CRIMINOLOGY'S ASTROLOGY: An Evaluation of Public Health Research on Firearms and Violence by PAUL H. BLACKMAN, Ph.D. Institute for Legislative Action National Rifle Association A paper presented at the annual meeting of the Academy of Criminal Justice Sciences, Denver, Colorado, March 13-17, 1990 Both astronomy and astrology are concerned with precise descriptions of the course and relationships between the stars and planets, with both interested in predicting the changing positions of the heavenly planets, and both able to assist in navigation due to their concerns for accuracy (Boorstin, 1983:14-21, 264). Similarly, criminology and "public health professionals" (Doctors of Medicine and Masters of Public Health = MD/MPHs) are concerned with accurate data regarding the causes of at least some deaths from external causes. Similarly, astrology, like public health, was an interdisciplinary study, part of which was involved with medicine (Boorstin, 1983:305). They differ in that astrology attempts to predict the future whereas public health professionals want to use their data to change the future. They share, however, a predisposition to believe that the citation of these figures and relationships -- mortality data and trends or the movements of the stars and planets -- prove something. Unfortunately, at least when the public health professionals address the issue of firearms and violence, they do not. All they produce are numbers and rates, with some percentages; these numbers sometimes include trends, with no particular meaning. All that count are deaths involving firearms. No other factors are measured. Sometimes, they are mentioned; and in areas where the MD/MPHs have some knowledge, they will give precise lists of the important factors they excluded, such as mental illness, alcoholism, etc., with regard to suicide. But the assumption is that the numbers alone justify the statement of the preconceived belief. Nothing will thwart that belief. If the numbers do not fit, then other, unknown, factors have been at play keeping the stars from acting properly. Description alone is inadequate; the numbers must be read, and the assumption is that if MD/MPHs read the numbers, society should accept their statements, preconceived though they be. Science has been identified as that which can be disproved. If the facts do not fit in a belief system, on the other hand, the facts are altered, or some explanation is forthcoming as to why the belief system still holds. So it is with astrology, and so it is with MD/MPH studies of firearms and violence. No disproof is possible; no proof is necessary. The mere statement is believed enough to make firearms and violence a public health problem, subject to epidemiological analysis, with prescription forthcoming -- however unrelated to the lists of figures given. As astrology and astronomy both sought data on the universe, so the so-called public health professionals and criminologists both seek data on firearms-related violence. There the similarity ends. Criminologists want to understand the nature of the problems and to test possible solutions. MD/MPHs want to state data and prescribe a predetermined solution, changing only as some of the fads discussed by the media change. The only sense, thus far, in which MD/MPHs are in any position to aid criminologists in their scientific efforts is that they are in a better position to get mortality and morbidity data (death and injury) related to firearms and other weapons -- the external items in the International Classification of Diseases (ICD) -- as collected for the Department of Health and Human Services, improved. It is theoretically possible that MD/MPHs could learn from social scientists and improve their research. There is nothing in the literature to date indicating that they care to, and precious little indicating they would be capable of it even if they cared to -- any more than astrologers could any longer learn from, and in turn assist, astronomers. PUBLIC HEALTH PROFESSIONALS APPROACH FIREARMS AND VIOLENCE Since firearms ownership and use by different groups in the community have different social meanings, their study has long been appropriate to sociology. Similarly, since a significant minority of violent criminals misuse firearms, other citizens use them for protection, and even accidents and suicides are at least investigated by police to determine causes of death, criminologist have long, and appropriately, studied the issue of firearms and violence. The question whether firearms involvement in violence should invite the attention of so-called public health professionals -- MD/MPHs, occasionally also possessing or assisted by a J.D. or Ph.D. -- depends upon whether firearms-related violence constitutes a serious and preventable medical problem. If the violence is not preventable, then, while treating gunshot wounds is relevant to the practice of medicine, the firearms issue is not relevant to "public health." For the most part, the issue of preventability is ignored in the public health literature on firearms; occasionally (e.g., Browning, 1976) it is assumed: "Violence by handguns, specifically that leading to homicide, can be formulated as a public health problem and as an area for the practice of preventive medicine." Largely ignoring the issue of prevention, the literature finds a public health issue based on the magnitude of the number of cases, usually by adding together four firearms-related codes from the ICD categories of externally-inflicted causes of death (E): accidents, suicides, homicide and legal interventions (although some MD/MPHs subtract legal interventions and war- related deaths), and deaths involving firearms where it is undetermined whether the shooting was accidental, suicidal, or the willful killing of another. "Gunshot fatalities are an important public health problem" (Morrow and Hudson, 1986). And "There is unquestionably a need to treat this public health matter with as much urgency as any dread disease" (AMA Council on Scientific Affairs, 1987). By simply and repeatedly asserting that it is a public health problem, medical attention is justified as far as MD/MPHs are concerned. MD/MPHs are called on not merely to study the problem, but to influence public policy on the matter, since the "lack of a stable visible constituency for change has often prevented [public] support from being translated into public policy. By emphasizing that firearms are a major public health problem, concerned health professionals can reverse this trend" (Wintemute et al., 1987a:377). Not only does the medical profession insist that the number of firearms-related deaths make it a public health issue, by counting only deaths, dismissing virtually anything not fatal as not subject to reliable measurement (Kellermann and Reay, 1986b), firearms become more important to medicine than to criminology, since there are clearly more firearm-related deaths than firearms-related willful homicides. The medical profession is not necessarily interested only in deaths (mortality), but non-fatal injuries (morbidity) are more difficult to measure -- particularly since the profession wishes to combine all types of gun-related deaths together, and they are apt to be associated with different ratios of injury-to-death. Generally, the ratios discussed are two (Martin et al., 1988) or 3-5 to one (Christoffel, 1989), although the American Academy of Pediatrics has put the figure at 10 to one, and an imaginative Garen Wintemute has put the ratio at about 40 to one for accidents (1988), and Martin et al.(1989) have said the ratio simply cannot be known. Having asserted that guns are a public health problem, the question then becomes, how should that problem be addressed? The answer indicates that the medical profession has some difficulty comprehending its own terminology. Paul Stolley, president of the American College of Epidemiology, said: "Homicide is not a disease, but it is a public health condition whose primary cause is possession of guns -- and it could be considered on epidemic because of the high incidence in certain populations" (Journal of the American Medical Association, 261:675, Feb. 3, 1989). A Firearms Epidemic? Dr. Stolley is not the only person who seemingly does not know what an epidemic is. Louis Sullivan, Secretary of Health and Human Services, said there is an "epidemic" of homicides involving black men (Washington Post, Dec. 10, 1989). And others have compared firearms-related deaths to other things which might be treated like epidemics. "Imagine the medical and general public concern if there were 11,124 deaths per year from typhoid fever, botulism, or mercury poisoning" (Browning, 1976:1199). Noting that handguns were involved in 474 deaths involving persons 13-18 and 156 for those under twelve, Schetky (1985) said: "Any disease that produced such an alarming number of deaths would receive prompt attention from the medical profession and the government, as did the recent outbreaks of toxic shock syndrome, legionnaires' disease, and acquired immuno-deficiency syndrome." The more cautious MD/MPHs try to limit the discussion of epidemics to particular groups. "Firearms are a major public health problem in the United States, as they have been for most of this century. For females as a group, and for male teenagers and young adults, true epidemic conditions now exist; their mortality rates are much higher than at any time previously...for much of our population a true epidemic of firearm deaths now exists" (Wintemute, 1987). But it does not, whether one is discussing the population as a whole, or particular groups. For the population as a whole, one might argue that our high homicide rates are endemic, but that is not the same as -- indeed, is closer to the opposite of -- epidemic. An epidemic requires -- besides a contagious disease, of course -- dramatically higher rates in a given population. The Centers for Disease Control (CDC) recently announced that influenza had reached epidemic proportions because, for two straight weeks, flu deaths accounted for 7«% of all deaths compared to the normal 3«% for pneumonia and influenza (Washington Post, Jan. 26, 1990, p. A3). And a CDC researcher was quoted as dismissing the idea that there is an AIDS epidemic, suggesting instead a series of "subepidemics" at the regional level (Washington Times, Aug. 8, 1989), at a time when the National Center for Health Statistics (NCHS) Monthly Vital Statistics Report (Aug. 2, 1989) indicated that deaths from AIDS had risen from a rate of 5.8 for the first quarter of 1988 to 8.4 for the first quarter of 1989. There have been no dramatic rises in overall firearm-related deaths, or firearm- related deaths among various subgroups -- blacks, females, teenagers and young adults, etc. -- which would justify the term "epidemic." There have been fluctuations in firearm-related deaths, but no dramatic changes. Even among blacks, the overall homicide rate has been declining for several years, with the current increases merely producing stability over the past several years (CDC, 1986; FBI, 1989; Dept. of Commerce, 1989; National Safety Council, 1989). Treating a Non-Epidemic Epidemiologically. Even if there is no epidemic of firearms-related violence, either overall or among certain population groups, this does not mean that the issue cannot properly be studied epidemiologically. As Stolley noted, epidemiologists study the "distribution and causes of disease and health conditions in the human population" (Harrisburg Patriot-News, Jan. 7, 1989). And "Available state public health data sources and standard epidemiological techniques can presently be employed to identify the geo-demographics of at-risk populations and stimulate further research toward improving our understanding of the factors involved" (Alexander et al., 1985:167). The epidemiological approach would look at various factors which might be associated with higher, or lower, incidents of particular ailments. Just as epidemiologists might note whether scurvy, malaria, or other ailments were associated with different diets, insects, living conditions and the like, regardless of whether the incidence of such disease was high or low, stable or fluctuating, so epidemiologists could look at firearms-related morbidity and mortality and look to see what related factors there are. Unfortunately, the times when MD/MPHs look at firearms, all they see are firearms, even if other trends should catch their eyes, and raise their concerns. When Boyd (1983) looked at suicide trends, he should have noted increasing suicide among teenagers and young adults and shown concern about the cause, since there was no overall rise in suicide which could be related to firearms availability. Similarly, a truly epidemiological approach comparing Seattle, Washington, to Vancouver, British Columbia (Sloan et al., 1988) would have shown that the homicide rates were dramatically higher among the unstable Asian population in Seattle than in the more stable one in Vancouver, and highest among the blacks of Seattle. At that point, further study might have found other clues to what caused there to be higher homicide rates based on such factors as poverty, education, nutrition, drugs, alcoholism, or any of dozens of factors. The epidemiological approach would be concerned with both trends and with factors associated both with higher and lower levels of death. The MD/MPH approach to firearms, however, miss all of those factors for a number of reasons. First, by combining the types of firearm-related deaths, explanatory factors are confused. The different death rates among the different ethnic groups are minimized by combining high elderly- white suicide rates with high young-black homicide rates. Second, all factors except firearms are simply ignored, or presumed comparable in the groups studied -- either expressly (Sloan et al., 1988) or implicitly. Third, in looking at firearms, there is no examination of those not "afflicted." MD/MPHs would have to look both at the healthy and the unhealthy to find the differences between the two. They have wholly failed to show the slightest interest in the matter. In theory, of course, the medical profession could improve in its epidemiological approach to the issue of firearms and violence. There are a number of reasons not to be optimistic that such improvements will occur in the foreseeable future. LIMITATIONS ON THE MEDICAL APPROACH TO THE PROBLEM OF FIREARMS-RELATED VIOLENCE One problem with the medical profession is that MD/MPHs approaching the gun issue look almost exclusively at works appearing in medical journals. Citations outside the public health profession are rare, with anti-gun groups (Schetky, 1985; Martin et al., 1989) or general-interest publications (Ruben and Leeper, 1981) as apt to be relied upon as criminological ones. Some of the results are absurd, with a public health study from 1968 relied upon for accidental-death figures in 1987 (Wintemute et al., 1987a) to reach an apparent ratio of injury to death of about forty to one -- when most guesses ranged from 2- to 10-1 (Martin et al., 1988; Jagger and Dietz, 1986). Similarly, relying upon two medical studies, Frierson (1989) noted: "Sixty to 70 percent of Americans keep firearms in their homes, and in some Western states, the figure rises to more than 90 percent." Frierson goes on to suspect that the increasing rate of suicide with guns by women in such western states as Nevada and Wyoming is because of increasing acceptance of gun ownership by women there. In fact, of course, gun ownership by women has always been highest in rural areas -- for sport -- with increasing ownership for protection associated with increasing urbanization, crime, and female-headed households (e.g., Lizotte, 1981). Frierson would appear to be undermining his effort to associate increased gun ownership with increased suicide, if the suicide is up most in areas where ownership has always been high rather than where it is rising. Additionally, in forgetting to cite anyone but themselves, calculations about the seriousness with which the issue of firearms-related violence is taken by the government are flawed. Jagger and Dietz (1986) cite only the number and expense of studies of firearms funded by the National Institutes of Health in 1983 in wondering if anyone cares about the problem. Although they made references to the CDC, there was no indication that they knew that National Institute of Justice, or other Justice Department funding, existed. In addition to citing only one another, medical journals rely upon MD/MPHs in reviewing articles for publication. The issue of "peer review" and the refereeing of medical articles is a controversial one (Altman, 1987; Hamilton, 1989) even when the articles relate only to medicine. Medical politics and the "Old Boy Network" affect whose articles are printed; MD/MPH reviewers generally are told the author(s) of the paper they are peer-reviewing. Peers, of course, do not have the data sources used by the authors and must rely upon their honesty and accuracy for data collected (Hines and Randal, 1989). And reviewers would appear to be rather careless even in checking the data available. Other readers have found simple arithmetic errors in published articles (Stolinsky, 1984; Kleck, 1987), and Kellermann and Reay (1986a) cited two publications for the statement that less than two percent of homicides were legally justified, neither of which gave any such figure. The most widely-cited medical publication, the New England Journal of Medicine (NEJM) has been forced to back off its claims of how wonderfully its peer-review system worked (Hines and Randal, 1989), in part because of an article praising two drugs to suppress premature heartbeats in patients with heart attacks, which were responsible for an estimated 1500 deaths. NEJM defended itself, saying "I think the publication of even studies that turn out to be wrong is valuable to society," and blaming "naive, non-critical interpretation" of its studies (Hamilton, 1989). If NEJM is careless in its peer-reviewed medical studies, the problems are exacerbated when it turns to its "special articles" on areas outside the strict parameters of biomedicine. Misinterpreting and Misunderstanding Data: Peer reviewers or referees, after all, are only as good as the peers writing. And there is little indication that MD/MPHs writing outside purely medical area can understand the works of criminologists, pollsters, and others. Either that, or they are indifferent as to whether their statements are true or false when they cite outside their own field. Ignoring the numbers of robberies and assaults reported to the FBI, but anxious to make firearm-related incidents sound higher than they are -- or were, using peak-crime year 1980 -- Jung and Jason (1988:515) told their peers, "Of the over 13 million crimes committed throughout the nation during 1980, firearms were used in 43% of the robberies and 24% of the aggravated assaults." Teret and Wintemute (1983:347) interpreted FBI data to say that 45% of homicides were committed impulsively. And Kellermann and Reay (1986b), responding to criticism that their data counted only deaths to conclude that firearms were less often used for protection than misused, and attempting to show that surveys supported their conclusion, wrote: "In 1978, both the National Rifle Association and the National [sic] Center for the Study and Prevention of Handgun Violence sponsored door-to-door surveys. Both included questions regarding firearms and violence in the home...Taken together, these two polls suggest that guns kept in homes are involved in unintentional deaths or injuries at least as often as they are fired in self-defense." In fact, the NRA-sponsored survey, while it asked about protective uses of firearms, and whether the firearm was fired, did not ask about the location of the incident, and did not ask any questions about accidents. The survey commissioned by the Center did not ask whether protective uses of guns involved their being fired, nor where accidents occurred, although it did ask where protective incidents occurred (the majority occurred outside the home). The Center's protective-use questions were asked only of persons who owned handguns for protection. The Kellermann and Reay conclusion is refuted by Kleck (1988). Unfortunately, the MD/MPHs who study firearms do not seem much better at interpreting the mortality data at which they should be more expert. With virtually no data separating handguns from long guns, and with relatively small figures for all firearms, Christoffel concludes that "Handgun injury remains a major cause of morbidity and mortality in American society, particularly among young persons" (1989). The Mayo Clinic (1988) reported that "Upwards of 35,000 Americans die of gunshot wounds each year...Firearms are involved in two-thirds of homicide cases..." And the CDC's Patrick O'Carroll chimed in with: "Clearly, if three-fourths of homicides are caused by firearms, we have to look at their role" (JAMA, 261:675, Feb. 3, 1989). In fact, of course, firearms have been involved in about 60% of homicides during most of the 1980s, and the figure for firearm-related deaths has been fairly stable at 30-33,000. Part of the problem with data comes from an effort to emphasize children with a rather fluid definition. "Almost 1,000 children die each year from unintentional gunshot wounds. Most of those deaths occur in the home of the child" (Teret and Wintemute, 1983:346). Their statement, based on the 1980 Accident Facts from the National Safety Council (NSC), is only true if one counts as children all persons 24-years-old and younger; for those 14 and below, the total was 360. And, if one uses the higher figure, their second statement loses some credibility, since 58% of deaths occur in some home - - not necessarily that of the child; Christoffel (1989) says about 50% of youthful gun- accident victims die in their own homes. Projecting, for those 14 and under, "most" is about 36%; for those 24 and under, "most" is about 29%. Ignorance of Laws and Ballistics: There are two areas where the MD/MPHs who are studying the firearms issue have their studies flawed by ignorance important to their work. Not all MD/MPH studies are calling for legislation regarding firearms -- indeed, some appear to be trying to have changes occur which do not depend for their success on compliance by the firearms equivalent of the "nut behind the wheel" (Wintemute et al., 1987a). Some of the studies, however, and other pronouncements, make it clear that public health professionals are studying or calling for the study of the effects of gun laws, and others are calling for more restrictive gun legislation. For those MD/MPHs, ignorance of the law is important. Stolley, of the American College of Epidemiology, called for legislation for the sake of studying its effectiveness (Harrisburg Patriot-News, Jan. 7, 1989): "What we feel is necessary is some sort of convincing test of handgun control....The college recommends experiments in banning handguns be sponsored by communities to determine whether the expected decline in homicide rates actually occurs....The point we are making is that there has never been a test....It has never been tested. If it is true that a handgun ban reduced homicide, then it will have an enormous public health benefit." Several communities -- most prominently Washington, D.C. -- have adopted the experiment Stolley says has never been tried, with any "enormous public health benefits" thus far unnoticeable; D.C.'s gun-related homicide rate has risen 260% since enacting into law a Stolley-type experiment. Similarly, two prominent CDC-sponsored studies published in NEJM (Sloan et al., 1988 and 1990) attempted to compare American and Canadian jurisdictions with different gun laws, presumably with an eye toward determining whether the more restrictive gun laws in Canada were associated with lower rates of homicide and suicide. Both studies misstated somewhat the laws affecting Seattle/King County (Washington State and U.S. federal law) and Vancouver and its metropolitan area (Canadian national law). The most seriously ignored aspect was in the second study, of suicide, where age groups were studied and most of the emphasis was on the age groups below the age of 25. Although, in general, the laws of the U.S. are less restrictive than those of Canada, acquisitions of handguns and of long guns are both lawful at a younger age in Canada than in the U.S. --18 vs. 21 for handguns; 16 vs. 18 for rifles and shotguns. Perhaps the most serious such flaw occurred in Boyd (1983), since his study is now considered by MD/MPHs reviewing the firearms issue as virtually definitive proof that increased firearms availability is associated with increased levels of suicide and that restrictive gun laws would curb such suicide. Boyd's study, however, covered the period from 1953 to 1978, and found that firearms involvement was up (particularly among young persons). Ignored by Boyd was the fact that the period studied included adoption of the most significant federal gun legislation ever adopted, the Gun Control Act of 1968 (Stolinsky, 1984), and the most widespread adoption or extension of state and local firearms laws -- especially aimed, as Boyd recommended, at handguns -- in American history, covering most of the population with some form of restrictions. Many of the laws adopted were aimed, among other things, at restricting the access of persons under certain ages -- including parts, at least, of the age group for which Boyd found suicide rising. More puzzling -- and more dangerous, whether the MD/MPHs are recommending legislation or practicing medicine -- is medical ignorance of forensic medicine and wound ballistics. Wintemute (1987:536) has written that "Concealability, ease of use, and magnitude of force delivered combine to make handguns uniquely lethal," compared to long guns. The statement is nonsense -- particularly with regard to Wintemute's especially hated types of handguns: semi-autos are more complicated to use than revolvers and some types of rifles and shotguns, are less apt to use expanding hollowpoint ammunition (which tends to make more damaging wounds); and so-called "Saturday Night Specials" tend to be lower in caliber and power than larger handguns. And, in general, handguns, at close range, would be third -- after shotguns and rifles -- in the amount of damage to tissue caused by the projected missiles. The more recent target of MD/MPHs looking at firearms are semi-automatic versions of military-style rifles, rifles designed in military calibers to wound rather than kill enemy soldiers, since killing a soldier puts one man out of commission, whereas wounding him puts out of commission all who are needed to remove the victim from the battlefield and care for him. Worse than merely calling for wrongheaded legislation, however, is the serious medical threat posed by the MD/MPHs who misunderstand wound ballistics; doctors who actually treat patients may follow their own and others' rhetorical descriptions as emergency-room advice. Haughton writes of an "increasing number of semiautomatic assault rifles on the street. They cause more -- and more serious -- injuries than...other rifles...(1989:24). Similarly, Claude Cadoux said he was tired of searching through the mangled internal organs of young people looking for bullets pumped into their bodies from semiautomatic assault weapons, and "It's ugly business, and we deal with it every day. It's about bones being shattered and blood oozing out of wounds" (American Medical News, Feb. 23, 1990, p. 13).I. In fact, military-style ammunition -- particularly the AK-47's 7.62x39mm. round -- tend to make wounds similar to medium-bore handgun wounds and less serious than larger rifle ammunition (Fackler et al., 1988; Fackler, 1989). The more serious threat is that misunderstanding wound ballistics may become a self-fulfilling prophesy by emergency-room personnel. If they believe the misinformation, they are apt to treat the wounds as if they did more damage, and actually aggravate injuries by manhandling organs searching for damage which -- aside from the manhandling -- had not occurred (Fackler, 1986; Fackler, 1988).II. Unfortunately, many surgeons prefer the rhetoric of describing emergency-rooms in terms of Viet-Nam, prefer to claim declining death rates due to improved medical care while ignoring increased deaths from medical and surgical misadventures and complications, and modestly underestimate the amount of medical malpractice while seeking more media practice. MEDICINE AND THE MEDIA While theoretically concerned with the huge numbers of deaths involving firearms, most actual medical studies are based on rather small-scale studies, whether involving firearms or any other topic. In theory, medical research in the thousands of poorly regulated medical journals is supposed to be a discussion of preliminary, interim analysis. Doctors are supposed to be noticing something based on some of the persons they have treated, and to publish the findings to learn if others have had similar experiences or if the hypothesis based on the initial small sample can be readily explained away and a tentative theory demolished. Partly for this reason, many studies are very modest in their presentation, with warnings in the fine print about the limited nature of the study. Unfortunately, such reticence is not recognized in the medical journals which publish such studies -- on firearms and other issues -- nor by others in the field; the authors themselves rarely emphasize the interim nature anyplace except the full study. (Full studies on firearms in medical publications are long if they exceed 3-5 pages.) It is hard to take seriously the interim nature of reports if no one so treats the findings. And small-scale findings quickly become all but definitive as the news media cover medicine, and MD/MPHs cover the firearms issue. The worst of the publications -- with the best lines to the general news media -- is the New England Journal of Medicine. Discounted by NIH researchers and academic physicians for its old-boy network and lack of serious controls (confidential communications), NEJM once made a media splash with a new finding on AIDS, based on two subjects. Its recent study -- denouncing oat bran with such coverage as to have millions of dollars worth of impact on the food industry -- was based on studying 20 nutritionists (Washington Post, Jan. 31, 1990, p. E5). The NEJM article studying homicide in Seattle and Vancouver was perceived universally as saying that the Canadian gun law was instrumental in explaining Vancouver's lower homicide rate. Yet, when challenged, since the study did not notice if the situation had changed since the gun law took effect, the authors insisted that "The intent of our article was not to evaluate the effect of the Canadian gun law..." (Sloan et al., 1989:1217). Similarly, when challenged on other articles praising a gun law, followed by editorial praise of both the study and the law, NEJM's editor misleadingly asserted that "Both the article about the law and my editorial on the subject described the results as 'preliminary' and were appropriately cautious in drawing conclusions" (Relman, 1986). It is hard to claim caution when articles and editorials are sent to the nation's news media on the Monday before Thursday publication. The best that can be said for NEJM's approach is that their "Special Articles," which "venture into general topics like economics and social pathology," may be less carefully reviewed than the "Original Articles" on "strictly biomedical subjects." If so, NEJM's using its medical credentials to push "Special Articles" outside its area of expertise, hoping its medical reputation enhances acceptance of its "Special Articles," is similar to Nobel-winning physicist William Shockley hoping his reputation in that field would enhance the acceptability of his views as an amateur geneticist. There is no particular reason to grant that favor. It would be more appropriate to discount -- as likely to have come from NEJM, whether on medical or quasi-medical material -- any study receiving widespread coverage in the news media on Wednesday evening or Thursday. Unfortunately, NEJM and firearms-related topics are not the only ones with such breadth, even when announced as preliminary and not directly related to the topic studied. Most recommendations on cholesterol intake and desirable levels for the young and the old, for the healthy and unhealthy, for men and for women, are based on studies of unhealthy middle-aged males (Consumer Reports, 1990). An evaluation of MD/MPH studies of firearms and violence begins wondering why medical research standards are not brought into play when doctors address the issue, and concludes fearing they have been. Closer to the MD/MPH approach to firearms, a study of 30 survivors of firearm- related suicide attempts, indicating that for some of those the act was impulsive (Peterson et al., 1985) has become the near-definitive support for the view that suicides with firearms are not serious attempts, where the seriousness determined the means chosen, but are impulsive, where success was determined by the chance of firearm availability (Wintemute et al., 1988). In addition to the general proclivity to attract media attention to articles on the firearms issue, NEJM may accompany such articles with editorials which, if anything, go beyond the article. Boyd (1983) noted one study which distinguished handguns from firearms in suicide. As it happens, the study was of an urban county (where guns owned are disproportionately handguns) in the mid-1970s involving a total of 35 subjects. NEJM nonetheless editorialized: "In view of these facts, writes Boyd, perhaps restriction of the availability of handguns would result in fewer Americans' killing themselves. His data are convincing" (Hudgens, 1983). If NEJM thinks the problem is "naive, non-critical interpretation" of studies, its editorials encourage that problem (Hamilton, 1989). Press releases represent an additional way to go well beyond the scope of the small studies MD/MPHs specialize in. When the NCHS and CDC released a small study on firearms-related deaths among 1-19 year olds, its press release emphasized the upper limit noted, calling them children, and misled the public about starting at age one. As a result, the misleading impact was that firearms were involved in 11% of the deaths, rather than the 4% for persons 0-19 (Fingerhut and Kleinman, 1989). In addition, the HHS press release used the study to promote a film on youth violence, even though the film itself had almost nothing to do with firearms, but was instead a study of the way some persons seriously worked to reduce the levels of violence among teenagers in a variety of locales. The greatest misuse of press releases, to go beyond the scope of a study, is done by Wintemute. In releasing a study where firearms safety training is dismissed as a way to reduce deaths because too small a percentage of gun-related deaths are accidental, the Wintemute press release (March 10, 1987, for Wintemute et al., 1987a) called for design changes in guns to make them harder to fire accidentally and easier to determine whether they were loaded. Although the study dealt with firearms as a whole, noting there was little information on the type of firearms, the press release called particularly for bans on concealable firearms, on firearms readily convertible to full-auto fire, and noted that guns in the home were more likely to be used to kill a friend or relative rather than being used for protection -- although the study neither investigated those issues nor cited any investigation by anyone else. A press release for a suicide study (Wintemute et al., 1988) not only went beyond the scope of the study, but appeared, if anything, to contradict the assertions in the study. "And, perhaps the researcher's most controversial opinion: 'People who own guns may be more likely to be self-destructive than those who don't'" (Crime Control Digest, July 4, 1988, p. 7). The study's conclusion -- also without much basis for support, certainly nothing in the study itself -- implied the opposite, that suicide attempts were spontaneous, passing actions made permanent only by the off-chance of firearms ownership; that, absent a firearm, the attempt would have been unsuccessful and not likely repeated. The impulse study cited involved a total of 30 subjects; Wintemute et al. did not replicate or attempt to replicate or refute the results. Some of the media enthusiasm for studies do not involve editorials or press releases going beyond the scope of small studies. Nor do the studies go beyond factual descriptions -- except that the studies, and the media coverage, state data with shock. An MD/MPH survey on teenager access to firearms found the percentages rather high and reported the matter as "startling" (Raleigh News & Observer, Jan. 4, 1990), even though there was no relationship between the findings and any harm. Similarly, a study of firearms and death among children found no shocking trends, but merely presented the data as shocking on their own (Fingerhut and Kleinman, 1989). Occasionally, the percentages are said with shock, although that depends entirely upon the tone of the article. If one looks at accidental deaths, a certain percentage will involve different locations or different age groups; any such result can be made shocking by use of adjectives, even if no trends, correctives, etc., are discussed or evaluated. LIMITATIONS IN THE PUBLIC HEALTH APPROACH Assumptions vs. Factors: Caveat Lector: While most criminologists take into account other factors which might influence what is being measured, the MD/MPH approach generally involves either assuming similarities between compared subjects, or, more commonly, warning readers at the end what has been ignored and what has been assumed. A good way to tell what MD/MPHs think is clearly related to a topic studied - - homicide, suicide, etc. -- is to look at the final caveats, where the reader is warned that the study may not apply any other time or place, or may be unreliable for a number of reasons. Ruben and Leeper (1981:275) announced their study of homicide ignored the place of the killings, motives, the role of alcohol or drugs, the characteristics of the perpetrators, and the education and marital status of the victim. The editorial accompanying Boyd's study of suicide (1983), began by noting the role of "depression and chronic alcoholism" and "schizophrenia and drug addiction" in suicide, before endorsing the Boyd study where those roles were ignored and only firearms were considered (Hudgens, 1983). The Vancouver/Seattle homicide comparison noted that the gun ownership data might not be reliable -- significant for something suggesting a relationship between ownership or availability and homicide rates -- acknowledged that Seattle and Vancouver might be different and thus not comparable, and noted that the Seattle area might not be projectible to the rest of the United States (Sloan et al., 1988). Their suicide study warned that they were ignoring such suicide-related factors as alcoholism, mental illness, and unemployment; noted that the area might not be comparable to the rest of the United States -- especially since gun use in suicide was lower; noted that the suicide data might have been flawed; and again noted that the gun ownership rates between the Seattle and Vancouver metropolitan areas might not have been measured comparably (Sloan, et al., 1990). The MD/MPH approach is to warn but to make projections anyway. So a short-term drop in suicide by gas in one place suggests a projectible long-term drop in gun-related suicides if gun availability is reduced (Boyd, 1983). In fact, the suicide by gas study seems non-projectible even to gas (Clarke and Lester, 1989:ch.2). Warning that Sacramento may not be typical of the nation, Wintemute et al., nonetheless inferred local firearms ownership based on the NORC pacific regional survey data (1988). Some of the assumptions made in lieu of study are unconvincing. A study of gun injuries in San Francisco made three "assumptions....The third is that the cost of hospital care is similar at [San Francisco General Hospital] and at U.S. hospitals as a whole. While these are not perfect assumptions, we feel that they are valid..." (Martin et al., 1988:3050). Most persons -- except, perhaps, MD/MPHs -- are aware that San Francisco is among the most expensive cities in the country. It has the highest per capita income of any major American city, and California's cost per hospital stay is higher than in any other state (U.S. Department of Commerce, 1989:104, 451). In terms of health-care costs for employers, the San Francisco metropolitan area is the nation's highest, about 18% above the U.S. average (Washington Post, Jan. 30, 1990, p. C9). Christoffel (1989) assumed "No one can believe that our Founding Fathers, in crafting the Second Amendment, intended to leave American children as vulnerable to firearm violence as they are today." In fact, the Founding Fathers begot more than a Constitution and a Bill of Rights, penned rights for the protection of adults, and lived in a society where, due to firearms technology, firearms were more apt to be stored loaded, and had fewer safety devices. Constitutionally, the view is that "the government may not reduce the adult population...to...only what is fit for children" (Sable Communications v. FCC, 1989, reaffirming Butler v. Michigan, 352 U.S. 380, 1957). Sloan, et al. (1988) simply assumed Seattle and Vancouver were similar based on such simplistic measures as the rough estimate by police of the clearance rate for homicides, the sentence established by law for unlawful carrying of firearms, and some aggregate economic data. For their follow-up study of suicide (1990), they switched from cities to metropolitan areas, and added to the proof of cultural comparability -- previously demonstrated mostly by proximity -- that six of the nine most popular television shows in one area were among the nine most popular in the other as well. Such comparisons could establish the cultural similarity of some American rural areas with high school musical and theatrical performances to big cities with symphony orchestras and first-run theaters. The CDC, representing the MD/MPH approach, said "The paper by Sloan et al. ...applies scientific methods to examine a focus of contention between advocates and opponents of stricter regulation of firearms, particularly handguns" (Mercy and Houk, 1988). Criminologist Gary Kleck wrote: "As a criminologist writing a book on the role of firearms in violence, I was disappointed to read the recent article on this topic by Dr. John Sloan and his colleagues. In a field already plagued by poor research, this study used a research design so primitive that it would have been regarded as inadequate by criminologists forty years ago, and arrived at conclusions which, even under the most generous standards of methodological adequacy, cannot be legitimately derived from the evidence" (unpublished critique). He told National Public Radio's "All Things Considered" (Dec. 16, 1989): "The research was worthless. There isn't a legitimate gun control expert in the country who regarded it as legitimate research. There were only two cities studied, one Canadian, one U.S. There are literally thousands of differences across cities that could account for violence rates, and these authors just arbitrarily seized on gun levels and gun control levels as being what caused the difference. It's the sort of research that never should have seen the light of day." Since one of the issues concerning MD/MPHs is not just firearms, but guns owned or acquired for protection, it would be interesting to know to what extent the guns whose fatal misuse is studied by public health professionals are owned for that purpose. Again, the tendency is to simply make the assumption, with recommendations made as if the assumption were accurate. "Certainly a majority of children under 15 are killing one another with guns found in the home...that parents brought into the home for protection" (Christoffel in Aurora Beacon News, Oct. 23, 1989). And Kellermann and Reay (1986a:1557) concluded that "The advisability of keeping firearms in the home for protection must be questioned," even though "our files rarely identified why the firearm involved had been kept in the home. We cannot determine, therefore, whether guns kept for protection were more or less hazardous than guns kept for other reasons" (1986a:1559). They assumed protection based on surveys showing that three-fourths cite protection as one reason for having a gun, although the same surveys cite protection as the primary reason only one-quarter of the time; firearms accidents tend to have seasonal fluctuations tied to the hunting season (Morrow and Hudson, 1986); and the only study expressly looking for motivation for gun ownership found protection clearly the reason for the availability of firearms in about 10% of the cases studied (Wintemute et al., 1987b:3108). Another odd assumption of the MD/MPH students of the gun issue is that firearms restrictions are not aimed at human behavior. Although the press release called for reducing the availability of firearms, Wintemute et al. (1987a:374) said they were looking for "preventive strategies beyond those addressing the behavior of persons actually involved in shootings. It may well be that here, as elsewhere in medicine and public health, the most effective measures do not attempt to modify the behavior of those to be protected." Similarly, Boyd (1983) found a "decline in the availability of firearms" by law similar to a "decrease in carbon monoxide content in domestic gas." Factors Forgotten and Figures Fudged: Ironically, the most forgotten factor in MD/MPH studies -- even though the data are frequently presented -- is ethnic variations. It is ironic since one of the espoused goals of the CDC in encouraging MD/MPH studies is to explain, with the goal of reducing, the extremely high rate of black-on-black homicide and the high rate of Hispanic homicide. Instead, the tendency is to ignore these in discussion in favor of focusing on firearms. Sloan et al. (1988) found two cities where the non-Hispanic Caucasian homicide rates were practically identical, the Oriental rates very different (lower for Vancouver's Orientals than for the whites; higher in Seattle), and the black homicide rate extremely high. They ignored the ethnic difference and focused on firearms. (They tried the same approach in the follow- up study of suicide, but when ethnic differences were unable to show differences in firearms use, they switched to age-groups. Sloan et al., 1990.) Similarly, after presenting charts showing tremendous variations by race, regardless of age-group or sex, Ruben and Leeper (1981:276) said, "the methods of analysis used in this study, however, show even more clearly than before the magnitude of the contribution of firearms to these deaths." Other efforts are made to highlight firearms instead of other factors and to make things seem worse than the data would properly show. Boyd (1983) used a low point of suicide as the basis for comparison. (He also added figures incorrectly and used a small-scale and old suicide study as evidence that 83% of firearm-related suicides involve handguns. None of this was caught, apparently, by the peer reviewers, but only by an ordinary reader. Maxwell, 1984). More or less favorable firearms-related death trends are fudged a bit with the explanation that things would be a lot worse but for improved medical care (Fingerhut and Kleinman, 1989; Frierson, 1989). The MD/MPHs like to use only mortality data and then to evade its limitations. Undermining the assumption somewhat is the fact that injury-deaths from medical/surgical complication/misadventure are up, that there are no data showing non-fatal injuries up (except an absurd claim by the American Academy of Pediatrics showing a 300% rise) from emergency-room surveillance, and there appears to be a serious problem with mistreatment of gunshot wounds based on myths of ballistics wounding (Fackler, 1986 and 1988). The pediatricians found an interesting way to increase the number of gun-related deaths and injuries (Christoffel, 1989). They assumed some 3,000 handgun-related suicides by teenagers each year (compared to the 1,300 firearm-related suicides actually reported to the NCHS), and assumed that the 3-5 non-fatal injuries for each fatal injury - - estimated from accidents and assaults -- applied equally to suicide attempts (while saying suicide attempts with a handgun were successful 90% of the time). The suicide figure is obtained by assuming that some deaths were misclassified -- perhaps to spare the parent -- even though there simply are not enough accidents and miscellaneous deaths for misclassification to make much difference -- unless the assumption is that some motor vehicle accidents, say, were really handgun-related suicides. The result is a total of 4,500 deaths -- far more than NCHS reports -- and 13,500-22,500 non-fatal injuries, although the fairly stable figure reported by the Consumer Product Safety Commission's NEISS (National Emergency Injury Surveillance System), counting all injuries to children 14 and under as accidental, was stable at under 4,000 during the late 1970s and early 1980s. At least the American Academy of Pediatrics made an effort to estimate injuries. Most MD/MPH studies count only deaths. The only widely reported study of injuries insisted on fudging a bit both regarding to who pays for injuries and how many there are. The study (Martin et al., 1988) estimated about 62,000 firearms-related injuries requiring ambulance transport or hospitalization, and that over 85% of the costs are not covered by insurance, and thus "taxpayers pay most of the costs associated with firearms injuries." Once it was noted that the 62,000 figure for injuries was very similar to that estimated by criminologist Kleck (Blackman, 1989), the response was that the study "does not include the large number of firearm injuries that do not require hospitalization (for which no estimate is available)" (Martin et al., 1989). Unfortunately, that means the only serious study to look at injuries announces no idea how many there are, how much they cost to treat, or who pays for it. The initial study, suggesting 85% taxpayer cost was untrue, since it included unreimbursed hospital expenses as if they were taxpayer dollars, which is only the case for some hospitals; for others, it simply lowers the profit margin or reallocates other revenues. And, of course, the unestimated but privately treated minor injuries, while raising slightly the cost of injuries, lower similarly the percentage involving tax dollars. Macro to micro to macro: The MD/MPH approach is to claim a public health problem demanding epidemiological study based on the large number of firearm-related deaths and non-fatal injuries (mortality and morbidity). Their studies, however, almost never cover anything but mortality, finding morbidity difficult to measure. This does not necessarily mean that the smaller number of deaths means other problems cannot be cited; MD/MPHs assume large numbers of injuries, which they assume are climbing faster than mortality rates (when those are rising), based on presumed improvements in medical care. Even survey research is dismissed as a source of larger-scale data (Kellermann and Reay, 1986b). Having taken the lower figure of deaths, the studies then get even smaller in sample size -- with no such restriction on the conclusions. Comparing just two cities, Sloan et al. (1988) concluded that "Our analysis of the rates of homicide in these two cities suggests that the modest restriction of citizens' access to firearms (especially handguns) is associated with lower rates of homicide....[A] more restrictive approach to handgun control may decrease national homicide rates." When challenged on the small sample -- two cities, representing less than ¬% of the U.S. population, for example, they responded that they were studying a large sample, some six million person-years (Sloan et al., 1989:1217). One could similarly look at a single kidney and announce one was studying millions and millions of nephrons. Other studies tend even more to jump from small samples to large conclusions. Indeed, with the public health approach justified by combining suicides, homicides, and accidents, to reach about 33,000 deaths annually, most of the studies single out only one of the three elements -- with accidents among the most popular. In addition, while most victims of firearm-related violence are in their late teens or adulthood, studies emphasize children, whose deaths from unnatural causes are relatively few. Interest in the firearms issue thus arises from large numbers, but the portion actually studied account for small percentages for those figures. In addition, the studies are often very small in scale. The figure of 83% of gun- related suicides involving handguns coming from a sample of 35; the assumption that suicide is spur-of-the-moment based on a sample of 30 (Wintemute et al., 1987b; Peterson et al., 1985). And the small-scale studies are made smaller by excluding items which might be covered: Two studies concluding that guns are rarely used for protection limited their studies to shootings in the home (Rushforth et al., 1975; Kellermann and Reay, 1986a). The earlier study also limited the shootings to accidents and justifiable shootings of strangers; at least the later work included other self-defense killings and suicides to broaden the base a little. The emphasis on children raises the emotional issues, but not the scientific ones. Firearms are involved in about 1.2% of the deaths of children 14 and under (NCHS, 1988:Table 1-25), outnumbered by non-gun homicides, pneumonia, bicycle and pedestrian deaths, heart diseases and cancers. To enlarge the tiny studies involving children, the term is broadened to include persons up to the age of 19 -- that is, persons who would be tried as adults should they commit murder, and whose executions have been upheld by the U.S. Supreme Court (Christoffel, 1989).III. Teret and Wintemute (1983) apparently even raised the "children" definition to age 24. Rankings, Rhetoric, and Relativity: Perhaps the single leading justification given for studying the firearms issue and for calling for restrictions on private access to firearms, in the MD/MPH literature, is that firearms are the nth leading "cause" of death, or of injury death and/or of premature death, either nationally, at a state level, or among a particular age and/or ethnic and/or gender group (e.g., Wintemute et al., 1987a; Alexander et al., 1985; Wintemute, 1987; Schetky, 1985). Some of the age groups appear carefully crafted, as with black males 15-34 and black females 15-24 (Wintemute et al., 1987a). Occasionally, matters are complicated by referring to suicide or homicide overall rather than to firearms-related deaths; similarly, sometimes the age groups are left vague -- "young children," "college students" (Schetky, 1985). Rarely, except where the categorization has made firearms rank first or second (either overall or among injury-related deaths) in an age/ethnic/gender group, is it said what "causes" are ahead of firearms. There is nothing, in itself, wrong with taking the four ICD categories and combining them to show the number of firearms-related deaths -- either for rhetorical or scientific purposes. There is something wrong with trying to announce that something is the nth leading cause of death, either relying exclusively upon the ICD or not, particularly when the item is firearms. The reason for singling out firearms here is that firearms-related deaths -- as a total -- are among the most accurate and complete listed by the NCHS based on the ICD categories. Most medical examiners have no difficulty distinguishing a gunshot wound, or wounds, from others, even if they are not sure what type of firearm or what actual cause of death (homicide, suicide, accident) involved a firearm as the means. Besides the four gun-related ICD categories (the three with known causes plus undetermined whether accident, homicide, or suicide), only a few "late effects" deaths -- for which there is no weapons breakdown in the ICD's "external" causes categories -- would be missing. On the other hand, vague causes of death are quite common among cancers, heart disease, respiratory disease, and numerous other ICD categories (Wintemute et al., 1987a; NCHS, 1988:Table 1-25). Any listing of "causes" of deaths which includes any cause approximating firearms -- at about 1«% of the nation's annual deaths -- probably actually exceeds firearms in reality, and fails to in national statistics because of inaccurate and incomplete reporting and/or knowledge on the part of attending physicians and medical examiners. In addition, there is an artificiality to ranking causes of death, especially once four different ICD subcategories involving firearms have been added. At that point, the issue is what to do with remaining batches of subcategories, and how to combine or divide other groupings of categories. In an effort to demonstrate how many serious ailments firearms-related deaths approached, Wintemute separated out breast cancer from other cancers (1987). There is no limit to how categories can be combined or separated.IV. Without pushing the ICD divisions too much, combinations could reasonably put firearms-related deaths for 1986 (NCHS, 1988:Table 1-25) in 19th place behind: acute myocardial infarction, coronary atherosclerosis, acute but ill-defined cardiovascular disease, chronic obstructive pulmonary diseases and the like, unspecified cardiovascular disease, pneumonia, cardiac dysrhythmias, cancer and other ailments of the genital organs, cancer of digestive organs and peritoneum (other than liver, kidney and colon), cancer of the colon, cancers of other and unspecified cites, motor vehicle accidents, other diseases of the anterior arteries and capillaries, lymphatic and hemotopoietic cancer, breast cancer, heart failure, liver disease (including cancer and cirrhosis), diabetes, kidney disease (including cancer). It is likely some causes of death close to firearms would be higher were some of the vague references included; certainly, by 1989, infectious and parasitic diseases (which would now include AIDS, with the ICD categories 042-044), which totalled 31,623 in 1986, would surpass firearms, and put them in 20th place -- all without lowering the number of firearm-related deaths -- from Wintemute's 7th place. Compounding the problem of taking any such rankings seriously is that the categories of the ICD leave out other contributory factors -- generally thought of as causes of death -- which surpass all but heart disease and cancer. One of them, tobacco, may surpass cancer, if the estimates of 200-400,000 annual death include appropriate subtractions from the cancer categories. In addition, alcohol is often credited as a cause of 100-200,000 deaths, but is spread out in unspecified ICD categories, including many of the injury-related deaths. Diet -- especially high fat and low fiber -- has been mentioned as the "cause" of one-third of cancer deaths, which would mean that diet (including ICD category E904, starvation, and the few thousand ICD deaths from vitamin and other nutritional deficiencies and the thousand from obesity) is the "cause" of about 150,000 deaths (NCHS, 1988; Washington Post/Health, Jan. 30, 1990, p. 9). In addition, the Environmental Protection Agency estimates that radon is the second leading cause of deaths from lung cancer, and is blamed for 10- 40,000 deaths -- putting it on a par with handgun-related deaths and possibly with firearm-related deaths as well (Consumer Reports 1989:623). In addition, a new study suggests thousands may die annually as a result of medical malpractice (Washington Post, March 1, 1990, p. A3). Depending upon the accuracy and validity and details of the study, and adding to the already-admitted (and presumed underreported) injury-related deaths from medical/surgical complications/misadventures, and the overdoses of prescription drugs (NCHS, 1988), and the homicides and suicides caused by prescription drugs (see Citizens Commission on Human Rights International, 1989), it is possible that the medical profession is responsible for more deaths than firearms are involved in. Indeed, if Fackler's fears are justified (1986 and 1988), some of the firearms-related deaths might be part of the malpractice-related deaths. With fewer than 2.5 million doctors and nurses (U.S. Dept. of Commerce, 1989:97), it might be easier to attempt to "control" them than to "control" 60-65 million owners of 180-200 firearms. A lack of relativity helps as well to elevate numbers -- and percentages -- rhetorically without any scientific benefit. When Fingerhut and Kleinman (1989) announced that firearms were involved in the deaths of 11% of children -- carefully excluding 0-1 year olds because there were too many deaths (about two-thirds of those occurring prior to age 15) and too few involved firearms -- it was unclear that the actual numbers of person who die during that period is quite low relative to deaths as a whole. When ABC's Peter Jennings, relying on MD/MPH data, announced that firearms were one of the leading causes of death among persons under 65, it was not clear that the age- grouping excluded over 70% of American deaths, and still covered under 5% of American deaths. Figures without comparison can be misleading. When Martin, et al. (1988) calculated that treating gunshot wounds cost almost $500-million, and projected that to $1-billion for uncalculated expenses, they failed to announce that the figure amounted to ¬% of the $391-billion in medical expenditures for the year in question (U.S. Dept. of Commerce, 1989:92).V. That figure also compares to just $10-billion spent on prevention, via screening, etc., to prevent nine chronic and partially preventable diseases which annually account for over half of all American deaths: stroke, diabetes, obstructive lung disease; lung cancer; breast cancer; cervical cancer; colorectal cancer; and cirrhosis (Washington Post/Health, Feb. 13, 1990). Similarly, when Christoffel (1989) complained that one-sixth of pediatricians surveyed had treated a child with a gunshot wound in the previous year, she gave no data on percentage who had treated bicycle, motor vehicle, or other injuries. Such data, without any comparative data, are meaningless, and useful only for rhetorical purposes, not as part of a study by MD/MPHs or social scientists. Worse, perhaps, is when comparative data are given, but misleading and meaningless, as when Schetky (1985) wrote that "The risk of being a victim of a violent crime exceeds that of being affected by divorce, dying of cancer, or being in a traffic accident." The "violent crime" figure -- although not made clear by her -- was based on victimization surveys, where firearms are involved in fewer than 10% of such crimes, and the comparison was to dying from cancer. The statement with regard to traffic accidents is almost certainly untrue. Overall, apples-and-orange comparisons are as bad as no comparative data at all. Risks and Rewards: Public Health's "Social Cost" Theory: One implicit -- and occasionally explicit (Martin et al., 1988) -- MD/MPH basis for determining what should be done regarding firearms and violence is what is called the "social cost" theory, a cost-benefit analysis which serves as the flip-side of the welfare state. The welfare state generously acknowledges that not everyone can afford what they need and so provides the essentials; the flip-side is that since the taxpayer/government is supplying medical care which may be required by individual actions, the government is also entitled to restrict individual action, that an individual's actions involve more than simply the individual who may be injuring himself. At its extreme, the "social cost" theory was practiced in 1984, where Big Brother benevolently forced citizens into exercise programs which, presumably, reduced the subjects' needs for medical care. The same idea has been used to analyze firearms and their possible regulation. Martin et al. noted that most of the medical costs they measured came from tax dollars. The statement was not entirely true, but certainly a majority of the medical services whose actual costs they measured came from tax dollars.VI. The authors then noted: "The findings of the study have important implications for legislators. When considering laws that would restrict the availability of firearms, elected officials must be aware that the issue is not simply one of individual rights, since taxpayers pay most of the costs associated with firearm injuries." The same sort of reasoning is used to justify requirements for motorcycle helmets and motor-vehicle seat belts. Similarly, the cost-benefit analysis has led the government to abandon the idea of requiring seat belts for school buses since relatively few lives would be saved at relatively great expense for the belts, and has led some in the medical profession to oppose massive testing and treatment for high cholesterol because the cost of the tests and the treatment would render but unjustifiably small improvements in the death rate from related heart disease (Consumer Reports, 1990). There are some flaws with the MD/MPH use of "social cost" theory with regard to firearms. First, all firearm-related deaths, and nothing but firearm-related deaths count in the "social cost" side of the ledger evaluating firearms freedom. There are a few exceptions, such as the assumed injuries. And some homicides involving law- enforcement or the military are not counted (Wintemute et al., 1987a). The exclusion of those deaths involving government employees might be justified if attempting to measure the costs of private gun ownership, but there is no indication that accidents to and suicides by law-enforcement officials are excluded from the debit side of the "cost" figure. In addition, all lives are of equal value. The death of a criminal counts as much as that of a law-abiding citizen. This is appropriate for the medical profession, who are bound by historical tradition to treat all patients regardless of their moral culpability. Nonetheless, the loss to society of drug-traffickers or drug-users shot assaulting honest citizens, may not be as great as the loss of a law-abiding taxpayer. Worse, to the extent lives are not equal, the MD/MPH approach is apt to give greater weight to the lives of criminals. This is achieved by measuring age-adjusted death rates, where the younger the dead person, the more life-years are lost (Browning, 1976). In late 1988, the nation was told how an eight-year-old girl dialed "911" to get help when a 19-year-old burglar broke into their house and began stabbing her 36-year-old father. The burglar was shot to death by the 14-year-old brother of the girl. Under age- adjustment, the shooting cost society about 45 years to firearms, while the father's death by stabbing would only have cost about 25 years to knives. Overall, comparing the ages of homicide victims and arrestees for murder and non-negligent manslaughter, a killer's life is worth about 5-10 life-years more than his victim's, on average (Washington Post, Dec. 22, 1988; FBI, 1989). Since those over 65 are excluded for all such calculations, if an 80-year-old woman prevents a 20-year-old from bludgeoning her to death by shooting him to death, the option was between 45 life-years lost of firearms and no effect on the age-adjusted death rate from non-gun homicides (Browning, 1976). If some year, 10,000 otherwise innocent victims of attempted murders were, instead, lawfully to kill their assailants, law enforcement officials might note a sudden drop in the rate of violent and property crimes; MD/MPHs, however, would note an increased premature loss of 50-100,000 life-years. The "social cost" theory leaves out all benefits from the use of firearms -- except occasionally by including on the benefit side the killings of criminals in the home (Kellermann and Reay, 1986a). Other justifiable killings are not counted. Non-fatal uses of firearms are not counted, even if they prevented injuries -- by gun or other implement -- which might otherwise have required medical attention and hence "social costs" (Blackman, 1989; Kleck, 1988). And the benefits to society from the sporting use of firearms, or the psychological benefits of feeling safer, and the like, cannot be measured using the deaths-only approach -- even if supplemented to deaths-plus- medical-costs only. And, in a dynamic society, there is no way to measure either the benefits or costs of imposing the sort of restrictive legislation the MD/MPHs envision. From the point of view of saving lives, after all, prohibition worked in the 1920s: The death rate from cirrhosis dropped about twice as much as the homicide rate rose. Most persons do not think the lives saved were worth the other, medically unmeasured, social costs. The problem is partly that the MD/MPHs studying the firearms issue fail to see benefits which they could easily see in other contexts. Popular publications suggesting that handguns pose "risks" similarly find that two of the riskiest of activities or technologies are surgery and x-rays, ranked fifth and seventh (Colburn, 1986) and seemingly moving up from eighth and ninth places in the 1970s (Dun's Review, Sept. 1979, p. 53). Similarly, MD/MPHs know the benefits of drugs, even if a small percentage of drugs exaggerate rather than diminish the problem for which they are prescribed -- with occasional disastrous results, including mass shootings by Laurie Dann and Joseph Wesbecker (Citizens Commission on Human Rights International, 1989). At worst, the "social cost" theory is a prescription for benevolent medical fascism, with prescribed nutrition, exercise, and other activities, along with proscribed sports, foods, and other items and activities. At best, it is a theory in need of some better tools of measurement, thus far lacking in the literature on firearms and violence. Premature, Pre-Determined, and Preposterous Prescriptions: To a large extent, the MD/MPH studies merely present data on the numbers and/or rates of death involving firearms among particular groups, perhaps over time. They make no effort to find the root causes of the accidents, suicides, and homicides, nor to discuss any factor aside from firearms -- although age, sex, and ethnicity will appear in the tables and accompanying text. And the policy recommendation amounts to little more than "do something," or "let's do more research to support doing something" (e.g., Jagger and Dietz, 1986). Others stick with vaguer proposals, merely calling for legislative action (Martin et al., 1988) or general restrictions. Hence, while acknowledging they cannot "prove a cause and effect" of accessibility and homicide, Ruben and Leeper (1981:276) conclude: "While efforts continue in the search for other elements related to the problem of homicide, there is ample evidence now for the obvious first step in attempting to prevent many of these deaths -- meaningful regulation of unnecessary and irresponsible access to firearms. When?" And "While current firearm fatality levels demonstrate the need for intervention, many prominent strategies...are embroiled in controversy....While public debate over an appropriate legislative response continues, public health involvement ...should not be postponed" (Alexander et al., 1985:167). Some of the proposals leap from firearms data -- with little or no information on handgun involvement -- to calling for restrictions on handguns. "In view of these facts, writes Boyd, perhaps restriction of the availability of handguns would result in fewer Americans' killing themselves. His data are convincing" (Hudgens, 1983). In presenting data on deaths among 1-19 year olds, the Department of Health & Human Services noted the data on all gun-related deaths were, in part, to support the policy "objective to reduce the number of handguns in private ownership" (Fingerhut and Kleinman, 1989:6). And the press release (U.Cal./Davis's Sacramento office, March 10, 1987) accompanying the Wintemute, et al. (1987a), study on firearms deaths of all kinds -- with no breakdown on the type of firearm involved -- called for a ban on handguns. None of these recommendations is either justified by the data given, nor takes into account the issue of whether substitution of long guns for handguns would make the situation worse (Kleck, 1984; Wright and Rossi, 1986: ch. 12). In theory, MD/MPHs, studying objectively, could at least determine which types, calibers and gauges of guns cause the most and least tissue damage. Unfortunately, to date, they have ignored the issue or misunderstood it. Other proposals address specific types of handguns -- curiously emphasizing the goal of making firearms more child-proof while rejecting the idea of firearms safety instruction. Wintemute, et. al. (1987a:377) rejected education because such a small proportion of gun-related deaths were accidents. And Christoffel (1989) said, "There is no evidence that public education alone -- in or out of the schools -- has any impact on gun violence." Yet Wintemute, et al., simultaneously called for more safety features on firearms, such as indicators to tell if the firearm is loaded -- something clearly aimed at accident prevention. And their press release called for a ban on semi-autos which could readily be converted to full auto. The proposal is a bit unclear. If a semi-auto is too readily convertible to full-auto now, the Bureau of Alcohol, Tobacco and Firearms will treat it as a machine gun, with registration -- by May 19, 1986 -- required, or its possession prohibited. To the extent any information is available on firearms used in homicide, suicide, or accidents -- and there was none in the article accompanying the press release -- revolver-caliber handguns predominate over semi-automatic pistols among criminals, who express a preference for ease of use (which would also favor revolvers) (Wright and Rossi, 1986: ch. 8), and revolvers are disproportionately involved in accidents involving children (Wintemute et al., 1987b:3108). Rifles and shotguns of all kinds are rarely used in homicides (FBI, 1989), but there is no information regarding action-type for those used. While there is a great deal of public discussion of semi-automatic firearms, particularly those with modern military styling, there have been no MD/MPH studies on the topic. Nonetheless, Garen Wintemute, who testified on the public-health necessity for a ban on so-called "Saturday Night Specials" in Maryland in 1988, also testified on the public- health need for a ban on military-style semi-automatic rifles in California in 1989. For neither topic are there any MD/MPH data. Going from the alleged massive public health problem involving about 33,000 firearm-related deaths per year, the most commonly repeated policy prescriptions approach accidents -- thus excluding about 97% of gun-related deaths -- particularly those involving children -- thus reducing the goal to about 1%. From then on, some of the proposals address improvements in the guns actually involved in some of those incidents, while others address firearms which are involved in none of those accidents, since they does not exist. Thus, the AMA Council of Scientific Affairs wants to "(5) Encourage the improvement or modification of firearms so as to make them as safe as humanly possible; (6) Encourage nongovernmental organizations to develop and test new, less hazardous designs for firearms;" (1987:16).VII. The AMA would also like to "ban the manufacture and importation of non-metallic, not readily detectable weapons, which also resemble toy guns" (1987:15). And the American Academy of Pediatrics similarly warns against such guns, coming in designer colors to please women but making them more easy to mistake as a toy (Christoffel, 1989). "Within the year, handguns made largely of plastic may be widely available at a relatively low cost. Children are likely to encounter these handguns, which are promoted by their manufacturer as 'dishwasher safe' and by others as 'particularly attractive for women to use as a self-defense weapon.' Because of their composition and light weight, these firearms may resemble toys even more closely than do those now on the market. Before they do, their unique potential for aggravating the problem we have described should be considered" (Wintemute et al., 1987b). Non-metallic handguns do not exist (with the possible exception of disposable firearms in the possession of some of the world governments' secret agencies). The prototype a Florida designer envisions is for a 10-shot, full-auto 40mm. grenade launcher weighing 19 pounds and with the bulk of several shotguns, for use by the American armed forces (Counter-Terrorism, 1987). To the extent polymer or "plastic" technology would allow a handgun to be made without metal, there would be nothing in the MD/MPH literature to show a problem or potential problem with non-detectability; among the matters ignored by MD/MPHs are data on guns taken into secure area for any purpose. The notion that women would seek out designer colors for something to be carried or kept generally concealed is interesting, but, in fact, designer colors do not require plastic guns; such unusually colored firearms have, in the past, been commercial failures. It is unclear what makes MD/MPHs think a woman, wishing a firearm with which to deter a rapist, will seek out a firearm which would look to potential rapists, as much as to children, like a water pistol. Perhaps the most ludicrous of suggestions, from the American Academy of Pediatrics (press release, Sept. 7, 1989), was that firearms be redesigned to have a brief delay in firing. Such guns would certainly make trap and skeet more challenging, and would likely increase the percentage of game animals wounded rather than killed, but any benefit is hard to contemplate. With some guns designed to fire instantly (law enforcement would demand it) and others designed for delay, the reaction of shooter and potential victim would, at this time, be equally difficult to contemplate. Officers facing a firearm with a delay expected, could be expected -- on pain of lawsuits or punishment for unnecessary discharge of their firearms -- to attempt literally to dodge bullets. One improper response to a firearm failing to discharge a bullet when the trigger is pulled is to look at the gun, which might involve pointing it at someone where it was initially pointed at a criminal or target. For protection, such a delay could be dangerous. The counterbalancing benefits were not spelled out by the AAP. In sum, aside from the unassailable assertion that an unacceptably large number of firearms-related deaths requires study with the desired result of improved policies aimed at reducing the figure, the MD/MPH studies have contributed nothing to intelligent policy making on the firearms issue, or to the issues of suicide and homicide. To the extent the studies have done nothing but calculate homicide rates among various groups in the community, generally with small samples, with explanatory factors generally ignored, there is little reason to expect improvement in the future. PUBLIC HEALTH CONTRIBUTIONS TO THE STUDY OF FIREARMS AND VIOLENCE It is difficult to criticize the methodology of MD/MPHs studying the issue of firearms and violence, since, by an large, there is none. The studies simply prepare charts on gun-related, perhaps compared with non-gun-related, deaths of various kinds and among various subgroups in a population. There are occasionally breakdowns involving more detailed circumstances, but those are largely limited to either a particular type of gun-related death (accidents, e.g., in Morrow and Hudson, 1986) or a particular location (the home the gun belonged in, Kellermann and Reay, 1986a). Trends over time may be presented, with the suggestion -- but no evidence -- that firearms availability was increasing. There may be the assertion that a gun law was working or would work, but generally with no clear understanding of the state of existing law. Nonetheless, although largely missed by the MD/MPHs studying them, the studies' data have produced some interesting items worthy of further study. And the nature of some of the studies -- small-scale, involving the opportunity for detailed investigations of particular firearms-related incidents -- suggests an area for further work. The studies indicate that accidents are a diminishing problem, accounting for 12% of firearms- related deaths since the 1920s, but only of 3-5% recently, with actual rates of accidental deaths down dramatically (Wintemute, 1987). Although it is generally asserted that firearms are owned for protection, with tests of risks presupposing such a motivation, studies of accidents indicate a greater tie between hunting seasons and accident rates (Morrow and Hudson, 1986). More detailed studies on the nature of accidents, particularly in the home, might help to verify this and test the notion whether it is protective or sporting arms which are more subject to accidental abuse, as well as the involvement of alcohol and drugs in such accidents. A small-scale study of women who survived suicide attempts with a gun -- albeit with twice as many subjects as one deemed near-definitive for the theory that firearms are chosen on impulse rather than because of a serious determination to kill oneself -- found that women who used firearms in suicide attempts were less apt to succeed than were men, lending some small support to the theory that women make suicide gestures more than men -- even when they choose a firearm for the attempt. And, although in some ways inconsistent with the previous statement, women using firearms unsuccessfully "have a closer epidemiological resemblance to completed suicides...than attempted suicides...." Such a conclusion supports the notion that the motivation to commit suicide led to the weapon being chosen rather than simply the chance of selecting a gun (Frierson, 1989). Clearly, more research is needed on that topic, and, since it requires fairly detailed investigation of small numbers of suicides, it may be the sort of research MD/MPHs could perform. MD/MPHs, because of their close ties to medical examiners and coroners, may be in better positions than criminologists to study the details regarding large numbers of homicides. Even though an anti-gun study, for example, Rushforth et al. (1977:533) found that about 10% of homicides were self-defense shootings by civilians, and that civilian self-defense shootings outnumbered those by law enforcement officers by a margin of three to one (see Kleck, 1988). Similarly, their ties and interest could provide the impetus for more detailed identification of the types of firearms actually involved in mortal injuries. In addition, MD/MPHs could encourage more efforts to include homicides and suicides in the "diseases" for which cross-tabulations are sometimes provided. The medical profession makes some effort to determine which persons who die of one disease -- say, strokes -- have been treated for various heart diseases, cancers, or the like. Such data can provide clues to interrelationships of ailments. In the same way, MD/MPHs could make some effort to determine which deaths whose proximate causes are homicide (including mercy killing) or suicide are related to -- and more properly seen as deaths caused by -- cancer, heart disease, AIDS, depression, Alzheimer's disease, and the like. Most significantly, were the MD/MPHs properly to use their epidemiological approach, the trends toward more firearm-related deaths involving persons under the age of 25, in both homicides and suicides, and the extremely high rates of homicide among black youths, would suggest areas where more extensive and realistic efforts should be made to determine the nature of such deaths and possible preventive measures. Of course, a realistic epidemiological effort would have to take into account the fact that firearms ownership, being higher among whites and less restrictive among persons over the age of 21, some other explanatory factors than firearms would have been sought. MD/MPHs may be in a position to study some possibilities. Past performance would not make one hopeful of sudden improvements in the criminological research skills of MD/MPHs. Although most of their work has been done while criminological work has been around as a guide, it has been largely ignored, distorted, or misunderstood. The only effort by MD/MPHs to interact with the attempt to learn from criminologists have been spurred by the CDC, whose efforts have not yet been thoroughly evaluated, but who have sponsored two of the worst efforts by MD/MPHs to study firearms and violence (Sloan, et al., 1988 and 1990). That failure would diminish hopes for improvement any time soon, but in an age when one cannot even "trust a communist to be a communist," who knows what skills and qualities MD/MPHs may acquire? REFERENCES ALEXANDER, G.R., et al. (1985) "Firearm-related Fatalities: An Epidemiological Assessment of Violent Death." American Journal of Public Health 75 (February): 165-168. ALTMAN, L.K. (1987) "Scandals Point to Weakness in Review Process." New York Times, July 21. BLACKMAN, P.H. (1989) "Correspondence." JAMA (Journal of the American Medical Association) 261 (May 12): 2637-2638. BOORSTIN, D.J. (1983) The Discoverers. New York: Random House. BOYD, J.H. (1983) "The Increasing Rate of Suicide by Firearms." NEJM 308 (April 14): 872-874. BROWNING, C.H. (1976) "Handguns and Homicide: A Public Health Problem." JAMA 236 (Nov. 8): 2198-2200. CDC (CENTER FOR DISEASE CONTROL) (1986) Homicide Surveillance: High Risk Racial and Ethnic Groups -- Blacks and Hispanics, 1970 to 1983. Atlanta. CHRISTOFFEL, K.K. (1989) Testimony before the House Select Committee on Children, Youth and Families on "Children and Guns." American Academy of Pediatrics (June 15). CITIZENS COMMISSION ON HUMAN RIGHTS INTERNATIONAL (1989) A Summary of the Psychiatric History of Selected Murderers and Mass Murderers. CLARKE, R.V. and D. LESTER (1989) Suicide: Closing the Exits. New York: Springer- Verlag. COLBURN, D. (1986) "You Bet Your Life: Weighing the Risks in an Age of Uncertainty." Washington Post/Health (May 21): 13-15. CONSUMER REPORTS (1989) "Radon: The Problem No One Wants to Face." Consumer Reports 54(October): 623-25. CONSUMER REPORTS (1990) "Forget About Cholesterol?" Consumer Reports 55(March): 152-57. COUNTER-TERRORISM (1987) "Analysis: Plastic Weapons and the Threat" and "Special Report: Plastic Weapons Development." Counter-Terrorism 2(Nov. 2): 4-6. FACKLER, M.L. (1986) "Ballistic Injury." Annals of Emergency Medicine 15(Dec. 12): 1451-1455. FACKLER, M.L. et al. (1988) "The Wound Profile: Illustration of the Missle-tissue Interaction." Journal of Trauma 28 Suppl. (January): S21-S29. FACKLER, M.L. (1988) "Wound Ballistics: A Review of Common Misconceptions." JAMA 259 (May 13): 2730-2736. FACKLER, M.L. (1989) "Wounding Pattern of Military Rifle Bullets." International Defense Review 1/1989:59-64. FBI (1989) Crime in the United States, 1988. Washington, D.C.: GPO. FINGERHUT, L.A. and J.C. KLEINMAN (1989) Firearm Mortality Among Children and Youth. Advance Data #178. NCHS (Nov. 3). FRIERSON, R.L. (1989) "Women Who Shoot Themselves." Hospital and Community Psychiatry 40(August): 841-843. HAMILTON, J. (1989) "Medical Journals Can Make You Sick." Washington Post/Health (Oct. 31): 7. HAUGHTON, J.G. (1989) "Doctors Should Be Fighting to Ban Guns." Medical Economics (Aug. 21): 24, 27, 30. HINES, W. and J. RANDAL (1989) "Peer Reviewers Debate Policing Medical Journals." Washington Post/Health (May 16): 11. HUDGENS, R.W. (1983) "Editorial: Preventing Suicide." NEJM 308: 897-898. JAGGER, J. and P. E. DIETZ (1986) "Death and Injury by Firearms: Who Cares?" JAMA 255 (June 13): 3143-3144. JUNG, R.S. and L.A. JASON (1988) "Firearms Violence and the Effects of Gun Control Legislation." American Journal of Community Psychology 16: 515-525. KELLERMANN, A.L. and D.T. REAY (1986a) "Protection or Peril?: An Analysis of Firearm-Related Deaths in the Home." NEJM (June 12): 1557-1560. KLECK, G. (1984) "Handgun-Only Control: A Policy Disaster in the Making," pp. 167- 199, in D.B. Kates, Jr. (ed.), Firearms and Violence: Issue of Public Policy. Cambridge, Mass.: Ballinger. KLECK, G. (1987) "Letters to the Editor." American Journal of Public Health 77 (April): 513. KLECK, G. (1988) "Crime Control Through the Private Use of Armed Force." Social Problems 35(February): 1-21. LIZOTTE, A.J. (1981) Letter to the Hon. Phil Crane, U.S. House of Representatives, Dec. 11. MARTIN, M.J. et al. (1988) "The Cost of Hospitalization for Firearm Injuries." JAMA 360(Nov. 25): 3048-3050. MARTIN, M.J. et al. (1989) "Correspondence." JAMA 261(May 12): 2638. MAXWELL, S.L., Jr. (1984) "Correspondence." NEJM 310(Jan. 5): 46-47. MAYO CLINIC (1988) "Guns in the Home: Statistics Shoot Down Theory that Guns are Best Way to Protect Yourself." Mayo Clinic Health Letter 6(Nov.): 6. MERCY, J.A. and V.N. HOUK (1988) "Firearm Injuries: A Call for Science." NEJM 319(Nov. 10): 1283-1284. MORROW, P.L. and P. HUDSON (1986) "Accidental Firearms Fatalities in North Carolina, 1976-1980." Am. J. Public Health 76(Sept.): 1120-1123. NCHS (NATIONAL CENTER FOR HEALTH STATISTICS) (1988) Vital Statistics of the United States, 1986. Vol. II: Mortality, Part A. Hyattsville, Md.: U.S. Dept. of Health and Human Services. NATIONAL SAFETY COUNCIL (1989) "Firearms Safety Fact Sheet," for 1989 Congress and Exposition, Chicago. PETERSON, L.G. et al. (1985) "Self-Inflicted Gunshot Wounds: Lethality of Method versus Intent." Am. J. Psychiatry 142: 228-231. RELMAN, A.S. (1986) "Correspondence." NEJM 315(Dec. 4): 1484-1485. RUBEN, E.R. and J.D. LEEPER (1981) "Homicide in Five Southern States: A Firearm Phenomenon." Southern Med. Journal 74(March): 272-277. RUSHFORTH, N.R. et al. (1975) "Accidental Firearm Fatalities in a Metropolitan County." Am. J. of Epidemiology 100: 499-505. RUSHFORTH, N.R. et al. (1977) "Violent Death in a Metropolitan County." NEJM 297(Sept. 8): 531-38. SCHETKY, D.C. (1985) "Children and Handguns: A Public Health Concern." Am. J. on Diseases of Children 139(March): 229-231. SLOAN, J.H. et al. (1988) "Handgun Regulations, Crime, Assaults, and Homicide: A Tale of Two Cities." NEJM 319(Nov. 10): 1256-1262. SLOAN, J.H. et al. (1990) "Firearm Regulations and Rates of Suicide: A Comparison of Two Metropolitan Areas." NEJM 322(Feb. 8): 369-373. STOLINSKY, D.C. (1984) "Correspondence." NEJM 310(Jan. 5): 47. TERET, S.P. and G.J. WINTEMUTE (1983) "Handgun Injuries: The Epidemiologic Evidence of Assessing Legal Responsibility." Hamline Law Rev. 6(July): 341-350. U.S. DEPT. OF COMMERCE (1989) Statistical Abstract of the United States. Washington, D.C.: GPO. WINTEMUTE, G.J. (1987) "Firearms as a Cause of Death in the United States." J. of Trauma 27: 532-536. WINTEMUTE, G.J. et al. (1987a) "The Epidemiology of Firearm Death Among Residents of California." Western J. of Medicine 146(March): 374-377. WINTEMUTE, G.J. et al. (1987b) "When Children Shoot Children: 88 Unintended Deaths in California." JAMA 257(June 12): 3107-3109. WINTEMUTE, G.J. et al. (1988) "The Choice of Weapons in Suicide." Am. J. of Public Health 78(July): 824-826. WRIGHT, J.D. and P.H. ROSSI (1986) Armed and Considered Dangerous: A Survey of Felons and Their Firearms. New York: Aldine de Gruyter. _ 1It is odd for physicians and surgeons to think they can tell the action type of a firearm used to inflict injury based on the ammunition recovered or the wound inflicted. Even supposing treatment to be so slow-paced that ammunition is carefully examined, the rounds normally used in semi-autos may also be used in non-semi-auto firearms. It also presumes firearms expertise to be even greater where in fact is it less. Recent data tapes on firearms traced by the Bureau of Alcohol, Tobacco and Firearms -- many initiated by that agency empowered to enforce the nation's firearms laws and offer technical advice on such matters -- identify .30-'06 "shotguns," .25 caliber "revolvers," and 12-gauge "rifles." While most any firearm can be made in a "wildcat" caliber, the greater likelihood is that someone holding a firearm in his hands was incapable of accurately reading the label. It is unlikely that MD/MPHs do better looking at wounds or spent bullet fragments. 2A Washington, D.C. advertising tabloid, The City Paper, Feb. 2, 1990, covering the work of two emergency room physicians, graphically described poking around and lifting internal organ after internal organ in the vain search for serious tissue damage in a clean bullet wound -- along with an accompanying article with a Fackler warning about the dangers of assuming there must be more damage, and manhandling organs searching for it. _ 3An odd result of the definition spurred the Alliance for Survival, Los Angeles, to ask children to trade their toy guns for teddy bears, and to ask further for the children to contribute essays on why they prefer to play with teddy bears more than guns. The effort was in response to the shooting by police of a 19-year-old playing with a Lazer Tag gun. It would be interesting to read the essays of the 19-year-old "boys" explaining their preference for teddy bears (Family Circle Magazine, Nov. 28, 1989). _ 4This was recently demonstrated when Ann Landers ranked firearms fourth as a cause of death in America (March 7, 1990). To achieve such a ranking, one must first subcategorize motor vehicle accidents as the ICD allows (striking other vehicle, striking bicycle, striking pedestrian, striking other objects, etc.). Then pneumonia and chronic pulmonary diseases (each with about 2« times as many deaths annually as firearms) can be added together with all other respiratory ailments. All she needs to do then to achieve her goal is to categorize strokes and other cerebrovascular and circulatory diseases as heart disease, and determine whether diabetes is more like cancer, heart disease, or respiratory disease, and firearms, involved in 1«% of the nation's deaths, rank fourth. It might be easier just to subcategorize motor vehicles and then rank firearms second, to death by natural causes (ICD categories 001- 799). _ 5An estimate for the costs of medical care -- plus work -- from alcohol-related injuries is approximately $80 billion (Washington Post/Health, Jan. 30, 1990, p. 19). _ 6Part of the problem is that Martin et al. actually measured only about half of what they projected medical costs to be, something under $500 million, of which about two-thirds came from tax dollars. They then projected another $500 million from unmeasured expenses, assuming (with or without justification) that the figure held. Since they later (Martin et. al., 1989) announced there were other firearm injuries besides those measured, of unknown quantity and unknown cost, the taxpayer share is also unknown. _ 7The Council's 16 pages of text end with a semi-colon.