Poster #36 - Perspectives on emergency plan and dependency：Practical implications for vulnerable groups in China

• In Event: Saturday Poster Luncheon

Abstract

Emergency response plans are intended to provide structured guidance for crisis actions. However, their effectiveness often diminishes when applied to vulnerable populations, such as older adults, patients, and children, whose limited mobility and special needs render conventional planning assumptions inadequate. Despite an increasing recognition of this challenge, existing research has largely focused on single-stage improvements (e.g., evacuation routes or supply distribution) or static vulnerability indices. This focus has resulted in a critical gap: how different stages of the emergency plan life-cycle influence each other, and which dependencies determine whether a plan ultimately succeeds or fails in protecting vulnerable groups. This gap persists, in part, due to the absence of systematic analytical frameworks capable of disentangling the causal hierarchy and dependency structure embedded in plan content.
This study addresses this gap by developing a multi-stage dependency evaluation framework grounded in the full life-cycle of emergency planning, covering formulation, compilation, approval, training, drills, evaluation, and revision. The framework operationalizes three analytical dimensions: (1) causal centrality, measuring the extent to which a factor influences or is influenced by others; (2) hierarchical level, distinguishing root causes, deep-seated factors, and surface-level outcomes; and (3) dependency strength, quantifying the conditional probability that a deficiency in one factor propagates to another. To implement this framework, we integrate Decision Making Trial and Evaluation Laboratory (DEMATEL) to quantify causal relationships, Interpretive Structural Modeling (ISM) to construct a multi-level hierarchical structure, and fuzzy Bayesian networks to infer dependency strengths under expert uncertainty.
This study examines four major fire incidents involving vulnerable populations in China: Changfeng Hospital, Dongguan Nursing Home, Guoan Elderly Apartment, and Lankao Adoption Center. Using accident investigation reports and systematic content analysis, we extract 19 evaluation indicators across eight emergency plan stages. Based on pairwise influence coding by domain experts and subsequent defuzzification, we obtain three key findings. First, non-standardized emergency linkage procedures and response mechanisms, along with inadequate resource and capacity assessments, exhibit the highest centrality (3.637 and 3.569 respectively), marking them as core systemic factors. Second, hierarchical analysis shows that insufficient emphasis on plan review, approval, and revision serves as root causes; they propagate through underlying factors such as the lack of targeted evacuation plans and ritualized drills, ultimately manifesting in surface-level deficiencies like inadequate supplies and untrained personnel. Third, dependency analysis reveals strong couplings, particularly between specific staff identification and standardized procedures (r=0.695), and indicates that emergency linkage, staff coverage, and critical-stage drills exhibit high dependency on other elements.
These findings indicate that the effectiveness of emergency response plans for vulnerable populations is not determined by any single stage, but rather by hierarchical causal chains and cross-stage dependencies. This study contributes to the field of crisis management by diagnosing structural and relational weaknesses in plans. Furthermore, It provides empirical evidence that enhancing the adaptability necessitates targeted interventions on root-cause factors, rather than merely implementing superficial amendments to plan content. The policy implications include the designation of named evacuation assistants, regular unscripted drills during critical stages, and the establishment of institutionalized audit mechanisms for plan revision.

Author

• Baoyan Duan, Beihang University
  Presenting Author