An Informative Failure to Replicate—the Type D is Distressed: a Comment on Meyer et al.

The Meyer et al. [1] article significantly adds to the list of recent failures to find that the Type D (“distressed”) personality—a composite of high negative affectivity and high social inhibition—predicts future cardiac events. After a string of positive studies, mainly from the Denollet research group [2], in which Type D was associated with cardiac deaths and other events, null findings have recently been reported. This is not the first commentary about Type D; Smith [3] and Coyne and de Voogd [4] preceded me, and their trenchant observations about previous Type D studies no doubt influenced the way Meyer and colleagues designed the study and analyzed their data. Both commentaries criticized the practice of identifying Type Ds as those who score above the median on continuous measures of both negative affectivity (NA) and social inhibition (SI). The frequency of cardiac deaths or other adverse events of “Hi-His” was then statistically compared with the other three quadrants (created by median splits)—“Hi-Low,” “Low-High,” and “Low-Low” groups—combined. These procedures assume that there are “personality types,” despite evidence from the individual differences literature suggesting that both of these trait dimensions are normally distributed. Further, splitting subjects into dichotomous categories based on continuous ratings reduces statistical power and precision. Finally, Smith [3] observed that the three versus one contrast that has served as past proof for the toxicity of Type D could be because of an additive effect of either NA or SI alone or two additive first-order effects without an interaction (see ref. [4] for other statistical possibilities). Meyer et al.'s study is notable because it uses both Denollet's analytic strategy and more appropriate procedures—maintaining the continuous nature of the NA and SI ratings and entering their main effects and the interaction. Neither approach found Type D predicted adverse events, whereas traditional risk factors, such as age and smoking, did. Also, continuous scores for SI were positively related to survival (!), whereas NA showed a trend toward more adverse events. These opposite trends for NA and SI may help to explain why Type D has not fared well in recent studies. As a construct, Type D does not appear to cohere in structure or outcomes. Why should people who experience elevated distress and are inhibited about expressing it be distinctively vulnerable. In contrast, evidence from psychosomatic medicine suggests persons who are hostile and display their antagonism are at elevated risk of cardiac disease [5]. The search for psychological trait predictors of illness will continue. Constructs such as conscientiousness [6, 7] and depression [8], which have been more fully informed by theory and assessment of individual differences and clinical psychopathology, have proven to be more robust and reliable than Type D in predicting health outcomes. Constructs cobbled together without input from these venerable areas seem doomed to the fate of the Type A coronary-prone personality.