Table 1.
Baseline patient and VA Medical Center (VAMC)-level characteristics within each service-specific cohort by receipt of low-value preoperative testing before propensity score weighting
| EKG study cohort† | CXR study cohort† | |||
|---|---|---|---|---|
| Low-value EKG group (n=49,820) |
Comparison group (n=586,004) |
Low-value CXR group (n=28,865) |
Comparison group (n=710,140) |
|
| Patient-level characteristics | ||||
| Age, year, mean (SD) | 63.2 (12.4) | 63.5 (13.7) | 64.4 (13.0) | 64.5 (13.7) |
| Male sex, n (%) | 45,269 (90.9) | 532,878 (90.9) | 26,655 (92.3) | 653,626 (92.0) |
| Race/ethnicity, n (%) | ||||
| Non-Hispanic White | 36,060 (72.4) | 415,344 (70.9) | 19,602 (67.9) | 510,843 (71.9) |
| Non-Hispanic Black | 9,843 (19.8) | 118,794 (20.3) | 6,342 (22.0) | 136,961 (19.3) |
| Hispanic | 2,625 (5.3) | 35,460 (6.1) | 2,183 (7.6) | 43,022 (6.1) |
| Non-Hispanic/multiracial | 1,292 (2.6) | 16,406 (2.8) | 738 (2.6) | 19,314 (2.7) |
| VA priority group‡, n (%) | ||||
| 1 | 18,409 (37.0) | 228,698 (39.0) | 10,884 (37.7) | 276,583 (39.0) |
| 2–6 | 24,911 (50.0) | 276,651 (47.2) | 14,278 (49.5) | 332,328 (46.8) |
| 7–8 | 6,500 (13.1) | 80,655 (13.8) | 3,703 (12.8) | 101,229 (14.3) |
| Number of Elixhauser conditions§, mean (SD) | 1.7 (1.8) | 1.5 (1.6) | 1.9 (1.9) | 1.5 (1.6) |
| Driving distance to nearest VA facility, miles, mean (SD) | 14.8 (14.7) | 14.9 (14.6) | 14.3 (14.1) | 15.0 (14.5) |
| VAMC-level characteristics | ||||
| Academic affiliation, yes, n (%) | 48,881 (98.1) | 574,075 (98.0) | 28,577 (99.0) | 696,434 (98.1) |
| Facility size, total # outpatient visits FY18, mean (SD) | 768,038 (366,899) | 802,318 (395,800) | 852,268 (389,269) | 809,367 (400,211) |
| Census region, n (%) | ||||
| Northeast | 5,932 (11.9) | 7,4691 (12.8) | 3,092 (10.7) | 92,170 (13.0) |
| Midwest | 12,677 (25.5) | 133,995 (22.9) | 5,825 (20.2) | 162,437 (22.9) |
| South | 19,154 (38.5) | 256,116 (43.7) | 14,164 (49.1) | 310,237 (43.7) |
| West | 12,057 (24.2) | 121,202 (20.7) | 5,784 (20.0) | 145,296 (20.5) |
| Rurality, n (%) | ||||
| Large metropolitan | 22,682 (45.5) | 623,481 (45.0) | 14,849 (51.4) | 319,720 (45.0) |
| Small metropolitan | 17,488 (35.1) | 210,327 (35.9) | 9,267 (32.1) | 254,668 (35.9) |
| Micropolitan | 5,557 (11.2) | 64,652 (11.0) | 2,644 (9.2) | 77,975 (11.0) |
| Noncore rural | 4,093 (8.2) | 47,544 (8.1) | 2,105 (7.3) | 57,777 (8.1) |
| Facility complexity level‖, n (%) | ||||
| High complexity | 41,687 (83.7) | 478,725 (81.7) | 25,548 (88.5) | 548,240 (82.3) |
| Medium complexity | 4,857 (9.8) | 54,428 (9.3) | 2,064 (7.2) | 65,279 (9.2) |
| Low complexity | 3,276 (6.6) | 52,851 (9.0) | 1,253 (4.3) | 60,621 (8.5) |
EKG electrocardiogram, CXR chest radiograph
†With the exception of driving distance for the EKG study cohort and age and sex for the CXR cohort, the low-value service groups and comparison groups had statistically significant differences in characteristics (p<0.05). Absolute standardized mean differences were <0.10 for all variables after applying propensity score weights (Appendix Table 3 and Appendix Table 5)
‡Veterans are assigned to 1 of 8 priority groups at VA enrollment based on service-connected illnesses, era of Service, and socioeconomic status determined by means testing. Priority group determines level of co-payment. Priority groups are condensed for presentation in the table based on similarity of copays between groups but were included separately in our models
§Individual Elixhauser conditions shown in Appendix Table 3 and Appendix Table 5
‖The complexity rating is based on a VAMC’s patient volume, number and breadth of physician specialists, patient case mix, intensive care unit capabilities, and degree of teaching and research
Missing values for race/ethnicity (4% missing), VA priority group (<0.1%), driving distance to nearest VA facility (0.1%), census region (0.8%), rurality (0.1%) generated using single imputation