Table 2.

Mixed effects logistic regression of iatrogenic fistula among all births with fistula

		Likely iatrogenica (“likely” + “probably” + “definitely”)			Probably iatrogenicb (“probably” + “definitely”)			Definitely iatrogenicc (“definitely”)
		n = 791			n = 618			n = 524
Fixed-effect covariatesd	n	Odds ratio	95% confidence interval		Odds ratio	95% confidence interval		Odds ratio	95% confidence interval
Date of fistula development
1994–1999	1478	(Ref)	(Ref)		(Ref)	(Ref)		(Ref)	(Ref)
2000–2004	1870	1.21	0.96	1.52	1.25	0.97	1.61	1.42	1.08	1.88
2005–2009	1211	1.56	1.22	2.00	1.72	1.31	2.25	1.99	1.49	2.67
2010–2017	910	1.94	1.48	2.54	1.90	1.41	2.56	2.20	1.60	3.04
Previous abdominal surgery
No	4965	(Ref)	(Ref)		(Ref)	(Ref)		(Ref)	(Ref)
Yes	501	3.00	2.41	3.73	2.81	2.23	3.53	3.38	2.66	4.30
Parity
1	2762	(Ref)	(Ref)		(Ref)	(Ref)		(Ref)	(Ref)
2	721	2.60	1.92	3.53	2.31	1.63	3.26	2.06	1.43	2.97
3–5	1177	2.98	2.18	4.07	2.82	1.99	3.99	2.37	1.64	3.44
6+	805	4.80	3.32	6.94	4.50	3.01	6.75	3.77	2.44	5.82
Age at fistula development
11–19	2020	(Ref)	(Ref)		(Ref)	(Ref)		(Ref)	(Ref)
20–24	1422	1.79	1.31	2.44	2.02	1.42	2.87	2.04	1.41	2.94
25–29	874	2.62	1.84	3.74	2.82	1.89	4.22	2.62	1.72	4.01
30–34	557	2.64	1.78	3.93	2.91	1.87	4.54	2.59	1.61	4.16
35+	583	2.46	1.62	3.75	2.56	1.60	4.09	2.46	1.49	4.05
Random-effect covariate
Country	5469	0.20	0.06	0.65	0.15	0.04	0.56	0.27	0.08	0.90

^a“Likely iatrogenic” includes: Type III ureteric injuries, Type I VCVF with live baby, Type I vault or Type I VCVF in women with clear obstetric fistula in another location, Type I vault fistula, and Type I VCVF with stillbirth

^b“Probably iatrogenic” includes: Type III ureteric injuries, Type I VCVF with live baby, Type I vault or Type I VCVF in women with clear obstetric fistula in another location, and Type I vault fistula

^c“Definitely iatrogenic” includes: Type III ureteric injuries, Type I VCVF with live baby, and Type I vault or Type I VCVF in women with clear obstetric fistula in another location

^dObservations used in regression = 5451