|
Abstract:
|
Researchers are studying the degree to which many outcomes of interest have changed during the COVID-19 pandemic. Our team was interested in whether Emergency Medical Services in Guilford County, NC administered more doses of naloxone (a life-saving opioid overdose reversal drug) during the first 29 weeks of the pandemic than in the past. Such a change can be quantified simply as a growth rate relative to a comparison period. However, a growth rate cannot convey whether the observed change is historically anomalous. A large change may not be noteworthy if changes of similar magnitude occurred often in the past. Therefore, we adapted the difference-in-difference (DID) technique to generate a quasi-control group of past changes. Using a rolling time window, we calculated the growth rate for every 29-week period in the available historical data (2014-20). This produced a quasi-control group of 270 past changes. We found that the growth rate of naloxone administrations during the pandemic was > 91 percent of past periods. This adaptation of DID can be applied to any outcome for which researchers want to understand the degree of change during a specific period relative to the past.
|
