Figure 2.
The bar graph compares the number of inter-encounter intervals that were collapsed by windowing versus those that remained separate. The horizontal axis represents the length of intervals in days, divided into three categories: less than 14 days, 14 to 28 days, and 28 days or more. The vertical axis shows the number of intervals. There are three sets of vertical bars, each set containing two bars: one blue and one orange. The blue bars represent intervals collapsed by windowing, while the orange bars represent intervals that remained separate. For intervals less than 14 days, there are approximately 853 intervals, with a small portion collapsed by windowing. For intervals between 14-28 days, there are approximately 887 intervals, again with a small portion collapsed by windowing. For intervals of 28 days or more, there are approximately 11,253 intervals, with the majority remaining separate. The graph indicates that most intervals exceeded 28 days, representing independent episodes.
Distribution of inter-encounter intervals and impact of windowing methodology. Inter-encounter intervals grouped by length. Blue indicates intervals collapsed into an ongoing episode by windowing. Orange indicates intervals that remained separate. Windowing collapsed 1,435 of 12,993 intervals (11.0%), reducing the episode count from 16,008 to 14,624 (8.6%).