Vehicles engaged in road traffic today are equipped with GPS tracking devices on a routine basis. Naturally, any sort of position tracking of vehicles (and thus persons!) raises issues of privacy and data security. There are, however, application domains where such concerns are outweighed by the benefits that can accrue through the use of tracking technology.
One of these domains is fleet management, where large vehicle fleets, for instance, of delivery services or trucking companies, can be monitored. In this way, the movements of vehicles are optimized to reduce fuel consumption and at the same time speed up delivery times. Similarly, in the health sector, GPS tracking allows to guide and optimize dispatching of ambulances. Alternatively, mobility analytics allows to study the accessibility of health care services for a particular population, and thus improve the equity in the delivery of such services.
In the EU Horizon 2020 project Track&Know 14 partner groups - technology partners, academic partners, and application partners - joined forces to develop technology and computational methods that enable big mobility data analytics applications such as the above.
Long travel distance can be a barrier to diagnosis and treatment
Among the tasks addressed by the GIS unit at GIUZ was the development of an application for optimizing the locations and spatial accessibility of health care services. We collaborated with the Royal Papworth Hospital, which provides testing and treatment for 'obstructive sleep apnoea' (OSA), a major cost-intensive public health concern. Up to 85 % of the OSA population are estimated of going undiagnosed. Poor accessibility of OSA facilities due to long travel distance can be a barrier to diagnosing and treating patients in the UK.
Several risk factors are known to be associated with OSA, e.g. age and obesity. We incorporated them into a model showing the uneven distribution of the population potentially at risk of suffering from OSA. Combining the risk model with the travel times allows to highlight areas of deprivation. Typically, these areas tend to be more rural and/or peripheral, with a combination of a population that is older than average, has more weight problems, and is living further away from centers on the countryside. This leads to higher rates of undiagnosed and untreated OSA, and thus to a poorer health condition.
Re-arrangement of facilities would bring substantial efficiency gains
Given these maps of the status quo, the question was then how the access to OSA testing facilities could be improved. We developed an optimization model that computed new, or additional facility locations such that the overall travel distances and times could be minimized for a given population in demand of health care. For instance, for the risk population a mere 10 facilities would suffice to obtain the same demand-weighted average travel distance from patients' homes as for the currently 27 OSA testing facilities! Conversely, optimally re-arranging the same number of 27 facilities, the average travel distance could be almost halved to 10.2 km instead of 19 km. This dramatic gain in efficiency is largely due to the fact that the current network of OSA testing facilities has grown historically, without much consideration of its spatial layout.
Of course, one might argue that it is unrealistic to completely re-arrange all facilities, because after all they represent physical infrastructure with trained staff. Therefore, the optimization model also allows adding few but optimally placed facilities, while keeping the existing facilities fixed. Using this approach, adding 5 facilities to the 27 existing ones still leads to a reduction of the demand-weighted average travel distance from 19 km to 12.6 km, or one third.
More equitable and greener: A classical win-win situation
Through this optimization approach, health care service provision could be made more equitable: The additional health care facilities are placed in the currently underserviced, deprived areas with a high population that is potentially at risk. In addition, they could be made greener: Reduced travel distances not only lead to better service accessibility, but also to less fuel consumption and CO2 production. A classical win-win situation.
Our application has found great interest from other health care providers within the NHS and beyond. To a good deal, this was also due to the interactive dashboard front-end developed by our research assistant Livio Brühwiler, which allows to vividly demonstrate the functionality and impact of such an evidence-based approach. Relying on spatial data enables more equitable and greener health service planning and management.
Webinar presenting the Track&Know Health Care Service Optimization application: talks start at 16:55; interactive dashboard demo at 40:10 to 50:50.
Hubler, Loris (2019): Facility location modeling scenarios to optimize obstructive sleep apnoea health facilities in the East of England. MSc Thesis, Dept. of Geography, University of Zurich.
Track & Know is a Horizon2020 project and an acronym for "Big Data for Mobility Tracking Knowledge Extraction in Urban Areas". Track & Know aims to introduce innovative software stacks and toolboxes addressing new emerging cross-sector markets related to automotive transportations and urban mobility, e.g. healthcare management.
GIUZ is one of the project partners and collaborates among others with the Royal Papworth Hospital (RPH). RPH is part of the UK National Health Service system and the leading center for respiratory diseases in the East of England. The case of their obstructive sleep apnoea testing facilities was used to study the provision of health care services to the population.