Table A3. Overview of the characteristics of the case–control studies considered and their methods for correcting for self-selection bias.

Study/reference year	Age group studied	Controls	Self-selection adjustment	Unadjusted odds ratio of breast cancer mortality in women attending screening vs those who did not (95% CI) age groups not reported where combined figure available	Adjusted odds ratio of breast cancer mortality in women attending screening vs those that did not (95% CI)
Gullberg et al, 1991	Not reported	Matched by age	Adjustment considered but not made	0.42 (0.22–0.78)	N/A
Palli et al, 1989	40–69 at diagnosis	5 Matched by year of birth and residence	Adjustment has been attempted using proxy variables	Age 40–49: 0.63 (0.24–1.64) Age 50–69: 0.51 (0.29–0.89)	Age 40–49: 1.46 (0.36–5.93) Age 50–69: 0.35 (0.14–0.85)a
Puliti et al, 2008	Residents aged 50–74	4 Matched by date of birth and residence	Adjustment using women in areas where the breast screening was not underway.	0.46 (0.38–0.56)	0.55 (0.36–0.85)b
Collette et al, 1984	50–64 at study entry	3 Matched by year of birth and residence	Adjustment considered but not made	0.30 (0.13–0.70)	N/A
Miltenburg et al, 1998	50–69 at first invitation	3 Matched by year of birth, invitation and residence	Adjustment considered but not made	0.54 (0.37–0.79)	N/A
Verbeek et al, 1984	35–65 at first invitation	5 Matched by year of birth	Adjustment considered but not made	0.48 (0.23–1.00)	N/A
Verbeek et al, 1985	35–65 at first invitation	5 Matched by year of birth	Adjustment considered but not made	0.51 (0.26–0.99)	N/A
Van Dijck et al, 1996	65–92 at last invitation	5 Matched by age and invitation	Adjustment considered but not made	0.56 (0.28–1.13)	N/A
Broeders et al, 2002	35+ at first invitation	5 Matched by age, invitation and residence	Adjustment considered but not made	Age 40–49: 0.84 (0.30–2.29) Age 50–59: 0.65 (0.30–1.42) Age 60–69: 0.65 (0.31–1.28) Age 70–79: 0.70 (0.32–1.54)	N/A
Van Schoor et al, 2010	40–69 at invitation			Age 40–49: 0.50 (0.30–0.82) Age 50–59: 0.54 (0.35–0.85) Age 60–69: 0.65 (0.38–1.13)	N/A
Van Schoor et al, 2011	50–69 at invitation	5 Matched by residence	Adjustment using the Duffy method	—	0.65 (0.49–0.87)
Paap et al, 2010	50–75 at invitation	1 Matched by year of birth and residence	Adjustment using the Duffy method	0.30 (0.14–0.63)	0.24 (0.10–0.58)b
Otto et al, 2012	49–75 at first invitation	1 Matched by invitation and year of birth	Adjustment using women in same area before breast screening.	0.45 (0.37–0.54)	0.51 (0.40–0.66)b
Friedman and Dubin, 1991	40–64 at entry	4 Matched by survival and year of birth	Adjustment considered but not made	0.75 (0.60–0.93)	N/A
Norman et al, 2007	40–64 at entry	Matched within original random control group	Adjustment considered but not made	Age 40–49: 0.89 (0.65–1.23) Age 50–64: 0.47 (0.35–0.63)	N/A
Elmore et al, 2005	40–65 at diagnosis	1 Matched by age, risk and health plan	Adjustment considered but not made	0.92 (0.79–1.08)c	0.92 (0.79–1.08)c
Moss et al, 1992	45–64 at entry	5 Matched on age at entry to trial and residence	Adjustment considered but not made	0.51 (0.27–0.98)d	N/A
Fielder et al, 2004	50–74 at diagnosis	5 Matched on year of birth and GP Practise	Adjustment using the Duffy method	0.49 (0.36–0.66)	0.75 (0.49–1.14)b
Allgood et al, 2008	50–70 at diagnosis	2 Matched on year of birth	Adjustment using the Duffy method	0.35 (0.24–0.50)	0.52 (0.32–0.84)b
Roder et al, 2008	45–80 at death of cases	3 Matched on date of birth	Adjustment has been attempted using socio-economic status		0.59 (0.47–0.74)
Gabe et al, 2007	⩾40 at death	3/4 Matched by date of birth and residence	Adjustment using the Duffy method	0.59 (0.41–0.84)	0.65 (0.39–1.09)b

Abbreviations: CI=confidence interval; IMPACT, DOM, CARE=Women's Contraceptive and Reproductive Experiences Study; HIP=health insurance plan; OR=odds ratio; TEDBC=trial of early detection of breast cancer.

Odds ratios for age groups have been presented where an overall odds ratio is not available.

^aAt least 2 examinations and last screen <30 months.^bCorrected for self-selection bias. ^cCombined risk groups, screening by mammogram. ^dOutcome of the analysis conducted solely in the screening district.