There is widespread scepticism concerning the reliability and validity of psychoanalytic judgements of patient–therapist transactions. We predicted that (a) in reviewing the initial part of 14 videotaped assessment interviews with borderline and dysthymic subjects, dynamic psychotherapists would agree in their ratings of psychoanalytically relevant characteristics of subjects' interpersonal relations; (b) inter-correlations among the ratings would conform with those expected from psychoanalytic descriptions of ‘paranoid-schizoid’ and ‘depressive position’ states of mind; and (c) these ratings would differentiate between borderline and dysthymic groups.

Six trained psychotherapists who were blind to the design of the study, independently rated qualities of interpersonal relatedness during the first 30 minutes of each interview, on & 30-item ‘personal relatedness profile’

There was satisfactory interrater reliability in judgements among the raters, and evidence that the items were interrelated. There was also & significant difference between the two subject groups.

It is possible to make reliable psychoanalytic judgements about qualities of interpersonal relatedness. Moreover, there is evidence that paranoid–schizoid and depressive positive aspects of psychological functioning do constitute & meaningful constellation of clinically grounded phenomena.

We investigated whether family history had prognostic significance in depression in a study which addressed some of the methodological shortcomings of previous studies.

We collected family history data on a consecutive series of 89 patients admitted with RDC major depression, blind to the outcome of the proband. This comprised 116, 283 and 120 first-degree relatives examined with the SADS–L, FH–RDC and case note data, respectively. The outcome of 74 of these probands (83%), previously categorised into four operationally defined groups, was then examined.

A positive family history of severe psychiatric illness (i.e. a relative with a history of either a psychosis, hospitalised depression or suicide) was associated with poor outcome in the proband. This association persisted after controlling for variable family size, age structure and gender. As family history was correlated with neither Kendell's neurotic/psychotic index nor the probands neuroticism score, an individual with high scores an all three would have a greatly increased chance of having a poor outcome.

A family history of severe psychiatric illness in a first-degree relative may be useful as one of the vulnerability factors for predicting poor long-term outcome in depression.

This study investigated three questions with major implications for suicide prevention: the sensitivity of the history of previous suicide attempt(s) as an indicator of suicide risk, the time interval from a preceding suicide attempt to the fatal one, and switching of suicide methods by those eventually completing suicide.

The lifetime history of suicide attempts and the methods the victims (n=1397) used were examined in a nationwide psychological autopsy study comprising all suicides in Finland within a 12-month research period in 1987–1988.

Overall, 56% of suicide victims were found to have died at their first suicide attempt, more males (62%) than females (38%). In 19% of males and 39% of females the victim had made a non-fatal attempt during the final year. Of the victims with previous attempts, 82% had used at least two different methods in their suicide attempts (the fatal included).

Most male and a substantial proportion of female suicides die in their first suicide attempt, a fact that necessitates early recognition of suicide risk, particularly among males. Recognition of periods of high suicide risk on the grounds of recent non-fatal suicide attempts is likely to be important for suicide prevention among females. Subjects completing suicide commonly switch from one suicide method to another, a finding that weakens but does not negate the credibility of restrictions on the availability of lethal methods as a preventive measure.