Originally Posted on PCG’s Health Policy News on April 19, 2022.
Author: Jessica Lang
Recently PCG’s Health Innovation, Policy, and Information Technology team has been trying to make sense of trends in claims data during 2020-2021 in Utah, where we are conducting an independent evaluation of a Medicaid demonstration. In reviewing this data, we often wind up speculating about how the COVID-19 pandemic affected healthcare utilization. We know elective procedures and most routine medical services were suspended for a time in 2020, and national data confirms that overall, people in the United States got a lot less healthcare in 2020. But how much less, and for how long? And how are the patterns of Utah specifically different from other parts of the U.S.?
One available indicator is community mobility data—location data aggregated from smartphones that is used to track and predict crowd behavior. Apps use this data to generate notices like “this location is typically most crowded at 4-6pm.” Google has long gathered this data from users of its apps, and since the beginning of the pandemic, has made the data available at the state and county level to support public health and research efforts.
I made this chart using Google’s community mobility data as a way of understanding more about how the stay-at-home order period played out in Utah. It shows how much people in Utah went to various types of locations relative to a pre-pandemic baseline period. Above 0 on the Y axis means more movement to that type of establishment; below 0 means less.
View a larger version of the chart here.
Several patterns are apparent:
- When the Utah governor issued a voluntary stay-at-home directive on March 27, 2020, there was a sharp spike in grocery and retail/pharmacy shopping followed by a 40% drop in retail, a smaller drop in grocery, and a nearly 50% decrease in trips to transit stops and workplaces.
- Grocery/pharmacy has fluctuated but mostly been within 15% of baseline since the spike in late March.
- Retail returned to baseline around April 2021 and stayed there until a 10% dip in early 2022.
- Transit stations followed a similar pattern, but returned to baseline a little later: by July 2021.
- The workplaces line shows a dip around each holiday/three-day weekend (this is a 7-day average, but is affected by having fewer workdays in a week), but has mostly slowly trended up, remaining at about 20% below baseline through March 2022.
A key caveat is that the baseline Google uses is the median value, for the corresponding day of the week, during the five week period from January 3 through February 6, 2020. This does account for routine differences between weekdays and weekends, but not for seasonal changes. For example, this chart doesn’t show the line for parks, as people mostly go to the park in the warmer months, and the baseline period was winter. Fluctuations in mobility around the winter holiday period in both years may be related to increased holiday travel, followed by increased COVID-19 cases and resulting caution, but the trends are hard to interpret without a pre-pandemic holiday baseline. Moreover, the world has changed in two years, and the baseline may be an outdated reference point. Is the current lower level of workplace attendance reflecting ongoing concerns about the pandemic, or a long-term shift to remote work?
Methodological limitations aside, it is not obvious what these trends means for people getting routine healthcare. Still, this data is a useful input. At a very general level, during a period where people are avoiding transit stations, we could expect to see fewer people getting mammograms. Community mobility data could be valuable for states looking to understand the “return to normal,” or after-action reviews of the public health emergency period. Furthermore, comparisons of counties and states could help illuminate the impact (or lack thereof) of stay-at-home orders or other policies. As the data is already aggregated for privacy protection, this resource can be shared freely, and applied to a wide range of questions.
Data from Google LLC “Google COVID-19 Community Mobility Reports”.
https://www.google.com/covid19/mobility/ Accessed: March 31, 2022.