Having a relapse of schizophrenia or recurrent psychosis is feared by patients, can cause social and personal disruption and has been suggested to cause long-term deterioration, possibly because of a toxic biological process.

To assess whether relapse affected the social and clinical outcomes of people enrolled in a 24-month randomised controlled trial of antipsychotic medication dose reduction versus maintenance treatment.

The trial involved participants with a diagnosis of schizophrenia or recurrent, non-affective psychosis. Relapse was defined as admission to hospital or significant deterioration (assessed by a blinded end-point committee). We analysed the relationship between relapse during the trial and social functioning, quality of life, symptom scores (Positive and Negative Syndrome Scale) and rates of being in employment, education or training at 24-month follow-up. We also analysed changes in these measures during the trial among those who relapsed and those who did not. Sensitivity analyses were conducted examining the effects of ‘severe’ relapse (i.e. admission to hospital).

During the course of the trial, 82 out of 253 participants relapsed. There was no evidence for a difference between those who relapsed and those who did not on changes in social functioning, quality of life, symptom scores or overall employment rates between baseline and 24-month follow-up. Those who relapsed showed no change in their social functioning or quality of life, and a slight improvement in symptoms compared to baseline. They were more likely than those who did not relapse to have had a change in their employment status (mostly moving out of employment, education or training), although numbers changing status were small. Sensitivity analyses showed the same results for those who experienced a ‘severe’ relapse.

Our data provide little evidence that relapse has a detrimental effect in the long term in people with schizophrenia and recurrent psychosis.

Irritability is a common symptom in children and adolescents, often resulting in referral to mental health services and is associated with depression. Depression in adolescents and adults at familial risk of, and with depression, is associated with reduced risk-taking on the Cambridge Gambling Task (CGT) particularly when the chance of winning is high. However, little is known about risk-taking in irritability. This study tests the hypothesis that increased irritability is longitudinally associated with later risk-taking behaviour on the CGT; specifically, that increasing irritability is associated with lower risk-taking when the chance of a favourable outcome is high.

We conducted a longitudinal study of the biological offspring of parents of children with depression (n = 337). Irritability, the exposure, was measured at wave one using the Child and Adolescent Psychiatric Assessment (CAPA). The primary outcome was risk-taking to obtain reward at varying probability ratios (6:4, 7:3, 8:2 and 9:1) measured by the Cambridge Gambling Task (CGT) at waves two and three. We investigated the longitudinal association between irritability at wave one and average risk-taking at each ratio across waves two and three using multi-level models. The extent to which risk-taking according to probability ratio varied with irritability was tested with interaction terms. We ran univariable models and then multivariable models.

In univariable (n = 207; Coef. 0.006, 95%CI −0.011–0.023, p = 0.470), and fully adjusted (Coef. 0.011, 95%CI −0.007–0.029, p = 0.213) models there was no evidence of a main association between irritability and risk-taking on the CGT. There was evidence of an interaction between irritability and risk-taking ratio (p = 0.019). In fully adjusted models including the interaction, a one-point increase in irritability was associated with relatively higher risk-taking at the less favourable ratios (6:4 – 0.018 (95%CI −0.002–0.037) and 7:3 – 0.015 (95%CI −0.005–0.035)) relative to the more favourable ratios (9:1 – 0.001 (95%CI −0.019–0.021) and 8:1 – 0.011 (95%CI −0.008-0.031)).

We found no evidence of relationship between irritability and subsequent risk-taking on the CGT overall. However, there was some evidence that those with higher irritability were relatively more risk-taking when less likely to win compared with when a favourable outcome was more likely. These findings warrant further investigation of the association between prior irritability and later depression in a larger community cohort. If prior irritability and depression are both associated with risk-taking, this strengthens the case for focussing on risk-taking as a potential target for preventive intervention.

The prevalence of emotional problems, such as depressive and anxiety disorders, increases sharply during adolescence. There is evidence for familial clustering of mental health problems during early childhood and adulthood, however no studies have investigated whether adolescent mental health problems cluster within families. This study tests the hypotheses that i) emotional problems during adolescence cluster within families, and that ii) conduct, peer and hyperactivity problems, prosocial behaviour and overall emotional and behavioural difficulties during adolescence also cluster within families.

We used cross-sectional data from a nationally representative survey of UK households, collected between 2019 and 2021, with 4,088 participants aged 10–16 years. Analyses included 1,241 participants who had complete outcome data and complete data on all covariates of interest. The Strengths and Difficulties Questionnaire (SDQ) was used to examine emotional problems, as well as conduct, peer and hyperactivity problems, prosocial behaviour and total difficulties. Multilevel modelling was used to: estimate clustering of i) emotional problems and ii) conduct, peer and hyperactivity problems, prosocial behaviour and total difficulties, within families, after adjusting for several individual- and family-level covariates associated with adolescent mental health problems (including individual and family demographics, school and sibling bullying, quality of parent-child relationship, parent mental health and parent romantic relationship satisfaction).

After adjusting for known covariates of adolescent mental health problems, there was substantial clustering of adolescent emotional problems (ICC: 0.439; CI95%: 0.36–0.52; SE: 0.042) and overall adolescent emotional and behavioural difficulties (ICC: 0.417; CI95%: 0.34–0.50; SE: 0.043) within families. There was also evidence of clustering of adolescent peer problems (ICC: 0.374; CI95%: 0.28–0.48; SE: 0.051), hyperactivity (ICC: 0.332; CI95%: 0.25–0.42; SE: 0.044), prosocial behaviour (ICC: 0.263; CI95%: 0.18–0.37; SE: 0.048) and conduct problems (ICC: 0.232; CI95%: 0.14–0.35; SE: 0.053) within families after adjustment.

We found strong evidence that adolescent emotional problems cluster within families even after accounting for individual- and family-level covariates which are associated with adolescent mental health problems. Over 40% of the variation was accounted for at the family level. This indicates how the contextual characteristics of the family environment may influence the mental health of young people. As such, social policy aiming to prevent or improve the mental health of young people should focus on family context.

Conduct disorder carries significant individual and societal repercussions. Despite heightened risk-taking and challenges in adapting to changing probabilities of choice outcomes being linked to maladaptive behaviours such as conduct disorder, no study to date has examined the association behind childhood decision-making and adolescent conduct disorder. This study seeks to address this gap by exploring the longitudinal association between these two variables. Understanding the mechanisms underlying conduct disorder could help with developing new preventive interventions.

We used data from the Millennium Cohort Study, a nationally representative UK cohort; participants included those with complete data on exposure, outcome and confounding variables (n = 7,237). The exposure, childhood decision-making at 11 years was measured using the Cambridge Gambling Task risk-taking and risk-adjustment measures. The outcome, a binary measure of adolescent conduct disorder was created using items from the risky and antisocial behaviour interview sections at age 17. We used logistic regression to examine the association between childhood decision-making and adolescent conduct disorder and adjusted for relevant confounders.

The univariable model showed that at age 11, each 20-point increase in risk-taking score increased the odds of conduct disorder behaviour at age 17 by 32% (OR = 1.32, 95% CI 1.18–1.44, p < 0.0001). In the multivariable model, there was strong evidence that a 20-point increase in risk-taking at 11 years was associated with 18% higher odds of conduct disorder behaviour at 17 years (OR = 1.18, 95% CI 1.05–1.33, p = 0.005). There was no evidence that this association differed by sex. Risk adjustment at 11 years showed no association with conduct disorder behaviours at age 17 both in the univariable model (OR = 0.96, 95% CI 0.88–1.06, p = 0.440) and the multivariable model (OR = 0.96, 95% CI 0.88–1.06, p = 0.433).

We found that risk-taking at 11 years was associated with conduct disorder behaviour at 17 years. If causal, our findings suggest that risk-taking might be a potential mechanism underlying adolescent conduct disorder behaviours. This may be useful in informing the design of preventive strategies, such as encouraging positive risk-taking in children and discouraging negative risk-taking behaviours.

Irritability is common and easily identified in childhood. It is transdiagnostic and a common reason for referral to mental health services. Irritability which does not decrease during early childhood is associated with adolescent depression. We hypothesised that irritability would be associated with increased risk-taking overall but reduced risk-taking in response to loss.

We used data from the Millennium Cohort Study, a population-based cohort of 18,552 children born in 2000–02. We examined whether irritability at 3, 5 and 7 years is associated with risk-taking on the CGT using multilevel mixed effect generalised linear models (MEGLMs). We also calculated the change in irritability between 3–7 years for each participant using multilevel mixed models. We then examined the association between this change measure and risk-taking on the CGT using MEGLMs. Analyses were adjusted for a broad range of confounders.

We found that children whose irritability did not decrease as would be expected from 3 to 7 years were more likely to stake a higher number of points per trial on the CGT at 11 years. This increase was most evident when the previous trial had been won. Irritability at 7 years was associated with staking a higher number of points per trial on the CGT (coefficient 0.52, 95%CI −0.04–1.08, p = 0.067) in fully adjusted model, whereas irritability at 3 and 5 years were not (3 years – coefficient 0.02, 95%CI -0.62–0.65, p = 0.961; 5 years – coefficient 0.14, 95%CI −0.45–0.73, p = 0.641). There was evidence of an interaction between irritability at seven years and whether the previous trial was won (p = 0.014). Childhood irritability which did not decrease between 3–7 years was associated with staking a higher number of points per trial on the CGT (coefficient 1.36, 95%CI 0.44–2.28, p = 0.004); there was evidence of an interaction between change in irritability and whether the previous trial was won (p = 0.056).

This is the first longitudinal population-based study examining the relationship between changes in irritability during early childhood and risk-taking behaviour measured by the CGT. Our findings illustrate that irritability in children is characterised by an increase in risk-taking at age 11 years, reflecting differences in how children behave in relation to rewards and losses based on prior irritability. Further understanding of how the processes such as risk-taking which link childhood phenotypes such as irritability, relate to future mental health, may enable the development of new interventions focussing on reactions to rewards and losses.

Seclusion is a restrictive practice that many healthcare services are trying to reduce. Previous studies have sought to identify predictors of seclusion initiation, but few have investigated factors associated with adverse outcomes after seclusion termination.

To assess the factors that predict an adverse outcome within 24 h of seclusion termination.

In a cohort study of individuals secluded in psychiatric intensive care units, we investigated factors associated with any of the following outcomes: actual violence, attempted violence, or reinitiation of seclusion within 24 h of seclusion termination. Among the seclusion episodes that were initiated between 29 March 2018 and 4 March 2019, we investigated the exposures of medication cooperation, seclusion duration, termination out of working hours, involvement of medical staff in the final seclusion review, lack of insight, and agitation or irritability. In a mixed-effects logistic regression model, associations between each exposure and the outcome were calculated. Odds ratios were calculated unadjusted and adjusted for demographic and clinical variables.

We identified 254 seclusion episodes from 122 individuals (40 female, 82 male), of which 106 (41.7%) had an adverse outcome within 24 h of seclusion termination. Agitation or irritability was associated with an adverse outcome, odds ratio 1.92 (95% CI 1.03 to 3.56, P = 0.04), but there was no statistically significant association with any of the other exposures, although confidence intervals were broad.

Agitation or irritability in the hours preceding termination of seclusion may predict an adverse outcome. The study was not powered to detect other potentially clinically significant factors.

Selective serotonin reuptake inhibitors (SSRIs) are first-line pharmacological treatments for depression and anxiety. However, little is known about how pharmacological action is related to cognitive and affective processes. Here, we examine whether specific reinforcement learning processes mediate the treatment effects of SSRIs.

The PANDA trial was a multicentre, double-blind, randomized clinical trial in UK primary care comparing the SSRI sertraline with placebo for depression and anxiety. Participants (N = 655) performed an affective Go/NoGo task three times during the trial and computational models were used to infer reinforcement learning processes.

There was poor task performance: only 54% of the task runs were informative, with more informative task runs in the placebo than in the active group. There was no evidence for the preregistered hypothesis that Pavlovian inhibition was affected by sertraline. Exploratory analyses revealed that in the sertraline group, early increases in Pavlovian inhibition were associated with improvements in depression after 12 weeks. Furthermore, sertraline increased how fast participants learned from losses and faster learning from losses was associated with more severe generalized anxiety symptoms.

The study findings indicate a relationship between aversive reinforcement learning mechanisms and aspects of depression, anxiety, and SSRI treatment, but these relationships did not align with the initial hypotheses. Poor task performance limits the interpretability and likely generalizability of the findings, and highlights the critical importance of developing acceptable and reliable tasks for use in clinical studies.

This article presents research supported by NIHR Program Grants for Applied Research (RP-PG-0610-10048), the NIHR BRC, and UCL, with additional support from IMPRS COMP2PSYCH (JM, QH) and a Wellcome Trust grant (QH).

Although suicide bereavement is associated with suicide and self-harm, evidence regarding mechanisms is lacking. We investigated whether depression and substance use (alcohol and/or other drugs) explain the association between partner suicide bereavement and suicide.

Linkage of nationwide, longitudinal data from Denmark for the period 1980–2016 facilitated a comparison of 22 668 individuals exposed to bereavement by a partner's suicide with 913 402 individuals bereaved by a partner's death due to other causes. Using causal mediation models, we estimated the degree to which depression and substance use (considered separately) mediated the association between suicide bereavement and suicide.

Suicide-bereaved partners were found to have a higher risk of suicide (HRadj = 1.59, 95% CI 1.36–1.86) and of depression (ORadj 1.16, 95% CI 1.09–1.25) when compared to other-bereaved partners, but a lower risk of substance use (ORadj 0.83; 95% CI 0.78–0.88). An increased risk of suicide was found among any bereaved individuals with a depression diagnosis recorded post-bereavement (ORadj 3.92, 95% CI 3.55–4.34). Mediation analysis revealed that depression mediated 2% (1.68%; 95% CI 0.23%–3.14%; p = 0.024) of the association between suicide bereavement and suicide in partners when using bereaved controls.

Depression is a partial mediator of the association between suicide bereavement and suicide. Efforts to prevent and optimize the treatment of depression in suicide-bereaved people could reduce their suicide risk. Our findings might be conservative because we did not include cases of depression diagnosed in primary care. Further work is needed to understand this and other mediators.

This paper investigates whether age of onset of depression, duration of the last episode, number of episodes, and residual symptoms of depression and anxiety are associated with depression relapse in primary care patients who have been on long–term maintenance antidepressant treatment and no longer meet ICD10 criteria for depression.

An observational cohort using data from ANTLER (N = 478), a double-blind placebo-controlled trial. The primary outcome was time to relapse using the retrospective CIS-R. Participants were followed for 12 months.

Primary outcome was available for 468 participants. Time to relapse in those with more than five previous episodes of depression was shorter, hazard ratio (HR) 1.84 (95% confidence interval [CI] 1.23–2.75) compared to people with two episodes; HR 1.57 (95% CI 1.01–2.43) after adjustment. The residual symptoms of depression at baseline were also associated with increased relapse: HR 1.05 (95% CI 1.01–1.09) and HR 1.06 (95% CI 1.01–1.12) in the adjusted model. There was evidence of reduced rate of relapse in older age of onset group: HR 0.86 (95% CI 0.78–0.95); HR attenuated after adjustment HR 0.91 (95% CI 0.81–1.02). There was no evidence of an association between duration of the current episode and residual anxiety symptoms with relapse.

The number of previous episodes and residual symptoms of depression were associated with increased likelihood of relapse. These factors could inform joint decision making when patients are considering tapering off maintenance antidepressant treatment or considering other treatments to prevent relapse.

Antidepressants have been proposed to act via their influence on emotional processing. We investigated the effect of discontinuing maintenance antidepressant treatment on positive and negative self-referential recall and the association between self-referential recall and risk of relapse.

The ANTLER trial was a large (N = 478) pragmatic double-blind trial investigating the clinical effectiveness of long-term antidepressant treatment for preventing relapse in primary care patients. Participants were randomised to continue their maintenance antidepressants or discontinue via a taper to placebo. We analysed memory for positive and negative personality descriptors, assessed at baseline, 12- and 52-week follow-up.

The recall task was completed by 437 participants. There was no evidence of an effect of discontinuation on self-referential recall at 12 [positive recall ratio 1.00, 95% CI (0.90–1.11), p = 0.93; negative recall ratio 1.00 (0.87–1.14), p = 0.87] or 52 weeks [positive recall ratio 1.03 (0.91–1.17), p = 0.62; negative recall ratio 1.00 (0.86–1.15), p = 0.96; ratios larger than one indicate higher recall in the discontinuation group], and no evidence of an association between recall at baseline or 12 weeks and later relapse [baseline, positive hazard ratio (HR) 1.02 (0.93–1.12), p = 0.74; negative HR 1.01 (0.90–1.13), p = 0.87; 12 weeks, positive HR 0.99 (0.89–1.09), p = 0.81; negative HR 0.98 (0.84–1.14), p = 0.78; ratios larger than one indicate a higher frequency of relapse in those with higher recall].

We found no evidence that discontinuing long-term antidepressants altered self-referential recall or that self-referential recall was associated with risk of relapse. These findings suggest that self-referential recall is not a neuropsychological marker of antidepressant action.

Eating disorders are stigmatised. Little is known about whether stigma has decreased over time and which groups hold more stigmatising beliefs.

To explore whether stigma towards eating disorders has changed between 1998 and 2008 and whether it varies by sociodemographic characteristics.

We used the Office for National Statistics Omnibus surveys 1998 and 2008. As outcomes, we selected four questions eliciting participants’ views on issues of blame and ability to recover, and compared their mean scores across eating disorders, depression and alcohol dependence in both years. We used multivariable linear regressions to investigate associations between sociodemographic characteristics and each stigma domain.

In total, 2720 participants had data on all variables of interest. Compared with 1998, in 2008 stigmatising views towards eating disorders improved. In both years, participants believed it was easier to recover from eating disorders than depression or alcohol dependence. Respondents believed people with eating disorders were more to blame for their condition than those with depression, but less than those with alcohol dependence. Men, those with less formal education, and those from ethnic minority backgrounds were more likely to place greater blame on individuals for their mental illness. Men were more likely than women to think it was possible to recover from an eating disorder.

Stigmatising attitudes towards people with eating disorders have improved over time, but are still greater than those observed for other mental illnesses. Improving eating disorder mental health literacy could help to reduce these negative views and lead to improved quality of life, greater help-seeking and better prognosis.

Catatonia, a severe neuropsychiatric syndrome, has few studies of sufficient scale to clarify its epidemiology or pathophysiology. We aimed to characterise demographic associations, peripheral inflammatory markers and outcome of catatonia.

Electronic healthcare records were searched for validated clinical diagnoses of catatonia. In a case–control study, demographics and inflammatory markers were compared in psychiatric inpatients with and without catatonia. In a cohort study, the two groups were compared in terms of their duration of admission and mortality.

We identified 1456 patients with catatonia (of whom 25.1% had two or more episodes) and 24 956 psychiatric inpatients without catatonia. Incidence was 10.6 episodes of catatonia per 100 000 person-years. Patients with and without catatonia were similar in sex, younger and more likely to be of Black ethnicity. Serum iron was reduced in patients with catatonia [11.6 v. 14.2 μmol/L, odds ratio (OR) 0.65 (95% confidence interval (CI) 0.45–0.95), p = 0.03] and creatine kinase was raised [2545 v. 459 IU/L, OR 1.53 (95% CI 1.29–1.81), p < 0.001], but there was no difference in C-reactive protein or white cell count. N-Methyl-d-aspartate receptor antibodies were significantly associated with catatonia, but there were small numbers of positive results. Duration of hospitalisation was greater in the catatonia group (median: 43 v. 25 days), but there was no difference in mortality after adjustment.

In the largest clinical study of catatonia, we found catatonia occurred in approximately 1 per 10 000 person-years. Evidence for a proinflammatory state was mixed. Catatonia was associated with prolonged inpatient admission but not with increased mortality.

Psychotic experiences are reported by 5–10% of young people, although only a minority persist and develop into psychotic disorders. It is unclear what characteristics differentiate those with transient psychotic experiences from those with persistent psychotic experiences that are more likely to be of clinical relevance.

To investigate how longitudinal profiles of psychotic experiences, created from assessments at three different time points, are influenced by early life and co-occurring factors.

Using data from 8045 individuals from a birth cohort study, longitudinal profiles of psychotic experiences based on semi-structured interviews conducted at 12, 18 and 24 years were defined. Environmental, cognitive, psychopathological and genetic determinants of these profiles were investigated, along with concurrent changes in psychopathology and cognition.

Following multiple imputations, the distribution of longitudinal profiles of psychotic experiences was none (65.7%), transient (24.1%), low-frequency persistent (8.4%) and high-frequency persistent (1.7%). Individuals with high-frequency persistent psychotic experiences were more likely to report traumatic experiences, other psychopathology, a more externalised locus of control, reduced emotional stability and conscientious personality traits in childhood, compared with those with transient psychotic experiences. These characteristics also differed between those who had any psychotic experiences and those who did not.

These findings indicate that the same risk factors are associated with incidence as with persistence of psychotic experiences. Thus, it might be that the severity of exposure, rather than the presence of specific disease-modifying factors, is most likely to determine whether psychotic experiences are transient or persist, and potentially develop into a clinical disorder over time.

This study aimed to investigate general factors associated with prognosis regardless of the type of treatment received, for adults with depression in primary care.

We searched Medline, Embase, PsycINFO and Cochrane Central (inception to 12/01/2020) for RCTs that included the most commonly used comprehensive measure of depressive and anxiety disorder symptoms and diagnoses, in primary care depression RCTs (the Revised Clinical Interview Schedule: CIS-R). Two-stage random-effects meta-analyses were conducted.

Twelve (n = 6024) of thirteen eligible studies (n = 6175) provided individual patient data. There was a 31% (95%CI: 25 to 37) difference in depressive symptoms at 3–4 months per standard deviation increase in baseline depressive symptoms. Four additional factors: the duration of anxiety; duration of depression; comorbid panic disorder; and a history of antidepressant treatment were also independently associated with poorer prognosis. There was evidence that the difference in prognosis when these factors were combined could be of clinical importance. Adding these variables improved the amount of variance explained in 3–4 month depressive symptoms from 16% using depressive symptom severity alone to 27%. Risk of bias (assessed with QUIPS) was low in all studies and quality (assessed with GRADE) was high. Sensitivity analyses did not alter our conclusions.

When adults seek treatment for depression clinicians should routinely assess for the duration of anxiety, duration of depression, comorbid panic disorder, and a history of antidepressant treatment alongside depressive symptom severity. This could provide clinicians and patients with useful and desired information to elucidate prognosis and aid the clinical management of depression.

According to the cognitive neuropsychological model, antidepressants reduce symptoms of depression and anxiety by increasing positive relative to negative information processing. Most studies of whether antidepressants alter emotional processing use small samples of healthy individuals, which lead to low statistical power and selection bias and are difficult to generalise to clinical practice. We tested whether the selective serotonin reuptake inhibitor (SSRI) sertraline altered recall of positive and negative information in a large randomised controlled trial (RCT) of patients with depressive symptoms recruited from primary care.

The PANDA trial was a pragmatic multicentre double-blind RCT comparing sertraline with placebo. Memory for personality descriptors was tested at baseline and 2 and 6 weeks after randomisation using a computerised emotional categorisation task followed by a free recall. We measured the number of positive and negative words correctly recalled (hits). Poisson mixed models were used to analyse longitudinal associations between treatment allocation and hits.

A total of 576 participants (88% of those randomised) completed the recall task at 2 and 6 weeks. We found no evidence that positive or negative hits differed according to treatment allocation at 2 or 6 weeks (adjusted positive hits ratio = 0.97, 95% CI 0.90–1.05, p = 0.52; adjusted negative hits ratio = 0.99, 95% CI 0.90–1.08, p = 0.76).

In the largest individual placebo-controlled trial of an antidepressant not funded by the pharmaceutical industry, we found no evidence that sertraline altered positive or negative recall early in treatment. These findings challenge some assumptions of the cognitive neuropsychological model of antidepressant action.

Cognitive mechanisms that characterize or precede depressive symptoms are poorly understood. We investigated cross-sectional and longitudinal associations between risk taking to obtain reward and adolescent depressive symptoms in a large prospective cohort, using the Cambridge Gambling Task (CGT). We also explored sex differences.

The Millennium Cohort Study (MCS) is an ongoing UK study, following the lives of 19 000 individuals born 2000/02. The CGT was completed at ages 11 (n = 12 355) and 14 (n = 10 578). Our main exposure was the proportion of points gambled, when the odds of winning were above chance (risk-taking to obtain reward). Outcomes were emotional symptoms (Strengths and Difficulties Questionnaire, SDQ) at age 11 and depressive symptoms (short Mood and Feelings Questionnaire, sMFQ) at age 14. We calculated cross-sectional and longitudinal associations, using linear regressions.

In univariable models, there was evidence of cross-sectional associations between risk-taking and SDQ/sMFQ scores, but these associations disappeared after we adjusted for sex. Longitudinally, there was weak evidence of an association between risk-taking and depressive symptoms in females only [a 20-point increase in risk-taking at age 11 was associated with a reduction of 0.31 sMFQ points at age 14 (95% CI −0.60 to −0.02)]. At both time-points, females were less risk-taking than males.

We found no convincing evidence of a relationship between risk-taking to obtain reward and depressive symptoms. There were large sex differences in risk-taking, but these do not appear to contribute to the female preponderance of depressive symptoms in adolescence.

The Patient Health Questionnaire (PHQ-9), the Beck Depression Inventory (BDI-II) and the Generalised Anxiety Disorder Assessment (GAD-7) are widely used in the evaluation of interventions for depression and anxiety. The smallest reduction in depressive symptoms that matter to patients is known as the Minimum Clinically Important Difference (MCID). Little empirical study of the MCID for these scales exists.

A prospective cohort of 400 patients in UK primary care were interviewed on four occasions, 2 weeks apart. At each time point, participants completed all three questionnaires and a ‘global rating of change’ scale (GRS). MCID estimation relied on estimated changes in symptoms according to reported improvement on the GRS scale, stratified by baseline severity on the Clinical Interview Schedule (CIS-R).

For moderate baseline severity, those who reported improvement on the GRS had a reduction of 21% (95% confidence interval (CI) −26.7 to −14.9) on the PHQ-9; 23% (95% CI −27.8 to −18.0) on the BDI-II and 26.8% (95% CI −33.5 to −20.1) on the GAD-7. The corresponding threshold scores below which participants were more likely to report improvement were −1.7, −3.5 and −1.5 points on the PHQ-9, BDI-II and GAD-7, respectively. Patients with milder symptoms require much larger reductions as percentage of their baseline to endorse improvement.

An MCID representing 20% reduction of scores in these scales, is a useful guide for patients with moderately severe symptoms. If treatment had the same effect on patients irrespective of baseline severity, those with low symptoms are unlikely to notice a benefit.

Funding. National Institute for Health Research.