We use cookies to personalise content and ads, to provide social media features and to analyse our traffic. We also share information about your use of our site with our social media, advertising and analytics partners who may combine it with other information that you’ve provided to them or that they’ve collected from your use of their services. You consent to our cookies if you continue to use our website. Term of use & Privacy Policy

SHIELD
Safety human incident & error learning database

Background

SHIELD (Safety Human Incident & Error Learning Database) Open Data Repository is a core element of the SAFEMODE project. This database and its taxonomy enables systematic analysis and collection of Human Factors elements in safety occurrences (incidents/accidents) in transportation, especially for aviation and maritime operations. Data queries using SHIELD will provide feedback to system and operation designers, to safety management, and to quantification of human components in safety risk models.

The database and taxonomy in SHIELD supports the goal of understanding how Human Factors aspects contribute to safety occurrences. This includes answering questions like what happened, how did it happen, why did it happen, how can it be prevented.

SHIELD stores occurrence reports that are entered by users. The collection and analysis of this data is based on a taxonomy that has been developed for SHIELD within the SAFEMODE project. The taxonomy describes the elements in the database and their relations. The database is structured according to three levels:

  1. Occurrence facts, describing facts of the occurrence, such as location, date, injuries, type of vehicle, involved actors, and contextual conditions;
  2. Occurrence assessment, describing the classification of the severity and the types of huma factors that contributed to the occurrence;
  3. Safety positive actions and learning, describing actions of human operators and/or technical systems that prevented occurrences getting worse (accident prevention) or that limited the consequences of an accident (consequences mitigation), and describing lessons learnt following a safety occurrence, e.g. changes in system design, training or procedures.

Entering Occurrence Reports into SHIELD

Occurrence reports can be entered into the database according to the Occurrence Facts taxonomy in SHIELD. There attributes are defined that describe facts of safety occurrences. SHIELD uses a limited set of the most relevant attributes. They refer to the involved human actors and the contextual conditions that are relevant for the performance of these actors. When entering a report specific information fields (attributes) are asked to be filled out. The occurrence data that needs to be filled in includes:

SHIELD stores occurrence reports that are entered by users. The collection and analysis of this data is based on a taxonomy that has been developed for SHIELD within the SAFEMODE project. The taxonomy describes the elements in the database and their relations. The database is structured according to three levels:

  • General data (e.g., headline, domain: aviation/maritime, report dates, etc.),
  • Occurrence & Narrative (e.g., date & time, Location of vehicle, damage,narrative, etc.),
  • Category & Severity (e.g., severity: accident/serious incident/etc.; category: collision/ capsizing/etc.)
  • Vehicles & Operation (e.g., vehicle category, vehicle specs, operator, operation phase, etc.)
  • Actors (e.g., type, role, qualification, experience, etc.)
  • Context (e.g., Precipitation, Visibility, weather, traffic density, etc.)
  • Prevention & Learning (e.g., accident prevention, consequences mitigation, safety learning)

The last point in the list Prevention & Learning gives input on lessons learned and improvements made from which other organizations can learn and profit from. The attribute accident prevention includes a description of ‘what saved the day?’. Including any action of a human operator and/or technical system that prevented the situation from getting worse. Consequence mitigation refers to any action of an human operator and/or technical system that limited the consequences of an accident. Safety learning describes lessons that have been learnt following the safety occurrence, and changes in design, procedures and organisation that have been adopted as a result.

SAFEMODE HF Taxonomy

The structure of the SAFEMODE HF Taxonomy is based on the NASA Human Factors Analysis and Classification System (NASA-HFACS), which was developed by NASA for space missions. In SHIELD Human Factors elements in the occurrence reports are classified on four main layers:

  • Organization
  • Supervision
  • Preconditions
  • Acts

Organization The decisions, policies or methods adopted at an organizational level that affect both the supervisory and individual operator performance. This level includes: Culture/Climate, Safety Management, Resources, Economy and Environment. Within each category exist specific items that describe factors that contribute to an occurrence. For example, within the category Culture/Climate exists the item “Safety Culture”. This item is chosen when a poor safety culture was identified within the organisation. This item may be assessed formally through a survey or inferred from other idiosyncratic sources of information in the organisation.

Supervision The decisions or policies of the supervisory chain of command that directly affect the practices, conditions or individual actions, resulting in errors or unsafe situations. This level includes the categories: Failure to Correct a Known Problem, Inadequate Supervision, Planned Inappropriate Operations, Supervisory Deviations from Procedures. Within each category exist specific items that describe influencing factors. For example, within the category Failure To Correct a Known Problem exists the item “Inadequate Personnel Management”. This item is used when a supervisor failed to identify an operator who exhibits recognizable risky behaviours or failed to institute the necessary remedial actions, leading to unsafe operator performance.

Precondition The environmental factors or conditions of individual operators affecting human performance and contributing to errors or deviations from authorized procedures. This level includes the categories: Physical Environment, Technological Environment, Communication, Team/Group, Self-imposed Stress. Within each category exist sub-categories and items that describe specific influencing factors. For example, the Environment Factors category includes the sub-category Physical Environment and one of the items within that sub-category is “Vision Affected By Environment”, referring to environmental conditions that affect the operator’s vision thereby contributing to an error.

Acts The decisions or policies of the supervisory chain of command that directly affect the practices, conditions or individual actions, resulting in errors or unsafe situations. This level includes the categories: Failure to Correct a Known Problem, Inadequate Supervision, Planned Inappropriate Operations, Supervisory Deviations from Procedures. Within each category exist specific items that describe influencing factors. For example, within the category Failure To Correct a Known Problem exists the item “Inadequate Personnel Management”. This item is used when a supervisor failed to identify an operator who exhibits recognizable risky behaviours or failed to institute the necessary remedial actions, leading to unsafe operator performance.

“SHIELD is useful for various purposes. The developed SHIELD taxonomy identifies HF elements during occurrence investigations. The database is useful to make comparisons and analysis of important HF elements.”

What can SHIELD be used for?

Based on the SAFEMODE HF taxonomy developed for SHIELD safety occurrences can be analysed and causes can be identified. The categorised and analysed data within SHIELD supports organizations and users in different activities. The following list describes possible uses of SHIELD:

  • Benchmarking and understanding commonalities between industries and limits of human performance on certain task
  • Understanding common incident factors and human error classes
  • Developing a deeper understanding of human failure and success, and the contributing factors therein related to the operational environment
  • Finding similarities between different industries
  • Performing a thematic analysis, so that patterns and common underlying errors can be identified
  • Learning from the results of the HF analysis to improve design of systems, improve design of procedures, improve design of training, and improve selection process of human candidates for a function in the operational field.
  • SHIELD provides a set of case studies/examples that serve as input to simulator training and studies
  • SHIELD supports system design
  • SHIELD enables cross analysis among different reports (e.g. links with working environment, corporate culture, HMI, legal framework)

SHIELD can answer different questions. For example, when developing a new system or operation, SHIELD can be used to assess the following questions:

  1. What went wrong with similar systems/operations?
  2. What were key Human Factors elements?
  3. What were the lessons learned?

Within SHIELD the user then selects vehicle category and operation type. Following this selection the user can obtain information on occurrence categories & severities, contributing Human Factors aspects (Acts / Preconditions / Supervision / Organization) and safety learning and changes.

SHIELD will be developed into a software tool, and included in a suite of digital SAFEMODE tools - named eHURID