Psychiatry and mental health

Epidemiological studies suggest that up to 50% of the population will develop a mental illness over the course of their lifetime1. For severe mental disorders such as depression and anxiety, estimates of lifetime prevalence lie between 20% to 25%. It is important to note that these estimates vary widely per country and population2, to a large degree due to availability of resources and access to mental health care3–5. For instance, a meta-analysis of reported lifetime prevalence in a global context found that countries in the Global South recorded a lifetime prevalence sometimes nearly 10 percentage points lower than countries in the Global North2, likely reflecting underdiagnosis in these communities and poor mapping of concepts developed in the Global North onto mental health care in the Global South6,7. The uniform application of diagnostic criteria combined with an unequal access to mental health care poses a major challenge not only to health care, but by extension to mental health research also8.

Diagnosis of psychiatric conditions is complicated by the large degree of heterogeneity in clinical presentation9,10. The variability in clincal presentations can vary so widely that the same diagnostic framework applied multiple times to the same patient (test-retest reliability) can reach different conclusions each time11. A meta-analysis on test-retest reliability of diagnostic classifications in the DSM12 found that the test-retest reliability for some psychiatric conditions such as MDD and GAD was fairly low. This is to no small part due to the phenomenon that clinical categories might have overlapping symptoms which means that some symptoms can be attributed to a number of different underlying psychiatric conditions13. Inversely, patients diagnosed with the same psychiatric condition may have few symptoms in common14. A further complication of the treatment and research into mental health is that psychiatric conditions commonly display a high degree of comorbidity with other psychiatric conditions15,16. Some suggest that the rigid application of diagnostic criteria stipulated by the DSM and the ICD17 play a fundamental role in the prevalence of comorbidity in psychiatry18, while others imply a temporal association between first diagnosis and subsequent psychiatric diagnoses of any classification19, both of which complicate the study of mental health. Methodological choices and frameworks that may work well for clinical practice may complicate and impede research practice.

To address the influence of bias in mental health research introduced by methodological choices, some researchers have proposed alternative measures to the DSM and ICD classifications to classify individuals for use in scientific practice. One such measure is the creation of a common factor representing general psychopathology, termed the p-factor20 which allows individuals to be assessed on a continuous measure rather than the case-control dichotomy20,21. However, because of the limited specificity of the p-factor researchers have created the HiTOP consortium that builds on the p-factor approach by adding multiple levels of specificity collapsed into different continuous spectra22, which aims to provide a less dichotomous and more nuanced framework for classification of mental health symptoms and psychiatric conditions. Similarly, PCA and ICA offer a data-driven decomposition of features and symptoms relevant for mental health agnostic of current diagnostic categories which allows researchers to study these factors in a continuous framework without the need for hard thresholds and cut-offs23,24.

The heterogeneity in psychiatric symptoms also extends to the genetic domain. Psychiatric disorders are highly polygenic and heritable25–28. This is illustrated through a large number of significantly associated genetic variants where each variant shows a small effect size26,29. Due to the genetic heterogeneity, research incorporating molecular genetic methods is challenging. Instead, research into the genetics of psychiatric disorders has focused on identifying and analysing genome-wide significant loci30,31 and polygenic scores derived from a GWAS on the phenotype of interest32. However, these GWAS analyses require large samples with rigorous phenotyping procedures which demands substantial resources. Resouces of this scale are available mainly through large consortia such as the PGC33, which has yielded important insights in the genetic variants associated with for instance SCZ34, BIP35, and MDD36. Some of these consortia samples can then be further supplemented using resources such as the UKB37,38. A more detailed background on the area of psychiatric genetics is provided in Psychiatric genetics. Analysis of the loci identified through a GWAS may provide insight into the biological processes and potential drug targets associated with a phenotype, such as alterations in neurotransmission mechanisms in SCZ39, alterations in gene expression of genes associated with synaptic functioning and cell signaling in BIP40. Although translation of GWAS results to clinical practice is still in its early stages, for instance in the adoption of polygenic scores32,41 or targeting novel drug targets42,43, insights into the biology underlying psychiatric disorders made possible through such large-scale genome-wide studies have proven invaluable33,44.