Expected Variables Present Study Results III: PAQS Query 3

**NOTE** This is a single site analysis, but results for all sites are included in the output attached to this record for ease of review & reuse. General observations are recorded below, and details about specific institutions can be discovered in the raw output. - Most anomalies are generally observed in the time periods before 2015 and after 2020 - The pre-2015 anomalies are likely explainable by small patient populations in that time period and less stable data as EHRs were in the process of being onboarded in the broader medical community. - The later anomalies are observed less frequently. Some can be considered an artifact of the drop in utilization due to the pandemic and a slower return to baseline after the fact. Others are anomalies above expectation, which could be due to an influx of new data in recent years or a change in coding practices. - Some institutions are more likely to have anomalies in the time series if there is a large range between the earliest and latest data point. For example, if there is very low capture pre-2015, this can cause the better capture in recent years to appear as anomalous.