Impact of a continuing medical education meeting on the use and timing of urgent cancer referrals among general practitioners - a before-after study

Table 6 The CME-M impact on patients’ primary care interval (PCI)

	Reference group N = 199 practices				CME-M group N = 139 practices				Comparison between groups
	Before CME-M	After CME-M	Before vs. after		Before CME-M	After CME-M	Before vs. after
			OR0^a (95% CI)	P			OR1^a (95% CI)	P	OR1/OR0^a (95% CI)	P
All data, N patients	1,227	753	-		990	1,102	-		-
PCI, median days (percentiles: 75^th, 90^th)	1 (13, 43)	1 (12, 46)	-		1 (14, 42)	1 (17, 53)	-		-
PCI ≥ baseline 75^th percentile^c %	25.8	24.4	0.81² (0.63;1.04)	0.099	25.1	27.9	1.19^b (0.95;1.49)	0.135^b	1.47 ^b (1.06;2.03)	0.021
PCI ≥ baseline 90^th percentile^c %	10.1	10.6	1.12^b (0.79;1.57)	0.532	10.4	12.5	1.27^b (0.92;1.74)	0.146	1.13^b (0.72;1.78)	0.583
Easier to detect, N patients	887	515	-		685	774	-		-
PCI, median days (percentiles: 75^th, 90^th)	0 (8, 34)	0 (7, 35)	-		0 (7, 35)	0 (9, 41)	-		-
PCI ≥ baseline 75^th percentile^c %	26.2	23.9	0.87 (0.66;1.15)	0.345	27.2	29.7	1.18 (0.92;1.52)	0.187	1.35 (0.94;1.94)	0.101
PCI ≥ baseline 90^th percentile^c %	10.5	10.1	0.97 (0.66;1.44)	0.888	10.2	11.9	1.24 (0.87;1.77)	0.241	1.27 (0.76;2.12)	0.353
Hard to detect, N patients	284	191	-		261	274	-		-
PCI, median days (percentiles: 75^th, 90^th)	7 (26.5, 71)	5 (28, 79)	-		7 (27, 58)	8 (30, 80)	-		-
PCI ≥ baseline 75^th percentile^c %	25.0	25.7	1.12 (0.69;1.79)	0.653	25.7	30.3	1.35 (0.87;2.08)	0.180	1.21 (0.66;2.21)	0.543
PCI ≥ baseline 90^th percentile^c %	10.2	11.5	1.20 (0.62;2.34)	0.584	10.3	13.1	1.34 (0.73;2.46)	0.348	1.11 (0.48;2.59)	0.805

Primary care intervals were measured as median, 75^th and 90^th percentile and the proportion of patients who had the longest primary care interval before and after the CME-M. The analyses are also stratified on cancer detection difficulty. Practices are divided into CME-M group and reference group. An effect within a group is shown as an odds ratio (OR0: Reference group; OR1: CME-M group). Comparisons between groups are shown as a ratio of odds ratios
The suspected cancer types were divided into two groups based on cancer detection difficulty: 1) easier to detect included suspected cancer in kidney, bladder, breast, head and neck, female genitalia, nevus (melanoma), penis, testis and gastrointestinal and 2) harder to detect included suspected cancer in pancreas, liver and gall bladder, brain, lymph node and bone marrow, lung, prostate, connective tissue including fat, muscle and bones, and non-specific serious symptoms
Abbreviations: PCI primary care interval, CME-M continuing medical education meeting, OR Odds ratio, CI confidence interval
^aAdjusted for cluster, practice type, patient gender, age and comorbidity
^bIn addition, adjusted for cancer detection difficulty
^c75^th and 90^th percentiles of baseline measurements (before the CME-M) were used for each group as basis
Bold = significance level of p ≤ 0.05

ISSN: 2731-4553