Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Aug 1.
Published in final edited form as: Inj Prev. 2015 Oct 23;22(4):302–307. doi: 10.1136/injuryprev-2015-041856

Alcohol, drug and other prior crimes and risk of arrest in handgun purchasers: protocol for a controlled observational study

Garen J Wintemute 1, Philip H Kass 2, Susan L Stewart 3, Magdalena Cerdá 1, Paul J Gruenewald 4
PMCID: PMC4842158  NIHMSID: NIHMS773375  PMID: 26498316

Abstract

Background and objective

Alcohol abuse is common in the USA and is a well-established risk factor for violence. Other drug use and criminal activity are risk factors as well and frequently occur together with alcohol abuse. Firearm ownership is also common; there are >50 million firearm owners in the USA. This study assesses the relationships between alcohol and drug abuse and future violence among firearm owners, which no prior research has done.

Design and study population

This records-based retrospective cohort study will involve all persons who legally purchased handguns in California in 2001—approximately 116 000 individuals—with follow-up through the end of 2013.

Methods

The principal exposures include prior convictions for alcohol-related and drug-related offenses. The primary outcome measure is an arrest following handgun purchase for a violent Crime Index offense: homicide, rape, robbery or aggravated assault. Subjects will be considered at risk for outcome events for only as long as their residence in California can be established independently of outcome events. Covariates include individual characteristics (eg, age, sex, criminal history, firearm purchase history) and community characteristics (eg, demographics, socioeconomic measures, firearm ownership and alcohol outlet density). We will employ survival analytic methods, expressing effects as HRs.

Discussion

The results of this large-scale study are likely to be generalisable and to have important implications for violence prevention policies and programmes.

BACKGROUND

Alcohol abuse is common in the USA. (We use ‘abuse’ in this article to refer to the gamut of alcohol-related or drug-related problems.) In the 2013 Behavioral Risk Factor Surveillance System survey, 16.8% of respondents reported binge drinking in the previous 30 days (≥5 drinks on an occasion for men, ≥4 drinks for women) and 6.2% reported heavy drinking (>60 drinks over 30 days for men, >30 drinks for women).1 Acute alcohol intoxication and a history of alcohol abuse have well-established associations with violence, including firearm violence. For men, alcohol-attributable mortality from firearm violence approximates that for MVCs.2

Firearm ownership is also common in the USA; >50 million adults—approximately 35% of men and 11% of women—are estimated to own firearms.3 Firearm violence “poses a serious threat to the safety and welfare of the American public”,4 with 11 208 firearm homicides and 21 175 firearm suicides in 2013.5 The combined mortality rate from firearm suicide and homicide has changed little in more than a decade.6

Firearm owners may be more likely than others to abuse alcohol.7 Yet no longitudinal research has determined whether alcohol abuse is associated with risk for violence among firearm owners. This study will determine whether associations between individual alcohol abuse and community alcohol outlet density exist in a large population of legal purchasers of handguns.

Our principal hypotheses are as follows:

  • Among authorised purchasers of handguns in California, prior arrests and convictions for alcohol-related offenses, drug-related offenses and other non-violent misdemeanour crimes will each be associated with an increase in risk for violent and firearm-related criminal activity following handgun purchase.

  • Risk will increase both as the number of prior arrests or convictions associated with each of these three types of offense increases and as the number of types of offense that are present increases.

  • An increase in risk will persist after controlling for other known or potential individual risk factors (eg, demographics, firearm purchase history) and community characteristics (eg, demographics, socioeconomic status, firearm ownership, alcohol outlet density).

  • Alcohol outlet density will be independently associated with risk for violent and firearm-related criminal activity.

We next briefly review some of the most salient studies providing the foundation and context for our proposed research. A more detailed review has been published elsewhere.2

Alcohol abuse and risk for violence

Acute intoxication

In Bureau of Justice Statistics (BJS) surveys, approximately 37% of inmates convicted of violent crimes reported being under the influence of alcohol when those crimes were committed.8,9 Acute alcohol use or intoxication among homicide offenders ranges from 33% to 60% in toxicology studies.10,11 Acute alcohol use is associated with a substantial, dose-dependent increase in risk for suicide,12–14 which appears to be greatest for firearm suicide.15,16

History of abuse

Large-scale survey studies have found an association between prior alcohol abuse and an increased risk of violence against others generally, and intimate partner violence, child abuse and child neglect.17–19 In a multicity case–control study, persons who committed suicide were more likely to have had alcohol-related trouble at work or hospitalisations for complications of alcohol use.20 Prior alcohol abuse is associated with increased risk for a first suicide attempt.21

Driving under the influence offenses

Driving under the influence (DUI) is the most common of the alcohol-related offenses we will study.22 One persistent finding of DUI research is the relationship between the number of DUI offenses and the prevalence of other risk behaviours, including drug use and other criminal activity.23–26 One such study used military police records to identify arrests. Repeat DUI offenders were more likely than first offenders (or non-offenders) to have been arrested for a violent crime, concealed weapon carrying or drug possession.27

Alcohol outlet density

Alcohol outlet density affects both the quantity and circumstances of individual alcohol consumption.28 Alcohol outlet distribution, particularly of on-premise establishments like bars, taverns and pubs, is an independent predictor of the rate and distribution of violence in communities.29–32 Alcohol outlet density is also associated with other conditions related to crime and violence, such as neighbourhood social disorganisation and residential instability, and its effects must be assessed with these conditions in mind.31

Other key exposures

Drug abuse

In a 2004 BJS survey, >25% of inmates incarcerated for violent crimes reported committing those crimes while under the influence of drugs; >50% reported use in the preceding month.33 Homicide offenders have rates of acute drug use or intoxication of 33–38% in toxicology studies.10 In National Violent Death Reporting System suicide data for 2010, opiates, marijuana and cocaine were identified in 20.0%, 10.2% and 4.6%, respectively, of those tested.12

Large-scale surveys have found that, as with alcohol, pre-existing drug abuse is associated with an increased risk of violence against others, intimate partner violence, child abuse and child neglect.17,19 In surveys and a systematic review, drug abuse has been associated with increases in risk for subsequent suicide attempts and self-directed non-fatal violence.13,17,21

Criminal history

Risk for new criminal activity, including violence, is greatly increased among persons with a criminal history.4,34–42 A dose–response relationship exists; the more extensive the prior criminal record, the greater the risk of subsequent offending.43 Risk for violence is highly concentrated, with 6–10% of offenders committing >50% of violent offenses.37 After an index offense, risk for reoffending is highest initially and declines exponentially thereafter, approaching that for non-offenders after 10–20 years.44–46

Other factors

Risk for violence is strongly related to male sex and (for interpersonal violence) younger age.4,34–37,47 Community characteristics such as concentrated poverty and residential segregation are associated with risk for violence and are believed to operate by multiple mechanisms.36,48–51

Co-occurrence of key exposures

Our principal exposures will likely co-occur but have been demonstrated to be statistically separable in previous research; each may act directly and modify risk associated with the others. Mental illness research is particularly informative on this point. In a nationally representative survey, the probability of incident violent behaviour was low for persons reporting neither substance abuse nor violence in the past, increased for those reporting only substance abuse or violence, and increased still further for those reporting both.52 In a case–control study of suicide, alcohol abuse was associated with a far greater increase in suicide risk among persons with a history of violence than among persons with no such history.53

Findings specific to firearm owners

Alcohol abuse

No study has examined associations between alcohol abuse and future violence among firearm owners. While alcohol abuse has been linked to homicide and suicide risk when firearms are in the home, such studies are not restricted to firearm owners and do not assess risk for committing interpersonal violence.20,53–56

Firearm ownership itself may be associated with alcohol abuse, including binge drinking and heavy drinking.7,57–63 Moreover, firearm owners who abuse alcohol are more likely than others to engage in firearm-related risk behaviours, including behaviours likely to be associated with violence.7,62,63

Demographics and criminal history

Longitudinal studies of violent and non-violent criminal activity among handgun purchasers in California have shown large increases in risk for men, a direct relationship between risk and the extent of any criminal history, and an inverse relationship between risk and age.64–67

Firearm ownership

Firearm ownership itself has been associated with an increased risk of arrest57,68 or ‘trouble with the police.’60 Carrying a firearm in public has also been linked to arrest for a non-traffic offense62,68 and aggressive or hostile driving behaviour.69,70

METHODS

Overview

This is a records-based retrospective cohort study. The study population will comprise all persons who legally purchased handguns in California in 2001—approximately 116 000 individuals. The measurement period for capturing outcome events will begin with each subject’s first handgun purchase in 2001 and end on 31 December 2013. We will extract data on arrests and convictions both prior and subsequent to the inception date from subjects’ criminal histories, and data on other alcohol-involved events from driver’s licence records. The primary outcome event will be an arrest in California for a violent Crime Index crime: homicide, rape, robbery and aggravated assault. Subjects will be considered at risk for outcome events for only as long as their residence in the state can be verified independently of outcome events.

Background: buying a handgun in California

California requires almost all transfers of firearms—sales, including those involving a private-party seller; pawn redemptions; and gifts—to be processed by a licensed firearm retailer. Transfers at gun shows or involving internet brokers are included. Purchases cannot be completed without a driver’s licence or military identification card, which must be ‘swiped’ through a reader to enter personal information. California also requires centralised, permanent records for firearm transfers. A separate record, referred to as a Dealer’s Record of Sale (DROS), is generated for each firearm transferred. Records for handguns have been required since the early 1950s; rifle and shotgun records have been required since January 2014.

A required 10-day waiting period allows for a background check to determine whether the purchaser is a prohibited person. California prohibits purchases by individuals who are not prohibited under federal law: persons convicted of violent misdemeanours (eg, assault and battery), persons who pose an imminent danger to themselves or others and have been hospitalised for emergency psychiatric evaluation, and persons subject to temporary domestic violence restraining orders.

Data sources

DROS file

DROS records contain identifiers for the purchaser, the retailer and the firearm, and the date and time of the transaction. First-time purchasers receive a unique DROS-ID number, and their subsequent purchases are associated with that number. If a purchaser has a record in the state’s Criminal History System (CHS), his unique Criminal Identification and Information (CII) number is added to his DROS record. If a prospective purchaser is found to be a prohibited person, the purchase is denied and a permanent record is kept. DROS records from 2001 will identify our study population. We will use DROS records from 1991 to 2013 and the denied file to construct each subject’s handgun purchase history.

Criminal History System

CHS contains >10 million records; nearly all are compu-terised.71,72 Records include extensive personal identifiers, the CII number and numbers assigned by agencies such as the FBI. Each criminal justice action, such as a charged offense or conviction, generates a separate entry.

Offenses are recorded very specifically. Aggravated assaults involving firearms, for example, are charged under weapon-specific sections of the Penal Code. There are specific Penal Code sections for intimate partner and sexual violence. Drug offenses involving opioids and cocaine are charged separately from those involving methamphetamine.

Recording of arrests in CHS is rapid and complete.71 Conviction recording is less complete, particularly for recent offenses. This will have little impact on our baseline data, as 2001 is our inception year. It is one of several reasons for our use of arrests as outcome events. There are minor missing data conditions. For example, records for specified offenses relating to possession of small quantities of marijuana must be deleted after 24 months.

Department of Motor Vehicles files

The Driving Record File contains licensee identifiers and entries for ‘department actions’ such as DUI arrests and convictions and judgements by law enforcement officers that collision-involved drivers had consumed alcohol. These field judgements have a sensitivity of ≥75% and high specificity, and they are more common than DUI arrests.73–76 We will use department action data to improve ascertainment of DUI arrests and convictions and in a supplemental analysis will assess the effect on our findings of incorporating field judgements. We will use driver’s licence renewal and vehicle registration records to establish residence in California.

Formation of the study population

The study population will comprise all individuals with handgun purchase records in the 2001 DROS file. There are 160 619 records involving an estimated 116 675 distinct purchasers. The data suggest that 2001 was a typical year, with a sales volume and age and sex distributions similar to those for 1996–2005. For 2001, 98.8% of records include a California driver’s licence number; the remainder contain a military ID.

Data acquisition and management

Data will be provided as computer files. We will submit criminal records requests for all study subjects to CalDOJ, which will use its standard protocols for linking subjects to records. Tentative matches will be reviewed manually by us using all available data. We will categorise crimes using the FBI’s Uniform Crime Reports system77 and a separate classification scheme for offenses of particular interest: involving firearms, intimate partner or domestic violence, and so on.

We will request records for all subjects from the Department of Motor Vehicle (DMV). In the past, we have achieved a nearly 100% hit rate for subjects with driver’s licence numbers and a >50% hit rate for others.

We will geocode subjects’ addresses and include data from the 2000 Census on characteristics of subjects’ communities at baseline. Alcohol outlet locations have been obtained and geocoded; small-area densities can be computed. Local arrest rates are available from CalDOJ.

There is no generally available direct measure of the prevalence of firearm ownership. The best-validated and most widely accepted proxy is the proportion of all suicides that involve fire-arms (FS/S).78 These data are available at the county level. For most cities, however, FS/S is too sensitive to small-number variation. Firearm injury data are available from the California Office of Statewide Health Planning and Development,79 which provides public access to these data geocoded to the ZIP code level and will provide tabulations to Census block groups. We will assess these data as an alternative proxy measure for firearm ownership.

Follow-up

Only events occurring in California will be captured as outcome events, as criminal records for other states are not available for this purpose. We will, therefore, consider subjects to be at risk for outcome events for only as long as their status as alive and residing in California can be verified independently of outcome events.

We will capture outcome events beginning 10 days following the DROS record date using the first record in 2001 for subjects having more than one. This is the first day on which the handgun could have been acquired, given California’s waiting period. (Acquisition dates are not available.) The measurement period will end 31 December 2013.

To track subjects through our follow-up period, we will begin with linkage to DMV records and voter registration data.80–83 We will search the California Death Statistical Master File, which includes all deaths in California and deaths elsewhere in the USA of California residents. For subjects with incomplete follow-up, we will request credit agency reports and query the National Death Index. Records linkage procedures will vary with the data involved.64–67,84 Our follow-up period will extend beyond the measurement period in order to increase the percentage of subjects with follow-up for the entire measurement period.

Subjects with no follow-up will be excluded from longitudinal analyses. Data on outcome events will be collected for these subjects to allow for an estimate of the bias introduced by lack of follow-up.

Power analysis

The study is powered to detect differences in risk for the outcome measure of greatest interest, even though that outcome is uncommon, involving the prespecified subgroups of greatest interest, even when those subgroups are small. In our prior studies, when handgun purchasers with no criminal history served as the referent group, RRs for all outcomes among purchasers with a criminal history were ≥5.0.64,66 In this study, we will have >90% power to detect at the 0.05 level (two-sided) a HR <1.5 comparing handgun purchasers with a criminal history to those with no criminal history with respect to time to first arrest for a violent Crime Index offense. We will have >80% power to detect HRs <2.0 for most subgroups of purchasers, including those with any alcohol-related convictions and with only one conviction, and <2.5 for purchasers with multiple alcohol-related convictions versus those with no criminal history.

Analysis plan

The principal outcome measure will be an arrest for a violent Crime Index crime. Secondary outcomes will include arrests for violent crimes involving firearms; for intimate partner or domestic violence; for sexual assault; for violent crimes generally; and for crimes of the following types, considered separately: involving firearms, alcohol and drugs.

The principal independent variables relate to the existence, nature and extent of subjects’ criminal histories at the time of handgun purchase. The following baseline characteristics will be included as covariates in multivariable analyses:

  • demographics: age, race/ethnicity, and sex;

  • handgun purchase history from 1991 to the index purchase;

  • desistance: elapsed time since the most recent arrest,44–46 for all offenses and for offenses categorised as above.

Community characteristics: violent and non-violent crime rates, urban/rural status,85 location by region in the state,86 population size, demographic and socioeconomic variables, prevalence of firearm ownership and alcohol outlet density.29 We are capturing changes of address and will treat community characteristics as time-varying covariates in time-to-event models.

We will use standard descriptive methods to compare subjects with and without prior criminal histories on variables in common. Generalised estimating equations87 will be used to account for clustering within census tracts or ZIP codes in comparisons of community characteristics at baseline.88

The primary analyses will determine the association between the primary outcome measure and the independent variables pertaining to each of four key exposures: alcohol offenses, drug offenses, other misdemeanour crimes and alcohol outlet density.

We will estimate the probability of experiencing a first new arrest over time by the Kaplan–Meier method.89 The significance of differences in probability of arrest between study subgroups (eg, subjects with and without a criminal history) will be assessed by the log-rank test. Cox proportional hazards regression will be used to calculate univariate and adjusted HRs and 95% CIs.90 Models that include community characteristics will use a robust sandwich estimator91 to account for clustering. Influence diagnostics will be performed, and reliance on the proportional hazards assumption will be validated.92

For each independent variable, we will develop multivariable models by adding covariates in blocks: (1) individual demographic characteristics, (2) community demographic and socioeconomic characteristics, (3) crime rates and (4) firearm ownership prevalence. Parsimonious models will be developed. We will determine whether alcohol outlet density modifies the effects of each of the other independent variables. Models will be constructed that include both drug and alcohol offenses and their interaction, to assess their combined effect. Additional analyses will focus on predefined subgroups of the study population.

Secondary analyses will include crime rates, alcohol outlet density and handgun purchases during the period of observation as time-dependent covariates. We will include arrests (by offense type) as time-dependent covariates to develop models addressing desistance. We will develop conditional, total time-recurrent events models to study effects as additional arrests occur and as duration of follow-up increases.92–94

Limitations

DROS records do not constitute a registry—only acquisitions are recorded—so we cannot determine whether subjects dispose of firearms and perhaps become non-owners. DROS did not capture long-gun ownership during our study period, and our findings may not apply to individuals who own long guns but not handguns.

Records capture only events known to the relevant agencies. Failed detection and lack of reporting result in missed events. Nonetheless, it is clearly of value to learn whether recorded events are associated with differences in risk, particularly because records such as we use here might constitute the basis for future interventions based on our findings.

The records are for California only, which may limit pre-purchase histories for recent immigrants. One effect of this will be occasional misclassification of subjects as having no criminal history. This will likely be a source of bias towards the null.

As discussed, persons convicted of violent misdemeanours are prohibited from purchasing firearms in California (for 10 years following those convictions). Given the effects of co-occurring alcohol abuse and violence, this may reduce the size of effects associated with alcohol-related arrests and convictions in our data—again, a conservative bias.

DISCUSSION

This study will assess the complex relationships between alcohol abuse (as assessed by arrests and convictions for alcohol-related offenses), other important and co-occurring potential risk factors, and future violence among firearm owners, a large and unstudied population that is of particular interest for studies of firearm-related violence. The findings are likely to satisfy many of the Bradford Hill criteria for inference that associations in observational studies are causal.95 The association between alcohol abuse (and other key exposures) and future violence is likely to be strong. Such an association would be consistent with a large body of prior research. The temporality criterion is addressed by the longitudinal study design. A dose–response relationship is likely, and the association would certainly be plausible and coherent with the known natural history of alcohol abuse.

The study’s findings will likely have clear implications for interventions to prevent alcohol-related violence and should be generalisable as no attributes unique to the study population or setting are being examined.

The richness of the data allows us to address important subsidiary questions. One is how characteristics identified as risk factors for violence in this study population interact when multiple factors are present. Another is whether risk for future violence associated with alcohol abuse is modified by local alcohol outlet density. Still another is whether desistance matters: is risk for continuing offenders increased in comparison with prior offenders who have ceased offending?

As much of the violence attributable to mental illness results from co-occurring alcohol or drug abuse, the findings of the study are likely to have implications for preventing violence among the mentally ill and efforts to prevent suicide and self-harm.

Acknowledgments

The authors are grateful to Mona Wright, Pam Keach and Vanessa McHenry of the Violence Prevention Research Program for technical assistance.

Funding This study is funded by grant number 1R01AA023551-01A1 from the National Institute on Alcohol Abuse and Alcoholism.

Footnotes

Contributors

All authors contributed to the design of the study and provided critical editorial input. GJW drafted the manuscript.

Competing interests

None declared.

Ethics approval

This study has been approved by the Institutional Review Board of the University of California, Davis (654883-1).

Provenance and peer review

Not commissioned; internally peer reviewed.

References

  • 1.Centers for Disease Control and Prevention. Behavioral risk factor surveillance system: prevalence data and data analysis tools. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; [accessed 21 Aug 2015]. http://www.cdc.gov/brfss/data_tools.htm. [Google Scholar]
  • 2.Wintemute GJ. Alcohol misuse, firearm violence perpetration, and public policy in the United States. Prev Med. 2015;79:15–21. doi: 10.1016/j.ypmed.2015.04.015. [DOI] [PubMed] [Google Scholar]
  • 3.General Social Survey. Data analysis using SDA (1972–2010) National Opinion Research Center, University of Chicago; [accessed 13 Apr 2011]. http://www3.norc.org/GSS+website/ [Google Scholar]
  • 4.Institute of Medicine, National Research Council. Priorities for research to reduce the threat of firearm-related violence. Washington DC: The National Academies Press; 2013. [Google Scholar]
  • 5.Centers for Disease Control and Prevention. Web-based Injury Statistics Query and Reporting System (WISQARS) National Center for Injury Prevention and Control, Centers for Disease Control and Prevention; [accessed 14 Mar 2015]. http://www.cdc.gov/injury/wisqars/index.html. [Google Scholar]
  • 6.Wintemute GJ. The epidemiology of firearm violence in the twenty-first century United States. Annu Rev Public Health. 2015;36:5–19. doi: 10.1146/annurev-publhealth-031914-122535. [DOI] [PubMed] [Google Scholar]
  • 7.Wintemute GJ. Association between firearm ownership, firearm-related risk and risk reduction behaviours and alcohol-related risk behaviours. Inj Prev. 2011;17:422–7. doi: 10.1136/ip.2010.031443. [DOI] [PubMed] [Google Scholar]
  • 8.Beck A, Gilliard DK, Greenfeld L, et al. Survey of State Prison Inmates 1991. Washington, DC: Bureau of Justice Statistics; 1993. NCJ 136949. [Google Scholar]
  • 9.Rand MR, Sabol WJ, Sinclair M, et al. Alcohol and crime: data from 2002 to 2008. Washington DC: Bureau of Justice Statistics; 2010. [Google Scholar]
  • 10.Darke S. The toxicology of homicide offenders and victims: a review. Drug Alcohol Rev. 2010;29:202–15. doi: 10.1111/j.1465-3362.2009.00099.x. [DOI] [PubMed] [Google Scholar]
  • 11.Kuhns JB, Exum ML, Clodfelter TA, et al. The prevalence of alcohol-involved homicide offending: a meta-analytic review. Homicide Stud. 2014;18:251–70. [Google Scholar]
  • 12.Parks SE, Johnson LL, McDaniel DD, et al. Centers for Disease Control and Prevention (CDC) Surveillance for violent deaths—National Violent Death Reporting System, 16 states, 2010. MMWR Surveill Summ. 2014;63:1–33. [PubMed] [Google Scholar]
  • 13.Borges G, Loera CR. Alcohol and drug use in suicidal behaviour. Curr Opin Psychiatry. 2010;23:195–204. doi: 10.1097/YCO.0b013e3283386322. [DOI] [PubMed] [Google Scholar]
  • 14.Bagge CL, Lee H-J, Schumacher JA, et al. Alcohol as an acute risk factor for recent suicide attempts: a case-crossover analysis. J Stud Alcohol Drugs. 2013;74:552–8. doi: 10.15288/jsad.2013.74.552. [DOI] [PubMed] [Google Scholar]
  • 15.Kaplan MS, McFarland BH, Huguet N, et al. Acute alcohol intoxication and suicide: a gender-stratified analysis of the National Violent Death Reporting System. Inj Prev. 2013;19:38–43. doi: 10.1136/injuryprev-2012-040317. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Branas CC, Richmond TS, Ten Have TR, et al. Acute alcohol consumption, alcohol outlets and gun suicide. Subst Use Misuse. 2011;46:1592–603. doi: 10.3109/10826084.2011.604371. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Harford TC, Yi HY, Grant BF. Other- and self-directed forms of violence and their relationships to DSM-IV substance use and other psychiatric disorders in a national survey of adults. Compr Psychiatry. 2013;54:731–9. doi: 10.1016/j.comppsych.2013.02.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Afifi TO, Henriksen CA, Asmundson GJG, et al. Victimization and perpetration of intimate partner violence and substance use disorders in a nationally representative sample. J Nerv Ment Dis. 2012;200:684–91. doi: 10.1097/NMD.0b013e3182613f64. [DOI] [PubMed] [Google Scholar]
  • 19.Kelleher K, Chaffin M, Hollenberg J, et al. Alcohol and drug disorders among physically abusive and neglectful parents in a community-based sample. Am J Public Health. 1994;84:1586–90. doi: 10.2105/ajph.84.10.1586. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Rivara FP, Mueller BA, Somes G, et al. Alcohol and illicit drug abuse and the risk of violent death in the home. JAMA. 1997;278:569–75. [PubMed] [Google Scholar]
  • 21.Borges G, Walters EE, Kessler RC. Associations of substance use, abuse, and dependence with subsequent suicidal behavior. Am J Epidemiol. 2000;151:781–9. doi: 10.1093/oxfordjournals.aje.a010278. [DOI] [PubMed] [Google Scholar]
  • 22.Criminal Justice Statistics Center. CJSC arrests data file 2003–2012. Sacramento, CA: Office of the Attorney General, California Department of Justice; [accessed 21 Dec 2012]. http://oag.ca.gov/crime/cjsc/stats/arrests. [Google Scholar]
  • 23.McCutcheon VV, Heath AC, Edenberg HJ, et al. Alcohol criteria endorsement and psychiatric and drug use disorders among DUI offenders: greater severity among women and multiple offenders. Addict Behav. 2009;34:432–9. doi: 10.1016/j.addbeh.2008.12.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Freeman J, Maxwell JC, Davey J. Unraveling the complexity of driving while intoxicated: a study into the prevalence of psychiatric and substance abuse comorbidity. Accid Anal Prev. 2011;43:34–9. doi: 10.1016/j.aap.2010.06.004. [DOI] [PubMed] [Google Scholar]
  • 25.McMillen DL, Adams MS, Wells-Parker E, et al. Personality traits and behaviors of alcohol-impaired drivers: a comparison of first and multiple offenders. Addict Behav. 1992;17:407–14. doi: 10.1016/0306-4603(92)90001-c. [DOI] [PubMed] [Google Scholar]
  • 26.Dugosh KL, Festinger DS, Marlowe DB. Moving beyond BAC in DUI. Criminol Public Pol. 2013;12:181–93. [Google Scholar]
  • 27.Lucker GW, Kruzich DJ, Holt MT, et al. The prevalence of antisocial behavior among U.S Army DWI offenders. J Stud Alcohol. 1991;52:318–20. doi: 10.15288/jsa.1991.52.318. [DOI] [PubMed] [Google Scholar]
  • 28.Gruenewald PJ, Remer LG, LaScala EA. Testing a social ecological model of alcohol use: the California 50-city study. Addiction. 2014;109:736–45. doi: 10.1111/add.12438. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Mair C, Gruenewald PJ, Ponicki WR, et al. Varying impacts of alcohol outlet densities on violent assaults: explaining differences across neighborhoods. J Stud Alcohol Drugs. 2013;74:50–8. doi: 10.15288/jsad.2013.74.50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Campbell CA, Hahn RA, Elder R, et al. Task Force on Community Preventive Services. The effectiveness of limiting alcohol outlet density as a means of reducing excessive alcohol consumption and alcohol-related harms. Am J Prev Med. 2009;37:556–69. doi: 10.1016/j.amepre.2009.09.028. [DOI] [PubMed] [Google Scholar]
  • 31.Gruenewald PJ, Freisthler B, Remer L, et al. Ecological models of alcohol outlets and violent assaults: crime potentials and geospatial analysis. Addiction. 2006;101:666–77. doi: 10.1111/j.1360-0443.2006.01405.x. [DOI] [PubMed] [Google Scholar]
  • 32.Treno AJ, Gruenewald PJ, Remer L, et al. Examining multi-level relationships between bars, hostility and aggression: social selection and social influence. Addiction. 2008;103:66–77. doi: 10.1111/j.1360-0443.2007.02039.x. [DOI] [PubMed] [Google Scholar]
  • 33.Mumola CJ, Karberg JC. Drug use and dependence, state and federal prisoners, 2004. Washington, DC: Bureau of Justice Statistics; 2006. NCJ 213530. [Google Scholar]
  • 34.National Research Council. Understanding and preventing violence. Vol. 1. Washington DC: The National Academies Press; 1993. [Google Scholar]
  • 35.National Research Council. Understanding and preventing violence, volume 2: biobehavioral influences. Washington DC: The National Academies Press; 1994. [Google Scholar]
  • 36.National Research Council. Understanding and preventing violence, volume 3: social influences. Washington DC: The National Academies Press; 1994. [Google Scholar]
  • 37.National Research Council. Firearms and violence: a critical review. Washington, DC: The National Academies Press; 2004. [Google Scholar]
  • 38.Beck A, Shipley B. Recidivism of prisoners released in 1983. Washington DC: Bureau of Justice Statistics; 1989. NCJ 116261. [Google Scholar]
  • 39.Maltz M. Recidivism. Orlando, FL: Academic Press, Inc; 1984. [Google Scholar]
  • 40.Tillman R. The size of the “criminal population:” the prevalence and incidence of adult arrest. Criminol. 1987;25:561–79. [Google Scholar]
  • 41.Cook PJ, Ludwig J, Braga AA. Criminal records of homicide offenders. JAMA. 2005;294:598–601. doi: 10.1001/jama.294.5.598. [DOI] [PubMed] [Google Scholar]
  • 42.Vittes KA, Vernick JS, Webster DW. Legal status and source of offenders’ firearms in states with the least stringent criteria for gun ownership. Inj Prev. 2012;19:26–31. doi: 10.1136/injuryprev-2011-040290. [DOI] [PubMed] [Google Scholar]
  • 43.Langan PA, Levin DJ. Recidivism of prisoners released in 1994. Washington DC: U. S. Department of Justice, Bureau of Justice Statistics; 2002. NCJ 193427. [Google Scholar]
  • 44.Blumstein A, Nakamura K. Redemption in the presence of widespread criminal background checks. Criminol. 2009;47:327–59. [Google Scholar]
  • 45.Kurlychek MC, Brame R, Bushway SD. Scarlet letters and recidivism: does an old criminal record predict future offending? Criminol Public Pol. 2006;5:483–504. [Google Scholar]
  • 46.Kurlychek MC, Brame R, Bushway SD. Enduring risk? Old criminal records and predictions of future criminal involvement. Crime Delinq. 2007;53:64–83. [Google Scholar]
  • 47.Piquero AR, Jennings WG, Diamond B, et al. A systematic review of age, sex, ethnicity, and race as predictors of violent recidivism. Int J Offender Ther Comp Criminol. 2015;59:5–26. doi: 10.1177/0306624X13514733. [DOI] [PubMed] [Google Scholar]
  • 48.Sampson R, Groves W. Community structure and crime—testing social-disorganization theory. Am J Sociol. 1989;94:774–802. [Google Scholar]
  • 49.Cerdá M, Tracy M, Ahern J, et al. Addressing population health and health inequalities: the role of fundamental causes. Am J Public Health. 2014;104(Suppl 4):S609–19. doi: 10.2105/AJPH.2014.302055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Cerdá M, Messner SF, Tracy M, et al. Investigating the effect of social changes on age-specific gun-related homicide rates in New York City during the 1990s. Am J Public Health. 2010;100:1107–15. doi: 10.2105/AJPH.2008.158238. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Ahern J, Cerdá M, Lippman SA, et al. Navigating non-positivity in neighbourhood studies: an analysis of collective efficacy and violence. J Epidemiol Community Health. 2013;67:159–65. doi: 10.1136/jech-2012-201317. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Elbogen EB, Johnson SC. The intricate link between violence and mental disorder. Arch Gen Psychiatry. 2009;66:152–61. doi: 10.1001/archgenpsychiatry.2008.537. [DOI] [PubMed] [Google Scholar]
  • 53.Conner KR, Cox C, Duberstein PR, et al. Violence, alcohol, and completed suicide: a case-control study. Am J Psychiatry. 2001;158:1701–5. doi: 10.1176/appi.ajp.158.10.1701. [DOI] [PubMed] [Google Scholar]
  • 54.Kellermann AL, Rivara FP, Somes G, et al. Suicide in the home in relation to gun ownership. N Engl J Med. 1992;327:467–72. doi: 10.1056/NEJM199208133270705. [DOI] [PubMed] [Google Scholar]
  • 55.Kellermann AL, Rivara FP, Rushforth NB, et al. Gun ownership as a risk factor for homicide in the home. N Engl J Med. 1993;329:1084–91. doi: 10.1056/NEJM199310073291506. [DOI] [PubMed] [Google Scholar]
  • 56.Karch DL, Logan J, McDaniel D, et al. Surveillance for violent deaths—National Violent Death Reporting System, 16 states, 2009. MMWR. 2012;61(SS-6):1–43. [PubMed] [Google Scholar]
  • 57.Diener E, Kerber KW. Personality characteristics of American gun-owners. J Soc Psychol. 1979;107:227–38. doi: 10.1080/00224545.1979.9922703. [DOI] [PubMed] [Google Scholar]
  • 58.Schwaner SL, Furr LA, Negrey CL, et al. Who wants a gun license? J Crim Justice. 1999;27:1–10. [Google Scholar]
  • 59.Miller M, Hemenway D, Wechsler H. Guns at college. J Am Coll Health. 1999;48:7–12. doi: 10.1080/07448489909595666. [DOI] [PubMed] [Google Scholar]
  • 60.Miller M, Hemenway D, Wechsler H. Guns and gun threats at college. J Am Coll Health. 2002;51:57–65. doi: 10.1080/07448480209596331. [DOI] [PubMed] [Google Scholar]
  • 61.Nelson DE, Grant-Worley JA, Powell K, et al. Population estimates of household firearm storage practices and firearm carrying in Oregon. JAMA. 1996;275:1744–8. [PubMed] [Google Scholar]
  • 62.Smith TW. 2001 National Gun Policy Survey of the National Opinion Research Center: Research Findings. Chicago, IL: National Opinion Research Center, University of Chicago; 2001. [Google Scholar]
  • 63.Casiano H, Belik S-L, Cox BJ, et al. Mental disorder and threats made by noninstitutionalized people with weapons in the National Comorbidity Survey replication. J Nerv Ment Dis. 2008;196:437–45. doi: 10.1097/NMD.0b013e3181775a2a. [DOI] [PubMed] [Google Scholar]
  • 64.Wintemute GJ, Drake CM, Beaumont JJ, et al. Prior misdemeanor convictions as a risk factor for later violent and firearm-related criminal activity among authorized purchasers of handguns. JAMA. 1998;280:2083–7. doi: 10.1001/jama.280.24.2083. [DOI] [PubMed] [Google Scholar]
  • 65.Wintemute GJ, Wright MA, Drake CM, et al. Subsequent criminal activity among violent misdemeanants who seek to purchase handguns: risk factors and effectiveness of denying handgun purchase. JAMA. 2001;285:1019–26. doi: 10.1001/jama.285.8.1019. [DOI] [PubMed] [Google Scholar]
  • 66.Wright MA, Wintemute GJ. Felonious or violent criminal activity that prohibits gun ownership among prior purchasers of handguns: incidence and risk factors. J Trauma. 2010;69:948–55. doi: 10.1097/TA.0b013e3181cb441b. [DOI] [PubMed] [Google Scholar]
  • 67.Wright MA, Wintemute GJ, Rivara FA. Effectiveness of denial of handgun purchase to persons believed to be at high risk for firearm violence. Am J Public Health. 1999;89:88–90. doi: 10.2105/ajph.89.1.88. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 68.Cook PJ, Ludwig J. Guns in America: results of a comprehensive national survey on firearms ownership and use. Washington DC: The Police Foundation; 1996. [Google Scholar]
  • 69.Miller M, Azrael D, Hemenway D, et al. ‘Road rage’ in Arizona: armed and dangerous. Accid Anal Prev. 2002;34:807–14. doi: 10.1016/s0001-4575(01)00080-x. [DOI] [PubMed] [Google Scholar]
  • 70.Hemenway D, Vriniotis M, Miller M. Is an armed society a polite society? Accid Anal Prev. 2006;38:687–95. doi: 10.1016/j.aap.2005.12.014. [DOI] [PubMed] [Google Scholar]
  • 71.Bureau of Justice Statistics. Survey of State Criminal History Information Systems, 2010. Washington DC: Bureau of Justice Statistics; 2011. [Google Scholar]
  • 72.California Department of Justice. The Adult Criminal Justice Statistical System (ACJSS) Longitudinal Data Base. Technical Manual. Sacramento, CA: Bureau of Criminal Statistics, California Department of Justice; 1985. [Google Scholar]
  • 73.Grossman DC, Mueller BA, Kenaston T, et al. The validity of police assessment of driver intoxication in motor vehicle crashes leading to hospitalisation. Accid Anal Prev. 1996;28:435–42. doi: 10.1016/0001-4575(96)00007-3. [DOI] [PubMed] [Google Scholar]
  • 74.Brick J, Carpenter JA. The identification of alcohol intoxication by police. Alcohol Clin Exp Res. 2001;25:850–5. [PubMed] [Google Scholar]
  • 75.Brubacher JR, Chan H, Fang M, et al. Police documentation of alcohol involvement in hospitalized injured drivers. Traffic Inj Prev. 2013;14:453–60. doi: 10.1080/15389588.2012.725879. [DOI] [PubMed] [Google Scholar]
  • 76.McCarty ML, Sheng P, Baker SP, et al. Validity of police-reported alcohol involvement in fatal motor carrier crashes in the United States between 1982–2005. J Safety Res. 2009;40:227–32. doi: 10.1016/j.jsr.2009.04.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 77.Federal Bureau of Investigation. Uniform crime reporting handbook. Clarksburg, WV: U.S. Department of Justice, Federal Bureau of Investigation; 2004. [Google Scholar]
  • 78.Azrael D, Cook PJ, Miller M. State and local prevalence of firearms ownership measurement, structure, and trends. J Quant Criminol. 2004;20:43–62. [Google Scholar]
  • 79.California Office of Statewide Health Planning and Development (OSHPD) Trends in firearm-related emergency department visits and hospitalizations in California, 1992 to 2010. Sacramento, CA: State of California, Office of Statewide Health Planning and Development; 2013. [Google Scholar]
  • 80.Hunt JR, White E. Retaining and tracking cohort study members. Epidemiol Rev. 1998;20:57–70. doi: 10.1093/oxfordjournals.epirev.a017972. [DOI] [PubMed] [Google Scholar]
  • 81.Ribisl KM. Minimizing participant attrition in panel studies through the use of effective retention and tracking strategies: review and recomendations. Eval Program Plann. 1996;19:1–25. [Google Scholar]
  • 82.Robinson KA, Dennison CR, Wayman DM, et al. Systematic review identifies number of strategies important for retaining study participants. J Clin Epidemiol. 2007;60:757–65. doi: 10.1016/j.jclinepi.2006.11.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 83.Weinberger M, Weinberger MH, Fineberg N, et al. Long-term follow-up of participants in clinical studies. J Clin Epidemiol. 2002;55:230–4. doi: 10.1016/s0895-4356(01)00437-1. [DOI] [PubMed] [Google Scholar]
  • 84.National Center for Health Statistics. National death index user’s guide. Hyattsville, MD: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 2013. [Google Scholar]
  • 85.United States Department of Agriculture. [accessed 17 Aug 2013];Urban-Rural Continuum Codes, 2003. http://www.ers.usda.gov/data-products/rural-urban-continuum-codes.aspx#.UubZwvvTnmQ.
  • 86.California Department of Social Services. The regions of California: recommended grouping of the counties for regional studies. Sacramento, CA: California Department of Social Services; 2002. [Google Scholar]
  • 87.Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics. 1986;42:121–30. [PubMed] [Google Scholar]
  • 88.Hubbard AE, Ahern J, Fleischer NL, et al. To GEE or not to GEE: comparing population average and mixed models for estimating the associations between neighborhood risk factors and health. Epidemiology. 2010;21:467–74. doi: 10.1097/EDE.0b013e3181caeb90. [DOI] [PubMed] [Google Scholar]
  • 89.Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457–81. [Google Scholar]
  • 90.Cox DR. Regression models and life tables. J Roy Stat Soc. 1972;34:187–220. [Google Scholar]
  • 91.Lin DY, Wei LJ. The robust inference for the Cox proportional hazards model. J Am Stat Assoc. 1989;84:1074–8. [Google Scholar]
  • 92.Hosmer DW, Lemeshow S. Applied survival analysis. New York, NY: John Wiley & Sons; 1999. [Google Scholar]
  • 93.Prentice RL, Williams J, Peterson AV. On the regression analysis of multivariate failure time data. Biometrika. 1981;68:373–9. [Google Scholar]
  • 94.Kelly PJ, Lim LL-Y. Survival analyses for recurrent event data: an application to childhood infectious diseases. Stat Med. 2000;19:13–33. doi: 10.1002/(sici)1097-0258(20000115)19:1<13::aid-sim279>3.0.co;2-5. [DOI] [PubMed] [Google Scholar]
  • 95.Hill AB. The environment and disease: association or causation? J R Soc Med. 1965;58:295–300. doi: 10.1177/003591576505800503. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES