Using Machine Learning to Predict Likelihood and Cause of Readmission After Hospitalization for Chronic Obstructive Pulmonary Disease Exacerbation

Figures & data

Table 1 Patient Demographics by Readmission Status

				Readmitted
Values Reported as n(%), Mean (± SD), or Median (IQR)	Total (n=3238)	Not Readmitted (n=2142)	Readmitted (n=1103)	Respiratory (n = 672)	Non-Respiratory (n=424)
Female	2007 (62%)	1306 (61%)	701 (64%)	427 (64%)	268 (63%)
Black	2917 (90%)	1891 (89%)	1026 (93%)^†	635 (94%)	386 (91%)^‡
Hispanic	37 (1%)	28 (1%)	9 (1%)	6 (1%)	3 (1%)
Age, years	64 (± 11)	65 (± 11)	64 (± 11)	63 (± 10)	65 (± 11)^‡
BMI, kg/m^2	26.7 (21.9, 33.3)	26.7 (21.9, 33.4)	26.7 (21.9, 33.0)	25.9 (21.3, 31.6)	27.7 (22.7, 34. 7)^‡
Elixhauser Criteria for Renal Failure*	524 (16%)	287 (13%)	237 (21%)^†	108 (16%)	129 (30%)^‡
Elixhauser Criteria for CHF*	1034 (32%)	576 (27%)	458 (42%)^†	235 (35%)	220 (52%)^‡
AHRQ Elixhauser Index	5 (0, 17)	3 (0, 14)	9 (0, 22)^†	8 (0, 19.3)	12 (2, 24)‡
HOSPITAL Score	2 (1, 4)	2 (1, 4)	3 (2, 4)^†	3 (2, 5)	3 (2, 4)
Length of Stay, days	3.5 (2.1, 5.6)	3.3 (2.1, 5.6)	3.7 (2.2, 5.7)	3.5 (2.2, 5.1)	3.8 (2.2, 6.1)

Notes: *Criteria fulfilled from previously documented admission prior to index admission; BMI = body mass index. ^†p < 0.05 when compared to not readmitted cohort; ^‡p < 0.05 when compared to respiratory cohort.

Abbreviations: CHF, congestive heart failure; AHRQ, Agency for Healthcare Research and Quality.

Figure 1 (A) ROC curves of 90-day readmission risk random forest model and HOSPITAL score. (B) ROC curve of 90-day readmission cause model.

Figure 2 Important variables in 90-day readmission risk model.

Abbreviations: LOS, length of stay; BMI, body mass index; BUN, blood urea nitrogen.

Figure 3 Important variables in 90-day readmission cause model.

Abbreviations: BUN, blood urea nitrogen; BMI, body mass index; LOS, length of stay.

Data Availability Statement

Since these data are highly granular and contains potentially sensitive patient information, public sharing of the data would breach the University of Chicago’s IRB protocol requirements. The data used in this study was accessed in a way that was compliant with relevant data protection and privacy regulations per the University of Chicago’s IRB (IRB #17-0332). Interested researchers may contact Mary Akel, University of Chicago, via email at [email protected] for data access requests.