Table 5.
QASC Europe FeSS management per country income status.
| High income countries | Middle income countries | p * | |||||
|---|---|---|---|---|---|---|---|
| Pre N = 2397, n (%) | Post N = 2199, n (%) | Pre/post changes (95% CI) a | Pre N = 1067, n (%) | Post N = 1058, n (%) | Pre/post changes (95% CI) a | ||
| Participating sites | 44 | 44 | 20 | 20 | |||
| Monitored and treated according to the combined FeSS Protocol b | 87 (3.6) | 773 (35) | 33% (21%, 46%) | 32 (3.0) | 355 (34) | 31% (19%, 43%) | 0.0365 |
| Monitored and treated according to the Fever Protocol c | 468 (20) | 1115 (51) | 31% (28%, 35%) | 139 (13) | 560 (53) | 39% (31%, 47%) | 0.0001 |
| Monitored and treated according to the Hyperglycaemic Protocol d | 525 (22) | 1178 (54) | 31% (28%, 33%) | 83 (7.8) | 524 (50) | 37% (28%, 46%) | <0.0001 |
| Monitored and treated according to the Swallow Protocol e | 835 (35) | 1404 (64) | 27% (25%, 30%) | 525 (49) | 774 (73) | 29% (21%, 37%) | 0.0045 |
Paracetamol, insulin administration and subsequent swallow assessment outcomes include only patients with a fever (Model N = 411 Pre: N = 477 Post), hyperglycaemic event (Model N = 512 Pre: N = 594 Post) or failed swallow screen (Model N = 404: N = 571 Post) within relevant time period respectively. The CIs for difference in proportions overlap due to methods of estimating marginal mean. Odds ratio confidence intervals do not overlap where p < 0.05 (data not shown).
Estimated marginal mean difference in proportion from mixed effects model, standard errors for confidence interval obtained using delta method.
Must meet d, e and f to be deemed as having been monitored and treated according to the combined FeSS Protocol.
Must meet all elements to be deemed as having been monitored and treated according to the Fever Protocol.
Must meet all elements to be deemed as having been monitored and treated according to the Hyperglycaemic (Sugar) Protocol.
Must meet all elements to be deemed as having been monitored and treated according to the Swallow Protocol. *Mixed effects logistic regression controlling for age, sex, NIHSS and correlation within site and country. Model omits patients with missing covariate data (Model N = 2814 Pre: N = 2734 Post).