Locating helicopter emergency medical service bases to optimise population coverage versus average response time

Abstract

Background: New South Wales (NSW), Australia has a network of multirole retrieval physician staffed helicopter emergency medical services (HEMS) with seven bases servicing a jurisdiction with population concentrated along the eastern seaboard. The aim of this study was to estimate optimal HEMS base locations within NSW using advanced mathematical modelling techniques.

Methods: We used high resolution census population data for NSW from 2011 which divides the state into areas containing 200-800 people. Optimal HEMS base locations were estimated using the maximal covering location problem facility location optimization model and the average response time model, exploring the number of bases needed to cover various fractions of the population for a 45 min response time threshold or minimizing the overall average response time to all persons, both in green field scenarios and conditioning on the current base structure. We also developed a hybrid mathematical model where average response time was optimised based on minimum population coverage thresholds.

Results: Seven bases could cover 98% of the population within 45mins when optimised for coverage or reach the entire population of the state within an average of 21mins if optimised for response time. Given the existing bases, adding two bases could either increase the 45 min coverage from 91% to 97% or decrease the average response time from 21mins to 19mins. Adding a single specialist prehospital rapid response HEMS to the area of greatest population concentration decreased the average state wide response time by 4mins. The optimum seven base hybrid model that was able to cover 97.75% of the population within 45mins, and all of the population in an average response time of 18 mins included the rapid response HEMS model.

Conclusions: HEMS base locations can be optimised based on either percentage of the population covered, or average response time to the entire population. We have also demonstrated a hybrid technique that optimizes response time for a given number of bases and minimum defined threshold of population coverage. Addition of specialized rapid response HEMS services to a system of multirole retrieval HEMS may reduce overall average response times by improving access in large urban areas.

Keywords: Bases; Coverage; Helicopter; Modelling; Optimization; Population; Response.

Fig. 1

The diagram at the top shows the population density of NSW as per the 2011 census data with the current MR-HEMS base locations in black and the CRRH E-HEMS base in blue. The diagram below indicates in green the population that can be reached within 45mins of activation from all existing base locations at A. Lismore, B. Lake Macquarie Airport (Newcastle), C. Westmead (blue), D. Bankstown Airport, E. Wollongong, F. Canberra, G. Orange, and H. Tamworth

Fig. 2

Greenfield MR-HEMS base locations for the MCLP and ARTM methods. a,b and c are the optimal MCLP solutions for 7, 8 and 9 bases respectively whereas d, e and f are optimal ARTM solutions for 7, 8 and 9 bases respectively

Fig. 3

MCLP fixed MR-HEMS base solutions. a and b are the replace one base and replace 2 base solutions respectively whereas c and d are the add one base and add two base solutions respectively

Fig. 4

ARTM fixed MR-HEMS base solutions. a and b are the replace one base and replace 2 base solutions respectively whereas c and d are the add one base and add two base solutions respectively

Fig. 5

Hybrid solutions showing the optimal base locations to provide the shortest average response time for a, minimum 97.75% coverage and b, minimum 98% coverage. CRRH (E-HEMS) base location in blue. Scenario A represents the optimal trade-off between population coverage and average response time. In scenario B the population coverage increases by only 0.25% whilst the average response time jumps by 30% (8mins)