Firearms and Suicide in the United States: Is Risk Independent of Underlying Suicidal Behavior?

Abstract

On an average day in the United States, more than 100 Americans die by suicide; half of these suicides involve the use of
firearms. In this ecological study, we used linear regression techniques and recently available state-level measures of suicide
attempt rates to assess whether, and if so, to what extent, the well-established relationship between household firearm ownership
rates and suicide mortality persists after accounting for rates of underlying suicidal behavior. After controlling for state-level
suicide attempt rates (2008–2009), higher rates of firearm ownership (assessed in 2004) were strongly associated with higher
rates of overall suicide and firearm suicide, but not with nonfirearm suicide (2008–2009). Furthermore, suicide attempt rates
were not significantly related to gun ownership levels. These findings suggest that firearm ownership rates, independent of
underlying rates of suicidal behavior, largely determine variations in suicide mortality across the 50 states. Our results
support the hypothesis that firearms in the home impose suicide risk above and beyond the baseline risk and help explain why,
year after year, several thousand more Americans die by suicide in states with higher than average household firearm ownership
compared with states with lower than average firearm ownership.

Full text

Original Contribution
Firearms and Suicide in the United States: Is Risk Independent of Underlying
Suicidal Behavior?
Matthew Miller*, Catherine Barber, Richard A. White, and Deborah Azrael
*Correspondence to Dr. Matthew Miller, Department of Health Policy and Management, Harvard School of Public Health, Room 305,
Kresge Building, 677 Huntington Avenue, Boston, MA 02115 (e-mail: mmiller@hsph.harvard.edu).
Initially submitted December 13, 2012; accepted for publication March 8, 2013.
On an average day in the United States, more than 100 Americans die by suicide;half of these suicides involve the
use of firearms. In this ecological study, we used linear regression techniques and recently available state-level mea-
sures of suicide attempt rates to assess whether, and if so, to what extent, the well-established relationship between
household firearm ownership rates and suicide mortality persists after accounting for rates of underlying suicidal behav-
ior. After controlling for state-level suicide attempt rates (2008–2009), higher rates of firearm ownership (assessed in
2004) were strongly associated with higher rates of overall suicide and firearm suicide, but not with nonfirearm suicide
(2008–2009). Furthermore, suicide attempt rates were not significantly related to gun ownership levels. These findings
suggest that firearm ownership rates, independent of underlying rates of suicidal behavior, largely determine varia-
tions in suicide mortality across the 50 states. Our results support the hypothesis that firearms in the home impose
suicide risk above and beyond the baseline risk and help explain why, year after year, several thousand more Ameri-
cans die by suicide in states with higher than average household firearm ownership compared with states with lower
than average firearm ownership.
firearms; guns; suicidality; suicide; suicide attempts
Abbreviation: BRFSS, Behavioral Risk Factor Surveillance System.
On an average day in the United States, more than100 Amer-
icans die by suicide; half of these suicides involve the use of
firearms (1). Suicide rates, both overall and by firearms, are
higher, in general, in places where household firearm owner-
ship is more common. By contrast, rates of suicide by methods
other than firearms are not significantly correlated with rates of
household firearm ownership (2,3). This pattern of higher
suicide rates in places where firearms are more readily avail-
able, driven by higher firearm suicide rates, has been reported
in ecological studies that have adjusted for several potential
confounders, including aggregate measures of psychological
distress, degree of urbanization, alcohol and illicit drug use
and abuse, poverty, education, and unemployment (2–6). House-
hold firearm ownership has also consistently been found to be
a strong predictor of suicide risk in studies that use individual-
level data. Every US case-control study, for example, has found
that the presence of a gun in the home is a risk factor for suicide
(7–20). In addition, the only large US cohort study to examine
the firearm-suicide connection found that suicide rates among
California residents who purchased handguns from licensed
dealers were more than twice as likely to die by suicide as
were age- and sex-matched members of the general popula-
tion, not only immediately after the purchase, but throughout
the 6-year study period (21). Here too, the increase in suicide
risk was attributable entirely to an excess risk of suicide with
firearms (21).
The following observations further support the plausibil-
ity that the association between firearms and suicide is real:
1) the association is robust to adjustment for measures of psy-
chopathology (7–16), 2) the risk extends beyond the gun owner
to all household members (14,15,21) and persists for years
after firearms are purchased (14,15,21), 3) the rates of psy-
chiatric illness and psychosocial distress are similar among
households with firearms versus those without firearms (15,
22–25), and 4) ecological studies of the firearm-suicide rela-
tionship, which are not subject to recall bias or to reverse
946 Am J Epidemiol. 2013;178(6):946–955
American Journal of Epidemiology
© The Author 2013. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of
Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
Vol. 178, No. 6
DOI: 10.1093/aje/kwt197
Advance Access publication:
August 23, 2013
by guest on July 25, 2016http://aje.oxfordjournals.org/Downloaded from

causation, yield associations similar to those observed in
individual-level studies. Nevertheless, the idea that the avail-
ability of firearms plays an important role in determining a
person’s suicide risk and a population’s suicide rate continues
to meet with skepticism, the most decisive objection being
that empirical studies to date have not adequately controlled
for the possibility that members of households with firearms
are inherently more suicidal than members of households with-
out firearms (26).
The current study takes advantage of recently available state-
level suicide attempt data to put to test, for the first time, the
hypothesis proffered by critics of the empirical literature that the
association between firearm ownership and suicide mortality
reflects unmeasured suicidal proclivities associated with firearm
ownership rather than an independent risk of death by suicide
conferred by accessto guns. We test this hypothesis by assessing
whether the association between state-level firearm ownership
and completed suicide is robust to simultaneously accounting
for suicide attempt rates.
MATERIALS AND METHODS
Suicide mortality data for each state were obtained through
the Centers for Disease Control and Prevention’s (Atlanta,
Georgia) Web-Based Injury Statistics Query and Reporting
System (1). Suicide data are grouped by firearm methods (Inter-
national Classification of Diseases, Tenth Revision, E codes
X72–X74) and nonfirearm methods (International Classifi-
cation of Diseases, Tenth Revision, E codes X60 –X71, X75–
X84, Y87.0, and U03). Analyses are based on mortality data
among adults aggregated over the 2-year period of 2008–
2009, corresponding to the time period for which contempo-
raneous suicide attempt data are available.
State-level data on the percentage of individuals living in
households with firearms (gun prevalence) were obtained
from the 2004 Behavioral Risk Factor Surveillance System
(BRFSS); 2004 is the most recent year for which state-level
estimates are available (27). The BRFSS, the world’s largest
telephone survey (with more than 200,000 adult respondents
annually), is an ongoing data collection program sponsored
by the Centers for Disease Control and Prevention, with all
50 states participating. A detailed description of the survey
methods used by BRFSS is available elsewhere (28–30).
Briefly, trained interviewers collect data on a monthly basis
by using an independent probability sample of households
with telephones among the noninstitutionalized US popula-
tion aged 18 years or older. All BRFSS questionnaires and
data are available on the Internet (www.cdc.gov/brfss). Firearm
ownership information was obtained by interviewers who began
the firearm section of the survey by first informing respon-
dents that “the next three questions are about firearms. We
are asking these in a health survey because of our interest in
firearm-related injuries. Please include weapons such as pistols,
shotguns, and rifles; but not BB guns, starter pistols, or guns
that cannot fire. Include those kept in a garage, outdoor stor-
age area, or motor vehicle.”Presence of firearms in the home
was assessed by asking respondents, “Are any firearms kept
in or around your home?”Firearm prevalence estimates
exclude respondents who did not know or refused to answer
the BRFSS firearm questions. Firearm ownership prevalence
data are also available from 2001 and 2002; the correlation
between the 2004 state-level measures of firearm ownership
and measures from 2002 and 2001 is nearly perfect (ρ=
0.98). In California, interviewers did not ask the household
firearms question in 2004; we substituted the firearm preva-
lence estimate from the 2002 questionnaire.
State-level measures of past-year suicide attempts were
obtained from the 2008 National Survey on Drug Use and
Health (31). The National Survey on Drug Use and Health is
a national- and state-level survey of a representative sample
of the noninstitutionalized US civilian population aged 12
years or older. The National Survey on Drug Use and Health
collects data on health risks related to the use of illicit drugs,
alcohol, and tobacco; initiation of substance use; substance
use disorders and treatment; health care; and mental health. The
report from which suicide attempt data were obtained asked
about suicidal thoughts and behaviors among adults only
(92,264 respondents aged 18 years or older in all 50 states in
the 2008 and 2009 surveys). Respondents were first asked, “The
next few questions are about thoughts of suicide. At any time
in the past 12 months, that is from [datefill] up to and includ-
ing today, did you seriously think about trying to kill your-
self?”If respondents answered that they had thought about
suicide, they were then asked, “During the past 12 months,
did you try to kill yourself?”Attempt data were also avail-
able separately for males, females, and adults aged 18–29
years.
Our primary outcomes were the rate of suicides per state
during the 2-year study period, the rate of suicides involving
firearms, and the rate of suicides involving methods other
than firearms. Standard linear regression analyses were under-
taken in conjunction with a bootstrapping process, in which
10,000 artificial data sets were generated to account for vari-
ability in the point estimates (as specified by the accompany-
ing survey-based 95% confidence intervals). This was achieved
by performing individual linear regressions in each artificial
data set and taking the 2.5th and 97.5th percentiles of effect
estimates as the boundaries for 95% confidence intervals. The
data sets were created through sampling from Gaussian dis-
tributions with means and standard errors provided by the
original survey estimates. Both unweighted analyses and anal-
yses weighted by the states’populations are shown. Subgroup
analyses are reported by using population-weighted regres-
sion results. The rate of suicidal acts, referred to herein assuicide
attempts unless otherwise noted, is the sum of suicidal acts that
are fatal (i.e., suicides) plus the far larger number of suicidal
acts that are nonfatal (i.e., past-year suicide attempts from the
National Survey on Drug Use and Health). Secondary strati-
fied analyses were performed for adult men, adult women,
adults aged 18–29 years, and adults aged 30 years or older.
To illustrate our main findings more concretely, we com-
pared suicide deaths and suicide attempts during our study
period in states that are the most extreme in their firearm prev-
alence. The group of high-prevalence states and the group of
low-prevalence states are matched so that the numbers of
person-years in the 2 groupings are approximately equal (i.e.,
the 16 states with the highest firearm prevalence are com-
pared with the 6 states with the lowest firearm prevalence).
Firearms and Suicide 947
Am J Epidemiol. 2013;178(6):946–955
by guest on July 25, 2016http://aje.oxfordjournals.org/Downloaded from

Similar mortality rate ratios were obtained when comparing
the 10 states that are most extreme in firearm prevalence (not
shown).
Our graphs depict the simultaneous effects of our 2 inde-
pendent variables of interest, firearm ownership and suicide
attempt rates, on our outcomes of interest, mortality from
suicide overall (Figure 1A), from suicide involving firearms
(Figure 1B), and from suicide involving all other methods
(Figure 1C). The visual discrepancy between a model that exam-
ines the crude relationship between household firearm own-
ership and suicide mortality (the solid regression line) and a
model that examines the simultaneous influence of suicide
attempt and firearm ownership rates on suicide mortality (a line
visualized through the cross marks) is a representation of how
much the relationship between firearm ownership and suicide
mortality is affected by the covarying contribution of suicide
attempt rates. The state-level measures of suicide mortality,
household firearm ownership, and suicide attempt rates that were
used in the primary analyses are provided in the Appendix.
RESULTS
Higher rates of firearm ownership are strongly associated
with higher rates of overall suicide and firearm suicide, but
not with nonfirearm suicide (Table 1). Suicide attempt rates
are not significantly associated with suicide mortality rates in
Figure 1. The visual discrepancy between the solid regression line (a model that examines the crude relationship between household firearm
ownership and suicide mortality, A) by firearm suicides, B) by nonfirearm suicides, and C) overall) and a model that examines the simultaneous
influence of suicide attempt and firearm ownership rates on suicide mortality (a line that can be visualized through the cross marks) is a represen-
tation of how little the association between firearm ownership and suicide mortality is affected by the covarying contribution of suicide attempt
rates.
948 Miller et al.
Am J Epidemiol. 2013;178(6):946–955
by guest on July 25, 2016http://aje.oxfordjournals.org/Downloaded from

unadjusted models (correlation coefficient = −0.08, P= 0.60)
(data not shown) or in models that control for firearm owner-
ship ( partial correlationcoef ficient = 0.02, P= 0.89) (Table 1).
Suicide attempt rates are also not significantly associated with
rates of household gun ownership (data not shown). The prev-
alence of household firearm ownership, which ranges from
10% to 66% across the 50 states, explains 67% of the varia-
tion in firearm suicide, 42% of the variation in overall suicide,
and less than 2% of the variation in nonfirearm suicide. By
contrast, suicide attempt rates, which range from 0.1% to
1.5%, explain less than 1% of the variation in rates of overall
suicide, firearm suicide, and nonfirearm suicide. Indeed, suicide
attempt rates are not significantly related to suicide mortality
rates overall or by method, even in crude comparisons. Adjust-
ment for suicide attempt data in regressions, therefore, has
little influence on the magnitude of the associations between
rates of firearm ownership and suicide mortality. For example,
the partial correlation coefficient relating rates of household
firearm ownership and suicide mortality in our primary anal-
ysis is 0.6 whether or not suicide attempts are included in
theregression(datanotshown).Likewise,regardlessofwhether
suicide attempt rates are included in regressions, the partial
correlation coefficient relating household firearm ownership
and firearm suicide is 0.8 (data not shown).
Regression analyses further quantify these relationships.
Suicide rates are, on average, 0.22 deaths (per 100,000 pop-
ulation) higher in states where firearm ownership rates are 1
percentage point higher (31). The relationship between firearm
ownership and suicide rates is entirely accounted for by the
relationship between firearm ownership and firearm suicides,
as reflected in a βcoefficient associated with firearm suicide
of 0.22 (95% confidence interval: 0.18, 0.27), which is virtu-
ally identical to that for overall suicide (β= 0.21, 95% confi-
dence interval: 0.14, 0.28); the βcoefficient relating firearm
ownership and nonfirearm suicide is essentially null (β=–0.02,
95% confidence interval: –0.05, 0.02). By contrast, suicide
rates are only slightly higher (0.30 deaths per 100,000 popu-
lation) in states where rates of suicide attempts were 1 per-
centage point higher (Table 1). Because suicide attempt rates
vary from 0.1 to 1.5 per 100,000 population, the maximum
influence of suicide attempt rates on the suicide mortality rates
observed across the50 states is small (0.30 × 1.4 = 0.42 deaths
per 100,000 population on a suicide rate scale that ranges from
8.6 to 28.9 deaths per 100,000 population). By contrast, because
firearm ownership prevalence ranges from 10% to 66%, the
corresponding (maximum) influence of household firearm
ownership on suicide rates is 0.22 × 56 = 12.3 deaths per
100,000 population. Weighted and unweighted analyses pro-
duce virtually identical coefficients for household firearm
ownership in relationship to suicide mortality (Table 1). Like-
wise, the relationship between variation in suicide attempt rates
and suicide mortality in both weighted and unweighted anal-
yses is similar, nonsignificant, and materially trivial (Table 1).
Secondary stratified analyses by sex and by age groupings for
which suicide attempt data were available produce patterns
similar to those in primary analyses (Table 2).
Although the aggregate number of people residing in the
16 high–gun ownership states and the 6 low–gun ownership
states is approximately equal, and the suicide attempt rates
are similar, almost twice as many adults (11,428) completed
Table 1. Association Between State-Level Measures of Household Firearm Ownership Prevalence
a
(2004), Suicide Attempt Rates
b
(2008–2009), and Suicide Mortality Rates
c
(2008–2009)
in the United States by Method of Suicide
Method of
Suicide
Partial Correlation Coefficient
d
Regression Coefficient
d
Household
Firearm
Ownership
Suicide Attempt
Rate
Household Firearm
Ownership
Suicide Attempt
Rate
Household Firearm
Ownership
Suicide Attempt
Rate
Correlation
Coefficient
P
Value
Correlation
Coefficient
P
Value
Regression
Coefficient 95% CI Regression
Coefficient 95% CI
Regression
Coefficient
Weighted by
Population
95% CI
Regression
Coefficient
Weighted by
Population
95% CI
Firearm 0.82 0.001 −0.04 0.79 0.22 0.18, 0.27 −0.33 −2.37, 2.06 0.22 0.17, 0.26 −0.56 −2.66, 2.18
Nonfirearm −0.13 0.38 0.08 0.55 −0.02 −0.05, 0.02 0.63 −1.52, 2.14 −0.02 −0.05, 0.02 1.14 −1.37, 2.54
Overall 0.65 0.001 0.02 0.88 0.21 0.14, 0.28 0.30 −3.40, 3.70 0.20 0.13, 0.26 0.58 −3.49, 4.15
Abbreviation: CI, confidence interval.
a
Household firearm ownership prevalence ranges from 10% to 66% (standard deviation, 14%).
b
Suicide attempt rates range from 0.1% to 1.5% (standard deviation, 0.5%). Suicide attempt rates are not significantly associated with household firearm prevalence.
c
Suicide mortality rates vary as follows: among adults, 8.6%–28.9%; firearm suicides, 2.0%–20.2%; and nonfirearm suicides, 4.9%–12.8%.
d
Models are simultaneously adjusted for household firearm ownership and suicide attempt rates.
Firearms and Suicide 949
Am J Epidemiol. 2013;178(6):946–955
by guest on July 25, 2016http://aje.oxfordjournals.org/Downloaded from

Table 2. Associations Among State-Level Measures of Household Firearm Ownership Prevalence
a
(2004), Suicide Attempt Rates
b
(2008–
2009), and Suicide Mortality Rates
c
(2008–2009) in the United States by Method of Suicide and Sex and Age Group
Covariates by
Subgroup Analysis
Firearm Suicide Nonfirearm Suicide Overall Suicide
Regression
Coefficient
d
95% CI Regression
Coefficient
d
95% CI Regression
Coefficient
d
95% CI
Stratified analyses of
2008–2009 suicide rates
Males
Household firearm
ownership
0.38 0.31, 0.45 −0.04 −0.08, 0.01 0.35 0.25, 0.45
Suicide attempt rate per
100 population
−0.25 −2.92, 2.34 0.79 −1.32, 2.00 0.54 −3.60, 3.63
Females
Household firearm
ownership
0.06 0.04, 0.08 0.00 −0.03, 0.03 0.06 0.03, 0.10
Suicide attempt rate per
100 population
0.07 −0.65, 0.75 0.29 −0.97, 1.30 0.37 −1.31, 1.71
Aged 18–29 years
Household firearm
ownership
0.23 0.18, 0.29 0.06
e
0.01, 0.11
e
0.29 0.21, 0.38
Suicide attempt rate per
100 population
−0.47 −1.23, 0.87 0.21 −0.93, 1.08 −0.25 −1.84, 1.66
Aged ≥30 years
Household firearm
ownership
0.21 0.15, 0.27 −0.05 −0.09, −0.00 0.16 0.08, 0.25
Suicide attempt rate per
100 population
−0.95 −2.63, 2.07 0.21 −1.78, 1.88 −0.74 −3.86, 3.50
Weighted stratified analyses of
2008–2009 suicide rates
Males
Household firearm
ownership
0.38 0.30, 0.44 −0.04 −0.08, 0.00 0.33 0.23, 0.43
Suicide attempt rate per
100 population
−1.25 −3.68, 2.89 1.17 −1.40, 2.28 0.08 −4.45, 4.46
Females
Household firearm
ownership
0.06 0.05, 0.08 0.00 −0.03, 0.03 0.07 0.03, 0.10
Suicide attempt rate per
100 population
−0.01 −0.64, 0.67 0.22 −0.94, 1.24 0.21 −1.23, 1.60
Aged 18–29 years
Household firearm
ownership
0.20 0.17, 0.24 0.04 −0.00, 0.07 0.24 0.18, 0.30
Suicide attempt rate per
100 population
0.29 −0.76, 0.94 0.85 −0.52, 1.12 1.14 −1.08, 1.83
Aged ≥30 years
Household firearm
ownership
0.20 0.15, 0.26 −0.04 −0.08, −0.00 0.17 0.08, 0.26
Suicide attempt rate per
100 population
−1.59 −3.22, 2.27 −0.10 −2.00, 1.97 −1.68 −4.67, 3.66
Abbreviation: CI, confidence interval.
a
Household firearm ownership prevalence ranges from 10% to 66% (standard deviation, 14%).
b
Suicide attempt rates range from 0.1% to 1.5% (standard deviation, 0.5%). Suicide attempt rates are not significantly associated with household
firearm prevalence.
c
Suicide rates vary as follows: among adults, 8.6%–28.9%; firearm suicides, 2.0%–20.2%; and nonfirearm suicides, 4.9%–12.8%.
d
Models are simultaneously adjusted for household firearm ownership and suicide attempt rates.
e
After exclusion of the South Dakota outlier, the regression coefficient is 0.04 (95% CI: −0.01, 0.08).
950 Miller et al.
Am J Epidemiol. 2013;178(6):946–955
by guest on July 25, 2016http://aje.oxfordjournals.org/Downloaded from

suicide in the high–gun ownership states compared with the
low–gun ownership states (6,038) (Table 3). This difference in
total suicides over a 2-year period is almost entirely attributable
to differences in firearm suicides (7,275 vs. 1,697), with vir-
tually no difference in the number of nonfirearm suicides
(4,153 vs. 4,341).
Figure 1illustrates the strong association between rates of
household firearm ownership and mortality from overall sui-
cide (Figure 1A) and suicide involving firearms (Figure 1B)
and the weak association between rates of firearm ownership
and suicide involving methods other than firearms (Figure 1C).
In addition, Figure 1illustrates that suicide attempt rates have
little influence on the relationship between firearm ownership
rates and mortality from suicide overall and from suicide by
firearms and correlate weakly with suicide by methods other
than firearms. For example, adjustment for suicide attempt
rates hardly moves the cross-hatches off the regression line
linking firearm ownership and firearm suicide rates (or off the
regression line linking firearm ownership to overall suicide
rates), providing a visual representation of how little the observed
association between suicide mortality and firearm ownership
depends on confounding by suicide attempt rates. These
visual renderings directly mirror results from linear regression
analyses.
DISCUSSION
As in previous empirical work from individual-level (7–
16,21) and ecological studies (2–6), we found that higher
rates of firearm ownership are associated with higher rates of
overall suicide and firearm suicide, but not with nonfirearm
suicide. Our finding that the firearm-suicide association per-
sists unabated after controlling for suicide attempt rates is
consistent with previous ecological work that controlled for
aggregate-level measures associated with suicidality, including
rates of major depression, serious suicidal thoughts, serious
mental illness, alcohol and drug dependence and abuse, urban-
ization, poverty, and unemployment (3,6,32). Moreover,
our finding that suicide attempt rates do not covary with either
firearm ownership or with rates of suicide suggests that, although
states with higher rates of firearm ownership may differ from
states with lower rates of firearm ownership, these differences
do not appear to play an important role in determining the fre-
quency with which people engage in suicidal behavior.
In the United States, where firearms are the method used
in more than 50% of all suicides and where roughly 1 in 3
homes contains firearms, even small relative declines in the
use of firearms in suicide acts could result in large reductions
in the number of suicides, depending on what, if any, method
would be substituted for firearms. Consider, for example, the
fact that more than 90% of all suicidal acts with firearms are
fatal, but suicidal acts with firearms constitute only 5% of all
deliberate self-harm episodes. In contrast, fewer than 3% of
all suicidal acts with drugs or cutting are fatal but, as a group,
such acts constitute approximately 90% of all attempts (33,
34). If even 1 in 10 of the approximately 22,000 persons who
attempted suicide with firearms in 2010 (the 19,932 who
died and the approximately 2,000 who survived) substituted
drugs or cutting, there would have been approximately 1,900
fewer suicide deaths. The potential for substantial reduction
in suicide rates is apparent in our comparison of suicides in
high–versus low–gun ownership states, where suicide attempt
rates are similar, but the rate of suicide is twice as high in high–
gun ownership states (with differences in mortality attributable
Table 3. Suicides and Suicide Attempts in US States with the Highest and Lowest Gun Ownership Levels, 2008–2009
Population Group by
State Gun Ownership
Level
Person-Years
No. of
Firearm
Suicides
No. of
Nonfirearm
Suicides
Total
No. of
Suicides
Population
With Suicidal
Acts,
e
%
95% CI
High–gun ownership states
a,b
All adults 62,383,037 7,275 4,153 11,428 0.41 0.18, 0.63
Adult men 30,273,657 6,263 2,905 9,168 0.38 0.16, 0.60
Adult women 32,109,380 1,012 1,248 2,260 0.44 0.17, 0.71
Adults aged 18–29 years 13,829,694 1,303 960 2,263 1.04 0.40, 1.67
Adults aged ≥30 years 48,553,343 5,972 3,193 9,165 0.24 0.09, 0.38
Low–gun ownership states
c,d
All adults 62,447,876 1,697 4,341 6,038 0.49 0.00, 0.98
Adult men 29,810,942 1,572 3,207 4,779 0.38 –0.04, 0.79
Adult women 32,636,934 125 1,134 1,259 0.60 –0.01, 1.21
Adults aged 18–29 years 13,335,648 219 778 997 0.97 –0.01, 1.94
Adults aged ≥30 years 49,112,228 1,478 3,563 5,041 0.26 –0.06, 0.58
Abbreviation: CI, confidence interval.
a
High–gun ownership states are Alabama, Alaska, Arkansas, Idaho, Iowa, Kentucky, Louisiana, Mississippi, Montana, Nebraska, North Dakota,
Oklahoma,South Dakota, Tennessee, West Virginia, and Wyoming.
b
In high–gun ownership states, 51% of adults live in households with firearms.
c
Low–gun ownership states are Connecticut, Hawaii, Massachusetts, New Jersey, New York, and Rhode Island.
d
In low–gun ownership states, 15% of adults live in households with firearms.
e
The percent of the population that engaged in fatal and nonfatal suicidal acts over the past year.
Firearms and Suicide 951
Am J Epidemiol. 2013;178(6):946–955
by guest on July 25, 2016http://aje.oxfordjournals.org/Downloaded from

entirely to differences in suicide by firearms), with a net excess
of approximately 6,000 suicides in high–gun ownership states
over a 2-year period.
Our study should be considered in light of several potential
limitations. First, our measure of firearm availability is house-
hold firearm ownership. This is a reasonable measure of expo-
sure because most firearm suicides involve firearms from the
victims’homes (15), but this measure does not provide poten-
tially important information about many characteristics of fire-
arm availability that may affect risk, such as how firearms are
stored. Second, our measure of suicide attempt rates is based
on survey responses, andwe were unable to control for the seri-
ousness or persistence of the suicidal intent that accompa-
nied the reported events. On the other hand, there is no a priori
reason to expect that the suicidal intent among attempters in
high–gun ownership states would be different (higher) than
the intent among attempters in low–gun ownership states.
Consistent with this expectation, prior work has failed to find
higher rates of mental illness, substance abuse or dependence,
or suicidal thoughts or attempts among people living in homes
with firearms compared with those living in homes without
firearms (22–25). Third, firearm prevalence data in primary
analyses come from the 2004 BRFSS (the latest year for which
nationally representative state-level data are available), whereas
mortality and suicide attempt data come from 2008–2009. The
effect of this temporal discrepancy on our results is likely to
be small because guns are highly durable and, as has been
observed previously, the cross-sectional pattern of household
firearm ownership tends to be remarkably constant over time
(35). Analyses that use mortality data from 2000–2004 and fire-
arm ownership data from 2002 or 2004, for example, produce
findings identical to the second decimal place (and substan-
tively indistinguishable from those reported by using mortality
data from 2008–2009 and firearm ownership data from 2004).
Fourth, our study used aggregate data for ourexposures and out-
comes of interest; as is always the case with analyses based on
aggregate data, drawing causal inferences about individual risk
factors can be problematic (36). However, our key finding—a
strong association between household firearm prevalence and
suicide by firearms (but not by other methods)—is likely to
reflect associations observable at the individual level because
findings from individual-level studies (15) indicate that the
majority of suicides by firearms occur in the decedents’homes
and involve firearms owned by the victims or other members
of the households.
Despite these limitations, our results support the hypothe-
sis established in prior individual-level work (7–16,21) that
the availability of lethal means is associated with risk of death
by suicide above and beyond the baseline risk of suicidal behav-
ior. Our study suggests that this additional risk is large, operates
across sex and age groups, and plausibly accounts for thou-
sands of deaths every year.
ACKNOWLEDGMENTS
Author affiliations: Department of Health Policy and Man-
agement, Harvard School of Public Health, Boston, Massa-
chusetts (Matthew Miller, Deborah Azrael, Catherine Barber);
Harvard Injury Control Research Center, Harvard School of
Public Health, Boston, Massachusetts, (Matthew Miller,
Deborah Azrael, Catherine Barber); and the Division of Epide-
miology, Department of Genes and Environment, Norwegian
Institute of Public Health, Oslo, Norway (Richard A. White).
This work was funded by grants from the Joyce Founda-
tion.
The views expressed in this paper are those of the authors
and not necessarily those of any funding body or others whose
support is acknowledged. The funders had no role in study
design, data collection and analysis, decision to publish, or
preparation of the manuscript.
Conflict of interest: none declared.
REFERENCES
1. Centers for Disease Control and Prevention. Web-based injury
statistics query and reporting qystem (WISQARS). Atlanta,
Georgia: Centers for Disease Control and Prevention, National
Center for Injury Prevention; 2005. http://www.cdc.gov/ncipc/
wisqars. (Accessed December 10, 2012).
2. Miller M, Azrael D, Barber C. Suicide mortality in the United
States: the importance of attending to method in understanding
population-level disparities in the burden of suicide. Annu Rev
Public Health. 2012;33(1):393–408.
3. Miller M, Lippmann SJ, Azrael D, et al. Household firearm
ownership and rates of suicide across the 50 United States.
J Trauma. 2007;62(4):1029–1034.
4. Miller M, Azrael D, Hemenway D. Firearm availability and
suicide, homicide, and unintentional firearm deaths among
women. J Urban Health. 2002;79(1):26–38.
5. Miller M, Azrael D, Hemenway D. Household firearm
ownership and suicide rates in the United States.
Epidemiology. 2002;13(5):517–524.
6. Miller M, Hemenway D, Azrael D. Firearms and suicide in the
northeast. J Trauma. 2004;57(3):626–632.
7. Bailey JE, Kellermann AL, Somes GW, et al. Risk factors for
violent death of women in the home. Arch Intern Med.
1997;157(7):777–782.
8. Brent DA, Perper J, Moritz G, et al. Suicide in adolescents
with no apparent psychopathology. J Am Acad Child Adolesc
Psychiatry. 1993;32(3):494–500.
9. Brent DA, Perper JA, Allman CJ, et al. The presence and
accessibility of firearms in the homes of adolescent suicides.
A case-control study. JAMA. 1991;266(21):2989–2995.
10. Brent DA, Perper JA, Goldstein CE, et al. Risk factors for
adolescent suicide. A comparison of adolescent suicide
victims with suicidal inpatients. Arch Gen Psychiatry. 1988;
45(6):581–588.
11. Brent DA, Perper JA, Moritz G, et al. Firearms and adolescent
suicide. A community case-control study. Am J Dis Child.
1993;147(10):1066–1071.
12. Brent DA, Perper JA, Moritz G, et al. Suicide in affectively
ill adolescents: a case-control study. J Affect Disord. 1994;
31(3):193–202.
13. Conwell Y, Duberstein PR, Connor K, et al. Access to firearms
and risk for suicide in middle-aged and older adults. Am
J Geriatr Psychiatry. 2002;10(4):407–416.
14. Cummings P, Koepsell TD, Grossman DC, et al. The
association between the purchase of a handgun and
homicide or suicide. Am J Public Health. 1997;87(6):
974–978.
952 Miller et al.
Am J Epidemiol. 2013;178(6):946–955
by guest on July 25, 2016http://aje.oxfordjournals.org/Downloaded from

15. Kellermann AL, Rivara FP, Somes G, et al. Suicide in the
home in relation to gun ownership. N Engl J Med. 1992;
327(7):467–472.
16. Wiebe DJ. Homicide and suicide risks associated with firearms
in the home: a national case-control study. Ann Emerg Med.
2003;41(6):771–782.
17. Shah S, Hoffman RE, Wake L, et al. Adolescent suicide and
household access to firearms in Colorado: results of a
case-control study. J Adolesc Health. 2000;26(3):157–163.
18. Dahlberg LL, Ikeda RM, Kresnow MJ. Guns in the home and
risk of a violent death in the home: findings from a national
study. Am J Epidemiol. 2004;160(10):929–936.
19. Kung HC, Pearson JL, Liu X. Risk factors for male and female
suicide decedents ages 15–64 in the United States. Results
from the 1993 National Mortality Followback Survey. Soc
Psychiatry Psychiatr Epidemiol. 2003;38(8):419–426.
20. Kung HC, Pearson JL, Wei R. Substance use, firearm
availability, depressive symptoms, and mental health service
utilization among white and African American suicide decedents
aged 15 to 64 years. Ann Epidemiol. 2005;15(8):614–621.
21. Wintemute GJ, Parham CA, Beaumont JJ, et al. Mortality
among recent purchasers of handguns. N Engl J Med.
1999;341(21):1583–1589.
22. Ilgen MA, Zivin K, McCammon RJ, et al. Mental illness,
previous suicidality, and access to guns in the United States.
Psychiatr Serv. 2008;59(2):198–200.
23. Miller M, Barber C, Azrael D, et al. Recent psychopathology,
suicidal thoughts and suicide attempts in households with and
without firearms: findings from the National Comorbidity
Study Replication. Inj Prev. 2009;15(3):183–187.
24. Sorenson SB, Vittes KA. Mental health and firearms in
community-based surveys: implications for suicide prevention.
Eval Rev. 2008;32(3):239–256.
25. Betz ME, Barber C, Miller M. Suicidal behavior and firearm
access: results from the Second Injury Control and Risk
Survey. Suicide Life Threat Behav;41(4):384–391.
26. National Research Council. Firearms and Violence: A Critical
Review. Washington, DC: The National Academies Press; 2005.
27. Centers for Disease Control and Prevention. Behavioral risk
factor surveillance system, 2004 survey data. Atlanta, Georgia:
US Department of Health and Human Services; 2005.
http://www.cdc.gov/brfss. (Accessed December 12, 2012).
28. Centers for Disease Control and Prevention. Behavioral risk
factor surveillance system operational and user’s guide,
version 3.0. Atlanta, Georgia: US Department of Health and
Human Services; 2005. http://www.cdc.gov/brfss/pdf/
userguide.pdf. (Accessed November 11, 2012).
29. Mokdad AH, Stroup DF, Giles WH. Public health surveillance
for behavioral risk factors in a changing environment:
recommendations from the Behavioral Risk Factor
Surveillance Team. MMWR Recomm Rep. 2003;52(RR09):
1–12.
30. Denny CH. The Behavioral Risk Factor Surveillance
System. In: Blumenthal DS, DiClemente RJ, eds. Community-
Based Health Research: Issues and Methods. New York,
New York: Springer Publishing Company, Inc; 2004:
115–131.
31. Crosby AE, Han B, Ortega LAG, et al. Suicidal thoughts
and behaviors among adults aged ≥18 years–United States,
2008–2009. MMWR Surveill Summ. 2011;60(13):
1–22.
32. Hemenway D, Miller M. Association of rates of household
handgun ownership, lifetime major depression, and serious
suicidal thoughts with rates of suicide across US census
regions. Inj Prev. 2002;8(4):313–316.
33. Miller M, Azrael D, Hemenway D. The epidemiology of case
fatality rates for suicide in the Northeast. Ann Emerg Med.
2004;43(6):723–730.
34. Spicer RS, Miller TR. Suicide acts in 8 states: incidence and
case fatality rates by demographics and method. Am J Public
Health. 2000;90(12):1885–1891.
35. Azrael D, Cook PJ, Miller M. State and local prevalence of
firearms ownership: measurement, structure, and trends.
J Quant Criminol. 2004;20:43–62.
36. Piantadosi S. Invited commentary: ecologic biases. Am
J Epidemiol. 1994;139(8):761–764.
(Appendix follows)
Firearms and Suicide 953
Am J Epidemiol. 2013;178(6):946–955
by guest on July 25, 2016http://aje.oxfordjournals.org/Downloaded from

Appendix Table 1. Data on Suicide by Method in the United States (2008–2009), Suicide Attempts (2008–2009),
and Firearm Ownership (2004) by State
State
No. of
Suicides
per 100,000
Population
No. of Firearm
Suicides
per 100,000
Population
No. of Nonfirearm
Suicides
per 100,000
Population
Adults in
Households
With Firearms, %
Suicide
Attempts per
100 Population
Alabama 17.3 12.3 5.0 52 0.22
Alaska 28.8 18.4 10.4 60 0.23
Arizona 21.1 12.0 9.1 32 0.62
Arkansas 19.4 12.6 6.9 59 0.92
California 13.5 5.4 8.2 20 0.41
Colorado 22.6 11.1 11.6 35 0.42
Connecticut 11.3 3.7 7.6 18 1.01
Delaware 15.6 5.7 9.9 26 0.12
Florida 18.9 9.7 9.2 25 0.52
Georgia 14.5 9.4 5.1 40 0.11
Hawaii 14.5 3.1 11.4 10 0.21
Idaho 23.8 13.8 10.0 56 0.62
Illinois 11.9 4.2 7.7 21 0.51
Indiana 16.4 9.2 7.2 38 0.42
Iowa 15.7 7.2 8.5 46 0.32
Kansas 16.7 9.1 7.6 43 0.52
Kentucky 17.8 11.9 5.9 48 0.22
Louisiana 14.8 9.9 4.9 45 0.41
Maine 17.7 9.7 8.0 40 0.82
Maryland 11.8 5.8 6.0 22 0.31
Massachusetts 10.0 2.0 8.0 11 0.21
Michigan 15.1 7.4 7.7 41 0.81
Minnesota 14.3 7.0 7.3 41 0.81
Mississippi 17.5 12.3 5.1 55 0.42
Missouri 17.6 9.7 7.9 44 0.32
Montana 26.9 17.3 9.7 63 0.33
Nebraska 12.8 6.8 6.0 45 0.51
Nevada 25.4 14.5 11.0 34 0.53
New Hampshire 16.4 8.1 8.4 31 0.32
New Jersey 8.6 2.6 6.0 11 0.61
New Mexico 25.0 12.4 12.6 40 0.73
New York 9.2 2.8 6.4 19 0.41
North Carolina 16.0 9.3 6.7 39 0.52
North Dakota 16.5 10.0 6.4 56 0.32
Ohio 14.1 7.1 7.0 34 0.71
Oklahoma 19.9 12.4 7.5 46 0.42
Oregon 20.3 11.2 9.1 40 0.42
Pennsylvania 15.7 8.0 7.7 38 0.32
Rhode Island 13.3 3.7 9.6 12 1.51
South Carolina 16.6 10.2 6.4 43 0.42
South Dakota 18.9 9.9 9.1 60 0.42
Tennessee 19.6 12.2 7.4 47 0.32
Texas 14.6 8.5 6.1 37 0.61
Utah 21.9 10.9 11.0 45 0.52
Table continues
954 Miller et al.
Am J Epidemiol. 2013;178(6):946–955
by guest on July 25, 2016http://aje.oxfordjournals.org/Downloaded from

Appendix Table 1. Continued
State
No. of
Suicides
per 100,000
Population
No. of Firearm
Suicides
per 100,000
Population
No. of Nonfirearm
Suicides
per 100,000
Population
Adults in
Households
With Firearms, %
Suicide
Attempts per
100 Population
Vermont 17.9 10.0 7.8 44 0.52
Virginia 15.4 8.5 6.9 37 0.32
Washington 17.5 8.8 8.7 34 0.72
West Virginia 17.4 11.8 5.6 58 0.72
Wisconsin 16.5 7.8 8.7 43 0.52
Wyoming 26.9 19.7 7.3 66 0.53
Firearms and Suicide 955
Am J Epidemiol. 2013;178(6):946–955
by guest on July 25, 2016http://aje.oxfordjournals.org/Downloaded from