TABLE 2

Three Multivariable Logistic Regression Models Analyzing Variables Associated With 365-Day Readmission by Timing of Ambulatory Visit Attended (Variable of Interest)

OR (95% CI)
Model 1a (N = 613)
Attended postdischarge visit0.65 (0.43–0.97)
  1+ asthma hospitalization in previous year versus none3.15 (2.03–4.87)
  1+ ED visit in previous year versus none1.56 (1.05–2.30)
  PICU stay during index hospitalization versus none2.78 (1.70–4.53)
Model 2a (N = 597)
Attended routine visit1.4 (0.91–2.08)
  1+ asthma hospitalization in previous year versus none3.43 (2.17–5.42)
  Any previous PICU admission versus none1.91 (1.08–3.36)
  PICU stay during index hospitalization versus none2.42 (1.45–4.05)
Model 3a (N = 597)
Attended both visits0.89 (0.58–1.36)
  1+ asthma hospitalization in previous year versus none3.47 (2.20–5.49)
  Any previous PICU admission versus none1.84 (1.05–3.25)
  PICU stay during index hospitalization versus none2.54 (1.52–4.26)
  • Manual backward selection procedure was used. Bivariate analyses performed with n = 597 led to the same conclusions for model inclusion as in the analysis presented in Table 1. Sixteen children were removed from routine visit and both visit models because they were readmitted before they had the chance for routine follow-up. Variables that are significant (α <0.05) in final model are presented.

  • a Variables included in full model included the following: type of visit attended (variable of interest); age; sex; race and/or ethnicity; asthma hospitalization in the previous year; ED visit in the previous year; asthma scheduled visit in the previous year; previous PICU admission; asthma severity; exposure to smoke, dogs and/or cats, mice, cockroaches, or mold; visit with an asthma specialist in the previous year; PICU stay during the index hospitalization; controller medication prescribed at discharge; postdischarge visit scheduled; and index hospitalization APR-DRG severity of illness.