Previous studies (various designs) present contradicting insights on the potential causal effects of diet/physical activity on depression/anxiety (and vice versa). To clarify this, we employed a triangulation framework including three methods with unique strengths/limitations/potential biases to examine possible bidirectional causal effects of diet/physical activity on depression/anxiety.

Study 1: 3-wave longitudinal study (n = 9,276 Dutch University students). Using random intercept cross-lagged panel models to study temporal associations. Study 2: cross-sectional study (n = 341 monozygotic and n = 415 dizygotic Australian adult twin pairs). Using a co-twin control design to separate genetic/environmental confounding. Study 3: Mendelian randomization utilizing data (European ancestry) from genome-wide association studies (n varied between 17,310 and 447,401). Using genetic variants as instrumental variables to study causal inference.

Study 1 did not provide support for bidirectional causal effects between diet/physical activity and symptoms of depression/anxiety. Study 2 did provide support for causal effects between fruit/vegetable intake and symptoms of depression/anxiety, mixed support for causal effects between physical activity and symptoms of depression/anxiety, and no support for causal effects between sweet/savoury snack intake and symptoms of depression/anxiety. Study 3 provides support for a causal effect from increased fruit intake to the increased likelihood of anxiety. No support was found for other pathways. Adjusting the analyses including diet for physical activity (and vice versa) did not change the conclusions in any study.

Triangulating the evidence across the studies did not provide compelling support for causal effects of diet/physical activity on depression/anxiety or vice versa.

Little is known about the dose and pattern of moderate-to-vigorous physical activity (MVPA) to prevent depression. We aimed to assess the prospective association of dose and pattern of accelerometer-derived MVPA with the risk of diagnosed depression.

We included 74,715 adults aged 40–69 years from the UK Biobank cohort who were free of severe disease at baseline and participated in accelerometer measurements (mean age 55.2 years [SD 7.8]; 58% women). MVPA at baseline was derived through 1-week wrist-worn accelerometry. Diagnosed depression was defined by hospitalization with ICD-10 codes F32.0-F32.A. Restricted cubic splines and Cox regression determined the prospective association of dose and pattern of MVPA with the risk of incident depression.

Over a median 7.9-year follow-up, there were 3,089 (4.1%) incident cases of depression. Higher doses of MVPA were curvilinearly associated with lower depression risk, with the largest minute-per-minute added benefits occurring between 5 (HR 0.99 [95% CI 0.96–0.99]) and 280 (HR 0.67 [95% CI 0.60–0.74]) minutes per week (reference: 0 MVPA minutes).

Regardless of pattern, higher doses of MVPA were associated with lower depression risk in a curvilinear manner, with the greatest incremental benefit per minute occurring during the first 4–5 h per week. Optimal benefits occurred around 15 h/week.

Estimating the risk of developing bipolar disorder (BD) in children and adolescents (C&A) with depressive disorders is important to optimize prevention and early intervention efforts. We aimed to quantitatively examine the risk of developing BD from depressive disorders and identify factors which moderate this development.

In this systematic review and meta-analysis (PROSPERO:CRD42023431301), PubMed and Web-of-Science databases were searched for longitudinal studies reporting the percentage of C&A with ICD/DSM-defined depressive disorders who developed BD during follow-up. Data extraction, random-effects meta-analysis, between-study heterogeneity analysis, quality assessment, sub-group analyses, and meta-regressions were conducted.

Thirty-nine studies were included, including 72,371 individuals (mean age=13.9 years, 57.1% females); 14.7% of C&A with a depressive disorder developed BD after 20.4–288 months: 9.5% developed BD-I (95% CI=4.7 to 18.1); 7.7% developed BD-II (95% CI=3.2% to 17.3%); 19.8% (95% CI=9.9% to 35.6%) of C&A admitted into the hospital with a depressive disorder developed BD. Studies using the DSM (21.6%, 95% CI=20.2% to 23.1%) and studies evaluating C&A with a major depressive disorder only (19.8%, 95% CI=16.8% to 23.1%) found higher rates of development of BD. Younger age at baseline, a history of hospitalization and recruitment from specialized clinics were associated with an increased risk of developing BD at follow-up. Quality of included studies was good in 76.9% of studies.

There is a substantial risk of developing BD in C&A with depressive disorders. This is particularly the case for C&A with MDD, DSM-diagnosed depressive disorders, and C&A admitted into the hospital. Research exploring additional predictors and preventive interventions is crucial.

Adolescence is a critical period for preventing substance use and mental health concerns, often targeted through separate school-based programs. However, co-occurrence is common and is related to worse outcomes. This study explores prevention effects of leading school-based prevention programs on co-occurring alcohol use and psychological distress.

Data from two Australian cluster randomized trials involving 8576 students in 97 schools were harmonized for analysis. Students received either health education (control) or one of five prevention programs (e.g. Climate Schools, PreVenture) with assessments at baseline and 6, 12, 24, and 30 or 36 months (from ages ~13–16). Multilevel multinomial regressions were used to predict the relative risk ratios (RRs) of students reporting co-occurring early alcohol use and psychological distress, alcohol use only, distress only, or neither (reference) across programs.

The combined Climate Schools: Alcohol and Cannabis and Climate Schools: Mental Health courses (CSC) as well as the PreVenture program reduced the risk of adolescents reporting co-occurring alcohol use and psychological distress (36 months RRCSC = 0.37; RRPreVenture = 0.22). Other evaluated programs (excluding Climate Schools: Mental Health) only appeared effective for reducing the risk of alcohol use that occurred without distress.

Evidence-based programs exist that reduce the risk of early alcohol use with and without co-occurring psychological distress, though preventing psychological distress alone requires further exploration. Prevention programs appear to have different effects depending on whether alcohol use and distress present on their own or together, thus suggesting the need for tailored prevention strategies.

Tackling methods of suicide and limiting access to lethal means remain priority areas of suicide prevention strategies. Although mental health services are a key setting for suicide prevention, no recent studies have explored methods used by mental health patients.

To investigate associations between main suicide methods and social, behavioural and clinical characteristics in patients with mental illness to inform prevention and improve patient safety.

Data were collected as part of the National Confidential Inquiry into Suicide and Safety in Mental Health. We examined the main suicide methods of 26 766 patients in the UK who died within 12 months of contact with mental health services during 2005–2021. Associations between suicide methods and patient characteristics were investigated using chi-square tests and univariate and multivariate logistic regression.

Suicide methods were associated with particular patient characteristics: hanging was associated with a short illness history, recent self-harm and depression; self-poisoning with substance misuse, personality disorder and previous self-harm; and both jumping and drowning with ethnic minority groups, schizophrenia and in-patient status.

A method-specific focus may contribute to suicide prevention in clinical settings. Hanging deaths outside of wards may be difficult to prevent but our study suggests patients with recent self-harm or in the early stages of their illness may be more at risk. Patients with complex clinical histories at risk of suicide by self-poisoning may benefit from integrated treatment with substance use services. Environmental control initiatives are likely to be most effective for those at risk of jumping or drowning.

At the basis of many important research questions is causality – does X causally impact Y? For behavioural and psychiatric traits, answering such questions can be particularly challenging, as they are highly complex and multifactorial. ‘Triangulation’ refers to prospectively choosing, conducting and integrating several methods to investigate a specific causal question. If different methods, with different sources of bias, all indicate a causal effect, the finding is much less likely to be spurious. While triangulation can be a powerful approach, its interpretation differs across (sub)fields and there are no formal guidelines. Here, we aim to provide clarity and guidance around the process of triangulation for behavioural and psychiatric epidemiology, so that results of existing triangulation studies can be better interpreted, and new triangulation studies better designed.

We first introduce the concept of triangulation and how it is applied in epidemiological investigations of behavioural and psychiatric traits. Next, we put forth a systematic step-by-step guide, that can be used to design a triangulation study (accompanied by a worked example). Finally, we provide important general recommendations for future studies.

While the literature contains varying interpretations, triangulation generally refers to an investigation that assesses the robustness of a potential causal finding by explicitly combining different approaches. This may include multiple types of statistical methods, the same method applied in multiple samples, or multiple different measurements of the variable(s) of interest. In behavioural and psychiatric epidemiology, triangulation commonly includes prospective cohort studies, natural experiments and/or genetically informative designs (including the increasingly popular method of Mendelian randomization). The guide that we propose aids the planning and interpreting of triangulation by prompting crucial considerations. Broadly, its steps are as follows: determine your causal question, draw a directed acyclic graph, identify available resources and samples, identify suitable methodological approaches, further specify the causal question for each method, explicate the effects of potential biases and, pre-specify expected results. We illustrated the guide’s use by considering the question: ‘Does maternal tobacco smoking during pregnancy cause offspring depression?’.

In the current era of big data, and with increasing (public) availability of large-scale datasets, triangulation will become increasingly relevant in identifying robust risk factors for adverse mental health outcomes. Our hope is that this review and guide will provide clarity and direction, as well as stimulate more researchers to apply triangulation to causal questions around behavioural and psychiatric traits.