Bridging Maintenance Error Taxonomies and Safety Management Systems: A Structured Critical Review of Aviation Maintenance Human Factors

Technical Summary: Bridging Maintenance Error Taxonomies and Safety Management Systems

Problem Statement
Aircraft maintenance is a safety-critical function where errors often remain latent, interacting with flight operations or inspection failures to cause accidents. While the human factors literature has established that maintenance error is rarely a result of individual carelessness but rather a systemic interaction involving fatigue, workload, procedures, and organizational pressure, the field remains fragmented. Current research is divided into three disconnected streams: accident/occurrence evidence, maintenance-error taxonomies (such as MEDA and HFACS-ME), and proactive Safety Management Systems (SMS) research. Consequently, organizations lack a clear operational pathway to translate retrospective knowledge of past failures into forward-looking controls that prevent recurrence. The central unresolved issue is not whether maintenance error matters, but how to operationalize retrospective error classification within proactive safety management.

Methodology
This paper employs a structured critical review design, aligned with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) search and screening logic, though it explicitly avoids statistical meta-analysis due to the heterogeneity of the literature.

• Search Strategy: The review utilized keyword searches and backward citation tracing across major academic platforms (Scopus, Web of Science, ScienceDirect, etc.) covering literature from the late 1990s to 2026.
• Eligibility: Sources were selected based on direct relevance to aircraft maintenance, maintenance-induced events, error classification, safety climate, fatigue, and SMS-oriented prevention.
• Corpus: From an initial pool of 68 candidate records, 24 sources were retained for final synthesis. This corpus includes 14 peer-reviewed journal articles (serving as the primary empirical evidence base) and 10 foundational non-journal sources (including organizational accident theory, MEDA, and HFACS frameworks) used for conceptual grounding.
• Analysis: The synthesis compared convergence and divergence across three clusters: (A) accident and event evidence, (B) taxonomies and explanatory frameworks, and (C) proactive safety-management research.

Key Findings and Mechanisms
The review identifies that maintenance error is a systems phenomenon shaped by the interaction of the worker and the environment. The literature consistently highlights several recurring human-factor mechanisms:

1. Fatigue and Workload: Fatigue degrades checking behavior, memory, and judgment. It is influenced by rostering, overtime, and shift design, acting as a bridge between individual performance and organizational design.
2. Documentation Quality: Poor readability, ambiguous procedures, and fragmented technical information create error pathways by increasing reliance on memory and informal workarounds.
3. Training and Competence: Effective training must be tied to actual task competence and real-world conditions rather than regulatory checklist completion.
4. Communication and Safety Climate: Distributed work across shifts and specializations requires robust handovers. Safety climate is determined by whether leadership reinforces reporting of unsafe conditions rather than suppressing them.
5. Supervision and Organizational Pressure: Time pressure, resource shortages, and implicit production targets often normalize shortcuts, making deviations locally rational despite system risks.

Limitations of Current Taxonomies
While tools like the Maintenance Error Decision Aid (MEDA) and the Human Factors Analysis and Classification System-Maintenance Extension (HFACS-ME) are valuable for structuring investigations and moving beyond blame, the review identifies critical limitations:

• They are primarily retrospective, organizing past events without automatically predicting future risks.
• They depend heavily on the quality of reporting and a blame-free culture; under-reporting leads to incomplete risk data.
• They often identify contributing factors without forcing the organization to assign preventive controls, owners, or effectiveness measures.
• Without integration into SMS, these taxonomies risk becoming mere documentation exercises rather than prevention tools.

Proposed Framework and Contributions
The paper's primary contribution is a practical framework designed to bridge the gap between retrospective classification and proactive SMS. The framework proposes a five-step cycle to convert classification data into safety controls:

1. Data Collection: Gather maintenance event data (occurrence reports, audits, near misses).
2. Coding: Apply structured taxonomies (MEDA/HFACS-ME) to identify contributing factors.
3. Trend Analysis: Identify recurrent patterns (e.g., fatigue, documentation issues) rather than treating events as isolated.
4. SMS Integration: Transfer recurrent high-risk patterns into the SMS risk register and convert them into Safety Performance Indicators (SPIs).
5. Control and Monitoring: Assign preventive controls (e.g., roster redesign, procedure simplification) and monitor their effectiveness through audits and feedback loops.

The paper provides specific examples of translating classification patterns into SMS actions, such as linking "fatigue-related deviations" to "overtime monitoring" and "documentation deviations" to "readability audits."

Significance and Claims
The paper argues that the field must move beyond the creation of new taxonomies or the debate over accident severity statistics. The significance of this work lies in its assertion that classification knowledge must be connected to prevention. The authors claim that the most critical step in maintenance safety is no longer identifying what happened, but operationalizing how to prevent recurrence through a structured link between error reports, risk registers, and continuous monitoring. By treating taxonomies as the first step of a safety-management cycle rather than the final product, organizations can transform retrospective error knowledge into forward-looking, resilient safety systems. The paper concludes that future research should focus on strengthening the empirical bridge between retrospective classification and predictive safety management, including longitudinal studies and digital tools that unify reporting and risk management workflows.