Skip to main content

Table 2 Statistical data of the intra- and interobserver reliabilities of the different histological criteria

From: Development and reliability of the histological THROMBEX-classification rule for thrombotic emboli of acute ischemic stroke patients

Intra- and interobserver reliability

Rater (N)

Few

Moderate

Many

Number of platelets: κ [95% CI]

 R1 (14)

− 0.12 [− 0.31 to 0.07]

0.39 [− 0.08 to 0.87]

0.58 [0.05–1.00]

 R2 (21)

0.7 [0.31–1.00]

0.52 [0.16–0.87]

0.7 [0.4–1.00]

 R1–R2 (130)

0.51 [0.32–0.7]

0.23 [0.10–0.36]

0.19 [0.04–0.34]

Rater (N)

Stratus

Cirrus

Cumulus

Distribution of platelets: α [mean diff.; limit: lower–upper]

 R1 (14)

0.89 [0.19; − 0.18 to 0.55]

0.56 [− 0.19; − 0.75 to 0.36]

0.79 [0.01; − 0.44 to 0.46]

 R2 (22)

0.87 [− 0.02; − 0.46 to 0.42]

0.81 [0.06; − 0.41 to 0.53]

0.84 [− 0.04; − 0.47 to 0.39]

 R1–R2 (127)

0.62 [− 0.07; − 0.69 to 0.55]

0.56 [− 0.01; − 0.59 to 0.56]

0.62 [0.08; − 0.52 to 0.68]

Rater (N)

Fine

Coarse

Dense

Property of fibrin net: α [mean diff.; limit: lower–upper]

 R1 (14)

0.91 [0.01; − 0.25 to 0.27]

0.80 [− 0.01; − 0.33 to 0.30]

0.92 [0.01; − 0.28 to 0.30]

 R2 (23)

0.66 [− 0.14; − 0.59 to 0.31]

0.57 [0.15; − 0.35–0.65]

0.73 [− 0.01; − 0.50 to 0.47]

 R1–R2 (131)

0.73 [0.07; − 0.36 to 0.51]

0.45 [− 0.01; − 0.49 to 0.47]

0.8 [− 0.07; − 0.46 to 0.33]

Rater (N)

Few

Moderate

Many

Number of neutrophil granulocytes: κ [95% CI]

 R1 (14)

0.66 [0.25–1.00]

0.29 [− 0.21–0.78]

0.43 [− 0.10–0.96]

 R2 (23)

0.38 [0.03–0.72]

0.47 [0.13–0.81]

0.64 [0.27–1.00]

 R1–R2 (132)

0.38 [0.20–0.56]

0.22 [0.06–0.39]

0.35 [0.18–0.52]

Number of intact neutrophil granulocytes: α [mean diff.; limit: lower–upper]

 R1 (14)

0.71 [0.03; − 0.28 to 0.34]

 R2 (23)

0.43 [0.02; − 0.65 to 0.68]

 R1–R2 (132)

0.65 [− 0.03; − 0.47 to 0.41]

Number of intact red cells: α [mean diff.; limit: lower–upper]

 R1 (14)

0.94 [0; − 0.23 to 0.22]

 R2 (23)

1.00 [0; − 0.04 to 0.04]

 R1–R2 (132)

0.64 [0.06; − 0.48 to 0.60]

  1. R1, intraobserver reliability for rater I; R2, intraobserver reliability for rater II; R1-R2, interobserver reliability. The related number of cases is indicated in parentheses. Weighted Cohen’s kappa (κ) was calculated for categorically scaled criteria. CI = confidence interval. For the continuously scaled criteria, Cronbach’s alpha is used. The related Bland–Altman plots can be seen in the Additional file 1: Figs. S1, S2, S3. Numerical values indicating sufficient reliability are written in italics. Chosen tolerance levels: κ = 0.5; α = 0.6.
  2. κ < 0 = poor, 0 ≤ κ ≤ 0.2 = slight, 0.2 < κ ≤ 0.4 = fair, 0.4 < κ ≤ 0.6 = moderate, 0.6 < κ ≤ 0.8 = substantial, and 0.8 < κ ≤ 1 = (almost) perfect agreement.
  3. α < 0.5 = unacceptable, 0.5 ≤ α < 0.6 = poor, 0.6 ≤ α < 0.7 = questionable, 0.7 ≤ α < 0.8 = acceptable, 0.8 ≤ α < 0.9 = good, and 0.9 ≤ α = excellent.
Back to article page

Neurological Research and Practice

ISSN: 2524-3489

Contact us