The Importance of Somatic Symptoms in Depression in Primary Care

The Importance of Somatic Symptoms in Depression in Primary Care

Depression is highly prevalent. Approximately 20% of patients in primary care present with clinically significant depressive symptoms. In certain urban areas of the United Kingdom, up to 17% of the general population are affected. Despite improved awareness of the condition in primary care, depression remains difficult to diagnose initially, and the majority of cases are only recognized at subsequent consultations, sometimes several years after the patient’s initial visit.

In primary care, physicians require sophisticated consulting skills to enable them to differentiate a wide range of symptoms from a complex narrative in a short period of time. Symptoms of depression include classic psychological symptoms, such as low mood, loss of interest, poor concentration, and associated anxiety, and somatic symptoms, such as changes in appetite, lack of energy, sleep disturbance, and general aches and pains. The suspicion of depression is usually raised by the presence of psychological symptoms. However, in approximately two thirds of patients with depression, the clinical picture is domi- nated by somatic symptoms, such as lack of energy and general aches and pains, which patients frequently attribute to normalizing causes. As a result, many physicians become preoccupied with lengthy investigations into possible underlying organic disease rather than considering depression as a diagnosis. Indeed, depression is mostly difficult to recognize in patients who present with chiefly somatic complaints.

We conducted a review of the recent literature to ascertain the importance of somatic symptoms in depression in primary care, focusing particularly on their effects on the recognition of depressive symptomatology.

DATA SOURCES AND EXTRACTION

An electronic search was performed on the MEDLINE, EMBASE, and PsychLIT/PsychINFO databases to find articles published between 1985 and January 2004 that contained the following terms in the title (PsychLIT/ PsychINFO) and/or abstract (MEDLINE and EMBASE): depression, depressive, depressed AND physical, somatic, unexplained symptoms, complaints,problems; somatised, somatized symptoms;somatisation, somatization, somatoform, psychosomatic; pain; recognition, underrecognition; diagnosis, underdiagnosis; acknowledgment, underacknowledgment; treatment, undertreatment AND primary care, ambulatory care; primary physician; office; general practice; attribution,reattribution; and normalising, normalizing. Non–English-language publications and abstracts were not considered.

The results of the literature search were reviewed to select those of the correct type (i.e., those that focused on the somatic symptoms that are part of depressive disorder), based on the content of their title and abstracts (if available). Articles clearly about the coexistence of depression with a defined organic pathology were not included. Handsearching of citation lists was then performed on selected articles, and other information was included from the authors own knowledge of the literature, international guidelines, diagnostic tools, and relevant theses.

STUDY SELECTION

MEDLINE, EMBASE, and PsychLIT/PsychINFO database searches identified 2213, 1901, and 2462 citations, respectively. On the basis of the content of titles and abstracts, primary care and general population studies, case series, editorials, and review articles relating to somatic symptoms in depression were selected and reviewed in full. Handsearching of the citation lists in these publications identified several supplementary relevant articles beyond the limits of the literature search. Articles concerning somatic symptoms in depressed patients with organic diseases, such as cancer or arthritis, and patients with the clinical diagnosis of somatoform disorder or somatization disorder were excluded. More than 80 pertinent papers relating to somatic symptoms in depression were identified by this approach and form the basis of this review.

DATA SYNTHESIS

The language used in the medical literature to describe somatic symptoms in depression is both confusing and contradictory; hence, the broad scope of our literature search. In the interests of consistency, unless directly quoted from a reference, we chose to use the term somatic symptoms throughout this review to describe a range of symptoms that includes changes in appetite and libido, lack of energy, sleep disturbance, nonpainful somatic symptoms (e.g., dizziness, palpitations, dyspnea), and general aches and pains (e.g., headache, backache, musculoskeletal aches, and gastrointestinal disturbances). However, a large number of terms, including physical symptoms (problems or complaints), chronic painful physical conditions, medically unexplained symptoms, somatized symptoms, painful symptoms, somatization, somatoform symptoms, psychosomatic symptoms, and masked depression, have been used interchangeably over the years to describe the phenomenon of somatic symptoms and general aches and pains in depression.

The redundancy of terms used in the medical literature to describe the somatic symptoms of depression is a reflection of the complicated processes of clinical decision making and differential diagnosis in psychiatry and primary care and the differences between the 2 settings. It is important that somatic symptoms associated with depression should not be confused with somatoform disorders, which comprise conversion, somatization, hypochondriasis, and, in particular, somatization disorder, a chronic disorder characterized by a combination of pain and gastrointestinal, sexual, and pseudoneurologic symptoms. Differential diagnosis is confounded by the knowledge that there is a high prevalence of depression in patients with somatization disorder. Similarly, there are substantial levels of hypochondriacal, conversion, and somatizing symptoms in patients with depression. Indeed, results from several surveys suggest that depression, rather than somatoform disorders, may account for most of the somatization symptoms seen in primary care. Moreover, depressive disorders are common in patients with chronic pain, and pain is a frequent complaint in patients with depression.

While the linguistic constructs of somatization disorder and somatoform disorder are familiar and serviceable tools for the psychiatrist, these diagnostic classifications are less widely used in primary care, in which somatic symptoms, such as general aches and pains, are frequently described as “medically unexplained physical complaints.” Lack of training or expertise is not an adequate explanation for the discrepancy in language between these 2 levels of care. Making a diagnosis in primary care differs from secondary care in that the primary care physician is aware of a patient’s background and history. The categorical labels used by psychiatrists may, therefore, be inadequate for the needs of primary care physicians. Indeed, in primary care, patients present with individual, complex, and often poignant narratives, which encompass the domains of both mind and body, and are influenced by multiple social, economic, and other forces. In this setting, categorization can be seen to either trivialize or amplify a patient’s problems by removing the context.

Somatic Symptoms Are Prevalent in Depression in Primary Care

Patients with depression present with a combination of psychological and somatic symptoms. In primary care, somatic symptoms often dominate the clinical picture. During the second phase of the Depression Research in European Society II study (DEPRES II)—a pan-European survey of 1884 individuals previously receiving treatment for depression - 2 of the 3 most common symptoms reported in current depressive episodes were somatic. In another community study conducted by the World Health Organization (WHO),7 69% of participants (N = 1146) meeting criteria for major depression had approached their primary care physician on the basis of somatic symptoms alone, and more than half had multiple medically unexplained somatic symptoms. Similarly, during a retrospective examination of 685 primary care patients in Canada, 76% of patients who were diagnosed with depression or anxiety disorders (N = 75) had identified a somatic symptom as the primary reason for their initial visit to their primary care physician.8 More recently, a U.S. study conducted in 573 patients with depression reported that more than two thirds (69%) of the patients complained of general aches and pains of mild severity or above.

Somatic symptoms are more commonly reported by certain groups of patients with depression, including women, particularly pregnant women; the elderly; those earning a lower income; children; culturally diverse populations; patients with coexisting organic conditions; and the imprisoned. Certain cultural groups, including African Americans,29 have a tendency to mention somatic symptoms more frequently, or to focus more heavily on these symptoms when consulting their primary care physician. Culturally framed symptom interpretations, concepts of mental health, and social stigmas are chiefly responsible. In some countries, depression is seen as a moral or social problem, rather than a mental illness. Interestingly, the specific types of somatic symptoms reported by patients differ between cultures, reflecting cultural patterns of symptom significance - for example, abdominal distress, headaches, and neckaches are reported more frequently by patients with depression from Japan than those from the United States.

The high prevalence of somatic symptoms in depression poses the question, “Can somatic symptoms be considered to be clinical predictors of underlying depres- sion in primary care?” It is known that patients with somatic symptoms have a greater risk of developing depression. Similarly, patients with depression are more likely than their nondepressed counterparts to develop somatic symptoms in the long term. Furthermore, the greater the number of somatic symptoms, the greater the likelihood that an individual has depression. A recent study in 1143 Japanese white-collar workers found that the number of somatic symptoms identified on a 12-item somatic symptom checklist positively correlated with the prevalence of depression.39 Indeed, of 902 individuals who did not report any somatic symptoms, only 1 subject met criteria for major depression. Kroenke and colleagues also found that multiple somatic symptoms (i.e., 6 or more symptoms) were an independent predictor of depression and anxiety in a study of 500 adults attending a primary care clinic chiefly for somatic complaints. In an earlier study of 1042 consecutive outpatients screened for depressive disorders, discriminatory factors indicative of depression included sleep disturbance, fatigue, musculoskeletal complaints, and back pain.

Nevertheless, it is still important to remember that not all somatic symptoms reported in primary care indicate a possible depression diagnosis. It is essential that physicians continue to investigate organic pathologies as the source of these complaints. In the event of medically unexplained complaints, however, a psychological cause should be considered. Indeed, studies show that most so- matic symptoms reported by patients in primary care cannot be linked to an identifiable organic disease. Kroenke and Mangelsdorff demonstrated this succinctly during a retrospective review of 1000 patient records in which they examined the incidence, etiology, and outcome of 14 common somatic symptoms. Notably, these symptoms included 8 of the most common complaints reported in primary care. An organic cause for the symptoms was found in only 16% of cases; 10% of cases were thought to be of psychological origin; the rest were of unknown origin. Khan and colleagues made similar observations in a sample of 289 primary care patients with somatic symptoms; 48% of the symptoms were deemed to be of psychological or unknown origin.

The relationship between general aches and pains, organic disease, and depression was explored during a crosssectional telephone survey conducted by Ohayon and Schatzberg in 18,980 individuals across 5 European countries. In total, 748 participants (4%) in the study met criteria for major depressive disorder; of these, 43.4% had experienced headaches, gastrointestinal disturbances, and joint/articular, limb, or back aches, and 32.7% of patients had a coexisting organic condition. The organic condition, however, could explain the presence of pain in only one third of cases. These results suggest that if an organic condition presents that explains somatic symptoms, it is unlikely that the symptoms are somatized, even if depression exists; however, when an organic condition is present but not sufficient to explain the amount/quality of the somatic symptoms, depression may be playing a role.

Somatic Symptoms Increase the Burden of Depression

Somatic symptoms increase the already marked burden and disability associated with depression. Data from the U.S. National Household Survey a crosssectional community-based study of 1486 adults with major depression or dysthymia-found that patients with general aches and pains that included arthritic/rheumatic-like pain, back problems, and severe headaches (N = 938) had poorer physical and mental health status and reported more psychiatric distress than patients without general aches and pains. Ohayon and Schatzberg’s large pan- European cross-sectional study demonstrated that depressive moods were prolonged in patients with general aches and pains by, on average, approximately 6 months.

While few studies have examined the effect of somatic symptoms in depression on quality of life, what is known is that patients with depression who achieve full remission following treatment demonstrate greater improvements in physical functioning than nonresponders. Addressing both the psychological and somatic symptoms of depression would appear, therefore, to be necessary to achieve and maintain remission.

The increased burden of somatic symptoms in patients with depression leads to increased utilization of health care services and greater economic burden. In the U.S. National Household Survey, depressed patients suffering from general aches and pains made approximately 20% more visits to their health care providers each year than those without aches and pains. Interestingly, these patients were 20% less likely to see a mental health specialist than patients who did not report general aches and pains. Clearly, the burden of treating these patients falls heavily on the primary care health system. Luber and colleagues also found that the presence of somatic symptoms, including general aches and pains, was predictive of increased total ambulatory costs in 3481 elderly patients at 1 primary care practice. The number of somatic symptoms correlated with service utilization costs. The economic burden of somatic symptoms in depression also extends to employers. A claims-based study in the United States showed that medical costs were elevated 2.8- and 4-fold in depressed patients with backache and migraine, respectively.

Somatic Symptoms Decrease the Recognition of Depression

Although most primary care physicians are skilled at recognizing and treating depression, and most cases are eventually recognized, there is still some evidence of underrecognition and undertreatment, particularly at the initial clinic visit. While time constraints during consultations are doubtlessly a contributing factor, somatic presentation and failure to observe and respond to these cues during the patient interview are among the major reasons for underdiagnosis. Indeed, depression is less likely to be recognized in patients who present with somatic symptoms than in patients who present with predominantly psychological symptoms.

In 1985, Bridges and Goldberg reported that primary care physicians misdiagnosed more than 50% of psychiatric patients who presented with somatic symptoms. In 1993, during a retrospective examination of consecutive patients at 2 primary care clinics in Canada, Kirmayer and colleagues found that 78% of patients with major depression who had presented with a primary complaint of somatic symptoms had been misdiagnosed. The underlying reasons are complex, encompassing patient and physician characteristics, what patients say to their primary care physician, how and when they say it, and how the physician interviews the patient. Primary care physicians are often anxious not to miss a life-threatening organic condition, and those who are less confident in depression diagnoses will investigate somatic symptoms first—sometimes at length—before considering depression as the underlying cause.

In primary care, the depressed patient’s tendency to attribute unexplained somatic symptoms to a normalizing nonpathologic cause, rather than a psychological cause, is a principal driver in misdiagnosis. Physicians at 1 primary care practice in the United Kingdom failed to recognize depression (or anxiety) in 85% of patients with a normalizing attributional style, compared with 38% of patients with a psychologizing style. Furthermore, a questionnaire-based study conducted at 6 primary care practices in Australia found that depressed patients with an extreme normalizing style were 20 times less likely to receive a current depression diagnosis and 4 times less likely to receive a lifetime depression diagnosis compared to those with a low normalizing style. Interestingly, another U.K. study demonstrated that patients with more severe depression, which is recognized more frequently than mild to moderate forms in primary care, had higher psychologizing styles and lower normalizing styles.

Stigmatization surrounding mental illness can make discussion of psychological issues uncomfortable, leading patients to normalize their symptoms. This is particularly apparent during initial visits to primary care physicians, before an intimate relationship and a feeling of trust have been established. Notably, in the WHO primary care study, a somatic presentation occurred more frequently in patients who did not have an ongoing relationship with their primary care physician.

Patients with depression are also acutely aware of time constraints during primary care consultations, a factor that can lead them to self-restrict the time spent explaining their symptoms. Short consultation times, combined with “competing demand” between somatic and psychological symptoms and the fear of stigma attached to a depression diagnosis, interact to decrease the chance that the condition is even discussed. If mentioned at all, patients frequently wait until toward the end of primary care consultations to share psychological concerns. This is a critical determinant in misdiagnosis. In 1 study,61 physicians from 36 primary care practices in the United Kingdom were 5 times less likely to recognize depression when psychological symptoms were mentioned late in the consultation, compared with when psychological symptoms were mentioned within the first 4 symptoms.

These data imply that the recognition of depression is patientled. However, the way in which a primary care physician conducts the consultation and responds to the type and sequence of symptoms revealed by the patient also influences the likelihood that psychological symptoms are mentioned and ultimately, therefore, whether a diagnosis is made. Bucholz and Robins found that certain symptoms, such as loss of appetite or weight loss, and particular patient characteristics, such as being female or separated or widowed, appeared to encourage physicians to discuss depressive illness. A U.S. focus group study in which 21 primary care physicians considered approaches to depression diagnoses revealed that physicians tend to approach a depression diagnosis in 1 of 3 ways: by investigating somatic complaints first, by initially focusing on psychological symptoms, or by examining both psychological and somatic aspects in tandem. Patient characteristics and verbal, vocal (e.g., sighing), and postural cues determine which path is utilized. Some physicians are less likely to allow patients to express these cues. Closed, hypothesis-driven questioning, in particular, can suppress verbal cues given by the patients, discouraging them from revealing their psychological symptoms. A holistic and narrative approach that includes appropriately timed, open, and directive questions about psychological issues should be encouraged. Furthermore, physicians must apply equal diagnostic weighting to symptoms regardless of when they are mentioned.

Somatic Symptoms Complicate the Treatment of Depression

Patients with depression and somatic symptoms are harder to treat. Papakostas and colleagues showed that somatic symptoms were present in 95% of patients with treatment-resistant depression (N = 40) who had enrolled in a 6-week treatment study. Logistic regression analysis demonstrated that the number of somatic symptoms was a risk factor for further treatment resistance and tended to predict a poorer response to treatment. Indeed, the severity of somatic symptoms appears to be correlated to poor treatment response. Bair and colleagues used data from the ARTIST (A Randomized Trial Investigating SSRI Treatment) study-a randomized study with naturalistic follow-up conducted in the United States in 37 primary care clinics-to show that the severity of baseline general aches and pains could predict response to antidepressant treatment. More than two thirds of the depressed patients in this study reported general aches and pains of varying severity at baseline. Analysis of depression outcomes after 3 months of therapy with selective serotonin reuptake inhibitors revealed that patients with moderately se- vere aches and pains at baseline were 2 times less likely to respond to treatment. Patients with severe aches and pains at baseline were 4.1 times less likely to respond to treatment.

Interestingly, in the ARTIST study, residual general aches and pains of mild severity or above were present in 58% of patients with depression after 3 months of antidepressant treatment. Residual depressive symptoms are known for their association with poor outcome in depression. In a study of 60 patients treated to remission and then followed up for 15 months, Paykel and colleagues found that 19 patients had residual depressive symptoms; the most common residual symptoms were somatic, occurring in 18 (95%) of the 19 patients. Relapse occurred in 76% of patients with residual symptoms who were available for follow-up, compared with only 25% of patients without residual symptoms (10/40 patients). Indeed, patients with residual symptoms relapsed almost 3 times faster than those without.

A naturalistic long-term follow-up of these patients showed that subjects remitting with residual symptoms continued to have more depressive symptoms and impairment to their global, social, leisure, and work functioning over the long term. There was a trend toward earlier recurrence in patients with residual symptoms compared to those without; 42% and 56% of patients with residual symptoms recurred within 1 and 2 years, respectively, compared with 20% and 42% of patients without.

Implications of Somatic Symptoms in Depression: Can Training Help?

It is important that primary care physicians acquire specific skills for the recognition of depression. To date, there has been a tendency to focus on the psychological symptoms of depression rather than the somatic symptoms. Indeed, during a recent Australian primary care survey, only one quarter of physicians reported basing a diagnosis of depression on somatic symptoms (e.g., vegetative symptoms, malaise, and multiple consultations). Even if depression is recognized in patients with somatic symptoms, the focus on and severity of somatic symptoms can detract from a patient’s willingness to comply with treatment. Improved awareness of the importance of somatic symptoms in depression among primary care physicians, refined interviewing techniques, and training schemes that focus on teaching patients to reattribute somatic causality may help.

Perhaps one of the most important steps in ensuring the success of educational intervention is directing it to where it is most needed. Low prevalence rates for depression, high levels of medically unexplained somatic symptoms, and low antidepressant prescription rates are useful predictors of sectors of the primary health care system in which training may be warranted. With increasing numbers of patients with depression now having initial contact with a practice nurse rather than a physician and the implementation by the U.K. National Health Service (NHS) of NHS Direct (a nurseled service), training in recognition of depression is vital among this staff group; detection rates for depression have been shown to be low among practice nurses. Indeed, nurses’ confidence in dealing with depressed patients has been shown to improve following training.

Encouraging results have been seen in the past when educational initiatives were used to improve the recognition of depression in primary care. The most prominent data came from a program instigated by the Swedish Committee for the Prevention and Treatment of Depression during the 1980s on the island of Gotland. Improving primary care physicians’ knowledge of the diagnosis and treatment of depression by means of 2-day seminars led to improved recognition rates for depression, coupled with significant decreases in inpatient care, morbidity, mortality, and costs. Over a 3-year evaluation period, an overall economic benefit of $26 million was noted. The Defeat Depression Campaign, successfully implemented in the United Kingdom during the 1990s, was one of several international educational initiatives that used the Gotland study as a model. A primary care survey conducted in 1996 at the end of the campaign showed a positive impact of the national initiative, but also highlighted a need for supplementary local and practice-based training.

In contrast to the Gotland study, the Hampshire Depression Project, which used seminar-based education to improve knowledge of current best practice guidelines for depression, failed to increase the sensitivity or specificity of the recognition of depressive symptoms. Notably, the Hospital Anxiety and Depression Scale (HADS) was used to confirm the presence of substantial depressive symptoms in this study. Despite its widespread use, the HADS excludes somatic symptoms.

None of these studies focused specifically on education about somatic symptoms in depression as a means to improve recognition. Nevertheless, their results convey useful lessons for the design of future educational programs. Indeed, while pronounced effects on the recognition of depression were clearly evident following the short-term educational program in Gotland, improvements had reverted to baseline values within 3 years, illustrating the need to repeat educational initiatives every couple of years to maintain long-term effects.

Several studies have examined the effect of training primary care physicians in reattribution skills as a method of improving the recognition of depression. As discussed previously, in primary care, patients with depression tend to attribute somatic symptoms to normalizing causes. Teaching patients to reattribute somatic symptoms to psychological problems entails making patients feel understood (in particular, their beliefs about the cause of their symptoms), providing feedback on the results of their physical examinations and medical history while offering a tentative suggestion that somatic symptoms may be linked to psychological and lifestyle factors (i.e., changing the agenda), and then, if the patient seems willing to accept this suggestion, fully explaining the link between the somatic symptoms and the psychological cause. In essence, primary care physicians must try to find explanations compatible with the patient’s experience of illness that may change his or her belief about the cause.

Training primary care physicians in reattribution skills has been shown to improve interviewing and specific reattribution ability, leading to improvements in patient-doctor communication. In 1 study, separate cohorts of 103 and 112 patients visited 8 primary care physicians before and after the physicians had undergone an 8-hour reattribution skills training program. Patients reported greater satisfaction with the service they received (i.e., they felt that they had received the help they had wanted) and attributed psychological symptoms less to somatic causes when visiting primary care physicians who had undergone reattribution training. Overall, the technique was cost-effective. Although primary care costs did not change, the cost of referrals to secondary care, external health providers, and private health care decreased by 23%. A study of the effectiveness of a training course to educate primary care registrars in reattribution skills in somatizing patients is ongoing. The results of this study are awaited with interest.

DISCUSSION

In general, both patients and physicians appear to have a set agenda of issues to discuss during a primary care visit. Some problems are addressed, while others are left to subsequent visits, if addressed at all. When depression is present, somatic complaints often dominate the clinical picture, impeding the discussion of psychological complaints and thus masking the depression diagnosis. The failure to recognize depression in these cases is largely a consequence of patient-related barriers, in particular, normalizing attributional style; however, specific physician-related barriers, such as the inability to elicit psychological symptoms or respond to verbal/nonverbal cues during the consultation, also contribute.

Failure to recognize somatic symptoms, such as low energy, sleep disturbance, reduced appetite and libido, and general aches and pains, as components of depressive illness is associated with significant health care expenditure. Depressed patients with somatic symptoms usually feel a greater burden of disease and consequently tend to rely heavily on primary care services. In addition, in many cases, patients are subjected to costly and time-consuming investigations to determine whether organic conditions underlie their symptoms.

Traditionally used criteria for depression may be partly to blame for the lack of awareness that surrounds the importance of somatic symptoms in depression. Until their recent revision, DSM-IV criteria for major depressive disorder did not include any mention of somatic symptoms. Widely used rating scales, such as the HADS, were even refined during their development to exclude somatic symptoms in order that diagnosis should not be overcomplicated.

The American Psychiatric Association’s recent revision of DSM-IV criteria (DSM-IV-TR) to include somatic symptoms as a symptom of depression is indicative of an increasing awareness of the importance of somatic symptoms in depression. The new criteria refer to “excessive worry over physical health and complaints of pain (e.g., headaches or joint, abdominal, or other pains)” among the associated features of major depressive disorder. Indeed, guidelines currently in development at the National Institute of Clinical Excellence in the United Kingdom for the management of depression define the condition as low mood and a range of emotional, cognitive, physical (i.e., somatic), and behavioral symptoms.

A holistic approach to recognition is clearly necessary, and primary care physicians need to have a high index of suspicion for depression when faced with medically unexplained somatic symptoms, including general aches and pains and lack of energy. Educational initiatives that raise awareness of the full spectrum of symptoms in depression, as well as aiming to improve attitudes and consulting skills in primary care, should be of benefit.