The Delta Study – Prevalence and characteristics of mood disorders in 924 individuals with low mood: Results of the of the World Health Organization Composite International Diagnostic Interview (CIDI)

Abstract Objectives The Delta Study was undertaken to improve the diagnosis of mood disorders in individuals presenting with low mood. The current study aimed to estimate the prevalence and explore the characteristics of mood disorders in participants of the Delta Study, and discuss their implications for clinical practice. Methods Individuals with low mood (Patients Health Questionnaire‐9 score ≥5) and either no previous mood disorder diagnosis (baseline low mood group, n = 429), a recent (≤5 years) clinical diagnosis of MDD (baseline MDD group, n = 441) or a previous clinical diagnosis of BD (established BD group, n = 54), were recruited online. Self‐reported demographic and clinical data were collected through an extensive online mental health questionnaire and mood disorder diagnoses were determined with the World Health Organization Composite International Diagnostic Interview (CIDI). Results The prevalence of BD and MDD in the baseline low mood group was 24% and 36%, respectively. The prevalence of BD among individuals with a recent diagnosis of MDD was 31%. Participants with BD in both baseline low mood and baseline MDD groups were characterized by a younger age at onset of the first low mood episode, more severe depressive symptoms and lower wellbeing, relative to the MDD or low mood groups. Approximately half the individuals with BD diagnosed as MDD (49%) had experienced (hypo)manic symptoms prior to being diagnosed with MDD. Conclusions The current results confirm high under‐ and misdiagnosis rates of mood disorders in individuals presenting with low mood, potentially leading to worsening of symptoms and decreased well‐being, and indicate the need for improved mental health triage in primary care.


| INTRODUC TI ON
Major depressive disorder (MDD) and bipolar disorder (BD) are complex and debilitating conditions affecting 16.2% and 2.4% of the population worldwide, respectively (Merikangas et al., 2011;Kupfer et al., 2012), and are among the leading contributors to the global burden of diseases (GBD, 2020). Core symptoms of MDD include a pervasive and persistent disturbance of mood and loss of interest/ pleasure in most daily activities (Otte et al., 2016). Individuals can also experience impaired concentration and indecisiveness, as well as fatigue or low energy, disturbances to sleep and appetite, headaches, muscle tension, and general symptoms of pain (American Psychiatric Association, 2013). BD, on the other hand, is typically characterized by intermittent depressive and manic (BDI) or hypomanic (BDII) episodes. While depressive episodes in BD may be indistinguishable from those in MDD, (hypo)manic episodes can include elevated levels of energy, euphoric mood, irritability, and hypersexuality, as well as a reduced need for sleep (Einat, 2007). Diagnosing MDD and BD requires a comprehensive collection of symptom-and patient-level data, as well as information on family history, course of illness, and prior treatment response. The World Health Organization World Mental Health Composite International Diagnostic Interview (CIDI (Kessler & Üstün, 2004)) can be a useful tool for the assessment and differential diagnosis of MDD and BD. It also captures subthreshold forms of elevated mood, including subthreshold BD and MDD with subthreshold BD, with individuals presenting with fewer and/or a shorter duration of (hypo)manic symptoms.
Both MDD and BD typically start early in life and are associated with severe functional impairments, high morbidity and mortality, including premature death due to suicide (Bourne et al., 2013;Passos et al., 2016;Ösby et al., 2016). The economic burden of MDD and BD is also substantial; in England, direct and indirect annual costs are estimated at £7.46 billion (US $9.85 billion) for MDD and £5.25 billion (US $6.93 billion) for BD (McCrone et al., 2008). In spite of their prevalence and negative prognosis, the recognition and diagnosis of these conditions presents a significant challenge, particularly in the primary care setting. For instance, research has shown that general practitioners (GPs) initially misdiagnose 50% of MDD patients (Mitchell et al., 2009). Short consultation times coupled with the difficulties associated with diagnosing MDD, where any two individuals may have no symptoms in common (Olbert et al., 2014;Fried et al., 2014), means that many are not receiving the support they need. This is a particular issue, given that the vast majority of patients with MDD receive treatment and care solely in the primary care setting.
In the case of BD, 60% of individuals are initially misdiagnosed with MDD (Hirschfeld et al., 2003), with many having to wait 8 to 10 years before receiving a correct diagnosis (Bauer et al., 2018;Patel et al., 2015). This is due, in part, to the fact that the majority of individuals with BD seek help when they are experiencing depressive symptoms as opposed to when they are in a (hypo)manic state.
Furthermore, in most instances, depressive episodes precede a first (hypo)manic episode, and awareness of one's (hypo)manic symptoms is relatively low (Regeer et al., 2015). This poses a significant problem for the diagnosis, treatment, and management of BD, with these individuals likely to be treated with antidepressant monotherapy, which is frequently ineffective in treating bipolar depression.
Critically, the use of antidepressants without a concurrent mood stabilizer can trigger and exacerbate a hypo(manic) episode, resulting in prolonged suffering and, in some cases, suicide (Bowden, 2005).
Taken together, the careful evaluation and management of all patients presenting with depressive symptoms is warranted. Even in the absence of a (hypo)manic episode, individuals diagnosed with MDD should be closely monitored and managed. This is because symptoms of MDD are frequently the initial presentation of BD, with factors including greater depression severity, recurrent MDD, and psychotic symptoms associated with a later BD diagnosis (Holma et al., 2008). Other, patient-level risk factors that are indicative of the disorder comprise an earlier age of onset of depressive symptoms, being white, living alone, not being married, and being unemployed (Hirschfeld et al., 2005). While the collection of extensive symptom-and patient-level data may prove difficult in the primary care setting, where time is a luxury, digital technologies may offer an innovative, time-and cost-effective alternative to conventional, interview-based methods.
A comprehensive and careful appraisal of the characteristics of individuals presenting with depressive symptoms is likely to improve biological disease understanding, facilitate patient stratification, and allow for personalized treatment plans and strategies. To this end, we carried out the Delta Study (Olmert et al., 2020), which aimed to develop and validate diagnostic algorithms, based on an online mental health questionnaire and blood biomarker data, that would reduce the misdiagnosis of BD as MDD as well as achieve a more accurate and timely diagnosis of MDD in those presenting with depressive symptoms. We adapted voluntary response sampling and online advertising to meet study recruitment targets estimated from published reports Hantouche et al., 1998;Zuithoff et al., 2010;Sung et al., 2013). The present study examined the resulting prevalence and characteristics of mood disorders in the Delta Study population, determined using the CIDI, and their implications for clinical practice.

| Study design and participants
Data shown in the present work were collected as part of the Delta Study (International Registered Report Identifier RR1-10.2196/18453), an investigator-led study conducted by the Cambridge Centre for Neuropsychiatric Research (CCNR) at the University of Cambridge, which aimed to improve mood disorder diagnoses in participants presenting with depressive symptoms (Olmert et al., 2020). The primary objective of the Delta Study was to identify BD patients among patients diagnosed as having MDD. The secondary objective of the Delta Study was to identify patients with MDD among undiagnosed low mood individuals. To this end, three patient groups were recruited. The first group comprised patients with current depressive symptoms who had recently (≤5 years) (Ghaemi et al., 1999;Morselli et al., 2003) been diagnosed with MDD, the second group comprised participants with current depressive symptoms and no lifetime mood disorder diagnosis, and the third group comprised patients with current depressive symptoms and a previous lifetime BD diagnosis (Figure 1). The study was approved by the University of Cambridge Human Biology Research Ethics Committee (approval number HBREC 2017.11) and was con-

| Online mental health questionnaire
Upon enrolment, participants were asked to complete a purposebuilt online mental health questionnaire available through the Delta Study website. The online mental health questionnaire was developed in collaboration with experienced psychiatrists and a service user advisory group, and was based on existing structured diagnostic interviews as well as a range of mental health screening questionnaires. It consisted of 635 distinct questions divided into six modules: i) demographic information; ii) manic and hypomanic symptoms; iii) depressive symptoms; iv) personality traits; v) psychiatric history and vi) other psychiatric conditions. Participant wellbeing in the past two weeks was assessed using the Warwick-Edinburgh  (Tennant et al., 2007). The online mental health questionnaire was adaptive to answers given by participants, so that only relevant questions were asked, and the maximum possible number of questions asked to an individual was 382 (284 on average). Data collected from the online mental health questionnaire were used to identify participants qualifying for the study's primary and secondary objectives following a pre-defined study protocol. Among the identified participants, those eligible for the diagnostic interview had to: i) consent to providing a blood sample (self-collected dried blood spots) and completing a telephone diagnostic interview; ii) be free from blood-borne illnesses; and iii) have no previous diagnosis of schizophrenia.

| Diagnostic interview
Participants who successfully completed the online mental health questionnaire and returned a dried blood spot (DBS) sample were invited to complete the CIDI, version 3.0 (Kessler & Üstün, 2004) via telephone. The CIDI is a modular diagnostic tool which is widely used in epidemiological studies on mental health (Kessler et al., 2005). It shows good concordance with structured diagnostic interviews conducted by clinicians, such as the Structured Clinical Interview for DSM disorders (SCID; area under the receiver operating characteristic curve (AUC) = 0.75-0.87 for MDD and 0.93-0.97 for BDI and II) (Haro et al., 2006;Kessler et al., 2006), and high interrater and test-retest agreement (86%-100% for MDD and 87%-99% for BDI and II) (Wittchen, 1994). All interviewers conducting the CIDI received in-person training from an external CIDI-certified instructor, as well as internal training and mentoring. Only modules of the CIDI required for lifetime mood disorder diagnosis, that is the screening, depression, and mania sections, were implemented, resulting in six possible outcomes: BDI, BDII, subthreshold BD, MDD, MDD with subthreshold BD, and none, referred to as 'low mood' hereafter. We adopted voluntary response sampling, whereby the CIDI interviews continued until pre-specified study recruitment targets were met.

| Statistical analysis
Power calculations showed that a minimum of 300 participants with a recent diagnosis of MDD by a medical professional and 300 symptomatic participants with no baseline diagnosis of mood disorder were required. Additionally, we aimed to recruit 40 participants with a previous diagnosis of BD made by a medical professional to provide a validation group. Data processing and analysis were conducted in R version 4.0.2 (R Core Team, 2020). Group differences were tested for using the Kruskal-Wallis test for continuous variables and the chisquared test for categorical variables (R package 'tableone' (Yoshida et al., 2020)). Post hoc tests included pairwise group comparisons using Dunn's test (R package 'FSA' (Ogle et al., 2020)) for continuous variables and pairwise chi-squared tests (R package 'rcompanion' (Mangiafico, 2020)) for categorical variables, with Bonferroni correction for multiple comparisons. Figures were prepared in R and Inkscape version 1.0.

| RE SULTS
The study flow diagram is shown in Figure 1. To achieve study   (Table 3). The average time interval (±standard deviation) between the diagnosis of MDD and the diagnosis of BD was 5.4 ± 5.9 years, and the average duration of BD diagnosis was 7.3 ± 6.6 years.

| D ISCUSS I ON
The aim of the present study was to explore the prevalence and characteristics of mood disorders in the Delta Study population.
Furthermore, the finding that individuals with BDI were more likely to smoke in comparison to the MDD group is in line with Li et al. (Li et al., 2017). While we found no significant differences in depression severity between BD and MDD in those with no previous mood disorder diagnosis, participants with BDI previously diagnosed as MDD scored higher on the PHQ-9 relative to MDD and low mood individuals, potentially reflecting the effects of receiving ineffective treatment (e.g., antidepressant monotherapy) and the lack of support due to incorrect diagnosis. Similarly, individuals with BDII previously diagnosed as MDD exhibited elevated depression severity relative to the low mood group. These findings are largely in line with studies revealing a direct association between depression severity and BD (Strober & Carlson, 1982;Holma et al., 2008). In accordance with previous research (Hirschfeld et al., 2005) and, in part, with the results from the baseline low mood group, individuals with BD reported being younger when they experienced their first low mood episode relative to the low mood group, and individuals with BDI reported poorer quality of mental health and wellbeing.

| CON CLUS ION
Taken together, the key findings from the current research are 2-fold.
First, given that depressive episodes in BD and MDD patients with no previous mood disorder diagnosis were largely indistinguishable, standard screening practices must go beyond brief symptom-count checklists, such as the PHQ-9 (Kroenke et al., 2001), when assessing the symptoms of those presenting with depressive symptoms.
Indeed, a careful and comprehensive evaluation of mood states in these individuals is warranted, with a recent machine learning study demonstrating that self-reported symptoms of elevated mood and grandiosity, as well as increased talkativeness and recklessness, can offer excellent discriminatory performance when distinguishing between BD and MDD (Tomasik et al., 2021). While time is a premium in primary care settings, with 50% of the global population spending five minutes or less per visit with their primary care physician (Irving et al., 2017), a highly scalable, low-cost online mental health questionnaire has the potential to facilitate the identification of BD and MDD in those presenting with low mood. Second, in light of the potential for the misdiagnosis of BD, all individuals diagnosed with MDD should be closely monitored and managed, with antidepressant-induced (hypo)mania and non-response to antidepressant medication warranting specialized evaluation by a psychiatrist to rule out a BD diagnosis.

ACK N OWLED G EM ENTS
We are most grateful to all participants of the Delta Study for their generous contributions and making this work possible. We are also

PEER R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/brb3.2167.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data and scripts used in this study are available upon approval from the corresponding authors at sb209@cam.ac.uk or jt455@cam.ac.uk.
Data availability is subject to data license agreements between the