Developing a Model on Factors Affecting the Mode Choice in EMS

Abstract

Various studies has been conducted on the mode choice model for different kinds of situa tions such as for work travel mode choice, driver’s route choice behaviour, choice of freight transport mode etc for different regions of the world and different unique policies have been recommended (Yamamoto et al., 2002 ; Xie et al., 2003 ; Li et al., 2019) . But till now, no work has been conducted on the mode choice behaviour of the patients and bystanders in case of emergency medical situations and the factors that are responsible behind the choice of those various conventional and unconventional modes. Emergency Medical Situation is a very critical set of circumstances where a minimum room of error can cause huge damage and fatality. This study aims to work on this perspective, where the factors behind the choice of modes will be analyzed, a model i.e., utility function will be developed for each unconventional mode and then possible recommendation will be given on which the una voidable critical situation can be handled in a compact way. The data for this thesis project was collected from different zones of Bangladesh such as Dhaka, Rajshahi, Bogura, Bari shal etc. And the zones were divided into mega-city and sub-urban category. About 1954 data were collected through questionnaire survey where various factors contributed to the decision-making process for the patients and bystanders, including geographical location, severity of the medical condition, availability and accessibility of different transportation options, response time requirements, and resource allocation considerations. C4.5 algo rithm was used to find the most responsible factors for the choice of modes. As there are enough data for making a decision tree, the training and testing data-set gave satisfactory observations. Later Multinomial Logit Regression method was used to build up the utility functions of the unconventional modes taking ambulance as the base, where the significant factors were taken (Rob. P-value < 0.05). After the utility functions were generated, based on the factors and modes different interesting findings were allotted. Although the data-set might have been limited and there might be temporal and spatial bias and prejudice, still unique policies have been recommended which can be implied by the government for the betterment of the Emergency Medical Service