Skip to main content

Table 4 The accuracy of influenza coding (ICD10 +) among patients with PCR + : A sub-group analysis by selected characteristics and test result availability

From: Evaluation of the accuracy of diagnostic coding for influenza compared to laboratory results: the availability of test results before hospital discharge facilitates improved coding accuracy

  Test results received before discharge
(n = 2450)
Test results received after discharge
(n = 309)
PCR+ ICD10+ Sensitivity
(95% CI)
Adjusted ORa
(95% CI); P-value
PCR+ ICD10+ Sensitivity
(95% CI)
Adjusted ORa
(95% CI); P-value
Type of laboratory test         
 Rapid PCRb 1551 1498 96.6 (95.6–97.4) 2.1 (1.3–3.5); P = 0.01 20 6 30.0 (11.9–54.3) 1.6 (0.4–6.4); P = 0.46
 Multiplex PCR 899 782 87.0 (84.6–89.1) Ref 289 95 32.9 (27.5–38.6) Ref
Season         
 Influenza 2132 2003 94.0 (92.9–94.9) 1.5 (1.0–2.3); P = 0.04 220 67 30.5 (24.4–37.0) 0.7 (0.4–1.2); P = 0.19
 Non-influenza 318 277 87.1 (82.9–90.6) Ref 89 34 38.2 (28.1–49.1) Ref
Hospital         
 A 675 644 95.4 (93.5–96.9) 4.1 (2.0–8.5); P < 0.01 121 32 26.4 (18.8–35.2) 0.5 (0.2–1.4); P = 0.20
 B 472 419 88.8 (85.6–91.5) 1.6 (0.8–3.3); P = 0.16 94 29 30.9 (21.7–41.2) 0.6 (0.2–1.5); P = 0.28
 C 344 325 94.5 (91.5–96.6) 3.1 (1.4–6.6); P < 0.01 39 19 48.7 (32.4–65.2) 1.2 (0.5–3.2); P = 0.69
 D 483 468 96.9 (94.9–98.3) 3.7 (1.5–9.3); P < 0.01 10 2 20.0 (2.5–55.6) 0.2 (0.03–1.8); P = 0.17
 E 219 208 95.0 (91.2–97.5) 2.3 (0.8–6.3); P = 0.12 5 1 20.0 (0.5–71.6) 0.2 (0.02–3.6); P = 0.30
 F 126 125 99.2 (95.6–99.9) 12.9 (1.6–106.7); P = 0.02
 G 131 91 69.5 (60.8–77.2) Ref 40 18 45.0 (29.3–61.5) Ref
Setting where the test was ordered         
 EDc 1492 1410 94.5 (93.2–95.6) 1.4 (0.9–2.0); P = 0.06 192 69 35.9 (29.2–43.2) 1.9 (1.1–3.5); P = 0.03
 Inpatient 958 870 90.8 (88.9–92.6) Ref 117 32 27.4 (19.5–36.4) Ref
  1. aEach analysis was adjusted for patient age, Charlson comorbidity index, year of admission, hospital length of stay and three other variables in this table
  2. b111 patients who were ordered both rapid and multiplex PCR were included in this group for the purpose of this analysis
  3. cThis group included 22 patients who were ordered an influenza test both in ED and inpatient wards