Exhibit 5.

Predictors of arrest for violent crime in people with a gun-disqualifying mental health adjudication record or nondisqualifying short-term involuntary hold

Predictor	Odds ratio	95% confidence interval	p value
Consecutive months of community tenure	1.00	(1.00, 1.01)	0.044
Gun-disqualifying mental health adjudication/NICS reporting
Disqualified, before NICS reporting began (ref.)	—a
Disqualified, after NICS reporting began	0.47	(0.38, 0.58)	<0.0001
Not disqualified, before NICS reporting began	0.75	(0.65, 0.87)	<0.0001
Not disqualified, after NICS reporting began	0.57	(0.48, 0.68)	<0.0001
Primary psychiatric diagnosis
Major depression (ref.)	—a
Schizophrenia spectrum	1.09	(0.99, 1.19)	0.065
Bipolar disorder	1.22	(1.10, 1.34)	<0.0001
Substance use disorder
No co-occurring alcohol or drug use disorder (ref.)	—a
Any co-occurring alcohol or drug use disorder	1.63	(1.50, 1.78)	<0.0001
Age	0.97b	(0.97, 0.97)	<0.0001
Sex
Female (ref.)	—a
Male	1.77	(1.64, 1.92)	<0.0001
Race/ethnicity
White (ref.)	—a
Black/African American	1.44	(1.31, 1.57)	<0.0001
Hispanic/Latino	1.07	(0.97, 1.18)	0.157
Other race/ethnicity	0.87	(0.62, 1.24)	0.448

Source Florida administrative record data assembled for this study, 2002–11.

Notes Analytic model specifications: generalized estimating equations logistic regression for repeated measures with a lagged dependent variable, controlling for time and adjusting for nonindependence of intraperson observations. Excludes observation months in which individuals were incarcerated, hospitalized, or legally ineligible to access firearms because of a felony criminal conviction. Estimates are based on 3,665 arrests in 33,815 individuals over a combined 3,374,173 months of observation during 2002–11. NICS is National Instant Criminal Background Check System.

^aValue of reference category is 1.00.

^bThis odds ratio refers to the change associated with each year of increasing age.