Table 4

Results of the weighted ANOVAs of quality items on the lifetime prevalence estimates
Quality item k p+ 95% C. I. ANOVA results
pl pu
Target population described:
 Yes 21 0.416 0.346 0.489 QB(1) = 0.77, p = .380; R2 = 0.0
 No 9 0.358 0.260 0.469 QW(28) = 4942.26, p < .001
Non response described:
 Yes 5 0.486 0.401 0.572 QB(1) = 4.42, p = .036; R2 = 0.676
 No 20 0.384 0.343 0.427 QW(23) = 1162.44, p < .001
Response rate reported:
 Yes 27 0.403 0.343 0.467 QB(1) = 0.21, p = .648; R2 = 0.0
 No 3 0.358 0.202 0.551 QW(28) = 4910.99, p < .001
Direct reporting from subjects:
 Yes 26 0.411 0.344 0.480 QB(1) = 0.87, p = .352; R2 = 0.0
 No 4 0.326 0.190 0.498 QW(28) = 4853.39, p < .001
Validated instrument:
 Yes 26 0.400 0.339 0.464 QB(1) = 1.32, p = .250; R2 = 0.015
 No 3 0.517 0.332 0.698 QW(27) = 4754.11, p < .001
Delimitation of pain:
 Yes 20 0.472 0.413 0.532 QB(1) = 17.98, p < .001; R2 = 0.376
 No 10 0.266 0.205 0.338 QW(28) = 2192.08, p < .001
Specification of pain:
 Questionnaire 16 0.399 0.316 0.487 QB(1) = 0.00, p = .999; R2 = 0.0
 Interview 14 0.399 0.311 0.494 QW(28) = 4635.79, p < .001

k: number of studies. p+: mean prevalence. pl and pu: lower and upper confidence limits of the 95% confidence interval around the mean prevalence. QB: between-categories statistic to test the influence of the moderator variable on the prevalence rates. QW: within-categories statistic to test the model misspecification. R2: proportion of variance accounted for by the moderator variable. Five of the 12 quality items were not included in the table due to one of their categories being represented by one study only or by none (survey designed for studying prevalence rates, data collection was uniform, validated interview, validated exam, and memory period clear).

Calvo-Muñoz et al.

Calvo-Muñoz et al. BMC Pediatrics 2013 13:14   doi:10.1186/1471-2431-13-14

Open Data