Screening in Diabetes Care: Detecting and Managing Depression in Diabetes

The analysis of depression screening in diabetes according to the four criteria of the United Kingdom’s National Screening Committee shows that both screening tests and treatment options are available. However, results of the Cochrane meta-analysis about depression screening in primary care settings indicate that the implementation of depression screening needs a structured approach to link these two components. A stepped-care approach comprising verification of positive screening results, treatment options, assessment of response to treatment, and adaptation may carry favorable results with regard to reduction of depression as well as cost-effectiveness. The association between diabetes and distress has long been recognized. In 1685 Thomas Willis, a British physician, suggested that diabetes might be a consequence of prolonged sorrow. In the middle of the 20th century Alexander regarded diabetes as one of the seven major psychosomatic diseases. In more recent years these historical observations have been supported by growing empirical evidence of a special relationship between emotional distress and diabetes. A meta-analysis regarding depression and diabetes onset showed that the presence of depressed symptoms increased the risk of developing diabetes by 37%. However, the effect is bidirectional. Meta-analytic findings suggest that the comorbidity of depression and diabetes is frequent: approximately one third of diabetic patients report symptoms of depression, and a smaller group of10%of diabetic patients meet the criteria of a clinical depression. In diabetes care settings the recognition rate of depression in diabetic patients is disappointingly low, ranging between20%and50%. Even in more specialized diabetes care settings approximately 50% of depressed diabetic patients remain undetected. Thus, there are strong and compelling arguments in favor of depression screening in diabetes, and this is also recommended by several guidelines for diabetes care (Fig. 16.1). However, there are also arguments against depression screening. Studies analyzing the effectiveness of depression screening in primary care settings do not all support large-scale implementation of depression screening. Increasingly, there is a need to justify depression screening in different medical conditions with regard to its effectiveness and ethical and clinical implications and to specify whether screening as a routine or more selective case-finding is warranted.12 Screening for depression potentially exposes both false positives and true positives (but otherwise unrecognized cases) to stigmatization and potential discrimination by health insurance companies or employers.

Screening for Depression in Medical Settings: Are Specific Scales Useful?

There are two broad strategies for screening and quantifying depression in medical settings. The first approach is replying upon measures developed in psychiatric samples, and the second approach is to concede that symptoms are substantially different and to develop customized scales. Here we discuss the merits of several specific scales for measuring depression in physical settings and make the case for scales tailored to specific populations. A subsequent chapter (Babaei and Mitchell) will present a contrasting position. There are two broad strategies for screening and quantifying depression in medical settings. The first approach involves using measures developed in psychiatric samples and assuming that their relevance holds. The second approach is to concede that there are intrinsic limitations to extrapolating those ‘‘general’’ measures to medically ill populations. In the former case the hypothesis is that symptoms of depression are essentially the same when depression occurs with and without physical illness. In the latter case the hypothesis is the symptoms are substantially different. Pursuing the latter, there are two key concerns. Firstly, such an approach assumes some constancy to the nature of depression across differing psychiatric and medical settings. Depression, however, is difficult enough to define in psychiatric patient samples. Even ignoring the debate as to whether depression is viewed as comprising a set of subtypes or is best modeled along a continuum, quantifying clinical depression remains problematic, as detailed elsewhere in this book. Over the past few decades, clinical depression has most commonly been viewed as synonymous with major depression, but, as numerous studies have shown, comparable symptomatic distress and disability associated with major depression and minor depression—and even with subsyndromal depression—begs an obvious question: Can imposing a cutoff score on a dimensional measure of depression accurately distinguish true cases and true non-cases in a psychiatric sample? Further, assuming that a cutoff is derived with an acceptable classification rate, can we extrapolate decision rules derived from psychiatric samples to screen and quantify depression caseness in the medically ill? As measures that have been widely used for decades (such as the Zung and the Beck Depression Inventory) generate widely differing cutoff scores across psychiatric, general practice, and medical settings, there would appear to be quantitatively and possibly qualitative differences to the nature of depression in medical contexts, making general measure extrapolation problematic.

Technological Approaches to Screening and Case Finding for Depression

What are the strengths and weaknesses of computer-based and other automated methods of detecting depression? Two promising technologies make use of the Internet and speech recognition. Whatever technology is used, each method needs to be assessed rigorously using the same high standards that have been applied to pencil-and-paper tests. We are in the midst of a technological revolution that inevitably will transform psychiatric clinical practice. A consensus for routine depression screening is building, and at the same time methods by which it could be accomplished are emerging. The hope is that the right technology can provide an easy, inexpensive, valid, and reliable public health approach to depression screening. Computerized assessment is well accepted in diverse fields, and the use of Internet-based survey technology has grown exponentially. Issues regarding the strengths and limitations of computerized assessments are addressed regularly in the literature. For example, such assessments have been shown to improve data quality while at the same time reducing cost as well as the time to score, analyze, and report results. Increasingly, as depressive disorders have been recognized as highly prevalent with significant morbidity, multiple screeners using an array of technological advances have been developed (Table 8.1 lists selected studies). This chapter will review the technologies that are currently available for automated depression screening and will discuss them in terms of criteria that should dictate their adoption. The growing list of technologies can be classified on several dimensions. Perhaps themostimportant of these isadaptivevs.non-adaptive. Inanadaptive technology pioneered by the Educational Testing Service, a computer, using a preprogrammed algorithm, decides which question to ask next given the responses so far. Paper-and-pencil is the classical non-adaptive technology— everyone gets the same paper with the same questions in the same order. Technological modality is a second dimension. Currently available technologies include the phone, the Internet, and hand-held electronic devices. The phone can be split into several groups, including agent: computer-assisted telephone interview (CATI), speech recognition, and touch-tone. Phone can also be classified as inbound (the patient initiates the call to a toll-free number) or outbound (the system initiates the call).

Overview of Depression Scales and Tools

There have been a large number of depression tools published for the purposes of detecting depression or rating its severity. Choosing between them is difficult without adequate information on their validity, reliability, and acceptability. Recently, ever-shorter-version mood measures have been released. Is a shorter scale a better scale? It is important to study each method against our best standard and ideally compare scales head to head to judge the optimal scale for each situation. Clinicians and researchers have developed a bewildering number of tools for the assessment of depression. These are most often questionnaires designed to help elicit symptoms of depression for the purpose of screening, diagnosis, and monitoring progress (Textbox 2.1). Although we often use the terms screening, diagnosis, and case-finding interchangeably, in an epidemiologic sense screening refers to the attempted detection of disorder in those who had not sought testing or did not suspect they had a particular condition. Often a screening test is not usually intended to be diagnostic, in that those with suspicious findings may be referred for more definitive examination. The latter is perhaps better known as case-finding. This means a screening tool can favor negative predictive value (NPV) over positive predictive value (PPV) (see Chapter 5). In both screening and case-finding the test may be applied ‘‘routinely’’ to all cases, or selectively to those thought to be at high risk. A screening test applied to many individuals should be as simple as possible to retain high uptake, and positive results must be paired with an acceptable next step. A case-finding measure may be more involved but should still consider acceptability. Adoption of a test in clinical practice probably depends more on acceptability than accuracy. During the past five decades there has been a considerable effort to improve the methods used to detect and quantify depression (Textbox 2.2). Some scales, such as theCronholm-Ottosson Depression Scale, have fallen into obscurity,while others, such as the Hamilton Depression Rating Scale and the Beck Depression Inventory, have each been cited over 10,000 times. Given that there are so many similar depression scales, it is not surprising that clinicians have trouble choosing between them.

Screening for Depression in Perinatal Settings

Implementing screening in perinatal settings poses a potentially complex set of issues, but screening is nonetheless increasingly being recommended and even mandated. When should screening occur—during pregnancy, postpartum, or both? What instrument should be used? How acceptable is screening to mothers? What difference does screening make to the management of postpartum depression? This chapter presents an evidence-based approach to all aspects of perinatal screening. Over the past 20 years there has been considerable interest in psychiatric disorders arising during the course of pregnancy and following childbirth. Most of the attention has been focused on depressive disorders arising within the first 3 months to 1 year after childbirth, commonly referred to as postnatal or postpartum depression. Pregnancy was once thought to be protective against depressive symptoms; however, women are just as likely to experience depressive symptoms while pregnant as they are during the postpartum period. The mean prevalence of antenatal depression is between 10.7% and 12%, with increasing prevalence and severity through the second and third trimesters. This is comparable with the 10% to 15% of women who develop postpartum depression. While the DSM-IV official recognition of postpartum depression arising after childbirth is confined to a postpartum specifier for those episodes of major depression that have an onset within 4 weeks after delivery, increasing knowledge of depression during the antenatal period has given rise to its equally important early recognition and treatment. Whatever the specifier of postpartum depression in the DSM-IV, depression at this time has been granted considerable importance because of its potential adverse impacts upon child development and maternal morbidity and mortality; and because of the treatment challenges inherent in pregnant and breastfeeding women. Even though the consequences of postpartum depression have been recognized, the illness itself is frequently not identified; it has been estimated that between 50% and 75% of the women suffering from postpartum depression will have it identified and potentially treated. More recent work has focused attention on depression during the course of pregnancy, so-called antenatal depression. However, the validity of measuring depression during pregnancy and in the postpartum period is not clear, especially the boundary between depressive symptoms and clinically significant depressive disorder.

Screening for Depression in Cancer Care

There is an increasing awareness of the importance of screening for depression and distress in oncology settings. Researchers have devised quick and simple methods for assessing symptoms in a wide range of patients that are acceptable to both patients and providers, and introduced computerized systems that make it possible to quickly screen a large number of patients efficiently. A large body of data concerning implementation of screening in cancer care seems to suggest that screening can serve to stimulate discussions of psychosocial and mental health issues between patients and oncology staff, but whether screening affects patient outcomes is still unclear. As with many other medical populations, people suffering with cancer are susceptible to clinical depression. More so than many other illnesses, cancer is associated with a poor prognosis and is in many ways synonymous with fear. References to the ‘‘Big C’’ and hushed tones prevailed in the past when a diagnosis of cancer was discussed, and remnants of these attitudes are still prevalent in many communities and countries worldwide. Hence, beyond the burden of the tumor and the associated treatment, the psychological toll of cancer is significant. Two thorough reviews of the prevalence of depression in cancer have been published in the past few years, and in addition to several other general reviews, summaries are available with reference to specific types of cancer (prostate, pancreatic, advanced disease) and specific patient groups (children and the elderly). Taking methodologic issues into consideration, the point prevalence of major depressive disorder and depression symptoms comorbid with cancer is most commonly cited between 10% and 25%.2 This rate varies considerably depending on how depression is measured (standard clinical interview, questionnaire), how it is conceptualized, the criteria used to define depression, the types of patients assessed (cancer type, demographics, inpatients versus outpatients), and the point during the cancer treatment trajectory when assessment occurs. Massie summarized 88 papers investigating depression prevalence in cancer patients: the highest rates of depression were found in head and neck, pancreatic, breast, and lung cancer patients (up to 50% of all patients), with lower rates generally reported in colon cancer, gynecologic cancer, and lymphomas (rates from 8% to 25%). The cancers with higher rates of depression generally have less positive prognoses (pancreatic, lung) or involve disfiguring treatments (head and neck, breast), perhaps explaining these discrepancies.

Screening for Depression in Neurologic Disorders

Depression appears to be particularly common in several neurologic disorders, including epilepsy, stroke, dementias, Parkinson’s disease, Huntington’s disease, and multiple sclerosis. There is some evidence that the ‘‘depression’’ associated with each neurologic disorder is distinct in symptoms and course. This suggests it may be useful to have depression scales validated for each neurologic disorder, yet most instruments appear to yield comparable acceptable sensitivities and specificities. However, head-to-head comparisons of scales and implementation studies are needed to resolve this issue. Depressive disorders are a common psychiatric comorbidity of neurologic disorders, including epilepsy, stroke, dementias, Parkinson’s disease (PD), essential tremor, Huntington’s disease, migraines and multiple sclerosis (MS), to name the principal ones. It is typically assumed that depressive disorders are a complication of these neurologic disorders. However, data published in the past 15 years have suggested a bidirectional relation between depression and stroke, epilepsy, dementia, and PD. In other words, not only are patients with these neurologic conditions at greater risk of developing depression, but patients with depression are at greater risk of developing one of these disorders. Early identification of comorbid depressive disorders is of the essence given their negative impact on quality of life and the course and response to treatment of most of these neurologic disorders. Unfortunately, depression often goes unrecognized and hence untreated. Clearly the use of screening instruments by neurologists may help remedy this problem. Several caveats need to be considered, however. First, the clinical presentation of comorbid depressive disorders may differ in several ways from that of primary depression, such as in cases of depression in epilepsy. Second, several somatic and cognitive symptoms are common in primary depression and most neurologic disorders (ie, fatigue, poor concentration, and slow thinking). Thus, a higher score may be a reflection of such symptoms and not of a depressive episode per se. Third, most of the available screening instruments for depression were developed for primary mood disorders and hence may yield false-positive or -negative findings.

Screening in Cardiovascular Care

There is great interest in screening in cardiovascular settings but little evidence that implementation of screening will affect depression or cardiac outcomes despite the epidemiologic evidence that depression predicts cardiac events and mortality. Since this chapter was accepted, in October 2008 the American Heart Association (AHA) Working Group published a Scientific Advisory recommending that all patients with cardiovascular disease be screened for depression, although this recommendation was not based on a systematic review of the evidence. Several weeks after release of the Scientific Advisory, a systematic review of depression screening in cardiovascular care was published but did not find evidence that patients with cardiovascular disease would benefit from screening for depression. The authors of the review noted that no published trials have assessed whether screening for depression improves depressive symptoms or cardiac outcomes in patients with cardiovascular disease, suggesting that the recommendations of the AHA Scientific Advisory were premature. High rates of depression were first documented among patients with cardiovascular disease (CVD) in the late 1960s. Early research on depression in CVD focused on patients with acute myocardial infarction (AMI) and conceptualized depression as an acute reaction to a catastrophic medical event. In the 1990s, groundbreaking work by Frasure-Smith and colleagues demonstrated a connection between major depression during hospitalization for AMI and subsequent mortality. Since then, many other studies have identified major depression or depressive symptoms as risk factors for mortality and recurrent cardiac events among patients with AMI or unstable angina pectoris (together known as acute coronary syndromes [ACS]) even after controlling for other known risk factors, although not all studies have reported a significant association. Other studies have reported that depression among patients with ACS is related to decreased quality of life and poor adherence to secondary prevention behaviors, including smoking cessation, taking prescribed medications, exercising, and attending cardiac rehabilitation. Less research on the relationship between depression and mortality has been done in other CVD patient groups, although similar links have been reported in studies of patients with congestive heart failure (CHF), for instance.

Implementing Screening as Part of Enhanced Care: Screening Alone is Not Enough

There are conflicting conclusions and policy recommendations relating to the effects of screening on the outcome of depression, but what does the latest evidence suggest? Based on the best available information to date, it emerges that screening alone is not a sufficient intervention to improve the quality and outcomes of care for depression. What is less clear is whether screening is a necessary condition for enhanced and improved quality of care and, given additional components, to what extent screening programs can potentially improve quality of routine care. Depression is the most common mental health problem and is associated with decrements in functioning and quality of life comparable to other chronic physical diseases. The prevalence, chronicity, and burden of suffering are such that the World Bank has predicted that depression will become the second leading cause of global disability by 2020. The economic consequences of depression are also profound, with the healthcare costs, welfare costs, and losses to productivity amounting to £9 billion ($20 billion) in the United Kingdom3 and $53 billion in the United States. Depression is most commonly encountered in primary care and in hospital settings, yet it often goes unrecognized by healthcare professionals. This has led to calls to implement screening programs to aid in the detection and management of this problem. The rationale and evidence base to support screening for depression is the focus of the present book and is discussed extensively in other chapters (see Chapters 2, 4, and 9). In the United States, screening has shifted from being an intervention that was not initially supported in national policy recommendations to being one that is regarded as being of proven effectiveness. An evolution in thinking has occurred that places screening at the center of mental health policy and practice, and is based upon the general assumption that screening will logically lead to improvements in the quality and outcome of care. Some have termed this the screening– detection–treatment–improvement paradigm. Recently screening for common mental health problems in the United States has become the cornerstone of the president’s agenda to improve the mental health of the U.S. population.

Clinical Judgment and the Influence of Screening on Decision Making

How do clinicians arrive at diagnostic decisions? In most cases the decision is not made following formal criteria, but by intuition. In addition, routine interviews are often narrow and the feedback gleaned from patients is inadequate. Yet it is not clear if screening helps or hinders clinical judgment. It might be that only clinicians who have low confidence and interviewing and diagnostic skills are open to the use of and actually helped by diagnostic tools. To provide a theoretical framework for understanding why it is difficult for physicians to detect depression in primary care settings, a broad array of research in the mental health fields can be described. For example, more than 1,000 studies have been conducted on clinical judgment in the area of mental health practice, and the results from these studies can be used to illuminate the challenges physicians face in judging whether a patient is clinically depressed and can benefit from treatment. In this chapter, results on clinical judgment will be described. A second topic will also be briefly discussed. Results from research on clinical judgment would seem to indicate that screening should be of value. Yet, as noted in Chapter 7, stand-alone screening programs have added little or nothing to outcomes. Reasons for this unexpected result will be explored. Three topics will be discussed: (1) narrowness of interviews, (2) nature of patient feedback, and (3) the cognitive processes of clinicians. Depression goes undetected because in many cases physicians do not ask patients if they have symptoms of a depressive mood disorder.3 To place this in context, it can be noted that mental health professionals also often do not ask patients about important symptoms and behaviors. Failure to inquire about depression in primary care settings can be viewed in the broader context of failure to inquire about important symptoms and events in mental health settings. Research on clinical judgment has demonstrated that lack of comprehensiveness is often a problem for interviews made in clinical practice. For example, in one study,4 mental health professionals saw patients in routine clinical practice, and afterwards research investigators conducted semi-structured interviews with the patients. Remarkably, the mental health professionals had evaluated only about 50% of the symptoms that were recorded using the semi-structured interviews.