Table 2

Change in trends and levels of major complications following accreditation of level I and II centres*

Centres	Intercept		Preperiod trend		Change in level		Change in trend
Centres	(%)	95% CI	Estimate	95% CI	Estimate	95% CI	Estimate	95% CI
Level I
Centre 1 (n=4531)†
Crude	38.85	31.66 to 46.04	‡	‡	0.32	−6.78 to 7.42	–	–
Standardised	36.99	31.38 to 42.98	‡	‡	−1.49	−7.00 to 4.03	–	–
Centre 2 (n=5300)
Crude	21.21	17.45 to 24.97	0.21	0.07 to 0.36	−0.87	−6.05 to 4.30	−0.30	−0.47 to −0.13
Standardised	24.98	22.78 to 27.18	0.13	0.05 to 0.22	−1.37	−4.42 to 1.68	−0.25	−0.35 to −0.15
Centre 3 (n=4700)†
Crude	19.25	11.99 to 26.50	‡	‡	−1.03	−6.77 to 4.71	–	–
Standardised	19.76	15.13 to 24.39	‡	‡	−1.15	−4.74 to 2.44	–	–
All§
Crude	25.42	23.04 to 27.80	0.10	0.01 to 0.19	−0.34	−3.97 to 3.29	−0.14	−0.25 to −0.03
Standardised	27.75	25.70 to 29.79	0.07	0.00 to 0.14	−2.03	−4.79 to 0.73	−0.12	−0.20 to −0.03
Level II
Centre 4 (n=1930)
Crude	20.52	15.53 to 25.50	−0.13	−0.65 to 0.39	−2.99	−11.05 to 5.07	¶	¶
Standardised	22.17	18.65 to 25.69	−0.21	−0.58 to 0.16	−2.53	−8.21 to 3.15	¶	¶
Centre 5 (n=802)
Crude	12.43	7.82 to 17.05	1.26	0.78 to 1.75	−14.85	-−21.10, to 8.61	-−1.13	-1.75, to 0.51
Standardised	19.18	14.68 to 23.68	0.60	0.12 to 1.08	−9.60	−15.77 to −3.43	−0.63	−1.24 to −0.02
Centre 6 (n=1361)
Crude	20.26	15.16 to 25.37	−0.01	−0.59 to 0.57	0.03	−7.07 to 7.12	0.16	−0.49 to 0.82
Standardised	24.24	18.79 to 29.68	−0.34	−0.95 to 0.28	−0.34	−7.90 to 7.23	0.63	−0.07 to 1.32
Centre 7 (n=826)
Crude	35.17	22.69 to 47.65	−0.42	−1.72 to 0.88	−1.70	−20.83 to 17.44	‡	‡
Standardised	38.61	29.82 to 47.41	−0.77	−1.68 to 0.14	−1.01	−14.46 to 12.43	‡	‡
All§
Crude	20.75	16.86 to 24.64	−0.04	−0.45 to 0.36	−1.00	−6.91 to 4.91	0.09	−0.34 to 0.52
Standardised	22.77	19.93 to 25.61	−0.20	−0.49 to 0.10	−1.02	−5.29 to 3.25	0.21	−0.17 to 0.60
Paediatric centres**
Centres 8 and 9 (n=715)
Crude	17.16	10.46 to 23.86	−0.86	−1.87 to 0.16	−1.86	−8.01 to 4.29	1.92	0.14 to 3.69
Standardised	18.01	9.36 to 26.67	−0.82	−2.25 to 0.61	0.83	−3.72 to 5.40	0.84	−0.80 to 2.47

*n: total number of admissions during the study period. Estimates and 95% CIs are obtained from segmented regressions with autocorrelated errors. Rubin’s rule was used to combine the estimates from the 20 imputed data sets.
†Due to collinearity, we only modelled a change in level after accreditation for centre 1 and centre 3.
‡Non-linear trend modelled using restricted cubic splines.
§Risk differences and 95% CIs were obtained from a generalised estimated equation, with robust SEs.
¶Non-linear trendmodeled using quadratic terms.
**Estimates and 95% CIs were obtained from a generalised estimating equation, with robust SEs. The time unit was semester due to the small sample size, and we only adjust for age, gender, body region of the most severe injury, injury severity and mechanism.
RD, risk difference.