Table 2.

Univariate and multivariate metaregression for the odds ratio of missing data at the primary endpoint

Type of analysis	Covariates	Odds ratio	P-value	Confidence intervals	Adjusted P-valuea
Univariate regression	Items of data requested (per number of items doubling)	1.27	<0.001	1.13–1.42	<0.001
Time to primary endpoint (per number of days doubling)	1.13	0.001	1.06–1.21	<0.001
Type of outcome		0.02		—b
Symptomc	1.00	—	—
Psychospiritual	1.14	0.7	0.61–2.16
Quality of life	1.91	0.04	1.02–3.57
Other	0.60	0.06	0.36–1.01
	Exclude patients based on age/performance status/extent of disease	1.65	0.02	1.08–2.52	0.03
Multivariate regression	Items of data requested (per doubling number of items)	1.19	0.007	1.05–1.35	0.02
Time to primary endpoint (per doubling number of days)	1.09	0.02	1.02–1.17	0.008

Multivariate model P < 0.001, adjusted R² 21.0%, I² residual = 95.4%, Tau² 0.6.

^aAdjusted for multiple testing (Monte Carlo permutation test).

^bNot possible to determine.

^cReference category.