Commentary: Questionable premises, overadjustment, and a smoking/suicide association in younger adult men

The analysis of smoking/suicide associations in this edition of the International Journal of Epidemiology risks deceiving rather than informing readers. 1 The paper’s calculations and conclusions are highlighted, yet their basic assumptions seem incompletely disclosed, and likely false. The paper violates normal public health precautionary principles like full disclosure of contrary data, supporting stringent drug safety evaluations, and judging people innocent until proven guilty. The paper fails to highlight that smoking precedes and probably causes multiple suicide precursors that the researchers adjusted for, as discussed below. Thus the paper’s major premises are likely to be false and its adjustments misleading. Based on probably false premises, the paper judges smoking, a known drug addiction, stressor, and debilitator, likely to be innocent of all roles in suicide. With little evidence, this and a few prior papers judge smokers guilty of vague suicide ‘predispositions’. 2 As the paper suggests, unless a suicide risk factor ‘very likely preceded smoking,’ smoking may cause the risk factor and ‘it would be inappropriate to control for’ risk factors ‘“on the causal pathway” between smoking and suicide….’ 1 Such inappropriate and often misleading adjustments are called ‘overadjustment’. Overadjustment in the paper is very likely if smoking contributes to any of the many suicide risk factors that the authors adjusted for. As shown below, it is likely that the authors made multiple overadjustments since smoking probably plays a role in multiple suicide risk factors that the authors adjusted for. The paper’s thrust relies wholly on unsupported, probably incorrect, and sometimes less than fully disclosed assertions that smoking follows and produces no: loss of control, stress, psychopathology, dysfunction, destitution, debility, divorce, delinquency, or related suicide precursors. The facts show that smoking precedes and probably causes all of those possible suicide precursors. 3‐5 In particular, randomized trial, challenge-rechallenge, animal, cohort, and other studies show that smoking precedes and probably causes a good deal of psychopathology, 3,4 substance abuse, suicideprone biochemical changes, and losses both via second-hand smoke and the social exclusions addicts face. 6 Also, those stopping smoking had reduced suicides 7 and reduced suicide precursors compared with continuing smokers and prior measures in the same individuals, respectively. 8,9