Data in longitudinal randomised controlled trials in cancer pain: is there any loss of the information available in the data? Results of a systematic literature review and guideline for reporting

Table 2 Method of analysis

Primary pain outcome analysed	NRS^a or VAS^b measures	60 % (44/74)
	Change from baseline	24 % (18/74)
	Aggregated NRS or VAS	10 % (7/74)
	Dichotomised outcome	23 % (17/74)
Analysis of repeated measures	Cross sectional at each time-point	37 % (27/74)
Analysis of repeated measures	longitudinal	38 % (28/74)
Method of analysis
Cross sectional	t-test	44 % (12/27)
	ANOVA/ANCOVA	19 % (5/27)
	Non parametric test	36 % (10/27)
Correction for multiple testing due to repeated measures		15 % (4/27)
Longitudinal	Mixed Model	25 % (7/28)
	Repeated measure ANOVA	43 % (12/28)
	GEE	1 % (4/28)
	Area under the curve	7 % (2/28)
	Unknown/other	10 % (3/28)
Results presented
	Mean (SD) at each time-point per group	44 % (31/71^c)
	F values only (for ANOVA/ANCOVA)	29 % (5/17)
	Proportions	59 % (5/17)
	Hazard ratio, odds ratio (with CI)	23 % (4/17)
	None	17 % (12/71^#)

Only one study using longitudinal regression type analysis presented a parameter estimate for group effect
^a Numerical rating scale
^b Visual analogue scale
^c From 74 studies, three were protocols

ISSN: 1471-2407