Geographic Information Software Programs' Accuracy for Interfacility Air Transport Distances and Time

• Published in: Air medical journal 2014-07, Vol.33 (4), p.165-171
• Main Authors: Soulek, Jaron J., BS, Arthur, Annette O., PharmD, Williams, Emily, MD, Schieche, Christoph, MD, Banister, Noah, BS, Thomas, Stephen H., MD, MPH
• Format: Article
• Language: English
• Subjects: Air Ambulances; Emergency; Geographic Information Systems; Health administration; Humans; Internal Medicine; Observer Variation; Retrospective Studies; Software; Time Factors
• ISSN: 1067-991X; 1532-6497
• DOI: 10.1016/j.amj.2014.03.002

Abstract Introduction This study aimed to evaluate consistency/predictability of interfacility flight times (IFFTs) and accuracy of geographical information system (GIS) software packages for estimating IFFT. Methods This retrospective study conducted by a program using a Bell 206 assessed the first 1000 IF transports occurring on 137 “runs” (ie, referring-receiving hospital pairings) made at least twice. GIS IFFT estimates using Google Earth™ (GE) and ArcGIS™ (AG) were compared against actual IFFT using linear regression; univariate analysis included assessment of medians with 95% binomial exact confidence intervals (CIs). Interrater agreement for GIS was assessed with κ. Results GE and AG estimates fell, respectively, within 1 mile of actual in 136/137 runs (99%, 95% CI 96%–100%) and 130/137 runs (95%, 95% CI 90%–98%). GE- and AG-predicted IFFT strongly ( P < .001) correlated with, underestimating by about 2 minutes, actual IFFT (GE: r2 0.93, coefficient 0.98, 95% CI .97–1.00; AG: r2 0.93; coefficient 0.98, 95% CI .96–1.0). GE and AG had statistically equivalent (κ > .8), “almost-perfect,” interrater agreement. Conclusion IFFTs for same-run helicopter EMS transports in our rural state setting are characterized by little variability. GIS is highly accurate in predicting IF logistics, with public-domain GE performing as well as more expensive AG.