Discrepancies Identified Between Patient Travel Patterns, Geographic Market Definitions

Discrepancies between standard geographic market definitions and actual patient travel patterns exist, as highlighted by a research letter published today in Annals of Internal Medicine.¹

Researchers and policymakers require accurate health care market definitions to evaluate provider access, variation, and competition in costs and outcomes. They commonly rely on patient travel thresholds or geographic boundaries, like counties or ZIP codes. However, few studies have assessed how well these definitions reflect actual patient travel patterns, with the existing studies mainly analyzing hospital care.

Consequently, the researchers conducted a study to create national estimates of patient travel time across provider types and to assess how well typical geographic market definitions capture patient travel patterns. To do so, they used data spanning 2018 to 2021 from the Medical Expenditure Panel Survey (MEPS), a nationally representative survey of health care utilization among the US civilian, noninstitutionalized population.²

The researchers emphasized the need for more accurate metrics to evaluate health care access and provider competition. | Image Credit: Song_about_summer - stock.adobe.com

They extracted data on emergency department (ED) visits, office visits (primary care, specialty care, nonphysician), and inpatient stays (originating in the ED vs direct admissions, pregnancy- or birth-related, surgical, or other).¹ For each encounter, the researchers geocoded patient and provider addresses using ArcGIS.

Also, the researchers calculated travel time by car and determined whether visits fell within selected geographic boundaries by using the Open Source Routing Machine and OpenStreetMap data. Overall, all estimates employed sample weights using Stata.

The unweighted sample included 825,292 office visits, 23,334 ED visits, and 10,552 inpatient stays. The researchers found the median travel time to be 12.7 (IQR, 7.0-22.3) minutes for primary care and 17.1 (IQR, 9.5-30.8) minutes for specialty care. The median travel time was longer for those living outside metropolitan statistical areas (MSAs) especially for specialty care (15.9 minutes inside vs 41.8 minutes outside).

Additionally, the median travel time was 13.6 (IQR, 7.5-24.2) minutes for ED visits and 18.1 (IQR, 9.6-36.9) minutes for inpatient stays. The researchers noted that patients traveled longer for stays that did not originate in the ED, were pregnancy- or birth-related, or were surgical.

Overall, for ambulatory care visits (office and ED), 73.8% (95% CI, 72.3-75.4) were within the patient’s county, 81.5% (95% CI, 80.1-82.4) were within 30 minutes of home, and 93.3% (95% CI, 92.6-94.0) were within 60 minutes of home. Conversely, a lower percentage of inpatient stays fell within these boundaries as 63.3% (95% CI, 60.8-65.8) were within the patient’s county and 86.4% (95% CI, 84.7-88.1) were within 60 minutes of home.

Similarly, 60.5% (95% CI, 58.2-62.8) of inpatient stays were within a hospital service area (HSA) and 85.4% (95% CI, 83.7-87.1) were within a hospital referral region (HRR); however, 50.9% (95% CI, 49.1-52.8) of primary care visits were within a primary care service area (PCSA).

“Our results provide nationally representative measures of travel patterns based on exact patient and provider addresses,” the authors wrote. “These measures can inform choices of market definitions and provide national benchmarks for patient travel time.

The researchers acknowledged the limitations of using geographic market definitions to assess care access and competition. They explained that certain market boundaries, like PCSA and HSA, often exclude portions of care used by residents, potentially underestimating access and competition.

Additionally, defining markets by travel time thresholds can be problematic since the 95th percentile of patient travel time is much higher than the median, especially outside of metropolitan areas. Looking ahead, the researchers suggested market definitions that better align with actual care-seeking patterns.

“In settings where the required data are available, flexible, data-driven market definitions may be preferable as they sidestep challenges posed by geographic market definitions by incorporating observed care-seeking patterns,” the authors concluded. “...In settings where such approaches are not feasible, we provide benchmarks for standard geographic market definitions.”

References

1. Decker SL, Gourevitch RA, Meille G. Patient travel patterns and concordance with geographic market boundaries. Ann Intern Med. doi:10/7326/ANNALS-24-00857
2. Medical Expenditure Panel Survey. AHRQ. Accessed November 4, 2024. https://www.meps.ahrq.gov/mepsweb/