operations during night. The officer of the watch considered Sola TS as a part of a land-based installation. The officer did not get clear indication from supporting systems on the bridge that Sola TS was a moving ship. The workload on the bridge was high with many alarms, officers in training, and communication through VHF. No personnel fatalities, but the frigate sank. The cost of a new frigate is estimated to be 1400 Mill USD. The accident analysis benefits from the sensemaking perspective, the design of bridge systems did not support reliable sensemaking.

   3. Trait Interindividual Differences in the Effectiveness of Modafinil

      Jeroen Van Cutsem, Emilie Dessy, Martine Van Puyvelde, Olivier Mairesse, Xavier Neyt, Nathalie Pattyn

      Abstract

      Purpose: To optimize the usage of modafinil as a cognitive enhancer and increase safety and effectiveness, we sought to assess the trait-characteristic of modafinil-sensitivity.

      Methods: 11 healthy participants (age = 21 ± 2 yr) were tested on 2 separate occasions during which they were sleep deprived for one night. During one trial they received 2 × 200 mg modafinil (i.e. EXP); during the other they received 2 × a placebo-capsule (i.e. CON). Physiological (e.g. heart rate and blood pressure), subjective (e.g. sleepiness and mood state) and behavioral (e.g. psychomotor vigilance task) measures of sleep-wake regulation were followed up.

      Results: Both PVT performance and perceived sleepiness were significantly improved in EXP at 2AM and 4AM during the sleep deprivation-night, compared to in CON. Additionally, an ICC of 0.90 for the delta (CON – EXP) in non-adjusted reaction time was observed.

      Conclusion: Stable and robust interindividual differences in modafinil-sensitivity are clearly present across day and night.

   4. Improvement of Workflow Structure to Prevent Human Error

      Toru Nakata

      Abstract

      Design of workflow is one of crucial factors for human reliability. When a procedure of tasks is badly designed, more troubles related to human errors will happen. This paper claims three essential points on design of workflows for human operations. First, independence of components of workflow is crucial for test of faultlessness: each component should be verified without interference of others. The second key is conservation of partial achievements, which protects preliminary achievements from disturbances. The last point is pre-fixation of execution order in order to inhibit omission. This paper deals this problem with examples at practical scenes in the industry.

   5. Building a Trojan Horse: Third Phase in the Experiment/Research with City Information Modeling (CIM) and the Design Ethics

      Gonçalo Falcão, José Beirão

      Abstract

      The Trojan Horse experiment is now on his third (of possible ten) phase. It’s an ongoing research connecting two researchers and two different knowledge fields – Architecture and Design. This research project was planed by both researchers, crossing academic fields and using the design concept of the Trojan Horse to develop a common project. In this early stage (we’re reporting the third test) the investigation is still inquiring the system, searching for positive tests and accessing their consistency, relating them with the results attained in the previous phases (one and two) with cumulative replicability. The ultimate goal is to improve the level of evidence without distorting the simulation results; this is done by extending the method initially designed by increasing the repetition factor by the value of one, tunning procedures and verifying the outcomes, while minimizing noise introduced by the diverse requirements from different compliances. Computer analysis will also be tested in it’s competence in stopping - or at least mitigating - human failures. Early results point to the need for tools that can better comply with the challenges posed by a voracious scientific production environment that stress the need to have more rigorous and attentive analysis. Results until now seem to point that evaluation and enquiry still requires some level of human dialog with the system - and/or a big investment in AI - in order to secure high-end intelligent systems (and not only digitalized human procedures). Future work will provide possible answers to this problem – a hypothesis is given for the fifth phase.

   6. Symbolic Context Model for Resilience Engineering

      Daichi Mitsuhashi, Taro Kanno, Satoru Inoue, Daisuke Karikawa, Kohei Nonose, Kimitaka Asatani, Kazuo Furuta

      Abstract

      Resilience can be defined as the ability of a system to adapt to changing contexts. To develop a better understanding of the concept of resilience and describe system resilience, it is more important to focus on a context’s characteristics than to discuss the definition of resilience itself. This study proposes a model to comprehend and describe a context’s characteristics, consisting of six a priori concepts: space, time, entity, event, state, and modality. Based on the model, this study also presents a way to formalize context. To demonstrate the model’s utility and partial validity, this paper provides examples of the reinterpretation and formalization of existing key concepts relating to human factors and resilience engineering based on the proposed model.

3. Human Reliability Analysis

   1. Frontmatter

   2. Identification of Performance Shaping Factors Affecting Subsequent Human Actions for Dependence Assessment in Human Reliability Analysis

      Jooyoung Park, Ronald L. Boring

      Abstract

      Dependence assessment refers to a method for adjusting the failure probability for the following action by considering the impact of the preceding action in human reliability analysis (HRA). This paper introduces a method for dependence assessment using performance shaping factors (PSFs), which is a different approach compared to existing dependency methods based on the Technique for Human Error-Rate Prediction (THERP). To be specific, the method presented here focuses on PSFs affecting multiple human actions and suggests how to quantify dependencies between human actions based on mathematical models and if-then logic for the PSFs. In this paper, the fundamental approach is introduced with a comparison to existing dependency methods. This paper also describes how we have developed the method.

   3. Is Dependency in Human Reliability Analysis a Real Phenomenon? Refining the Dependency Concept Through Research

      Torrey Mortenson, Ronald Boring

      Abstract

      Dependency, specifically defined as the degree to which one erroneous action can impact subsequent actions, has been a fixture in many human reliability analysis (HRA) methods for the better part of forty years. Its incipience in the Technique for Human Error-Rate Prediction (THERP) was supported, seemingly, by examination of observed phenomena, but the documented evidence and link to proven psychological theories were absent. While HRA is typically a quantitative effort with qualitative inputs, we should be careful to ensure that HRA methods respect their psychological foundations lest the models constructed fail to measure the characteristics of human error accurately and sufficiently for the assessment of safety critical applications. This paper seeks to refocus attention on a foundational question in HRA—is dependence in a chain of human erroneous actions such that ‘error begets error’ a real phenomenon? Research is still needed to validate the construct and quantification of dependency.

   4. Formative vs. Summative Dependence in Human Reliability Analysis

      Ronald Boring, Jooyoung Park, Torrey Mortenson

      Abstract

      Dependence in human reliability analysis (HRA) is the concept that once an initial human error has occurred, subsequent human errors are more likely. The approach almost universally adopted in HRA was first introduced in the Technique for Human Error Rate Prediction (THERP). In THERP, calculated human error probabilities (HEPs) are adjusted for dependence in the final step of quantification. This final adjustment for dependence involves anchoring the HEP to a set of values corresponding to low through complete dependence. The effect is to increase the HEP. In this paper, we propose formative dependence, which occurs at the onset of quantification. We demonstrate that dependence should be considered earlier in the calculation process, because it is not actually necessary to calculate the HEP when it is subsequently overridden by the dependence anchors. By applying dependence first in the calculation, processing steps can be eliminated, making for more efficient analysis.

   5. Human Error of a System Based on FMEA

      Qing Xue, Binjue Wang, Minxia Liu

      Abstract

      Many studies and system accidents have shown that human factors, such as illegal operations, fatigue work, and weak theoretical knowledge, are the main reasons for system accidents or incidents. Under the background of Man-Machine-Environment system, it is undoubtedly of great practical significance to contemporary society to study the accidents caused by human errors in the system and find a countermeasure for effectively controlling human errors in the system. This article follows the qualitative and quantitative analysis and research ideas, using Failure Mode and Effect Analysis (FMEA) and fault tree analysis (FTA) to study the human error in the man-machine interface of a certain equipment system.

   6. Challenges and Opportunities of Applying the Human Reliability Analysis (HRA) Techniques for the Petrochemical Industry in China

      Renyou Zhang, Dingsen He, Peiyuan Shang, Jun Ge

      Abstract

      It is widely acknowledged that human error is one of the significant reasons to trigger over half of the accidents. So far, dozens of HRA approaches have been published for qualitatively and quantitatively assessment to human error. However, most of the HRA techniques are sourcing from the nuclear industry, so people from the Chinese petrochemical industry may not be familiar with those nuclear industry-based HRA methods. Although we deem that HRA techniques could provide considerable benefits to safety management, directly using those methods in the petrochemical industry may cause some unexpected issues, since HRA is a relatively new idea to the petrochemical industry in China. Therefore, this study conducts a well-designed questionnaire survey to indicate the opportunities and challenges of the future HRA application in the petrochemical plants of China. This survey indicates the attitudes, opinions, and the potential barriers of practicing HRA methods in the petrochemical industry. The people involved in the questionnaire survey are all from the Chinese petrochemical industry, including management team and front-line operator.

4. Human Error, Reliability, Resilience, and Performance in Aviation

   1. Frontmatter

   2. Is Our Human Factors Capability in Aviation and Maritime Domains Up to the Task?

      Barry Kirwan, Katarzyna Cichomska, Beatrice Bettignies-Thiebaux, Andrew Kilner

      Abstract

      It is often said by Human Factors professionals that Human Factors needs to be involved in design projects from the start. But how often is this feasible, and how often does it really occur in practice? And, perhaps most importantly, do design and operational organizations have the right people, processes and motivation in place to allow this to happen? As part of a European-funded project called SAFEMODE, which looks at the potential for cross-domain Human Factors learning in aviation and maritime domains, twelve organizations participated in a Human Factors Capability Profiling exercise. This paper gives an anonymized overview of the results, showing the range and differences in HF ‘maturity’ in the two domains. A key insight is that for both domains, there is a need for further uptake of validated HF methods that can inform design decisions at early design stages.

   3. Flying High. Voice Stress Analysis to Detect Pre-symptomatic Acute Hypobaric Hypoxia at 25000 Ft

      Martine Van Puyvelde, Emma Debecker, Xavier Neyt, Frederic Detaille, Wim Vanderlinden, Nathalie Pattyn

      Abstract

      Previous research showed that Acute Hypobaric Hypoxia (AHH) induced pre-symptomatic compensatory voice markers at 20000 ft but not at 25000 ft. These studies were conducted in an intermittent design with pauses in between. Since AHH evolves more rapidly at higher altitudes, compensation might have occurred in theses pauses. Hence, we studied voice reactivity of 16 subjects to AHH at 25000 ft in a hypobaric chamber in a continuous speech design. We analyzed fundamental frequency-range (F0-range) and voice onset time (VOT) in function of the hypoxic symptoms as indicated by the subjects. We did not find the pre-symptomatic compensation (i.e., decreased F0-range and VOT). We found increased F0-range and decreased VOT. We hypothesize that at higher altitudes, voice reactivity to AHH is the output of physiological processes of both compensation and control loss. Voice stress detection is a promising future tool and the role of breathing in regulatory compensation processes should be examined.

   4. Human Factors Analysis for a New Wake Vortex Air Traffic Alert

      Frédéric Rooseleer, Adriana-Dana Schmitz, Barry Kirwan

      Abstract

      Humans play an essential role in air transport safety. Yet the integration of Human Factors into design is not systematically comprehensive nor uniform. A Horizon Europe 2020 research project called SAFEMODE, aims to consolidate a HUman Risk-Informed Design (HURID) framework. Rather than take a piecemeal approach, focusing on either Human Factors, learning from incidents, or risk modelling, SAFEMODE integrates all of these into one framework for designers to use, whether developing improvements or new systems, and at varying stages of design from the early concept stage, to detailed design prior to deployment. One of the selected Case Studies supporting HURID validation is the design of a new Wake Vortex Air Traffic Alert for Cruise phase of flight. The paper presents the HURID application in support of the concept design.

5. Organizational Effects on Human Performance

   1. Frontmatter

   2. How Many Human Factors Influenced the June 30, 2013, Yarnell Hill Fire 19 Fatalities and Yet Were Never Investigated Nor Documented?

      Fred J. Schoeffler, Lance Honda, Joy A. Collura

      Abstract

      Wildland firefighting is inherently dangerous, fraught with simple mishaps to inevitable fatal outcomes for the unwary. Established, sound rules and guidelines will continue to work to keep Wildland Firefighters and Firefighters safe. Human factors, consistent across all work groups, are variously broken down into human errors, human failures, error chain(s), etc. making fatalities unavoidable. All we can do is reduce them. Nineteen Prescott FD Granite Mountain Hot Shot wildland firefighters and supervisors perished on the June 2013 Yarnell Hill Fire in Arizona. Inexplicably, they left their Safety Zone during explosive fire behavior. A Serious Accident Investigation Team found, “… no indication of negligence, reckless actions, or violations of policy or protocol.” The authors and others infer: the final, fatal link, in a long chain of bad decisions with good outcomes. Among other things, tunnel vision, auditory exclusion, goal fixation, non-critical thinking, destructive goal pursuit, Groupthink, and “Friendly Fire” are discussed.

   3. Continuous Stress and Consequence on C4t, Complexity, Risk, and Necessity of Leadership Level 5

      Salvador Ávila, Luiz Fernando Lopes de Souza, Lucas Menezes Pereira

      Abstract

      The stressful environment installed in the workplace within the industry or in areas where the population lives in cities, affects people in different ways. The vectors that promote this stressful environment can be physical, organizational, cognitive, social, and affective within a network of relationships that result in inclusion, or frustration, both movements of a dichotomous relationship between comfort and discomfort. The organizational-social and affective pressure for results within the expectation of a socio-affective bond with work will result in a greater or lesser cognitive load that affect psychological functions causing mind map instability. Some tests indicate the level of stability of these psychological functions inherent to the work environment and citizen. It is important to keep intact the perception of scenarios, level of attention, operations of memory, identification of patterns. This paper will treat a vacation family as a task and will present SARS Methodology and Tools.

   4. Professional Deformations of Bailiffs with Different Professional Efficiency

      Yana Korneeva, Anastasiya Korneeva, Anna Yurjeva

      Abstract

      The study purpose is to identify and describe the professional deformations types of bailiffs with different professional efficiency. Research methods are questionnaires and multivariate statistical analysis methods. The study used the author’s questionnaire for assessing the professional deformation types and work performance subjective assessment by law enforcement officers. The study involved 277 federal bailiffs service employees of the Arkhangelsk region. The bailiffs are characterized by a high level of such professional deformations as overcontrol, conservatism and authoritarianism. By means of a two-stage cluster analysis, all employees are divided into three groups according to the professional performance level (low, medium and high) according to their subjective assessment. Bailiffs with a low professional performance level are distinguished by a higher professional aggression, behavioral transfer, conservatism and learned helplessness level. Authority, overcontrol, and anxiety do not have statistically significant differences in groups, which may be due to their positive impact on employee performance.