Table 4. Estimated adjusted effects of platelet count at enrolment (baseline value) and ensuing serial platelet data (current value, % change from previous day) for each day of illness, on the subsequent development of DSS.

Associated discrimination (AUC) is also shown.

	Day 3 of illness		Day 4 of illness		Day 5 of illness
Number of events/sample size	59/877 (7%)		37/848 (4%)		15/823 (2%)
	OR (95% CI)	AUC	OR (95% CI)	AUC	OR (95% CI)	AUC
Baseline value, i.e. value on day 3 of illness [+ 10,000 cells/mm3]	0.91 (0.86, 0.96)	0.68	0.97 (0.91, 1.03)	0.66	1.01 (0.92, 1.10)	0.51
Current value [+ 10,000 cells/mm3]	0.91 (0.86, 0.96)	0.68	0.89 (0.83, 0.95)	0.72	0.83 (0.71, 0.94)	0.67
% change from previous day [+ 10%]	-	-	0.72 (0.61, 0.84)	0.74	0.68 (0.54, 0.85)	0.73

These analyses were performed on patients enrolled on day 3 of illness only.

For each day of illness, the sample size refers to the number of patients still at risk on that day (i.e. those without DSS up to that day), and the number of events refers to the number of patients who developed DSS on any subsequent day.

Each analysis is based on a logistic regression model with the development of DSS on subsequent days (yes/no) as the outcome, the specific aspect of platelet count dynamics (baseline or current value, or % change) as the main covariate and other predictors from the reduced logistic regression model at baseline (age, sex, history of vomiting, temperature, palpable liver) as additional covariates.

OR: Odds Ratio, 95% CI: 95% Confidence Interval, AUC: Area Under the ROC Curve (corrected for optimism by cross-validation).