TABLE 4.
Performance of Diagnostic Testing Based on CPG Specific Recommendations
Group | Hospitals, n (%) | Patients, n | % Received | P a | OR (95% CI) | P b |
---|---|---|---|---|---|---|
CBC count | ||||||
No CPG | 28 (68.3) | 13 265 | 50.3 | <.0001 | Ref | |
CPG: obtain test | 6 (14.6) | 2660 | 57.2 | 1.3 (0.7–2.4) | .37 | |
CPG: do not obtain | 6 (14.6) | 3340 | 39.4 | 0.6 (0.4–1.0) | .06 | |
CPG: test not mentioned | 1 (2.4) | 445 | 72.1 | 2.6 (1.9–3.5) | <.001 | |
Serum electrolytes | ||||||
No CPG | 28 (68.3) | 13 265 | 28.8 | .46 | Ref | |
CPG: obtain test | 2 (4.9) | 504 | 29.6 | 1 (0.5–2.1) | .92 | |
CPG: do not obtain | 0 (0.0) | — | — | — | — | |
CPG: test not mentioned | 11 (26.8) | 5941 | 28.0 | 1 (0.5–1.8) | .90 | |
Blood culture | ||||||
No CPG | 28 (68.3) | 13 265 | 41.8 | <.0001 | Ref | |
CPG: obtain test | 6 (14.6) | 2818 | 40.9 | 1 (0.5–1.7) | .90 | |
CPG: do not obtain | 5 (12.2) | 2916 | 24.9 | 0.5 (0.1–1.7) | .24 | |
CPG: test not mentioned | 2 (4.9) | 711 | 60.9 | 2.2 (1.1–4.4) | .03 | |
C-reactive protein | ||||||
No CPG | 28 (68.3) | 13 265 | 17.6 | <.0001 | Ref | |
CPG: obtain test | 2 (4.9) | 1235 | 50.6 | 4.8 (2.3–10.2) | <.0001 | |
CPG: do not obtain | 4 (9.8) | 1869 | 8.2 | 0.4 (0.1–1.8) | .24 | |
CPG: test not mentioned | 7 (17.1) | 3341 | 13.1 | 0.7 (0.3–1.9) | .50 | |
Viral testing (RSV, influenza) | ||||||
No CPG | 28 (68.3) | 13 265 | 28.0 | <.0001 | Ref | |
CPG: obtain test | 3 (7.3) | 954 | 38.7 | 1.6 (1–2.6) | .04 | |
CPG: do not obtain | 3 (7.3) | 2321 | 15.5 | 0.5 (0.2–1.5) | .19 | |
CPG: test not mentioned | 7 (17.1) | 3170 | 25.5 | 0.9 (0.4–2.1) | .76 | |
CXR | ||||||
No CPG | 28 (68.3) | 13 265 | 63.6 | <.0001 | Ref | |
CPG: obtain test | 8 (19.5) | 4374 | 60.5 | 0.9 (0.6–1.4) | .57 | |
CPG: do not obtain | 0 (0.0) | — | — | — | — | |
CPG: test not mentioned | 5 (12.2) | 2071 | 61.3 | 0.9 (0.6–1.4) | .66 |
RSV, respiratory syncytial virus; —, no patients.
P values comparing differences in proportion of patients receiving testing based on specific mention in the CPG are unadjusted and aggregate (combining across hospitals).
P values are adjusted for hospital clustering.