Human Reliability Analysis for Cardiopulmonary Resuscitation Process in Emergency Medicine Using a Modified Hybrid Method Based on the Markov Model and Fault Tree Analysis

In emergency departments (ED), human reliability assessment is essential for improving the quality of treatment and preventing medical accidents. A medical accident is expressed as an injury to a patient caused by the negligence of a doctor or nurse who is providing medical care. This study aimed to assess the human reliability in the cardiopulmonary resuscitation (CPR) process and recommend some comments to minimize human errors and improve patient safety.

The main factors in the CPR process (such as rate and depth of chest compression and rate of ventilation) are identified based on the American heart association (AHA) roles. Data were recorded during three months in the evening shifts in the ED and CPR room of Imam Reza Hospital in Mashhad, Iran. In total, 42 samples were collected, and a modified hybrid approach according to the fault tree analysis and Markov method was proposed for the analysis of CPR team (including emergency medicine, medical interns, and nurses) reliability in the resuscitation process. Finally, the important basic events (errors) were selected using the Boruta algorithm by R software.

An FTA-Markov-based hybrid method is considered to compute the human reliability in the CPR process. The obtained results from human reliability analysis using the sensitivity analysis via Boruta algorithm and the proposed hybrid method show that an interrupt between chest compression process for rhythm control, the cycle of CPR, the depth of chest compression, and the discussion about reversible causes are the most effective factors in the human reliability of CPR process.

The human reliability of the CPR process in the ED has been assessed using a hybrid method based on the FTA and Markov method for the first time. To improve the quality of treatment and prevent medical accidents during the CPR process, the main factors in the process are identified, and then, the proposed hybrid method is used to calculate human reliability.