Healthcare Training Institute - Quality Education since 1979
Psychologist, Social Worker, Counselor, & MFT!!

Section 17
The Need for a Bipolar Depression Rating Scale

Question 17 found at the bottom of this page
Answer Booklet | Table of Contents
Get PRINTABLE format of this page

Bipolar disorder is characterized by periods of mania and ⁄ or hypomania, generally interspersed between episodes of depression. The latter is its predominant mood state conferring the bulk of its associated burden and risk of suicide and yet diagnostic distinction is still based on the presence of mania. Like major depression, bipolar disorder is a common psychiatric illness that has been estimated in previous studies to have a lifetime prevalence of less than 1%, but recent studies indicate a lifetime prevalence closer to 6.4% (using criteria that include subsyndromal manic or hypomanic symptoms. However, it is often undetected with a third of patients waiting more than a decade after the onset of the illness before seeking help, and even then almost 70% are repeatedly misdiagnosed. The delay in detection and diagnosis occurs most commonly because of a misdiagnosis of unipolar depression. This is consistent with the fact that most patients with
bipolar disorder first present in the depressive phase of their illness. Although such misdiagnoses are understandable, its consequences both for the individual and those around them are costly. Therefore clinically it is important to be able to detect bipolar depression and discriminate between it and unipolar illness.

Increasingly, bipolar disorder is viewed as a spectrum that comprises a variety of overlapping syndromes. Patients with bipolar disorder that initially present with depressive episodes and are consequently incorrectly diagnosed as having unipolar disorder are often called “converters”. These patients tend to have a more volatile illness than unipolar patients, with only brief periods of being well. Furthermore, they are more likely to have a history of temperamental instability, a younger age of onset, an unstable interpersonal and occupational history and to have suffered postnatal episode. Thus a patient presenting with these clinical features should signal the possibility of bipolarity. Additional features that are thought to indicate bipolar depression include a family history of bipolar disorder, a premorbid hyperthymic personality, the occurrence of atypical depressive features, psychotic episodes, antidepressant-induced mania ⁄ hypomania, antidepressant wear-off and a lack of response to three or more antidepressant treatment trials. In practice, mixed states form a large subgroup of patients within bipolar disorder. However, despite the potential clinical significance of mixed features in cohorts with predominantly depressive presentations, there is little prevalence data for this subgroup. Like bipolar disorder, mixed states form a spectrum that extends from depressive features within mania to manic features within depression, with admixtures in between. The Diagnostic and Statistical Manual of Disorders, 4th edition (DSMIV), defines mixed states as the concurrent presence of a full manic and depressive syndrome, most likely the least common subtype. Indeed the most common manifestation of mixed states is the presence of a few features of one pole of the illness during the polar opposite phase further highlighting the restrictive nature of DSM-IV criteria.

The emergence of manic features in a dominantly depressive presentation is difficult to diagnose. Here there is an overlap with the concept of agitated depression. Patients with depressive mixed states tend to have low scores on mania rating scales and the hyperactivity they experience tends not to be goal-directed. Clinically meaningful signs and symptoms, such as changes in energy, neurovegetative symptoms and distorted cognitions tend to characterize the manic or hypomanic component of such mixed presentations. Conventional depression rating scales make no attempt to detect and measure these symptoms. Clinically the phenomenological separation of unipolar and bipolar depression should be underpinned by differences across other domains. The neuropsychological profiles for instance overlap considerably with patients in both groups demonstrating memory and executive functioning impairment. However, patients with bipolar depression have impaired sustained attention and poor immediate and delayed verbal recall, greater than that found in unipolar depressed patients. Similarly, preliminary functional neuroimaging studies are differentiating patterns of activation in unipolar and bipolar-depressed patients and across the phases of bipolar disorder. However, most of the findings are as yet tentative and await replication.

Depression rating scales
In order to be responsive to the needs of individuals with bipolar depression a well-devised scale must be used that is sensitive to the specific symptoms they experience. Existing scales can be categorized along a number of parameters including syndrome vs. symptom, diagnosis vs. severity, brief vs. comprehensive, Likert vs. visual analogue or other measurement styles, and trained observer vs. self-rating scales. In this paper we will focus briefly on the use of self-report scales with depressed individuals and then examine in greater detail the strengths and limitations of observer rating scales as it is the latter that has greater utility in treatment studies.

Self-report scales
Self-report scales like the Beck Depression Inventory (BDI) the Zung Self Rating Scale and the Internal State Scale allow the assessment of depressive symptoms in bipolar disorder. These are widely used, and are simple to administer across a number of clinical and subclinical populations. Unfortunately, self-report scales can lack reliability and have limited usefulness in patients with diminished concentration, poor motivation, fatigue, advanced age or limited reading skills. Factor analytic studies have found that the structure of the BDI is influenced by socio-economic and diagnostic characteristics of the patient sample. For instance, depressed subjects tend to exaggerate their negative feelings and somatic symptoms, a tendency that is more prominent in some subtypes of depression. In particular, patients that have been depressed for a long time have difficulty comparing their emotions when depressed to their feelings when well, as they may no longer have a benchmark for normality. In bipolar depression perhaps more so than unipolar disorder, insight is often compromised, further limiting the usefulness of self-report methodology. Self-report scales also suffer from inaccuracy because of interpretive errors and non-response. However, perhaps most significant of all is the fact that opportunities that emerge for discussion during an observer-rated interview are inevitably lost during self-report assessment. For these reasons in the present study the rationale for development of a clinician-rated measure for bipolar depression is discussed.

Observer-rated scales
Of the observer-rated scales, two of the best known are the Hamilton Depression Rating Scale and the Montgomery-Asberg Depression Rating Scale (MADRS). The 17-item Hamilton Depression Rating Scale (HAM-D) was developed in the 1960s so as to provide a structured measure of depression severity. Since then it has been adopted universally and is widely considered to be the “gold standard” of clinician-rated depression scales. Numerous studies have assessed its psychometric properties and confirmed its role with respect to rating depressive symptoms that are typical of unipolar depression. Not surprisingly it has also undergone several modifications with 21-item and 31-item versions attempting to capture a broader range of depressive symptomatology. However, the longer a scale becomes the more cumbersome it is to administer and score. This then detracts from its appeal and ultimately limits its utility. This is indirectly borne out by the MADRS, a 10-point observer rated scale designed to reflect change in the severity of depression especially during treatment. It too is easy to administer and score, and like the HAM-D against which it was validated, it has become a global standard. Total MADRS scores when assessing depressed patients demonstrate strong positive correlation with a number of other established instruments such as the Clinical Global Impressions Rating Scale for Severity and Improvement, and the Symptom-Checklist-90. However despite their ubiquitous use these scales have important limitations, particularly in bipolar disorder, that warrant careful consideration.

Firstly, these scales do not evaluate all aspects of depression and unfortunately some of the aspects omitted are particularly relevant to the assessment of bipolar disorder. For example the MADRS does not assess worthlessness, motor retardation or loss of pleasure ⁄ interest. The utility of these scales in the assessment of bipolar depression is hence diminished, especially as the construct of depression itself is not that of a homogeneous entity. Depression is clearly heterogeneous and multifaceted and the factors brought together to measure and evaluate depression vary from scale
to scale. Strong convergent validity across scales only indicates that they are measuring the same constructs, not that the full scope of constructs are covered, or that these are all of the most clinically relevant constructs. Importantly, there is no means of gleaning the clinical salience of the measure. For example, features that are not mood-specific, such as functional, genital or gastrointestinal-related symptoms, can contribute to a significant proportion of the total HAM-D score. This means that in some cases the severity of depression can be determined by items that may in fact be associated with something other than depression, resulting perhaps in total scores being a misleading index of severity, and even to misclassification. Weight loss for example does not necessarily equate with depression severity just as a reduced anxiety does not connote diminished suicidality. Discriminant validity is further confounded by comorbidity as shown in a study of elderly patients in which investigators found that high scores on eight somatic items from the HAMD exaggerated the total scores of depression when in fact these symptoms were evidence of concurrent medical illness. Scales weighted towards somatic symptoms risk reflecting side-effect profiles of, for example, the atypical agents, reflecting items measuring appetite, weight and sleep. Scales weighted to core symptomatology may be better able to track true change.

Another limitation is that atypical depressive symptoms such as hyperphagia and hypersomnia are not included in the 17- item HAM-D or the MADRS. These symptoms are disproportionately common in bipolar disorder, and the HAM-D only provides unidirectional measures of sleep and appetite, and is excessively weighted with regard to the former. The pertinence of these symptoms to the diagnosis of atypical depression is in itself a topic of some discussion however, such omissions clearly have implications for delineating depressive subtypes and capturing depressive symptomatology across phenotypes. In patients with bipolar disorder mixed states are relatively common with up to a third meeting criteria for a mixed episode and almost half having a lifetime history of mixed episodes. Of particular interest in the latter study is the finding that HAM-D rating failed to discriminate, both on individual items and total score, patients with a mixed episode from those with depression alone. None of the unipolar rating scales have any items that identify mixed state constructs, despite the frequency with which they occur and their clinical significance. The lack of a recognized instrument that measures depressive mixed episodes may add to their misidentification. Constructs including irritability, lability, increased speech and motor drive and agitation are typical of mixed states and merit concurrent assessment. Clearly, the HAM-D is not a “one size fits all” instrument and its use as such raises serious concerns as regards its symptom sensitivity and phenotype specificity. Items such as hypochondriasis and insight have been criticized in their ability to gauge depression severity and are not symptoms commonly described in bipolar cohorts. Furthermore, the value of rating weight loss in hospitalized patients has been questioned as hospital staff routinely strive to prevent weight loss. Indeed it is weight gain that is more frequently present in atypical depression and is more frequent in bipolar than unipolar individuals. Such contextual elements are also important in the rating of items such as sexual interest, which patients tend to underrate when away from their partners. Despite the introduction of structured guidelines, interrater reliability in the HAM-D has been found to be problematic with raters at different facilities being trained using separate guidelines and versions. Items such as those involving subjective evaluation are inherently difficult to interpret even for trained observers and the scoring of items is somewhat idiosyncratic. For instance, depressed mood rates a maximum of 4 compared
to 6 points for sleep disturbance.

Naturally, the robustness of the MADRS in rating overall depression and in particular change in total score is achieved at the expense of specificity for symptoms and its brevity that makes it so easy to administer sacrifices comprehensiveness. In common with the HAM-D the validity of some MADRS items has been questioned, namely sleep disturbance, reduced appetite and suicidal ideation. The MADRS also lacks discriminant validity to the extent that total scores equating to a diagnosis of Major Depression have been described when rating patients with bulimia nervosa. Recently it has been suggested that because the MADRS was developed to be maximally sensitive to change in treatment following the administration of older antidepressants (amytryptiline, clomipramine, mianserin and maprotiline) it cannot be assumed to be as sensitive to changes that occur in response to the newer classes of antidepressants. Clearly this has significant implications for its use in pharmaceutical research especially as there is little consensus with respect to a normative cut-off score for defining remission. Similar problems exist with the HAM-D resulting in different cut-off points being suggested for different disorders including stroke, Parkinson’s Disease and Alzheimer’s Disease. This further compounds the problem of interrater reliability. It seems that Hamilton’s instructions indicating that the scale was devised solely for the assessment of primary depressive illness have been largely forgotten. As awareness of differences between unipolar and bipolar depression increases it highlights the necessity of being aware of the limitations of commonly used assessment tools.

Conclusion
Continued investigation of the symptoms that individuals experience has allowed diagnosis and treatment of depression to progress, and the continuing expansion of the DSM and ICD classifications attests to the growing knowledge base underpinning the differentiation of subcategories of disorders. It is therefore necessary that more sophisticated means of assessing these symptom differences and their respective responses to change be developed. Scales developed specifically for particular populations are likely to better identify and weight symptoms that characterize that disorder, and to accurately track change. As discussed, bipolar patients often first present with depressive symptoms, which can be readily mistaken for unipolar depression. The corollary of this is that medication appropriate to unipolar depression is initiated and this is often detrimental to patients with bipolar disorder. Antidepressant treatment is generally less effective and risks precipitating mania, mixed states and rapid cycling as well as the additional burden of misdiagnosis for both patient and practitioner alike. A specific instrument more sensitive to the array of bipolar depressive symptoms would advance both research and clinical goals. With this in mind, we are currently developing and validating a new Bipolar Depression Rating Scale (BDRS), which will assess aspects of bipolar depression and mixed phase presentations.
- Berk, M et al; Scale Matters: The Need for a Bipolar Depression Rating Scale; Acta Psychiatrica Scandinavica; Sep2004 Supplement 422, Vol. 110, p 39

=================================
Personal Reflection Exercise #3
The preceding section contained information about the need for a bipolar depression rating scale.  Write three case study examples regarding how you might use the content of this section in your practice.

QUESTION 17
According to Berk, what is the main disadvantage caused by not having a universal bipolar depression rating scale? Record the letter of the correct answer the Answer Booklet.