Elsevier

Expert Systems with Applications

Volume 61, 1 November 2016, Pages 262-271
Expert Systems with Applications

Non-emergency patient transport services planning through genetic algorithms

https://doi.org/10.1016/j.eswa.2016.05.028Get rights and content

Highlights

  • NURA, a Non-Urgent transport Routing Algorithm is proposed in this work.

  • The proposed algorithm is able to generate detailed routes for ambulances.

  • NURA relies on a genetic algorithm, and it includes a scheduling algorithm.

  • A comparison between NURA and human experts’ solutions is presented.

  • NURA reduces the time spent by patients in ambulances and increases ambulance usage.

Abstract

Non-emergency Patient Transport Services (PTS) are provided by ambulance companies for patients who do not require urgent and emergency transport. These patients require transport to or from a health facility like a hospital, but due to clinical requirements are unable to use private or public transport. This task is performed nowadays mainly by human operators, spending a high amount of time and resources to obtain solutions that are suboptimal in most cases. To overcome this limitation, in this paper we present NURA (Non-Urgent transport Routing Algorithm), a novel algorithm aimed at ambulance route planning. In particular, NURA relies on a genetic algorithm to explore the solution space, and it includes a scheduling algorithm to generate detailed routes for ambulances. Experimental results show that NURA is able to outperform human experts in several real scenarios, reducing the time spent by patients in ambulances during non-emergency transportations, increasing ambulance usage, while saving time and money for ambulance companies.

Introduction

Route planning is a classic problem with remarkable importance in multiple environments, with a wide range of applications in the fields of Intelligent Transportation Systems (Di Lecce, Amato, 2011, Fontanelli, Bini, Santi, 2010, Liu, 2002), autonomous robotics (Latombe, 1991, Makhal, Raj, Singh, Chakraborty, Nandi, 2012), aerospace environments (Hui, Zhong, Weihua, 2008, Tulum, Durak, Yder, 2009) and military guidance and navigation systems (Lei, Qing, Zheng, Yu, 2010, Ruuben, Kreison, 2013, Zafar, Qazi, Baig, 2006).

An important application of route planning involves determining optimal routes for ambulances in both emergency and non-emergency transport services. Specifically, the Non-emergency Patient Transport Services (PTS) are provided by ambulance companies for patients who do not require emergency transport. These patients require transport to or from a health facility like a hospital, but due to clinical requirements are unable to use private or public transport.

Computing optimal routes for ambulances is a non-trivial problem that depends on the number and characteristics of the available ambulances, as well as their location. Several incompatibilities may arise due to the limited number of positions available in each ambulance, the equipment included, or depending on the legislation related to patient transport and minimum service conditions in each country. The main objective in this environment is, in general, to reduce the amount of time a patient spends in an ambulance which could be negative for their comfort and produce additional related health problems.

The Non-emergency Patient Transport Services Route Planning problem could be defined as the determination of the daily schedule for each available ambulance indicating the stops to be performed during the day, including the estimated time for the ambulance to arrive to each point of the route, and the patients that should be get on or off the ambulance at the stop. Most companies perform the service assignment by means of experts that are aware of the limitations of the system and the constraints that should be addressed in order to have a feasible solution, which is often a suboptimal one since human experts are not able to test enough combinations in an adequate time. The main planning unit in this problem is called service, which includes any single transportation of patients to or from a health center or a home address. For example, a return trip of a patient from his home to a hospital for a medical check would require two services: transferring the patient to the hospital, and another transportation to bring him home.

To solve the limitations of traditional systems, in this paper we propose the Non-Urgent transport Routing Algorithm (NURA), a route planning algorithm for non-emergency patient transport based on two main components: (i) an evolutionary algorithm (specifically, a genetic algorithm) to assign the services to be completed during a day to the set of available ambulances, and (ii) a scheduling algorithm based on solutions provided by human experts that, given the assigned services to a specific ambulance, determines the schedule for that ambulance including the times when the ambulance should pass through each point of the route, and ensures that the schedule provided is feasible.

Evolutionary Algorithms imitate the principles of natural evolution as a method to solve parameter optimization problems. They have been successfully used to solve various types of optimization problems (Greenwood, Lang, & Hurley, 1995), since they provide an optimal solution without checking all the possible solutions, reducing the execution time drastically. In this work, we compared the results obtained by our proposal with real planning obtained by human experts working in an existing ambulance company, and proved how our algorithm is able to provide better solutions, saving the time required by the experts.

This paper is organized as follows: Section 2 reviews the related work regarding non-emergency medical transport and the Vehicle Routing Problem (VRP). In Section 3 we present NURA, our proposed Non-Urgent transport Routing Algorithm which allows automatically obtaining a complete schedule for each available ambulance including all the stops to perform during the day. Section 4 introduces the structure and main parameters of the Genetic Algorithm (GA) used to explore the solution space. Section 5 presents the scheduling algorithm used to evaluate each solution. Section 6 shows the obtained results compared to those generated by human experts. Finally, Section 7 concludes this paper.

Section snippets

Related work

In this section, we are going to deal with some some approaches that are similar to the one we are introducing in this paper. This section is divided into two subsections: In the first one, we are going to mention how non-emergency medical transport has been faced by several authors when applied to different medical services around the world. In particular, we focus in the main problems that have been addressed in this field. In the second subsection we are going to review some approaches

NURA: Non-urgent transport routing algorithm

So far, genetic algorithms have been applied to different fields and applications. In particular, we previously proposed an approach focused on traffic accidents urgent sanitary resource allocation based on multi-objective genetic algorithms (Fogue et al., 2013), and a system able to reduce the emergency services arrival time by using vehicular communications and Evolution Strategies (Barrachina et al., 2014).

In this work, we propose the NURA, a route planning algorithm for non-emergency

Service assignment: genetic algorithm

Genetic algorithms are a sub-set of evolutionary algorithms, which are based on Darwinian theories of evolution. Given a population formed by individuals, natural selection due to the limited resources and environmental pressure increase the level of adaption of the individuals to their environment, i.e., the fittest individuals are able to survive and transfer their beneficial features to their offspring. The new individuals will compete again in the environment. These individuals are formed

Schedule generation algorithm

Obtaining the fitness value for a particular solution requires knowledge of the average travel time per patient, which can only be obtained if the daily scheduling of each ambulance is known. This is done by knowing the characteristics of the particular ambulances and the services that are assigned to them for the day. The genetic algorithm assigns a subset of services to each ambulance, and then a method to determine the schedule of each ambulance is required to compute necessary metrics such

Experimental results

In this section we evaluate the performance achieved with the NURA algorithm in different scenarios obtained from a real ambulance company. First of all, we will study the influence of several parameters in the efficiency of the genetic algorithm, in order to justify the values selected and presented in previous sections. Secondly, we will compare the solutions generated using NURA with those computed by human experts working in the ambulance company.

Several scenarios with different amounts of

Conclusions

Non-emergency patient transport is a rapidly growing area, especially due to demographic changes in today’s society. However, specific non-emergency medical transportation services are not often found, as they are mainly focused on emergency medical transport.

In this paper we presented NURA, a novel algorithm to generate Non-Emergency Patient Transport routes with the aim of reducing the time spent by the patients in the ambulances. Existing ambulance companies make use of human experts to

Acknowledgments

This work was partially supported by the Ministerio de Economía y Competitividad, Programa Estatal de Investigación, Desarrollo e Innovación Orientada a los Retos de la Sociedad, Proyectos I+D+I 2014, Spain, under Grant TEC2014-52690-R, and by the Government of Aragón and the European Social Fund (T91 Research Group).

References (30)

  • V. Di Lecce et al.

    Route planning and user interface for an advanced intelligent transport system

    IET Intelligent Transport Systems

    (2011)
  • S. Fontanelli et al.

    Dynamic route planning in vehicular networks based on future travel estimation

    2010 ieee vehicular networking conference (vnc)

    (2010)
  • M. Gibin et al.

    An exploratory cartographic visualisation of London through the Google Maps API

    Applied Spatial Analysis and Policy

    (2008)
  • G. Greenwood et al.

    Scheduling tasks in real-time systems using evolutionary strategies

    Proceedings of the third workshop on parallel and distributed real-time systems

    (1995)
  • I.M. Hains et al.

    Non-emergency patient transport: what are the quality and safety issues? A systematic review

    International Journal for Quality in Health Care

    (2011)
  • Cited by (23)

    • A branch-and-price algorithm for a routing problem with inbound and outbound requests

      2022, Computers and Operations Research
      Citation Excerpt :

      The literature on non-emergency transport of patients is relatively poor, the emergency case having attracted more attention. Fogue et al. (2016) studied a problem that shares many similarities with ours. The study focuses on the inconveniences of patients with wheelchairs or on stretchers.

    • Developing a data analytics platform to support decision making in emergency and security management

      2019, Expert Systems with Applications
      Citation Excerpt :

      In the case of medical incidents, the knowledge about their distribution can help to efficiently schedule resource allocation across the different municipalities and study the optimal location of medical centres close to urban areas and rural villages with higher frequencies of these types of emergencies. Among others, see the works of Huang, Kim, and Menezes (2010), Liu et al. (2016), Sakr and Elgammal (2016), Aringhieri, Carello, and Morale (2013) and Fogue et al. (2016). These articles helped to design the analysis of incidents according to the geographical distribution by municipalities, which provides to the emergency and security services with the information required to improve these processes.

    • Optimal multivariable conditions in the operation of an absorption heat transformer with energy recycling solved by the genetic algorithm in artificial neural network inverse

      2018, Applied Soft Computing Journal
      Citation Excerpt :

      Hakli and Uğuz [28] applied the GA for the automated partition of land in order to improve agricultural productivity. Fogue et al. [29] developed a GA with the objective of planning ambulance routes, demonstrating that it is possible to save time and money for ambulance companies. Park et al. [30] used a GA to maximize total profits in retail supply chains by optimizing replenishment times and quantities and vehicle routes.

    • A novel approach for automated land partitioning using genetic algorithm

      2017, Expert Systems with Applications
      Citation Excerpt :

      Thus, GA gets rid of local minima by providing a diversity of population. GA is increasingly used in many real-world problems and engineering, such as scheduling problems (Asadzadeh, 2015; Casas, Taheri, Ranjan, Wang, & Zomaya, 2016), designing packing systems (Thomas & Chaudhari, 2014), facility layout problems (Pourvaziri & Naderi, 2014), planning patient transport services (Fogue, et al., 2016), face recognition systems (Sukhija, Behal, & Singh, 2016), forest fire spread prediction (Denham, Wendt, Bianchini, Cortes, & Margalef, 2012), land use planning (Porta, Parapar, Doallo, Rivera, Sante, & Crecente, 2013) and filtering problems (Ar & Bostanci, 2016) etc. Due to their simple structure, ease of implementation and direct use in discrete problems, GA is used for this land partitioning problem.

    View all citing articles on Scopus
    View full text