Table 1:

Summary of characteristics of 34 studies that measured the proportion of hospital readmissions deemed avoidable

Variable	No. (%) of studies*
Study characteristics
Year of publication, median (IQR)	2000 (1993–2005)
No. of hospitals per study, median (range)	1 (1–234)
Conducted at single centre (n = 31)†	26 (83.9)
Conducted primarily in teaching hospitals (n = 28)‡	18 (64.3)
Index admission used as unit of analysis§	19 (55.9)
No. of index admissions, median (IQR) (n = 19)**	1289 (743–3050)
Follow-up period for readmission, mo, median (IQR)	2 (1–6)
No. of readmissions, median (IQR)	151 (75–313)
Type of patient
Medical	25 (73.5)
Surgical	13 (38.2)
Geriatric	11 (32.4)
Assessment of avoidability (n = 31)††
Information used for assessment
Index admission	25 (80.6)
Clinical visits between index admission and readmission	10 (32.3)
Readmission	27 (87.1)
Interviews with physician or nurse††	7 (22.6)
Interviews with patient or family††	9 (29.0)
Groups or factors included in assessment
Physician	28 (90.3)
Nurse	2 (6.5)
Patient	7 (22.6)
Social	16 (51.6)
System	5 (16.1)
Minimum no. of reviewers, median (range)	1 (1–3)
One reviewer only	17 (54.8)
Outcomes
No. of readmissions deemed avoidable, median (IQR)	35 (17–70)
% of readmissions deemed avoidable, median (IQR)	27.1 (14.9–45.6)
% of index admissions followed by an avoidable	2.2 (1.5–7.0)
readmission, median (IQR) (n = 19)

Note: IQR = interquartile range.

^*Unless stated otherwise.

^†Number of included hospitals not stated in three studies.^10,22,27

^‡The teaching status of included hospitals was not stated in six studies.^{10,18,22,27,30,33}

^§The unit of analysis was the readmission in the other 15 studies.

^**The denominator comprises the 19 studies in which the unit of analysis was the index admission.

^††Excludes data from the three studies based on administrative databases alone.^18,27,33