SynopsisPsychological factors have long been thought to play a contributing role in either the predisposition, onset or course of various physical illnesses. Recently, rapid advances in immunology have created interest in the interaction between psychosocial factors, behaviour and the immune system. This paper reviews some of the models proposed to explain the relationship between psychological variables and physical illness and presents evidence for a contribution of psychological factors to certain illnesses in which abnormalities in immunologic state are thought to be important. From a somewhat different perspective, animal studies have demonstrated complex effects of stress, on disease susceptibility. Recent human studies have demonstrated consistent immunologic changes in people undergoing acute naturally occurring psychological stress such as bereavement or an important examination. In humans, the effects of chronic stress may be different from acute stress, corresponding to the findings in animals. Abnormalities in immunologic functioning and physical illness are reviewed for different psychiatric disorders — depression, anorexia nervosa and schizophrenia; depression is the only disorder which consistently demonstrated immunologic changes. Possible mechanisms for the stress/immune-change relationship are suggested.

There were 22 bulimic women who participated in a double‐blind placebo‐controlled study of the effects' of imipramine hydrochloride in the treatment of bulimia over a 16 week period. Particpants receiving the active drug demonstrated a significantly greater reduction in purging (frequency of self‐induced vomiting plus the use of laxatives) at both the 6 and 16 week assessment periods. Depression was reduced to a significantly greater extent in those receiving the active drug at 6 weeks but not at the 16 week assessment. These findings, from the longest duration medication study in this condition, suggest that imipramine is an effective treatment for bulimia. However, only one‐third of the participants receiving imipramine had stopped purging by the end of the study. Thus, clinicians may need to add other approaches, such as cognitive‐behavioral treatment, to the management of this condition.

In this study, the predictive capacity of the Minnesota Multiphasic Personality Inventory-2 Posttraumatic Stress Disorder-Keane (MMPI-2 PK) scale was examined in a sample of trauma victims who experienced a serious workplace-related accident and subsequent injury. In keeping with a number of previous investigations, the PK scale was largely ineffective in identifying posttraumatic stress disorder (PTSD) beyond overall symptom and functional severity. In contrast, sets of clinical and content scales proved to be significant predictors of PTSD. These findings suggest that the PK scale is not particularly useful in detecting PTSD in civilian trauma samples. Clinicians might be best advised to use the MMPI-2 clinical and content scales in their assessment of PTSD in civilian patients presenting with a history of trauma.

P sy chia trists are known to have high rates of men tal dis orders (1-4), and a higher fre quency of both emo tional disor ders and sui cide has been re ported more among prac tis ing psy chia trists than among other medi cal prac ti tio ners (5,6). One ex pla na tion for these high rates is that medi cal stu dents who are at tracted to psy chia try may be prone to men tal ill ness early in life. Thus, it is this vul ner abil ity that con trib utes to their se lect ing psy chia try as a ca reer, thereby ac count ing for the high rates of emo tional dis tur bance. For ex am ple, Wal ton (7) re ported that high lev els of neu roti cism were as so ci ated with a posi tive at ti tude to psy chia try as a ca reer choice among medi cal stu dents, al though it is an em piri cally es tab lished vul ner abil ity fac tor for emo tional dis tur bance (8). An other ex pla na tion is that the spe cific stres sors as so ci ated with psychi at ric prac tice, com bined with per son al ity vul ner abil ity, lead to emo tional dys func tion.Evi dence sup ports this lat ter in ter pre ta tion. In a rela tively recent study, Deary and oth ers (3) com pared a sam ple of randomly se lected con sult ing psy chia trists (n = 39), who worked within the Na tional Health Serv ice in Scot land, with a group of 149 phy si cians and sur geons. Sev eral vari ables re lated to the stress pro cess, in clud ing per son al ity traits, cop ing strategies, psy cho logi cal dis tress, burn out, job stress, and work demand. Com pared with the other phy si cians and sur geons, psy chia trists re ported fewer clini cal work de mands and, as a group, did not re port more work-related stress than did the other phy si cians and sur geons. There were, how ever, sig nificant dif fer ences in the per son al ity char ac ter is tics, with psychia trists scor ing sig nifi cantly higher than the other phy si cians on the per son al ity di men sions of neu roti cism, openness-to-experience, and agree able ness and lower in consci en tious ness. Given that work de mands were less and reported work-related stress was lower, psy chia trists re ported higher work-related emo tional ex haus tion and de pres sion. The Cen tre for Ad dic tion and Men tal Health and De part ment of Psy chia try, Uni ver sity of To ronto Ad dress for cor re spon dence: Dr P Garfinkel, Presi dent and CEO, Cen tre for Ad dic tion and Men tal Health,