A mirror-image analysis of psychiatric hospitalisations among people with severe mental illness using Independent Supported Housing

Table 2 Mirror-image analysis of inpatient psychiatric hospitalisations

	Outcome	Mirror-image period	Sample size (n)	pre-index (n)	post-index (n)	IRR	95% CI
	Outcome	Mirror-image period	Sample size (n)	pre-index (n)	post-index (n)	IRR	upper	lower
Primary Analysis	No. of psychiatric admissions	utilisation period	144	68	28	0.41	0.27	0.64
		90 days	121	20	9	0.45	0.20	0.99
		180 days	90	27	10	0.37	0.18	0.77
		270 days	70	33	12	0.36	0.19	0.70
		365 days	47	21	9	0.43	0.20	0.94
	No. of psychiatric person-days (including censored stays)	utilisation period	144	2113	802	0.38	0.35	0.41
		90 days	121	787	149	0.19	0.16	0.23
		180 days	90	908	229	0.25	0.22	0.29
		270 days	70	939	233	0.25	0.21	0.29
		365 days	47	512	219	0.43	0.37	0.50
Sensitivity Analysis	No. of psychiatric admissions	utilisation period	120	26	21	0.81	0.45	1.44
		90 days	102	0	3	-	-	-
		180 days	74	7	7	1.00	0.35	2.85
		270 days	58	10	10	1.00	0.42	2.40
		365 days	42	6	9	1.50	0.53	4.21
	No. of psychiatric person-days (including censored stays)	utilisation period	120	1089	622	0.57	0.52	0.63
		90 days	102	225	30	0.13	0.09	0.20
		180 days	74	299	177	0.59	0.49	0.71
		270 days	58	369	176	0.48	0.40	0.57
		365 days	42	230	211	0.92	0.76	1.11

Primary and sensitivity analyses are conducted with whole utilisation periods and mirror-image periods of 90, 180, 270 and 365 days post- vs. pre- ISH initiation (index). The absolute number of outcome variables occurring in each pre- and post-index period are compared by IRR and are tested for significance by 95% CI. In sensitivity analyses, people were excluded if they had an admission within 90 days before index. Censored stays resulted from hospitalisations with admission before the observation period started
N Number of service users included in the analysis, IRR incidence rate ratio, CI confidence interval, ISH Independent Supported Housing

ISSN: 1471-244X