77 Gun availability and control in suicide prevention

Although countries differ in the most frequent methods of suicide, firearms are a preferred method in some countries. It is a method with high case fatality. Individual (case–control) studies and population studies show that restriction of firearms reduces suicide. This is especially evident in the young. Further, population studies show that gun control, such as strict licensing and restricted availability of firearms, is effective. Individual and population studies also found that one must especially control the availability of firearms to people with mental disorders. Canada's Criminal Law Amendment Act of 1977 (Bill C-51) illustrates the effect of legislating means restriction when one controls for confounding social and economic factors. The young show the most significant decrease in suicide rates, not substituting guns with other methods of suicide. Studies on firearms and suicide across the globe support the studies from Canada. Yet, there are researchers who espouse the opposite and the debate has been polemic. More research is needed to strengthen the conclusion on the positive effects of gun control laws worldwide.

One public health approach to suicide prevention is restricting the means of suicide: today there is prevailing consensus that this method can be effective (Stengel 1964; Kreitman 1976; Leenaars et al. 2000; World Health Organization 2002, 2006). Erwin Stengel (1964) was one of the first in the twentieth century to propose a public health approach as a means to decrease the incidence of suicide, noting, for example, that the detoxification of domestic gas (from coal gas with high carbon monoxide content to natural gas) might have reduced the suicide rates in nations where the switch had taken place. Subsequent research on the detoxification of domestic gas in England supported Stengel's proposal (Kreitman 1976). A comprehensive review of the research around the globe supports the approach more generally, often called controlling the environment (Leenaars et al. 2000). After careful analysis of all measures by an international team of researchers, headed by Bertolote (1993), a series of tactics to prevent suicide that had support in the empirical literature were proposed. Among the tactics were gun possession control, detoxification of domestic gas, detoxification of car emissions and control of toxic substance availability. Gun control legislation is often cited as the prototypical example of public health intervention to prevent suicide (Bertolote 1993; Leenaars et al. 1998; WHO 2006). Gun control is, for example, an excellent example because guns are a lethal method and easy availability can obviously facilitate death. One would predict that availability would increase suicide and that stricter gun laws would reduce the suicide rate. There are ample laws, for example, in Australia, Canada, New Zealand and the United States to allow for scientific investigation. In this chapter, a review is performed of the research on the association of the availability of the guns and suicide, and on the effect of gun control legislation or laws on suicide.

In a public health perspective, one can study the association between firearms and suicide at an individual and population level, sometimes called the ecological level. Each level has its limitations and benefits. The individual (case–control) empirical studies problem is finding reasonable controls and allowing for generalizations, often from small groups to larger groups in society. The population studies problem includes the ‘ecological fallacy’ when drawing conclusions about individuals from groups (aggregate) data. Each approach has its weaknesses and its strengths.

Availability of guns is difficult to measure; oftentimes only indirect findings are available, e.g. the accidental death rate by firearms and the percentage of firearms used for crimes such as murder (Lester 2000). There are, however, a few studies on actual firearm ownership, such as measured by gun licences. In the United States, measures of actual firearm ownership are available in nine regions (but not for all states). Markush and Bertolucci (1984) found that in the US the actual ownership was positively associated with suicides by shooting and the total suicide rate. Lester (1988a) replicated that ownership was positively associated with the firearm suicide rate. Lester (1988b) found that ownership of firearms in Australia was positively associated with suicide by shooting; Carrington and Moyer (1994a) reported that firearm suicide was associated with ownership in Canada; and Etzersdorfer, Kapusta and Sonneck (2006) concluded that the same was valid in Austria.  Killias (1993) and Lester (1996) both found a positive association between the percentage of households with guns and suicide by shooting in twelve nations. A longitudinal analysis of firearm availability in an array of countries—US, Switzerland, Finland, France, Canada, Sweden, Australia, the Netherlands, England and Wales, Scotland, Norway, Spain and New Zealand—support the association; the study confirms the same patterns and associations of households owning firearms and suicide (Ajdacic-Gross et al. 2006). The international data, further, showed that the proportion of firearm suicides decreased simultaneously with the decrease in the proportion of homes owning guns. Consequently, the following question surfaced among public health professionals: would restriction of the means reduce suicides and violence in general (Leenaars 2006, 2007)?

The case–control studies indicate that a gun in the home is significantly associated with a higher risk of suicide, especially among the young (Miller and Hemenwey 1999). Kellerman et al. (1992), in a frequently cited study, found that people who died by suicide were more likely to have a gun in the home. They studied all cases of suicide that occurred in the home during thirty-two months in two urban areas, Shelby County, Tennessee and King County, Washington. They controlled for sex, race, age range and confounding variables such as failure to graduate from high school, living alone, consumption of alcohol, current medication, and hospitalization. After controlling for the confounders, the presence of a gun was highly significant. The risk was higher when the gun was a handgun, loaded and kept in unlocked storage. Cummings et al. (1997) overcame the common problem of informational bias in case–control studies on suicide. They studied the actual handguns present in the home by analysing whether the purchase of a gun from a licensed dealer was associated with the suicides of 353 people using any method to kill themselves during the period 1980–1992. They controlled for social factors (confounders) such as income and education, and showed that the person who died by suicide was more likely to have had a gun and to have purchased a gun as compared to matched controls.

Not only studies across age groups, but also the study by David Brent on adolescents have shown similar results. Brent et al. (1988), in a pilot study, noted that a small group of adolescents who died by suicide (N = 27), compared to a psychiatric control group, had more often had firearms available in the home. In a larger-scale study (N = 47), Brent et al. (1991) showed that the presence and accessibility of firearms in the home of adolescents committing suicides were significant factors in the deaths. Guns were twice as likely to be in the home; the more guns, the greater the risk. Handguns did not differ from long guns with respect to risk. Alcohol was associated; the people who died by suicide with guns were more frequently intoxicated. A third case–control study (Brent et al. 1993), with a larger sample (N = 67), replicated the finding: the association between suicide and guns in the home is significant. Even if one controls for mental health disturbance (psychopathology), there is a particularly high risk. Subsequent case–control studies replicated the finding (Beautrais et al. 1996; Bailey et al. 1997). Brent et al. (1993) showed statistically that one can generalize the results of the individual cases to the community as a whole, not just for adolescents, but for the whole population (Miller and Hemenway 1999; Lester 2000). However, from a public health perspective, the individual case studies must be supplemented by population studies.

The research at the individual level showed that the association of guns and suicide is not simplistic. Psychopathology and drug/alcohol abuse, especially alcohol, are associated in the equation (Miller and Hemenway 1999). If a suicidal person, for example, has a mental disorder, the risk of the gun in that home greatly increases the risk for suicide.

Although countries differ in the most frequent method of suicide, firearms are a preferred method in a number of countries. This is true in Canada, for example. Gun control, by strict licencing and restricted availability of firearms, is often cited as the prototypical example of public health intervention to prevent suicide (Bertolote 1993; Leenaars et al. 1998).

An opportunity to study the preventive effects of legislative means restriction, e.g. gun control laws, on suicide is provided by Canada's Criminal Law Amendment Act of 1977 (Bill C-51), enforced since 1978. This Act required an acquisition certification for all firearms, restricted the availability of some types of firearms to certain types of individuals, set up procedures for handling and storing firearms, required permits for those selling firearms, and increased the sentences for firearm offences. Early commentators on the impact of this Act (Mundt 1990), reported little impact of the Act on firearm suicide in Canada, but presented only simple charts with no statistical analysis of the trends. Lester and Leenaars (1993, 1994) reported a comprehensive study on the preventive effects on suicide the Act had demonstrated in Canada. The results suggested that strict firearm control laws might well have been associated with changes in suicide rates due to the reduced use of firearms for suicide, and no overall switching from firearms to other methods for suicide. Lester and Leenaars' results suggest that Bill C-51 in Canada appeared to have had a significant positive impact, namely, lowering the rate of suicide by guns.

Subsequently, Leenaars and Lester (1997a) examined whether the restriction of firearms in Bill C-51 in Canada had a preventive impact for those of all ages or only for some ages. The results showed that the percentage of suicides using firearms decreased only for those aged 15–64, while it increased for those aged 65+. Thus, the general conclusion seems to be that the impact of making the gun control laws stricter in Canada on suicide was not apparent in those over the age of 65, but significantly so in the young. Shneidman and Farberow (1957) had already shown that the suicidal elderly have a stronger intent to die than younger adults. They are less ambivalent and there is a chronic course (Leenaars 2004). Younger suicides are generally more situationally disturbed and impulsive (Leenaars 2004). Perhaps the younger individual may be sufficiently delayed by increased difficulty in obtaining the preferred method for suicide so that the suicidal pain has passed by the time he or she has succeeded in obtaining the means for suicide. This argues for the waiting period hypothesis. Considering the elderly, the result is consistent with other research (Lester 1992) suggesting that education programmes, crisis intervention services and telephone crisis centres are less effective with the elderly than with young adults. Thus, much more thought must be given and research undertaken to devising tactics for preventing suicide in the elderly (Richman 1993).

There is evidence that men and women may respond differently to restricted access to lethal methods for suicide. A study by Leenaars and Lester (1996) was designed, therefore, to explore whether the 1977 gun control law in Canada had different effects on men and women in their use of guns for suicide. The results indicated that the passage of C-51 seemed to have had a greater preventive impact on suicide as well as on homicide in women than in men, possibly because more men switched suicide method. This phenomenon is called displacement. Carrington (1999) subsequently suggested that significant downward changes in death rates for suicide for both men and women occurred, with no evidence for switching method in both sexes. Recent re-evaluations found that the intervention had an impact on both men and women, probably because more men use firearms, but some men, indeed, showed displacement to other methods when committing suicide (Leenaars et al. 2003). Thus, public health approaches, such as gun control, may have limits in their preventive effects. The decrease of suicides among women is in agreement with other findings showing that suicidal women respond better to clinical intervention programmes, educational endeavours, use of telephone crisis centres, and receiving care for psychiatric disturbances (Lester 1995). Thus, it makes sense that this would be true with tactics of environmental control, but research to explain the phenomenon and for developing gender-specific interventions is lacking.

In a review of gun control studies, Stack (1998) pointed out that gun control studies rarely take into account other societal factors (confounders) which may influence the suicide rate—factors such as divorce, unemployment and the age structure of the population. A study by Leenaars and colleagues (2003) examined whether social changes might have had an impact on suicide rates from guns. Birth, marriage and divorce rates as measures of domestic integration—a social variable which Durkheim (1897) argued was critical in determining suicide rates—the unemployment rate (Platt 1984), median family income and the percentage of the males aged 15–24 years as a percentage of the total male population (a group with one of Canada's highest suicide rates) were studied. The results showed that the passage of Bill C-51 in Canada in 1977, introducing stricter gun control, appears to have been followed by a significant reduction in the suicide rate by firearms, even after controlling for some social variables. Thus, it appears that even if one controls for social variables, gun control succeeds in preventing suicides. There may, of course, have been social changes other than those considered that might be responsible for changes in the suicide rates, and the use of a different set of indicators such as alcohol and drug consumption the percentage of foreign-born, social class and education levels should be studied in the future, if and when the necessary data become available.

However, some researchers have disputed the conclusion that gun restriction has prevented suicide in Canada and elsewhere (Rich et al. 1990; Sloan et al. 1990; Kassirer 1991). Rich et al. (1990) reported, using a small sample, that stricter gun-control laws in Toronto were followed by decrease in suicide using guns, but at the same time an increase of suicide using other methods, like subway jumping, was observed. However, Carrington and Moyer (1994b) replicated the positive findings for gun control in Ontario. They also showed that most of Canada's provinces had either stable or decreasing rates of firearm suicides, following C-51 (Carrington and Moyer 1994a). Results on related phenomena of homicide (Leenaars and Lester 2001) and accidental deaths (Leenaars and Lester 1997b) hold equal promise.

There have been two subsequent laws in Canada, Bill C-17 and C-68. Bill C-17 in 1991 tightened the screening provisions for firearm acquisition (FAC). A more complete search of an applicant's personal and criminal records and 28-day waiting periods were required. Bridges (2004), following Leenaars and Lester's design, examined the impact of the bill and showed a significant decrease after passage of Bill C-17 in the rate of suicides involving firearms and the percentage of suicides using firearms. Bill C-68 was more an attempt at people control than gun control. It required people to register all shotguns and rifles. It was to be implemented in 1999, then 2001, then in 2003, and so on. The bill has encountered mounting problems and has yet to be enacted. The bill itself, unlike C-17, for example, is controversial in Canada and has met with significant opposition: the current government, in fact, introduced a bill on June 19, 2006 to rescind the bill. To date, the bill is still in limbo.

Farmers, hunters, indigenous people and many other people in Canada, as elsewhere around the world, have been against such registration of guns. They view this act as a control of common rights. Environmental controls, at least within public health, are known to be more effective than people controls. Simply stated, not all controls are accepted. This debate is not unusual about gun control worldwide (Leenaars 2006).

Beautrais, Fergusson and Horwood (2006) examined the impact of more restrictive firearm legislation (Amendment to the Arms Act 1992) in New Zealand on suicides involving firearms. The amendment restricted accessibility, required confidentiality checks of two references, and introduced more strict storage and safety requests. After the legislation, it was shown that firearm-related suicide decreased dramatically, especially in youth. Thus, the New Zealand study not only replicated the Canadian findings, but also supported the finding that the tactic was especially useful with young people, a high-risk group for suicide globally (Wasserman et al. 2005).

In the last twenty years or so, the positive effects of gun control laws have also been studied in the United States and Australia; these studies too support the research of Leenaars and Lester, the most comprehensive studies to date on any continent (Lester 2000). For example, in the US, Ludwig and Cook (2000) examined the association of guns and suicide with the implementation of the Brady Handgun Violence Protection Act and noted a decrease of suicide by firearms in older adults (age >55); when there was no reduction in suicide by all methods. The authors attributed this to interim waiting periods for firearms acquisition in the Brady Law, which has since been dropped in the permanent law. Maxwell and colleagues (1984), Medoff and Magaddino (1983) and Sommers (1984) have also reported that the restrictions in the selling and purchasing of guns played the most significant role in reduced suicide rates. These restrictions were especially important in the control of buying of guns by people with mental health disorders (psychopathology). A waiting period was also important (Medoff and Magaddino 1983). The various studies also offered some data on the fact that these impacts were true even if one controls for some social factors. The number of factors that were studied, however, were few. This warrants caution in over-generalizing, something Leenaars and his colleagues (2003) attempted to rectify. Thus, it is of note that waiting periods, like restriction of sale to some people such as those with mental disorders (psychopathology), may be important. Further studies need to explore what is effective not only at an individual level, but also at national and international levels. The science of controlling the environment for the means is at that stage now. Probably regional differences will emerge. For example, the US is known to be well above the average among industrialized nations in firearm violence (Hahn et al. 2003).

In Australia, after a 1996 firearm massacre in Tasmania in which thirty-five people died, the Australian government undertook gun law reforms. Legislation restricted the sale of semi-automatic shotguns and rifles to civilians. Gun registration was legislated and the prohibition of private sales was implemented. Chapman et al. (2006), using official statistics, showed that after the law there was an accelerated decline in firearm deaths, especially suicides. This Australian example once more showed that removing lethal means in the environment could be associated with a sudden and ongoing reduction in suicides and related phenomena of homicide and accidental deaths. This finding is consistent at both the individual and population levels of analysis.

Kleck (1991), Britt et al. (1996) and Kleck (1997) argued that gun control fails to reduce the rate of suicide and violence in general. Hemenway (1999) pointed out problems in Kleck's analysis. Also, Leenaars et al. (2003) addressed Kleck's arguments about statistics and designs at a population level, by using interrupted time series analyses and controls for possible confounders, and showed that the associations between the availability of guns and suicide hold. Cummings et al. (1997) showed the same at an individual level. The discussion, of course, will continue (Leenaars 2006).

The tactic of controlling the environment is more complex than gun control, although it is a prototypical example. Referring to gun control as a means for suicide prevention in Canada and in the US probably makes sense, because it is those countries' main method of suicide. It is a highly lethal method. However, in other countries, guns are not a frequent means. For example, in Cuba, firearms are the least frequent means for committing suicide in the nation: hanging is the most common for men and immolation the most frequent for women. The main suicide method in Lithuania is also hanging. This is true in Australia. In China and India, the common methods are poisoning by insecticides or pesticides and hanging. Hanging and drowning are the most popular in Ireland. Hanging is the most common method in Thailand, but the use of firearms is increasing. Firearm suicides have increased in Russia. The same are true in Columbia (WHO 2002). Hanging and the use of firearms are the main methods in South Africa. There is great variation in means of committing suicide and, although gun control may make little sense in some regions of the five continents, the tactic of controlling the environment may well have increasing applicability, as suggested in the WHO report (Bertolote 1993), and as studies on different continents show (Leenaars et al. 2002).

There are numerous studies on the impact of guns on suicide, such as firearm ownership, availability of guns, guns in the home and firearm storage, and gun control legislation (see Lester 2000; Leenaars 2007; Miller and Hemenway 1999 for reviews). Studies on the availability of guns show that if present, the rate of use for the means of suicide increases. Ownership shows the same pattern. Individual-level studies show that the more guns in the home, the more suicides, especially among the young. Population-level studies show the same patterns and associations. There are consistencies (commonalities) in the empirical studies, both individual and ecological. Most of the empirical studies, together with gun control law studies across the globe, show that availability of a potentially dangerous means affects the rate of use of the means for death (WHO 2006). Firearms are a method with high case fatalities, not only for suicide, but also for homicide.