Skip to main content
. Author manuscript; available in PMC: 2015 Feb 3.
Published in final edited form as: Health Aff (Millwood). 2013 Oct;32(10):1740–1747. doi: 10.1377/hlthaff.2013.0613

EXHIBIT 1.

Characteristics Of US Hospitals Before And After Matching For Analysis

Variable Before matching (number or mean)
After matching (number or mean)
Standardized difference in means (low versus high) (%)
Initial sample (n = 2,976) Analytic sample (n = 2,826) Low staffing (n = 1,413) High staffing (n = 1,413)
Registered nurse hours per adjusted patient day 6.4 6.5 5.1 8.0 164.1
Balancing score a a 7,170 7,170 0.0
COVARIATES
Number of beds 227 224 219 229 5.3
Resident/fellow physicians per bed 0.06 0.05 0.05 0.06 5.6
Operating profit margin −0.02 −0.02 −0.02 −0.02 0.9
Percent black patients 11.6 11.6 11.6 11.7 0.9
Percent Hispanic patients 2.3 2.4 2.4 2.4 0.5
Percent Medicaid mix 19.1 17.7 17.3 18.0 6.3
Socioeconomic status indicator −0.7 −0.6 −0.6 −0.6 1.0
Herfindahl-Hirschman Index 772 766 802 729 4.7
Technology level (number) 3.5
 High 1,049 1,025 498 527
 Low 1,927 1,801 915 886
Ownership (number) 0.4
 For-profit 578 575 289 286
 Not-for-profit 2,398 2,251 1,124 1,127
Geographic location (number) 2.3
 Urban 2,655 2,548 1,280 1,268
 Rural 321 278 133 145
Skilled nursing facility (number) 4.5
 Yes 485 360 167 193
 No 2,491 2,466 1,246 1,220

SOURCE Authors’ analysis of data from all 2,976 hospitals with data and analytic sample of 2,826 matched hospitals. NOTE Standardized differences represent differences in means in units of standard deviation between the high and low staffing group for each covariate expressed as a percentage.

a

The balancing score for the high and low staffing hospital groups is estimated through the matching process; thus, there are no values prior to matching.