BACKGROUND

This appendix sets out an approach to the classification of depression that was used in the development of the guideline update (including the analysis of the evidence and the development of recommendations) and will be of value in routine clinical use.

Depression is a heterogeneous disorder in which a number of underlying presentations may share a common phenomenology but have different aetiologies. Despite considerable work on the aetiology of depression including neurobiological, genetic and psychological studies, no reliable classificatory system has emerged that links either to the underlying aetiology or has proven strongly predictive of response to treatment. A number of classification systems/subgroupings have been used, including reactive and endogenous depression, melancholia, atypical depression, depression with a seasonal pattern/seasonal affective disorder and dysthymia. These have been based on varying combinations of the nature, number, severity, pattern and duration of symptoms, and in some cases the assumed aetiology. Over time pragmatic definitions have emerged, enshrined in the current two major classification systems, DSM–IV-TR (APA, 2000c) and ICD–10 (WHO, 1992). These have defined a threshold of severity of clinical significance with further classification in terms of severity (for example, mild, moderate or severe as adopted in DSM–IV with regard to major depressive disorder), duration and course of the disorder (for example, recurrent, presence of residual symptoms) and subtype based on symptom profile (for example, melancholic, atypical). Other aspects of depression such as response to treatment (for example, treatment resistant, refractory) and aetiology (for example, preceding life events) do not feature specifically in the classifications and lack accepted definitions, although are used in clinical practice. The classification has some use in describing likely outcome and course (Khan et al., 1991; Barrett et al., 2001; Sullivan et al., 2003; Blom et al., 2007; Jackson et al., 2007; Conradi et al., 2008; Holma et al., 2008; Van et al., 2008) although social support, social impairment or personality factors also need to be taken into account. Lower severity and duration of a depressive episode predicts, to some extent, a greater likelihood of spontaneous or earlier and eventual improvement whereas greater severity, chronicity and number of previous episodes predict a higher chance of subsequent relapse.

The lack of a highly reliable or valid classificatory system has significant and practical clinical consequences, particularly in primary care where the full range of depression presents. A major concern is whether depression should be classified using dimensions or categories. Categories help distinguish cases from non-cases, while dimensions help distinguish severe disorder from mild (Cole et al., 2008). Clinicians are often required to make a categorical decisions – for example to treat with antidepressants or not, to refer for further interventions or not – and consequently there can be pressure to interpret data on a single dimension in a categorical way, for example, treat or not treat based solely on a symptom severity rating (for example, a PHQ-9 score alone). This conflicts with the recognised need to take multiple factors/dimensions into consideration within a consultation, including the patient’s view on the cause of symptoms and acceptable treatment, and in the guideline update a major challenge has been to provide a useful categorisation that adequately captures the complexity.

CLASSIFICATION OF DEPRESSION AND NICE GUIDANCE

The approach adopted in the previous depression guideline (NICE, 2004a; NCCMH, 2004) was based on ICD–10 and rested on a dimensional approach based on a symptom count further elaborated by taking into account the presence of social role impairment and the duration of both symptoms and social impairment. The subsequent categorisation of depression into mild, moderate and severe has led to a number of concerns in practice. First this classification appears to have often been implemented with an emphasis on a symptom count alone with other important factors such as duration and social impairment ignored, although it should be noted that in general there is a relationship between the number of symptoms and severity of functional impairment (Faravelli et al., 1996). Second it implies that the different symptoms experienced are equivalent, although, in fact, symptom patterns may be important. Third, it does not take into account illness duration and course. This tendency may be exacerbated by the use of measures such as the Patient Health Questionnaire (PHQ-9; Kroenke et al., 2001) or Hospital Anxiety and Depression Scale (HADS; Zigmond & Snaith, 1983) under the Quality and Outcomes Framework (Department of Health, 2004).

A drawback inherent in using ICD–10 depression criteria is that most of the treatment research on which the guideline has to be based uses DSM–IV or previous, essentially similar, versions of DSM (DSM–III and DSM–III-R) criteria. As discussed below, the criteria are similar but not identical, and this has particular relevance for the ‘threshold’ of the diagnosis of a clinically significant depressive episode and therefore what are considered subthreshold depressive symptoms.

DIAGNOSIS OF A DEPRESSIVE/MAJOR DEPRESSIVE EPISODE

The criteria for diagnosing depressive episodes in ICD–10 and DSM–IV overlap considerably but have some differences of emphasis. In ICD–10 the patient must have two of the first three symptoms (depressed mood, loss of interest in everyday activities, reduction in energy) plus at least two of the remaining seven symptoms; while in DSM–IV the patient must have five or more out of nine symptoms with at least one from the first two (depressed mood and loss of interest). Both diagnostic systems require symptoms to have been present for at least 2 weeks to make a diagnosis (but can be shorter in ICD–10 if symptoms are unusually severe or of rapid onset). In both ICD–10 and DSM–IV the symptoms must result in impairment of functioning that increases with the episode severity. Table 143 compares the symptoms required in ICD–10 and DSM–IV.

Table 143

Comparison of symptoms of depression in ICD–10 and DSM–IV.

DETERMINING SEVERITY OF A DEPRESSIVE/MAJOR DEPRESSIVE EPISODE

Both ICD–10 and DSM–IV classify clinically important depressive episodes as mild, moderate and severe based on the number, type and severity of symptoms present and degree of functional impairment. Table 144 shows the number of symptoms required by each diagnostic system, which are less specific than DSM–IV. The prescriptive symptom counting approach of ICD–10 tends to lend itself to using symptom counting alone to determine severity.

Table 144

Number of symptoms required in ICD–10 and DSM–IV for a diagnosis of depressive episode/major depression (but note they also need assessment of severity and functional impairment to ascertain diagnosis and severity).

As ICD–10 requires only four symptoms for a diagnosis of a mild depressive episode, it can identify more people as having a depressive episode compared with a DSM–IV major depressive episode. One study in primary care in Europe identified two to three times more people as depressed using ICD–10 criteria compared with DSM–IV (11.3% versus 4.2%; Wittchen et al., 2001a). However another study in Australia (Andrews et al., 2008) found similar rates using the two criteria (6.8% versus 6.3%) but slightly different populations were identified (83% concordance), which appears to be related to the need for only one of two core symptoms for DSM–IV but two out of three for ICD–10. These studies emphasise that, although similar, the two systems are not identical and that this is particularly apparent at the threshold taken to indicate clinical importance.

DIAGNOSIS OF SUBTHRESHOLD DEPRESSIVE SYMPTOMS

Given how common milder forms of depression are, and the problems inherent in defining a ‘threshold’ of clinical importance because of the diagnostic system differences and the lack of any natural discontinuity identifying a critical threshold (Andrews et al., 2008), this guideline update has broadened its scope to include consideration of depression that is ‘subthreshold’, that is, does not meet the full criteria for a depressive/major depressive episode. A further reason is that subthreshold depression has been increasingly recognised as causing considerable morbidity and human and economic costs, is more common in those with a history of major depression and is a risk factor for future major depression (Rowe & Rapaport, 2006).

There is no accepted classification for this in the current diagnostic systems with the closest being minor depression, a research diagnosis in DSM–IV. At least two but less than five symptoms are required, of which one must be depressed mood or diminished interest. This includes ICD–10 depressive episode with four symptoms and, given the practical difficulty and inherent uncertainty in deciding thresholds for significant symptom severity and disability, there is no natural discontinuity between minor depression and mild major depression in routine clinical practice. There is however a danger of ‘medicalising’ distress by adopting minor depression as a discrete diagnosis, which would inevitably broaden the concept of depression. For this guideline update the GDG therefore use the term ‘subthreshold depressive symptoms’ to avoid this problem while providing a way of describing this part of the depressive spectrum.

Both DSM–IV and ICD–10 do have the category of dysthymia, which consists of depressive symptoms which are subthreshold for major depression but which persist (by definition for more than 2 years). There appears to be no empirical evidence that dysthymia is distinct from subthreshold depressive symptoms apart from duration of symptoms.

ICD–10 has a category of mixed anxiety and depression, which is less clearly defined than minor depression, and is largely a diagnosis of exclusion in those with anxiety and depressive symptoms subthreshold for specific disorders. Not unexpectedly it appears to be a heterogeneous category with a lack of diagnostic stability over time (Wittchen et al., 2001b; Barkow et al., 2004). For this reason it has not been included in this guideline.

DURATION

The duration of a depressive episode can vary considerably among individuals. The average course of an untreated depressive episode is between 6 and 8 months with much of the improvement occurring in the first 3 months, and 80% recovered by 1 year (Coryell et al., 1994). There is evidence to suggest that patients who do not seek treatment for their depression may recover more quickly than those who seek but do not receive treatment (Posternak et al., 2006). There is also some evidence to suggest that people who do not seek help have a shorter mean duration of depressive episode (Posternak et al., 2006).

Traditionally the minimum duration of persistent symptoms for major depression is 2 weeks and for chronic depression (or dysthymia) 2 years. These conventional definitions have been adopted in the absence of good evidence as there is only a modest empirical base for the minimum duration (for example, Angst & Merikangas, 2001) and none that we could find for the ‘cut-off’ between acute and chronic depression. As with severity, duration is better thought of as a dimension with a decreased likelihood of remission with increasing chronicity over a given time frame (Van et al., 2008). The conventional criteria are therefore better viewed as guides rather than cut-offs. It is likely that that the minimum duration after which therapy provides more benefit than occurs by spontaneous improvement is somewhat longer than 2 weeks (possibly 2 to 3 months, Posternak et al., 2006), but this has never been tested empirically. By 2 years it does appear that outcome is poorer, supporting consideration of chronicity in describing the disorder; nevertheless the point at which acute becomes chronic is not clear, and indeed may not be a meaningful question. There is some evidence that outcome is poorer after about 1 year (for example, Khan et al., 1991). However there seems little to be gained by redefining duration for the guideline as long as it is recognised that the conventional definitions are merely signposts to include consideration of duration in relation to outcome and need for treatment.

COURSE OF DEPRESSION

An influential model of the course of major depression proposes that the onset of an episode of depression consists of a worsening of symptoms in a continuum going from depressive symptoms through to major depression. Phases of improvement with treatment consist of response (significant improvement) to remission (absence of depressive symptoms) which if stable for 4 to 6 months results in (symptomatic) recovery, meaning that the episode is over (Frank et al., 1991). It is important to distinguish this use of recovery from more recent concepts related to quality and meaning of life in spite of continued symptoms. After recovery a further episode of depression is viewed as a recurrence to distinguish it from a relapse of the same episode. There has been no consensus as to how long a period of remission should be in order to be able to declare recovery; different definitions result in different definitions of episode length and time to full or subthreshold depressive recurrence (Furukawa et al., 2008). Therefore, in practice it can be difficult to distinguish between relapse and recurrence, particularly when people have mild residual symptoms. Follow-up studies of people with depression have shown that, overall, more time is spent with subthreshold depressive symptoms than major depression and there is a variable individual pattern ranging from persisting chronic major depression, through significant but not full improvement (partial remission), to full remission and recovery (Judd et al., 1998). DSM–IV defines full remission when there has been an absence of symptoms for at least 2 months. For partial remission, full criteria for a major depressive episode are no longer met, or there are no substantial symptoms but 2 months have not yet passed. DSM–IV specifies ‘with full inter-episode recovery’ if full remission is attained between the two most recent depressive episodes and ‘without full inter-episode recovery’ if full remission is not attained. In DSM–IV, therefore, separate episodes are distinguished by at least 2 months of not meeting major depression criteria, which is in contrast to the more stringent ICD–10 requirements of 2 months without any significant symptoms. There is therefore some ambiguity as to whether full remission is required to define separate episodes.

Nevertheless the number of episodes and degree of symptom resolution have important implications for considering the course of an individual patient’s depressive disorder. The risk of a further episode of major depression within a given time frame is greater with an increasing number of previous episodes (Solomon et al., 2000; Kessing & Andersen, 2005) and also if there has not been full remission/symptomatic recovery (Paykel et al., 1995; Kanai et al., 2003; Dombrovski et al., 2007). If someone presents with minor depressive symptoms it is therefore crucial to determine whether or not this directly follows an episode of major depression.

DEPRESSION SUBTYPES

Different symptom profiles have been described and are included in the classification systems. In DSM–IV, severe major depression can be without or with psychosis (psychotic depression) and there are specifiers that include melancholia, atypical features, catatonia, depression with a seasonal pattern (seasonal affective disorder) and post-partum onset. ICD–10 also provides specifiers for psychotic and somatic symptoms, the latter similar to DSM–IV melancholia. However, these subtypes do not form distinct categories (for example, Kendell, 1968; Angst et al., 2007) and they add a further complexity to the diagnosis of depression. The GDG judged that these specifiers were best considered where appropriate after the diagnosis of a depressive disorder is made and they are not discussed in detail here. Some specifiers, particularly psychosis and seasonal pattern depression, have potential treatment implications and are considered in the guideline update where evidence is available.

CLASSIFICATION OF DEPRESSION IN THE GUIDELINE UPDATE

The depression classification system adopted for the guideline update had to meet a number of criteria, notably the use of:

• a system that reflects the non-categorical, multidimensional nature of depression
• a system that makes best use of the available evidence on both efficacy and effectiveness
• a system that could be distilled for practical day-to-day use in healthcare settings without potentially harmful over-simplification or distortion
• terms that can be easily understood and are not open to misinterpretation by a wide range of healthcare staff and service users
• a system that would facilitate the generation of clinical recommendations.

These criteria led the GDG to adopt a classificatory system for depression based on DSM–IV criteria. When assessing an individual it is important to assess three dimensions to diagnose a depressive disorder – a) severity (symptomatology and social impairment), b) duration, and c) course – as linked, but separate, factors (see below). In addition there was recognition that a single dimension of severity was insufficient to fully capture its multidimensional nature.

As discussed above the following depressive symptoms require assessment to determine the presence of major depression. The symptoms need to be experienced to a sufficient degree of severity and persistence to be counted as definitely present. At least one core symptom is required; both core symptoms would be expected in moderate and severe major depression.

CLASSIFICATION IN RELATION TO DEPRESSION RATING SCALES AND QUESTIONNAIRES

Depression rating scales and questionnaires give ranges that are proposed to describe different severities of depression. Some of these were described in Appendix 13 of the previous guideline. In reconsidering this for the update it quickly became apparent, not only that there is no consensus for the proposed ranges, but also that the ranges in different rating scales and questionnaires do not correspond with each other. In addition there is a variable degree of correlation between different scales, which indicates that they do not measure precisely the same aspects of depression. When these factors are added to the need to consider more than symptoms in determining severity, and more than severity in considering diagnosis, the GDG was concerned not to perpetuate a spurious precision in relating scores in depression rating scales and questionnaires to the diagnosis or severity of depression, which must in the end be a clinical judgement.

Nevertheless it is necessary to try and translate trial evidence (which may only provide rating scales or questionnaire scores) into a meaningful clinical context as well as relating this guideline update to the previous guideline which used the APA (2000a) cut-offs. The change to DSM–IV-based diagnosis and the inclusion of minor depression (subthreshold depressive symptoms) in the update means that the descriptors of ranges previously given are no longer tenable. Table 145 gives the descriptors and ranges used in this guideline update, with the important caveat that these must not be taken as clear cut-offs or a short-cut to classify people with depression.

IMPLICATIONS OF THE PROPOSED CLASSIFICATION

An important implication is that symptom counts alone (for example, using the PHQ-9) should not be used to determine the presence or absence of a depressive disorder although this is an important part of the assessment. The score on a rating scale or questionnaire can contribute to the assessment of depression and rating scales are also useful to monitor treatment progress.

Another very important point to emphasise is that making a diagnosis of depression does not automatically imply a specific treatment. Making and agreeing a diagnosis of depression is a starting point in considering the most appropriate way of helping that individual in their particular circumstances. The evidence base for treatments considered in this guideline are based primarily on RCTs in which standardised criteria have been used to determine entry into the trial. Patients seen clinically are rarely assessed using standardised criteria reinforcing the need to be circumspect about an over-rigid extrapolation from RCTs to clinical practice.

Diagnosis using severity, duration and course (see above) necessarily only provides a partial description of the individual experience of depression. People with depression vary in the pattern of symptoms they experience, their family history, personalities, pre-morbid difficulties (for example, sexual abuse), psychological mindedness and current relational and social problems – all of which may significantly affect outcomes. It is also common for people with depression to have a comorbid psychiatric diagnosis, such as anxiety, social phobia, panic and various personality disorders (Brown et al., 2001), and physical comorbidity, or for the depression to occur in the context of bipolar disorder (not considered in this guideline). Gender and socioeconomic factors account for large variations in the population rates of depression, and few studies of pharmacological, psychological and other treatments for depression control for or examine these variations. This emphasises that choice of treatment is a complex process and involves negotiation and discussion with patients. Given the current limited knowledge about which factors are associated with better antidepressant or psychotherapy response, most decisions will rely upon clinical judgement and patient preference until there is further research evidence. Trials of treatment in unclear cases may be warranted but the uncertainty needs to be discussed with the patient and benefits from treatment carefully monitored.