Background. Chronic alcohol abuse is associated with a wide range of cognitive deficits. However, little is known about memory for real-life events (autobiographical memory) in non-amnesic alcoholic patients. The purpose of this study was to investigate (a) non-amnesic alcoholics' ability to recall specific autobiographical memories and (b) their subjective experience when they access specific memories.

Method. Twenty non-amnesic (without Korsakoff syndrome) recently detoxified alcoholics and 20 healthy controls completed the Autobiographical Memory Test (AMT), which assesses the frequency of specific (versus general) memories recalled in response to cue words, and the Memory Characteristics Questionnaire (MCQ), which assesses subjective experience (e.g. the amount of sensory and contextual details experienced) when remembering specific events.

Results. Alcoholic patients recalled specific memories less frequently and general memories more frequently than healthy controls. Nevertheless, when a specific past event was accessed, alcoholic patients subjectively experienced as many sensory and contextual details as controls.

Conclusions. These findings suggest that non-amnesic alcoholics have difficulties strategically accessing event-specific autobiographical knowledge, which might result from changes in frontal lobe function that are associated with alcoholism.

Objective. The substantial literature examining social reasoning in people with delusions has, to date, neglected the commonest form of decision making in daily life. We address this imbalance by reporting here the findings of the first study to explore heuristic reasoning in people with persecutory delusions.

Method. People with active or remitted paranoid delusions, depressed and healthy adults performed two novel heuristic reasoning tasks that varied in emotional valence.

Results. The findings indicated that people with persecutory delusions displayed biases during heuristic reasoning that were most obvious when reasoning about threatening and positive material. Clear similarities existed between the currently paranoid group and the depressed group in terms of their reasoning about the likelihood of events happening to them, with both groups tending to believe that pleasant things would not happen to them. However, only the currently paranoid group showed an increased tendency to view other people as threatening.

Conclusion. This study has initiated the exploration of heuristic reasoning in paranoia and depression. The findings have therapeutic utility and future work could focus on the differentiation of paranoia and depression at a cognitive level.

Background. Prefrontal cortical dysfunctions, including disturbances in adaptive context-specific behavior, have been reported in neuropsychological and brain imaging studies of schizophrenia. Some data suggest that treatment with antipsychotic medications may ameliorate these deficits.

Method. We investigated antisaccade performance in 39 antipsychotic-naive, first-episode schizophrenia patients who were re-evaluated 6 weeks after treatment initiation. A group of matched healthy subjects were examined at similar time-points. Patients and healthy individuals available for longer-term testing were re-assessed 26 and 52 weeks after initial testing.

Results. Before treatment, patients showed elevated rates of response suppression errors and prolonged latencies of correct antisaccades. Increased rates of antisaccade errors were associated with faster response latencies during a separate, visually guided saccade task, but only prior to treatment. Throughout the 1-year follow-up, patients progressively improved in their ability to voluntarily suppress context-inappropriate behavior. Although treatment assignment was by clinician choice, results of exploratory analyses revealed that patients treated with risperidone progressively planned and initiated correct antisaccades more quickly than patients receiving haloperidol.

Conclusions. Deficits in the voluntary control of spatial attention are exaggerated during acute episodes of illness, but remain an enduring aspect of prefrontal dysfunction in schizophrenia even after treatment. During acute illness, speeded sensorimotor transformations may compound these deficits and contribute to the heightened distractibility associated with acute psychosis. Continued improvement in task performance throughout the 1-year follow-up suggests that partial normalization of prefrontal cognitive functions resulting from antipsychotic treatment may have a longer and more gradual time course than the reduction of acute psychotic symptoms.

Background. In psychosis, the prime indicator of outcome has been relapse, but hospital readmission can no longer be used for this purpose. Researchers now require methods for assessing relapse that are objective, blind, reliable and valid. We describe the reliability and validity of such a technique using case-notes.

Method. Information from routine clinical notes of participants in the Lambeth Early Onset (LEO) study (less all references that would unblind the assessor) were recorded on a form divided into 1-month sections. Operational definitions of remission and relapse enabled clinicians to identify remissions and relapses blindly from the summary information. We calculated reliability regarding both the fact and the timing of remission and relapse. PANSS ratings at 6 and 18 months provided a measure of validity.

Results. The kappa value for the identification of remission by individuals ranged from 0·64 to 0·82, while that for consensus between paired raters was 0·56. The corresponding values for relapse were 0·57–0·59 and 0·71. Intra-class correlations for time to remission and to relapse were very high. Raters guessed correctly whether the participants came from the intervention or control group on 60–75% of occasions. Independent PANSS ratings were strongly related to the remission/relapse status of participants.

Conclusions. The reliability of the technique described here was moderate to good, its validity was good, and it provides a useful and timely addition to methods of evaluating remission and relapse in psychosis. On the basis of our experience, we recommend consensus rather than individual ratings.

Background. Panic disorder is a prevalent, often disabling, disorder among primary-care patients, but there are large gaps in quality of treatment in primary care. This study describes the incremental cost-effectiveness of a combined cognitive behavioral therapy (CBT) and pharmacotherapy intervention for patients with panic disorder versus usual primary-care treatment.

Method. This randomized control trial recruited 232 primary-care patients meeting DSM-IV criteria for panic disorder from March 2000 to March 2002 from six primary-care clinics from university-affiliated clinics at the University of Washington (Seattle) and University of California (Los Angeles and San Diego). Patients were randomly assigned to receive either treatment as usual or a combined CBT and pharmacotherapy intervention for panic disorder delivered in primary care by a mental health therapist. Intervention patients had up to six sessions of CBT modified for the primary-care setting in the first 12 weeks, and up to six telephone follow-ups over the next 9 months. The primary outcome variables were total out-patient costs, anxiety-free days (AFDs) and quality adjusted life-years (QALYs).

Results. Relative to usual care, intervention patients experienced 60·4 [95% confidence interval (CI) 42·9–77·9] more AFDs over a 12-month period. Total incremental out-patient costs were $492 higher (95% CI $236–747) in intervention versus usual care patients with a cost per additional AFD of $8.40 (95% CI $2.80–14.0) and a cost per QALY ranging from $14158 (95% CI $6791–21496) to $24776 (95% CI $11885–37618). The cost per QALY estimate is well within the range of other commonly accepted medical interventions such as statin use and treatment of hypertension.

Conclusions. The combined CBT and pharmacotherapy intervention was associated with a robust clinical improvement compared to usual care with a moderate increase in ambulatory costs.

Background. There have been conflicting reports on time trends in urban–rural differences in the incidence of schizophrenia. This study explored the potential time trends in these differences with regard to birth cohort and age at onset.

Method. Linking data from the Danish Civil Registration system with data from the Danish Psychiatric Central Register, a cohort born in Denmark from 1910 to 1986 was established (5·05 million people). Overall, 23051 people were classified with schizophrenia in 1970–2001.

Results. Urban–rural differences in schizophrenia risk may have existed for people born in Denmark since 1910, and have existed at a constant level for people born from 1945 to 1986. Males aged <20 years had a risk of 3·90 [95% confidence interval (CI) 3·28–4·65] associated with urban birth while males [ges ]20 years had a risk of 2·12 (1·98–2·27). Females <20 years had a risk of 2·49 (95% CI 2·01–3·09) associated with urban birth while females [ges ]20 years had a risk of 1·90 (95% CI 1·74–2·08). At age 46, 1·84% (95% CI 1·76–1·93) of males and 1·05% (95% CI 0·99–1·12) of females born in the capital area had developed schizophrenia, while 0·81% (95% CI 0·75–0·86) of males and 0·56% (95% CI 0·51–0·60) of females born in the rural area had developed schizophrenia.

Conclusions. There was no evidence of time trends in the urban–rural differences in the incidence of schizophrenia in Denmark, suggesting that the cause(s) responsible for these differences were not related to exposures that became more prevalent in urban areas over time. This finding is in contrast to findings from Finland and The Netherlands.

Background. Bright light and high-density negative air ion exposure are efficacious for winter and non-seasonal depression compared with a low-density negative ion placebo. Similarly, auditory stimuli improve mood in clinical populations. This study compared the short-term effects of bright light, an auditory stimulus, and high- and low-density negative ions on mood and alertness in mildly depressed and non-depressed adults.

Method. One hundred and eighteen subjects, 69 women and 49 men (mean age±S.D., 19·4±1·7 years), participated once across the year. Subjects were randomly assigned to one of four conditions: bright light (10000 lux; n=29), auditory stimuli (60 dB; n=30), or high-density (4·5×1014 ions/s flow rate; n=29) or low-density (1·7×1011 ions/s; n=30; placebo control) negative ions. Exposure was for 30 min on three consecutive evenings between 1900 and 2100 hours. Mood and alertness assessments, using standardized scales, occurred before, and 15 and 30 min during exposure. The Beck Depression Inventory classified subjects as depressed ([ges ]10; n=35) or non-depressed (<10; n=83).

Results. The three active stimuli, but not the low-density placebo, reduced depression, total mood disturbance (a global affect measure) and/or anger within 15–30 min. Neither testing season nor degree of depressive symptoms affected response to stimuli.

Conclusions. The auditory stimulus, bright light and high-density ions all produced rapid mood changes – with small to medium effect sizes – in depressed and non-depressed subjects, compared with the low-density placebo, despite equivalent pre-study expectations. Thus, these stimuli improve mood acutely in a student sample, including a subset with depressive symptoms.

Background. Recognition can be normalized in schizophrenia by providing patients with semantic organizational strategies through a levels-of-processing (LOP) framework. However, patients may rely primarily on familiarity effects, making recognition less sensitive than source monitoring to the strength of the episodic memory trace. The current study investigates whether providing semantic organizational strategies can also normalize patients' internal source-monitoring performance.

Method. Sixteen clinically stable medicated patients with schizophrenia and 15 demographically matched healthy controls were asked to identify the source of remembered words following an LOP-encoding paradigm in which they alternated between processing words on a ‘shallow’ perceptual versus a ‘deep’ semantic level. A multinomial analysis provided orthogonal measures of item recognition and source discrimination, and bootstrapping generated variance to allow for parametric analyses. LOP and group effects were tested by contrasting recognition and source-monitoring parameters for words that had been encoded during deep versus shallow processing conditions.

Results. As in a previous study there were no group differences in LOP effects on recognition performance, with patients and controls benefiting equally from deep versus shallow processing. Although there were no group differences in internal source monitoring, only controls had significantly better performance for words processed during the deep encoding condition. Patient performance did not correlate with clinical symptoms or medication dose.

Conclusions. Providing a deep processing semantic encoding strategy significantly improved patients' recognition performance only. The lack of a significant LOP effect on internal source monitoring in patients may reflect subtle problems in the relational binding of semantic information that are independent of strategic memory processes.

Background. Patients may present with cognitive impairment in the euthymic phase of affective disorder, but it is unclear whether the impairment is prevalent before onset of the illness. The aim of the present study was to examine the hypothesis that genetic liability to affective disorder is associated with cognitive impairment.

Method. In a cross-sectional high-risk case–control study, healthy monozygotic (MZ) and dizygotic (DZ) twins with (High-Risk twins) and without (the control group/Low-Risk twins) a co-twin history of affective disorder were identified through nationwide registers. Cognitive performance of 203 High-Risk and Low-Risk twins was compared.

Results. Healthy twins discordant for unipolar disorder showed lower performance on almost all measures of cognitive function: selective and sustained attention, executive function, language processing and working and declarative memory, and also after adjustment for demographic variables, subclinical symptoms and minor psychopathology. Healthy twins discordant for bipolar disorder showed lower performance on tests measuring episodic and working memory, also after adjustment for the above-mentioned covariables. The discrete cognitive impairment found seemed to be related to genetic liability, as the MZ High-Risk twins showed significant impairment on selective and sustained attention, executive function, language processing and working and declarative memory, whereas the DZ High-Risk twins presented with significantly lower scores only on language processing and episodic memory.

Conclusions. The hypothesis that discrete cognitive impairment is present before the onset of the affective disorder and is genetically transmitted was supported. Thus, cognitive function may be a candidate endophenotype for affective disorders.

Background. Although caffeine is the most commonly used psychoactive substance and often produces symptoms of toxicity and dependence, little is known, especially in community samples, about the association between caffeine use, toxicity and dependence and risk for common psychiatric and substance use disorders.

Method. Assessments of lifetime maximal caffeine use and symptoms of caffeine toxicity and dependence were available on over 3600 adult twins ascertained from the population-based Virginia Twin Registry. Lifetime histories of major depression (MD), generalized anxiety disorder (GAD) and panic disorder, alcohol dependence, adult antisocial behavior and cannabis and cocaine abuse/dependence were obtained at personal interview. Logistic regression analyses in the entire sample and within monozygotic (MZ) twin pairs were conducted in SAS.

Results. In the entire sample, measures of maximal caffeine use, heavy caffeine use, and caffeine-related toxicity and dependence were significantly and positively associated with all seven psychiatric and substance use disorders. However, within MZ twin pairs, controlling for genetic and family environmental factors, these associations, while positive, were all non-significant. These results were similar when excluding twins who denied regular caffeine use.

Conclusions. Maximal lifetime caffeine intake and caffeine-associated toxicity and dependence are moderately associated with risk for a wide range of psychiatric and substance use disorders. Analyses of these relationships within MZ twin pairs suggest that most of the observed associations are not causal. Rather, familial factors, which are probably in part genetic, predispose to both caffeine intake, toxicity and dependence and the risk for a broad array of internalizing and externalizing disorders.

Background. Epidemiological studies have found lower than expected prevalence of psychiatric disorders among disadvantaged race-ethnic minority groups in the USA. Recent research shows that this is due entirely to reduced lifetime risk of disorders, as opposed to persistence. Specification of race-ethnic differences with respect to clinical and social characteristics can help identify the protective factors that lead to lower lifetime risk among disadvantaged minority groups.

Method. Data on 5424 Hispanics, non-Hispanic Blacks, and non-Hispanic Whites came from the National Comorbidity Survey Replication, a nationally representative survey conducted with the World Mental Health version of the Composite International Diagnostic Interview. Race-ethnic differences in risk of disorders were compared across specific diagnoses, ages of onset, cohorts and levels of education.

Results. Both minority groups had lower risk for common internalizing disorders: depression, generalized anxiety disorder, and social phobia. In addition, Hispanics had lower risk for dysthymia, oppositional-defiant disorder and attention deficit hyperactivity disorder; non-Hispanic Blacks had lower risk for panic disorder, substance use disorders and early-onset impulse control disorders. Lower risk among Hispanics, relative to non-Hispanic Whites, was found only among the younger cohort (age [les ]43 years). Lower risk among minorities was more pronounced at lower levels of education.

Conclusion. The pattern of race-ethnic differences in risk for psychiatric disorders suggests the presence of protective factors that originate in childhood and have generalized effects on internalizing disorders. For Hispanics, but not for non-Hispanic Blacks, the influence of these protective factors has emerged only recently.

Background. We tested the hypothesis that depressive symptoms in healthy young adults would be associated with elevated levels of C-reactive proteins (CRP).

Method. We studied the association between depressive symptoms and CRP in 1201 young adults, as a part of the on-going population-based Cardiovascular Risk in Young Finns Study. Depressive symptoms were determined by responses to a revised version of Beck's Depression Inventory in 1992 and 2001. CRP and other known cardiac risk factors were measured in 2001.

Results. Higher depressive symptomatology in 1992 and in 2001 and their means score were related to higher CRP levels (B's range from 0·24 to 0·21, p<0·001). These relationships persisted after separate adjustments for various risk factors including sex, age, education, oral contraceptive use, dietary fat, physical activity, alcohol consumption, smoking status, LDL-cholesterol, HDL-cholesterol, systolic blood pressure and history of acute infectious disease. Adjustments for obesity and triglycerides levels, however, somewhat attenuated the relationship between depressive symptoms and CRP.

Conclusions. We concluded that higher levels of depressive symptoms are associated with higher levels of CRP, but this association may largely be attributable to obesity or triglycerides.

Background. There is a limited literature examining face affect recognition in offenders. In line with the Integrated Emotion Systems (IES) model, existing studies suggest a psychopathy-related deficit in sad/fear recognition. However, many of these studies have small samples, and few include a healthy control group in order to examine the contribution of criminality.

Method. We compared the performance of male criminals with dissocial (antisocial) personality disorder (PD group) and healthy male IQ-matched controls, on a morphed face affect-processing task and examined the relationship between psychopathy (assessed using the Psychopathy Checklist: Screening Version, PCL: SV), score and performance on this task in the PD sample.

Results. The PD group had a specific deficit in the recognition of sad facial affect that was present even at 100% expression intensity. This deficit could not be attributed to impulsive responding as the PD group generally had longer mean reaction times than healthy controls. Within the PD group, those with high scores on the PCL: SV were less accurate than low scorers at classifying sad facial affect. There was also a significant negative correlation between total psychopathy score and sad affect recognition accuracy. There were no specific relationships between affect recognition and the subcomponents of psychopathy.

Conclusions. The findings suggest that criminality/antisocial personality may be associated with a deficit in the recognition of aversive cues in others, and that this deficit is more severe in psychopathic offenders. The findings lend further support to the IES model.

Background. Prior population-based twin studies from two Anglophonic countries with relatively high rates of drug use – the USA and Australia – suggest that genetic factors contribute substantially to individual differences in the use, abuse and dependence of illicit psychoactive substances. Would these results replicate in Norway, a Nordic country with a low prevalence of illicit drug use?

Method. Lifetime use, abuse and dependence of five illicit drug categories (cannabis, stimulants, opiates, cocaine and psychedelics) were assessed at personal interview in 1386 complete young adult twin pairs ascertained from the Norwegian Institute of Public Health Twin Panel. Twin model fitting was performed using the Mx statistical package on three phenotypes: any lifetime use, endorsement of at least one DSM-IV symptom of abuse or dependence, and meeting DSM-IV criteria for abuse or dependence.

Results. Significant lifetime use of illicit substances (defined as use 10 or more times) was reported by only 6·4% of the sample. Meaningful analyses were possible for use of any substance and each of the five substances individually, but for symptoms or a diagnosis of abuse/dependence meaningful analyses were possible only for any substance and cannabis. Full twin models uniformly found twin resemblance to be due largely or entirely to genetic factors. Best-fit models for all analyses included only genetic and individual-specific environmental effects with heritability estimates ranging from 58% to 81%.

Conclusion. In accord with prior results from the USA and Australia, genetic factors appear to play an important role in the etiology of use and abuse/dependence of illicit drugs in Norway.

Background. The ability to inhibit inappropriate or unwanted actions is a key element of executive control. The existence of executive function deficits in schizophrenia is consistent with frontal lobe theories of the disorder. Relatively few studies have examined response inhibition in schizophrenia, and none in adolescent patients with early-onset schizophrenia (EOS).

Methods. Twenty-one adolescents with the onset of clinically impairing psychosis before 19 years of age and 16 matched controls performed a stop-signal task to assess response inhibition. The patients with EOS were categorized as paranoid (n=10) and undifferentiated subtypes (n=11). The undifferentiated group had higher levels of negative symptomatology. Stop-signal reaction time (SSRT) and go-signal reaction time (Go-RT) were analysed with respect to hand of response.

Results. The undifferentiated early-onset patients had significantly longer SSRTs, indicative of poor response inhibition, for the left hand compared to the paranoid early-onset patients and control participants. No differences existed for inhibitory control with the right hand. The three groups did not differ in Go-RT.

Conclusions. Our results indicate a specific lateralized impairment of response inhibition in patients with undifferentiated, but not paranoid, EOS. These findings are consistent with reports of immature frontostriatal networks in EOS and implicate areas such as the pre-motor cortex and supplementary motor area (SMA) that are thought to play a role in both voluntary initiation and inhibition of movement.

Background. Military samples provide an excellent context to systematically ascertain hospitalization for severe psychiatric disorders. The National Collaborative Study of Early Psychosis and Suicide (NCSEPS), a collaborative study of psychiatric disorders in the US Armed Forces, estimated rates of first hospitalization in the military for three psychiatric disorders: bipolar disorder (BD), major depressive disorder (MDD) and schizophrenia.

Method. First hospitalizations for BD, MDD and schizophrenia were ascertained from military records for active duty personnel between 1992 and 1996. Rates were estimated as dynamic incidence (using all military personnel on active duty at the midpoint of each year as the denominator) and cohort incidence (using all military personnel aged 18–25 entering active duty between 1992 and 1996 to estimate person-years at risk).

Results. For all three disorders, 8723 hospitalizations were observed in 8120136 person-years for a rate of 10·7/10000 [95% confidence interval (CI) 10·5–11·0]. The rate for BD was 2·0 (95% CI 1·9–2·1), for MDD, 7·2 (95% CI 7·0–7·3), and for schizophrenia, 1·6 (95% CI 1·5–1·7). Rates for BD and MDD were greater in females than in males [for BD, rate ratio (RR) 2·0, 95% CI 1·7–2·2; for MDD, RR 2·9, 95% CI 2·7–3·1], but no sex difference was found for schizophrenia. Blacks had lower rates than whites of BD (RR 0·8, 95% CI 0·7–0·9) and MDD (RR 0·8, 95% CI 0·8–0·9), but a higher rate of schizophrenia (RR 1·5, 95% CI 1·3–1·7).

Conclusions. This study underscores the human and financial burden that psychiatric disorders place on the US Armed Forces.

Background. We explored the applicability of recently proposed research criteria for mild cognitive impairment (MCI) in a memory clinic and changes in case definition related to which memory tests are used and the status of general cognitive function in MCI.

Method. A total of 166 consecutive GP referrals to the Cambridge Memory Clinic underwent comprehensive neuropsychological and psychiatric evaluation.

Results. Of 166 cases, 42 were excluded (significant depression 8, established dementia 29 and other disorders 5). Of 124 non-demented, non-depressed patients, 72 fulfilled Petersen's criteria for amnestic MCI based upon verbal memory performance [the Rey Auditory Verbal Learning Test (RAVLT)] and 90 met criteria if performance on verbal and/or non-verbal memory tests [the Rey figure recall or the Paired Associates Learning test (PAL)] was considered. Of the 90 broadly defined MCI cases, only 25 had pure amnesia: other subtle semantic and/or attention deficits were typically present. A further 12 were classed as non-amnestic MCI and 22 as ‘worried well’.

Conclusions. Definition of MCI varies considerably dependent upon the tests used for case definition. The majority have other cognitive deficits despite normal performance on the Mini-mental State Examination (MMSE) and intact activities of daily living (ADL) and fit within multi-domain MCI. Pure amnesic MCI is rare.

Objective. Recurrence risk estimates for schizophrenia are fundamental to our understanding of this complex disease. Widely cited estimates are from small/older samples. If these estimates are biased upwards, then the rationale for molecular genetic studies of schizophrenia may not be as solid.

Method. We created a population-based, Swedish national cohort by linking two Swedish national registers into a relational database (the Swedish Hospital Discharge Register and the Multi-Generation Register). Affection was defined as the lifetime presence of at least two in-patient hospitalizations with a core schizophrenia diagnosis.

Results. Merging the Swedish national registers created a population-based cohort of 7739202 individuals of known parentage. The lifetime prevalence of the narrow definition of schizophrenia was 0·407% and we estimated that one in every 79 extended Swedish families had been impacted by schizophrenia. The proportion of affected families with multiple affected members was 3·81%. Recurrence risk estimates for all relative types were strikingly similar to those reported in smaller and older studies. For example, we estimated λsibs at 8·55 [95% confidence interval (CI) 7·86–9·57] compared with a literature estimate of 8·6.

Conclusions. In the largest and most comprehensive sample yet studied, we confirm the accepted estimates of recurrence risks for schizophrenia, and provide more accurate estimates of recurrence risks of schizophrenia in relatives, an estimate of the familial impact of schizophrenia, and the multiplex proportion (essential for gauging the generalizability of findings from multiplex pedigrees). These data may be valuable for planning and interpreting genetic studies of schizophrenia.

Background. The evidence regarding the independent influence of duration of untreated psychosis (DUP) on rate and time to remission is far from unequivocal. The goal of the current study was to examine the role of predictors for rate and time to remission in first-episode psychosis (FEP).

Method. The differential effect of age, gender, age of onset, duration of untreated psychosis (DUP), duration of untreated illness (DUI), pre-morbid adjustment, co-morbid diagnosis of substance abuse and adherence to medication on the rate of and time to remission were estimated using a logistic and Poisson regression, and survival analysis respectively, in FEP patients.

Results. In a sample of 107 FEP patients 82·2% achieved remission over a period of 2 years after a mean of 10·3 weeks (range 1–72). Regression analysis, based on complete data on all variables of interest (n=80), showed status of remission to be positively influenced by better pre-morbid adjustment (RR 0·57, 95% CI 0·34–0·95, p<0·05), later age of onset (RR 1·09, 95% CI 1·05–1·13, p<0·0001), higher level of adherence to medication (RR 1·96, 95% CI 1·38–2·76, p<0·001) and shorter DUI (RR 0·99, 95% CI 0·997–0·999, p<0·005). Time to remission was influenced by age of onset (HR 1·04, 95% CI 1·00–1·08, p<0·04) and adherence to medication (HR 1·58, 95% CI 1·11–2·23, p<0·01).

Conclusions. Improving adherence to medication early in the course of treatment may be an important intervention to improve short-term outcome.

Background. Aggressiveness on the roads and/or anger behind the wheel are considered to be a major traffic safety problem in several countries. However, the psychological mechanisms of anger and/or aggression on the roads remain largely unclear. This study examines a large cohort of French employees followed over the period 1994–2001 to establish whether psychometric measures of aggression/hostility were significantly associated with an increased risk of an injury accident (I-A). An I-A was defined as a traffic accident in which someone was injured, that is required medical care.

Method. A total of 11754 participants aged from 39 to 54 years in 1993 were included in this study. Aggression/hostility was measured in 1993 using the French version of the Buss–Durkee Hostility Inventory (BDHI). Driving behaviors and I-A were recorded in 2001. Sociodemographic and alcohol consumption data were available from annual follow-up of the cohort. The relationship between aggression/hostility scores and I-A was assessed using negative binomial regression models with time-dependent covariates.

Results. The overall BDHI scoring was not statistically predictive of subsequent I-A: adjusted rate ratio (aRR) 1·02, 95% confidence interval (CI) 0·81–1·28, for participants with intermediate scores and aRR 1·25, 95% CI 0·98–1·61 for those with high scores, both compared to those with low scores. The only BDHI subscales found to be associated with I-A were ‘irritability’ (aRR 1·33, 95% CI 1·02–1·75 for participants with high scores) and ‘negativism’ (aRR 1·32, 95% CI 1·01–1·71 for participants with high scores).

Conclusion. Overall aggression/hostility personality traits did not predict I-A in this large cohort of French employees, suggesting that aggressiveness on the roads and/or anger behind the wheel extend beyond the individual's general propensity for aggression.