Spent most of yesterday replicating an analysis. That included estimating population-based network travel distance/time to the nearest & next-nearest acute care hospitals across the rural U.S. South.

Worth reiterating this point:

hospital service areas are not measures of patient access to care https://twitter.com/Arrianna_Planey/status/1319823653140353024?s=20

The consumerist orientation of health care (& subsequently, much of health services research) means that we have to be mindful to avoid conflating *location within health system market areas* with *access*.

Now, compare the map I made with this, which essentially models service areas post-closure based on proximity (pulled from https://doi.org/10.1016/j.jss.2020.08.055 ).

This is a fundamentally different orientation in focus/measurement. https://twitter.com/Arrianna_Planey/status/1320376831044968450?s=20

Incidentally, I am designing a course on this topic- conceptualizing and measuring healthcare access. I can't wait to teach this! https://twitter.com/Arrianna_Planey/status/1320380669139312640?s=20

One last one:

Here is a rough draft of a map. The purple lines are travel routes from rural origin points to the nearest hospital that exceed 60 minutes network travel time. If you drop the number to=>45 minutes, basically all of eastern Oklahoma (Indian territory) is affected.

I point this out, b/c 45 and 60 minutes are commonly used cutoff points for catchment area methods. This = pretty serious implications for modeling rural hospital service areas, & it means that these methods are even more divorced from measures of rural PT spatial access to care

I think this is worth examining, because there is an implicit association between "proximity" &"population" underlying spatial methods that model hospital/health system service areas. Accordingly, that means that people in rural places may not be counted as "populations."

The above is related to https://twitter.com/Arrianna_Planey/status/1227603253644419072?s=20