Skip to main content

Table 2 Distribution of subjects in clusters formed by the reference data (A) and the incomplete data sets (B–F)

From: Searching for the optimal measuring frequency in longitudinal studies -- an example utilizing short message service (SMS) to collect repeated measures among patients with low back pain

  Reference, A, the full data set, 26 weekly measures Kappa Agreement (Weighted Kappa)
Cluster 1 “Fast improvers” Cluster 2 “Normal improvers” Cluster 3 “Slow improvers” Cluster 4 “Indifferent”
B, first 8 weeks
 Cluster 1 N =11 N = 12 N = 0 N = 1 0.272 (0.548)
 Cluster 2 N = 5 N =29 N = 8 N = 0
 Cluster 3 N = 1 N = 9 N =9 N = 13
 Cluster 4 N = 0 N = 13 N = 6 N =12
C, first 13 weeks
 Cluster 1 N =15 N = 3 N = 0 N = 0 0.348 (0.611)
 Cluster 2 N = 0 N =23 N = 2 N = 5
 Cluster 3 N = 0 N = 33 N =7 N = 0
 Cluster 4 N = 2 N = 4 N = 14 N =21
D, first 8 weeks + monthly thereafter
 Cluster1 N =14 N = 4 N = 0 N = 0 0.618 (0.720)
 Cluster 2 N = 1 N =52 N = 8 N = 2
 Cluster 3 N = 0 N = 1 N =8 N = 2
 Cluster 4 N = 2 N = 6 N = 7 N =22
E, every other week
 Cluster 1 N =16 N = 15 N = 0 N = 0 0.642
 Cluster 2 N = 1 N =45 N = 7 N = 3 (0.823)
 Cluster 3 N = 0 N = 1 N =13 N = 0
 Cluster 4 N = 0 N = 2 N = 3 N =23
F, first 18 weeks
 Cluster 1 N =15 N = 1 N = 1 N = 0 0.611 (0.708)
 Cluster 2 N = 0 N =47 N = 3 N = 1
 Cluster 3 N = 0 N = 7 N =8 N = 1
 Cluster 4 N = 2 N = 8 N = 11 N =24
  1. Figures in bold show number of subjects in B–F that are classified in clusters with similar trajectories as A