TABLE 1.

Poisson regression analyses showing trends in treatments

	Poisson GLM					Poisson GLM with population size as offset
	Overall test	Period II vs I a		Period III vs II a		Overall test	Period II vs I a		Period III vs II a
	p‐value	IRR (95% CI) b	p‐value	IRR (95% CI) b	p‐value	p‐value	IRR (95% CI) c	p‐value	IRR (95% CI) c	p‐value
Bed rest for non‐specific low back pain	<0.001	0.47 [0.44 0.51]	<0.001	0.46 [0.42 0.50]	<0.001	< 0.001	0.44 [0.41 0.48]	<0.001	0.42 [0.38 0.45]	<0.001
Bed rest for hernia nuclei pulposi	<0.001	0.72 [0.68 0.76]	<0.001	0.52 [0.49 0.55]	<0.001	< 0.001	0.68 [0.65 0.71]	<0.001	0.48 [0.46 0.51]	<0.001
Discectomy for spinal stenosis	<0.001	0.61 [0.58 0.64]	<0.001	0.28 [0.24 0.32]	<0.001	< 0.001	0.57 [0.54 0.59]	<0.001	0.27 [0.23 0.31]	<0.001
Spinal fusion	<0.001	1.13 [1.04 1.22]	0.005	0.45 [0.40 0.51]	<0.001	<0.001	1.02 [0.95 1.09]	1.0	0.43 [0.38 0.48]	<0.001
Invasive pain treatment	<0.001	2.21 [2.06 2.37]	<0.001	0.66 [0.59 0.74]	0.001	<0.001	1.94 [1.83 2.07]	<0.001	0.58 [0.52 0.66]	<0.001

Abbreviation: GLM, generalized linear model, with robust standard estimates.

^a Cells contain incidence rate ratios (IRR), their 95% confidence intervals and p‐values for comparing the incidences in the two periods. P‐values for comparing periods are Bonferroni corrected p‐values (corrected for two pairwise comparisons).

^b Adjusted for age group, sex and their two‐way interaction.

^c Adjusted for age group, sex and their two‐way interaction, with logarithm of population size used as offset.