Kellermann AL, Rivara FP, Somes G, et al. Suicide in the Home in Relationship to Gun Ownership. N Engl J Med. 1992; 327: 467-72.
methodological and conceptual errors:
- an "adjustment" to eliminate suicide outside the home for the stated purpose of exaggerating the focus on guns
- ignored the vast body of data on suicide method substitution
- the authors virtually ignored their own data showing that factors, such as psychotropic medications, drug abuse, living alone, and hospitalization for alcoholism, have much higher correlations with suicide than guns
- failed to address the important social and ethical dilemma - how to reduce overall suicide rates
- ignored the role of failing health in the suicide of the elderly
In another effort to prove that guns in the home are a significant risk, Kellermann and his co-authors purported to examine certain correlates of suicide.  Though the authors' own data showed higher correlations between suicide and psychotropic medications, drug abuse, living alone, and hospitalization for alcoholism, the article focused on guns. [See Graph 9: -- "Kellermann's Suicide Odds Ratios"]
The authors' "adjustment" -- their word -- that eliminated the 30% of suicides outside the victim's home intentionally skewed the data towards their foregone conclusion. The authors candidly acknowledged their bias -- "Our study was restricted to suicides in the victim's home because a previous study has indicated that most suicides committed with guns occur there..." [emphasis added].
As Kleck's review  of the broad expanse of American and cross-cultural suicide literature shows, even if guns instantly evaporated from the US, universal access to nearly equally effective and accessible means of suicide -- hanging, auto exhaust, drowning, and leaping -- would likely interfere with an overall reduction in suicide. Evidence of such "method substitution" is extensive. Many cultures that have severe gun restrictions -- Japan, China, USSR, Germany, Luxembourg, Denmark, Belgium, Surinam, Trinidad, Tobago, Hungary, Rumania, Czechoslovakia, Yugoslavia, Finland, and Sweden -- have total suicide rates far exceeding the USA suicide rate. Many others -- Canada, Iceland, Bulgaria, Norway, and Australia -- exceed the USA suicide rate though not quite so dramatically.  [See Graph 10: "International Suicide Rates Comparisons"]
Guns are often portrayed as uniquely lethal as tools of suicide, yet, amongst tools of suicide, guns are neither uniquely available, uniquely lethal, nor causal of suicide.  [See Graph 11: "Suicide Method Lethality"] The authors' preoccupation with guns bypasses the real social dilemma, reducing the total suicide rate. Changing merely the method of death is an inadequate response to a grave social problem. Is suicide from hanging or auto exhaust so much more "politically correct" that research, particularly in these times of financial austerity, should focus on one instrumentality rather than on the common roots and prevention strategies?