Suicide Prevention and Intervention: Summary of a Workshop (2001)

much different experience from patients presenting with other rare maladies that often fascinate physicians.

The workshop attendees discussed case identification versus treatment of already identified cases as approaches to reducing suicide through primary care. Drs. Davis, Reynolds and Schulberg stated that studies show that 30 to 50 percent of cases of clinical depression goes undiagnosed. Dr. Brent discussed studies indicating that treatment of known cases was a more practical approach, since many of those undetected may remit without treatment. The workshop attendees also discussed that different approaches will likely be optimal for different age, gender, and ethnic and racial groups.

Research on Suicide Intervention in Primary Care. Dr. Schulberg noted five problems in research on suicide prevention via primary care. (1) There is a dearth of studies in this area in the US, with most of the data coming from countries with national health care systems such as Britain and Scandinavia. (2) Methods for case identification, vary across studies, with little or nothing known about the validity or reliability of many of them (psychological autopsy is an exception as a well-studied method). (3) Universal definitions of suicidal ideation and attempts, a necessity for cross-study comparisons, have not been adopted. (4) Variability of observational time period used to define suicide rate, ranging from 1 to 6 months, with rates ranging from 1.0 percent to greater than 7 percent across studies, hinders cross-studies comparisons. (5) Most research is retrospective. Ethical issues complicate prospective studies because study design often requires placing at-risk patients in control groups.

Dr. Schulberg discussed a current prospective study of suicide prevention in older adults through primary care, the “Prospect Study.” Dr. Schulberg highlighted “two premises at the heart of this [study].” First, effective treatments for depression do exist. Second, the primary care physician personally is not going to engage in all the necessary activities. This study is expected to generate information on whether or not treatment can occur within the primary care setting, and if and how the primary care physician’s role can be enhanced in suicide prevention. In the Prospect Study, patients are “randomized into the intervention arm and into the usual care arm,” which gives power to this study, rarely available in studies of suicide. All patients are screened multiple times across the study, and Dr. Schulberg noted that that depressive symptoms, not suicidality, is the key independent variable. Identified patients are offered individual psychotherapy and/or medications.

Dr. Schulberg summarized with three important points for treating suicidality in the primary care setting. (1) The primary care physicians’ awareness of suicidality does not necessarily lead to proper intervention. (2) Physicians need to be significantly more comfortable with asking about suicidal symptoms. (3) Possibly the most potent intervention deals with organization and fiscal structure of the physicians’ practice. Current constraints result in multiple disincentives to assess and/or treat suicidality.

Dr. David Hemenway discussed the impact of firearm suicides on the national suicide rate for the United States. Reducing access to methods has proven successful in impacting overall

suicide rates, as efforts in other nations including Japan and Great Britain have shown. He discussed research findings on firearm access and suicide, cross-national data, and issues of data collection.

Dr. Hemenway explained that the U.S. has the highest rate of firearm suicide of all 27 developed nations, whereas we have the 16^th highest rate of suicide. Over 50 percent of all suicides are by firearm in the U.S., according to Dr. Hemenway. Some reasons for the high firearm suicide rate according to Dr. Hemenway are the high number of handguns and less regulation of firearms than in other developed nations. The U.S. lacks a national licensing or registration system, and there are no national storage laws. There is also a large secondary market of gun sales.

Dr. Hemenway pointed out that one problem in doing cross-national studies is that there is a lack of good measures, especially of gun prevalence. One report he described looked at 5–14 year olds in the U.S. and compared their suicide rate with that of other industrialized countries. Easy accessibility of guns was found to be a risk factor for suicides in the U.S., compared to other countries.

Dr. Hemenway discussed data from two types of studies, case-controlled and ecological studies. The case-controlled studies on gun prevalence and suicide risk have revealed significant increases in suicide in homes with guns, even when other factors such as education, arrests, and drug abuse were controlled for. Dr. Hemenway and the workshop participants discussed criticisms of these studies made by other researchers in the field. Case-controlled studies have been criticized for not looking at all suicides, just those occurring in the home. Workshop participants qualified this criticism, explaining that this was true of the largest study done by Kellerman and colleagues, but not of other studies done by Brent and colleagues. Another criticism is that the respondents are not telling the truth. Another criticism is that the respondents are not telling the truth. Dr. Hemenway described a survey study done by Kellerman and colleagues, finding that people do tell the truth, but indicated that this research was not definitive. There is some evidence that women underreport the presence of a gun in the home. Dr. Hemenway described a study of an HMO population in Seattle, that found 25% of the suicides had purchased a handgun from a licensed dealer in the state of Washington, as compared to only 15% of controls (those who did not commit suicide).

Virtually all ecological studies on suicide and gun prevalence have shown a positive association between the two, according to Dr. Hemenway. Sometimes this association is statistically significant and sometimes not. One major obstacle in looking at the relationship between the two variables is that fact that good measures of gun availability are lacking. There are no national surveys of gun prevalence in the home. Other measures used instead include the strictness of gun laws, percent suicides with a gun, fatal firearm accident rate, and firearm ownership, considered to be the best standard.

Studies that have compared gun suicide rates between states with a high prevalence of guns and states with a low prevalence of guns, have consistently found that the high gun states have higher rates of suicides committed with firearms than low gun states. Almost twice as many people in the states with high gun prevalence commit suicide. The evidence suggests that there is little substitution of means.

Dr. Hemenway stressed the need for collecting data on fatal injuries, including firearm injury. While the U.S. has an excellent data collection system motor vehicle injuries, there is no comparable system for violent deaths such as homicide, suicide, and unintentional gun deaths. Currently, not much is known about suicides at the national level or even across states. Most information is gathered from death certificates and small area studies. A national surveillance system that informs us on circumstances such as the percentage of suicides committed with a handgun, how recently a gun was purchased, whether or not substance abuse was involved, or if there was a precipitating event would be extremely valuable to researchers, according to Dr. Hemenway.

Dr. Hemenway informed the workshop attendees that the Harvard Injury Center has begun collecting such data for a national violent death reporting system. With the pilot system, information on the blood alcohol content of those committing suicide is available, as well as information on drug use, location of death, etc. Dr. Hemenway noted that the Centers for Disease Control are hoping to develop further a national violent death reporting system, if they receive the funding from Congress.

Dr. Philip May commented that one variable that also needs to be examined in studies on gun prevalence and suicide risk is gun values and family culture regarding guns. Are family members educated in safety training with the guns in their household? Is the gun used for hunting or not?

Dr. David Brent described a study on handguns with highly compliant families of depressed adolescents. The families attended all interviews and therapy sessions and formed a good relationship with the clinicians. Only 1 in 4 of these families removed a firearm from their home when the clinician requested it. Dr. Brent emphasized the need to figure out how to separate guns from persons at risk of suicide.

Dr. Hemenway emphasized the need to get at storage habits of gun owners through surveys and research. He stressed that maintaining a violent death registry would not only provide extremely valuable data on suicide, but it could also provide data on homicide and unintentional gun deaths. Dr. Hemenway was successful in getting questions about guns and injury added to the National Comorbidity Survey. These questions will provide researchers with data previously unavailable, such as percentages of persons with mental illnesses who have guns in their household and whether or not these people are at increased risk for suicide. Dr. Hemenway also stressed the need to add questions examining the presence of guns in homes to longitudinal studies.