Knowledge and attitudes towards evidence-based medicine of mentors in general practice can be influenced by using medical students as academic detailers

Figures & data

Table 1. GP's baseline characteristics.

Variables	Intervention group (n = 49)	Control group (n = 49)
Sociodemographics
Female sex, n (%)	40 (81.6)	39 (79.6)
Age (years), median (IQR)	54 (45–57)	51 (46–55)
Work and scientific experience
Working experience in primary practice (years); median (IQR)	26 (16–31)	22 (17–27)
Specialization in family medicine	45 (91.8)	34 (69.4)
Master's or doctoral degree	17 (34.7)	4 (8.2)
Access to specialized EBM databases	25 (51.0)	18 (36.7)
Practice location and organization
Settlement type
Urban	33 (67.3)	31 (63.3)
Suburban	13 (26.5)	11 (22.4)
Rural	3 (6.1)	7 (14.3)
Type of practice
Concession or private	34 (69.4)	36 (73.5)
Health centre	15 (30.6)	13 (26.5)
Practice size (number of insured persons)	1700 (1330–1920)	1780 (1500–2006)
Patients examined daily, median (IQR)	65 (51–75)	65 (55–80)
Patients appointment bookings	21 (42.9)	28 (57.1)

IQR, interquartile range; EBM, evidence based medicine.

Table 2. Self-evaluated knowledge of GPs before and after an educational intervention presented as the percentage out of the maximum achievable score on six five-point scales ranging from ‘insufficient’ to ‘excellent’.

Sample	Before the intervention	After the intervention	Absolute difference	Relative difference	P	r
Whole sample
Intervention group (n = 49)	56 (20.4)	62 (18.1)	6 (12.1)	20 (46.9)	0.038	0.20
Control group (n = 49)	46 (15.0)	47 (16.4)	1 (11.8)	6 (27.9)
Stratified analysis: Ph.D. or specialization in family medicine
Intervention group (n = 37)	57 (21.1)	64 (18.1)	7 (12.2)	22 (51.1)	0.039	0.24
Control group (n = 34)	48 (6.7)	49 (18.3)	1 (12.7)	50 (26.1)
Stratified analysis: no scientific degree nor specialization
Intervention group (n = 12)	51 (18.1)	53 (15.6)	3 (11.8)	12 (29.2)	0.34	0.08
Control group (n = 15)	41 (9.4)	42 (10.0)	1 (9.9)	7 (32.4)

Data are presented as arithmetic mean (standard deviation).

P = Mann–Whitney U test, one-tail Monte Carlo statistical significance on the sample of 10 000 tables; r = standardized effect size calculated as Z/(√(n)).