Improving question formulation for use in evidence appraisal in a tertiary care setting: a randomised controlled trial [ISRCTN66375463]

Table 5 Question specificity by group assignment at baseline and follow-up.^*

Specificity	Baseline Frequency (%)	Follow-up Frequency (%)	Within-Group Pre-Post Comparison (p-value)†
Intervention (n = 17)
Primary Outcome‡			0.008
Properly specified	3 (17.7)	10 (58.8)
Mis-specified	14 (82.4)	7 (41.2)
Secondary Outcomes
Patient			0.028
Absent	5 (29.4)	2 (11.8)
Broad	9 (52.9)	5 (29.4)
Specific	3 (17.7)	10 (58.8)
Intervention			0.157
Absent	0 (0.0)	0 (0.0)
Broad	6 (35.3)	4 (23.5)
Specific	11 (64.7)	13 (76.5)
Comparison			0.014
Absent	11 (64.7)	5 (29.4)
Broad	0 (0.0)	0 (0.0)
Specific	6 (35.3)	12 (70.6)
Outcome			0.008
Absent	3 (17.7)	1 (5.9)
Broad	11 (64.7)	7 (41.2)
Specific	3 (17.7)	9 (52.9)
Control Group (n = 22)
Primary Outcome‡			--
Properly specified	5 (22.7)	5 (22.7)
Mis-specified	17 (77.3)	17 (77.3)
Secondary Outcomes
Patient			0.083
Absent	5 (22.7)	4 (18.2)
Broad	12 (54.6)	11 (50.0)
Specific	5 (22.7)	7 (31.8)
Intervention			0.564
Absent	2 (9.1)	1 (4.6)
Broad	10 (45.5)	11 (50.0)
Specific	10 (45.5)	10 (45.5)
Comparison			0.974
Absent	16 (72.7)	17 (77.3)
Broad	1 (4.6)	0 (0.0)
Specific	5 (22.7)	5 (22.7)
Outcome			0.961
Absent	3 (13.6)	4 (18.2)
Broad	13 (59.1)	12 (54.6)
Specific	6 (27.3)	6 (27.3)

^* Percentages may not sum to 100 due to rounding. † Two-sided Wilcoxon signed ranks test. ‡ A properly specified question was one that had an explicit statement for each of the four dimensions of question specificity. If a question had at least one dimension coded as absent, it was classified as mis-specified.

ISSN: 1472-6947