Does the type of abortion provider influence contraceptive uptake after abortion? An analysis of longitudinal data from 64 health facilities in Ghana

Table 3 Poisson regression estimates of main and interaction effects for postabortion contraceptive uptake and uptake of a long-acting and permanent rather than a short-acting contraceptive method

	Receipt of any postabortion contraception (n = 26,125)			Receipt of LA/PM for women who receive postabortion contraception (n = 17,456)
	Main effects	Interaction effects
	Uptake of any contraception	Uptake of any contraception by induced clients	Uptake of any contraception by PAC clients
	RR (95 % CI)	RR (95 % CI)	RR (95 % CI)	RR (95 % CI)
Type of provider
House Officer	1			1
Physician	0.98 (0.93, 1.02)			1.58 (1.37, 1.83)
Midwife	1.18 (1.13, 1.24)			1.46 (1.23, 1.73)
Age category
10-19		1	1	1
20-29		1.05 (1.03, 1.07)	0.92 (0.88, 0.95)	1.05 (0.97, 1.13)
30-39		1.11 (1.08, 1.13)	0.87 (0.83, 0.90)	1.52 (1.41, 1.64)
40-49		1.10 (1.05, 1.16)	0.87 (0.81, 0.95)	1.96 (1.75, 2.20)
Trimester
1st tri	1			1
2nd tri	0.98 (0.94, 1.02)			1.07 (0.96, 1.73)
Indication for abortion
Postabortion care				1
Induced abortion				1.35 (1.24, 1.46)

LA/PM, long-acting and permanent method; RR, risk ratio; CI, confidence interval
All models include a facility fixed effect, an interaction term between client age category and indication for abortion (induced or PAC), and are adjusted for provider type, client age category, trimester, and indication for abortion
We include columns for induced and PAC procedures for the estimated RR for each age category because the association between age and receipt of contraception is modified by whether the procedure is considered a PAC or an induced procedure ([see Additional file 1] for additional analysis of the interaction between age category and abortion indication)

ISSN: 1471-2458