Table 2.

 Regression analysis with annual change in rate of NHS civil involuntary admissions between 1988 and 2008 in England at subsequent year as dependent variable, and annual changes in non-secure mental illness bed provision, before and after adjusting for other variables in table

Annual change in	Annual change in civil involuntary admissions in subsequent year
	Unadjusted		Adjusted*
	B (SE)	P value	B (SE)	P value
Non-secure beds in given year	−0.55 (0.10)	<0.001	−0.49 (0.14)	0.005
Non-secure beds in subsequent year	−0.32 (0.14)	0.039	−0.08 (0.15)	0.610
Civil involuntary admissions in given year	0.42 (0.24)	0.100	−0.01 (0.18)	0.960

Outlier caused by change in method of recording involuntary admissions between 1995 and 1996 was excluded.

*Overall: F=8.5, r²=0.66, P=0.002.