 |
|
Somatoform Symptoms and Treatment Nonadherence in Depressed Family Medicine Outpatients
Robert Keeley, MD;
Marcia Smith, PhD;
John Miller, MD
Arch Fam Med. 2000;9:46-54.
ABSTRACT
| |
Objectives To examine whether somatoform symptoms, specifically symptoms of conversion, somatization, and hypochondriasis, are associated with side-effect reporting and treatment nonadherence in depressed family medicine outpatients, and to measure whether symptoms improve with pharmacotherapy.
Design Inception cohort study with 14-week follow-up.
Setting Inner-city family medicine residency clinic.
Patients Thirty-nine consecutive adults with major depressive disorder were asked to participate, and 30 consented.
Intervention Antidepressants for 14 weeks.
Main Outcome Measures The Personality Assessment Inventory (PAI) was administered before treatment. The PAI is a self-reported inventory compatible with the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, designed to measure a broad range of personality characteristics. After 14 weeks, the side-effect incidence and treatment nonadherence rates were determined, and 12 patients were readministered the PAI.
Results Depressed family medicine patients demonstrated trends toward elevated Somatic Complaints scale and conversion subscale scores and a lower Suicidal Ideation scale score relative to those of a standardized depressed psychiatric patient profile. Conversion and hypochondriacal symptoms were associated with side-effect reporting and treatment nonadherence. Somatization and hypochondriacal symptoms improved clinically and statistically during treatment for depression.
Conclusions Somatoform distress is a complex, common, and understudied phenomenon in primary care that can adversely affect the treatment of depression. Somatoform symptoms of conversion and hypochondriasis, but not somatization, were found to be risk factors for treatment nonadherence. Somatization and hypochondriacal symptoms may represent personality states that improve with pharmacotherapy, and conversion symptoms may be a personality trait resistant to medical treatment for depression.
INTRODUCTION
MAJOR depression is a common disorder that has been identified as a priority area by the Agency for Health Care Policy and Research because of underrecognition and undertreatment in the primary care setting.1-3 Approximately one third of depressed primary care patients treated by physicians in actual clinical practice stop using antidepressants within the first month of treatment,4-6 and only 75% will adhere to treatment after 90 days, even with a multifaceted, intensive intervention.7-8 Personality factors modulate medication tolerance and response in depressed patients. For example, somatic distress is associated with resistance to treatment9 and is potentially associated with treatment nonadherence.6, 10 Severity of baseline psychiatric comorbidity may predict improvement of psychiatric disease in primary care patients.11
Most depressed primary care patients present exclusively with somatic symptoms,12 referring to the presentation of psychological conflicts and issues as physical symptoms.13 A variety of syndromes characterized by somatic complaints, including conversion hysteria, hypochondriasis, and Briquet syndrome, have been described since the time of ancient Greek physicians and have taken on different clinical meanings over time.14 In addition, the somatic presentations of distress encountered in primary care are rarely equivalent to somatoform disorders from the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV).12, 15-19 Descriptive work on somatoform syndromes in primary care is needed,19 along with longitudinal studies to explore the interactions of somatic distress, personality factors, and illness experience.16
Our primary goal was to implement a personality assessment tool in a consecutive sample of depressed family medicine outpatients to explore the interaction between personality factors, particularly somatoform symptoms, and illness experience, specifically side-effect reporting and treatment nonadherence. Concomitantly, we examined the influence of pharmacotherapy for depression on the following 3 distinct types of somatoform symptoms described in DSM-IV15: conversion, somatization, and hypochondriasis.
Conversion symptoms refer to unusual, dramatic physiological symptoms that defy medical explanation. Most involve unusual sensory-motor problems, such as numbness or paralysis, or impairments in perception, such as blindness or deafness.14 Conversion symptoms have high comorbidity with depression and environmental stressors, but have not been studied extensively in family medicine patients, probably due to questions about the validity of the syndrome and whether it should be considered a diagnosis or a symptom.13
Somatization refers to routine physical complaints such as headaches, back problems, pain, or gastrointestinal tract ailments.14 An abridged somatization construct has been demonstrated to have a high prevalence in primary care, from 4.4% to 20%,18, 20 whereas somatization disorder has only a 0.3% to 0.7% prevalence.21 The abridged construct is related to low socioeconomic status, female sex, older age, and Hispanic ethnic background.18, 20
Hypochondriasis refers to a preoccupation with health and physical functioning and exaggerated fear of disease.14, 16 Hypochondriasis in DSM-IV is an uncommon condition in primary care, with a prevalence of about 3%, but frequently associated with major depression.22 Subthreshold hypochondriasis, or elevated health concerns, may occur in 7.7% of family medicine patients.17 Unlike somatization, hypochondriasis is unrelated to any demographic factors.22
Somatizing patients demonstrate less introspection, are more likely to deny psychological causes of somatic symptoms than psychosocial presenters of distress, and are more likely to discount common somatic symptoms by endorsing normal conditions or environmental circumstances.16, 23 Therefore, we postulated that patients with an elevated Somatic Concerns scale score would more often attribute common somatic complaints to environmental circumstances, specifically medication side effects, and would experience increased incidence and degree of side-effect reporting.10, 23-24 Since somatization is associated with abnormal illness behavior and poor recognition of psychiatric distress,16 we postulated that the somatization subscale score would correlate with treatment nonadherence. Based on a study in a sample of depressed psychiatric patients treated with fluoxetine hydrochloride25 for 8 weeks, in which hypochondriacal concerns were not associated with premature treatment dropout, we theorized that pretreatment measurement of hypochondriacal concerns would not be associated with treatment nonadherence. Finally, we theorized that effective pharmacotherapy for depression would result in no significant changes in somatic or hypochondriacal symptoms, because several well-designed studies view both constructs as personality traits9, 25-26 resistant to change with pharmacotherapy.
SUBJECTS AND METHODS
STUDY SAMPLE
Patients were recruited from a community-based inner-city family practice residency clinic in Denver, Colo. Patients starting antidepressant therapy for a diagnosis of depression were eligible. Attendees at this clinic are 45% Hispanic, 45% white, 5% black, and 5% other ethnicity. Participating physicians asked consecutive patients diagnosed as having depression to participate in the study and obtained informed consent.
Exclusion criteria included failure to complete the psychometric assay at study entry, organic mental disorders, English as a second language, serious suicidal risk, diagnosis of bipolar disorder, current treatment with an antidepressant, or any history of treatment with other types of psychotropic medications.
The diagnosis of depression was made by the treating physician who determined treatment plans, including choice of antidepressant, titration and dosing schedule, and follow-up time. All patients were observed by a resident physician (R.K.) together with a faculty family physician (J.M.). Treating physicians were allowed to modify the treatment plan throughout the study period.
MEASUREMENT STRATEGY
At baseline, all patients were administered the Personality Assessment Inventory14 (PAI), a 45-minute, self-administered, objective inventory of adult personality designed to measure the same diagnostic categories as DSM-IV. The PAI was developed and standardized for use by individuals older than 17 years and requires a fourth-grade reading level. The test includes 344 items that are answered on a 4-point Likert-type scale with anchors totally false, slightly true, mainly true, and very true. It includes 22 nonoverlapping scales: 4 validity, 11 clinical, 5 treatment consideration, and 2 interpersonal scales.
The PAI was chosen over other personality instruments such as the NEO–Five Factor Inventory27-28 and the Tridimensional Personality Questionnaire29-30 because the scales included on the PAI such as Depression and Antisocial Features are DSM-IV compatible. The scales on the NEO–Five Factor Inventory, including Extraversion and Openness to Experience, and on the Tridimensional Personality Questionnaire, including Novelty Seeking and Reward Dependence, are more theoretically related to DSM-IV.
The PAI validity scales include Inconsistency, Infrequency, Negative Impression, and Positive Impression, which assess factors that could distort the results of testing. Clinical syndromes assessed by the PAI were selected on the basis of stability over time, importance within the psychiatric literature, and significance in contemporary diagnostic practice.14 The items on the clinical scales directly reflect the phenomenology and symptomology of each clinical construct. The Depression scale sums 3 subscales assessing the affective, cognitive, and physiological aspects of depression, and correlates closely with other depression instruments, including the 17-item Hamilton Depression Rating Scale (r = 0.78)31 and the Beck Depression Inventory (r = 0.81).32 Depression scale answers and patient medical charts were reviewed to confirm that patients fulfilled criteria for DSM-IV major depressive disorder.15 The Somatic Complaints scale includes conversion, somatization, and health concerns, or hypochondriacal, subscales. The PAI scale and subscale raw scores are transformed to T scores derived from a sample of 1000 community-dwelling adults selected to match 1995 US census projections on the basis of race, sex, and age. The PAI T scores are calibrated for each scale to have a mean of 50 and an SD of 10. Mean and SD T scores have been published for a clinical population of 1246 outpatients obtained from psychiatric, general medical, substance abuse, and other professional settings and for a depressed population drawn mostly from psychiatric practices including 20% inpatients.33
FOLLOW-UP
After baseline assessment, subjects were treated pharmacologically for depression for up to 14 weeks. Data collection included a record of patient characteristics, including race, age, number of other medical problems and medications, medication types, dosing schedules, side effects, and time until first follow-up clinic visit. A retrospective medical chart review and 4-question telephone interview or mailed survey assessed side effects. The interview or survey asked the patient to describe which medications(s) were taken and for how long, to list reasons for stopping if antidepressant therapy was discontinued, and to describe side effects ascribed to the medications. After 14 weeks, study participants were asked to take a second PAI.
ANALYTIC STRATEGY
Published test-retest standard errors of measurement are useful when trying to evaluate the meaning of changes over time and add power to analyses of changes in PAI scales and subscales.33-34 Power analysis was performed using the published error of measurement for the Warmth scale of 4.8 T and the clinical SD of 13.5 T.14, 33 With a sample size of 14, it was determined that a minimal effect change of 4 T on the Warmth scale could be detected based on a 2-sided z test with an overall error rate of less than 0.05. The power is 96% when examining a medium effect of a T score of 6.75 in a sample size of 14.34
Mean and SD T scores were determined for the study sample and were compared with characteristics of published clinical and depressed patient populations. A z test was used to refute the hypothesis that the mean T scores in our study sample were statistically equivalent to T scores for the clinical and depressed populations.34 Patients adhering to the treating physician's treatment plan were assigned to the treatment completion subgroup. Patients unable to complete a therapeutic course of antidepressants due to intolerable side effects were assigned to the treatment nonadherence subgroup. The treatment completion and nonadherence subgroup PAI scale and somatization subscale scores were compared using an unpaired t test. Significance was accepted for an overall 2-tailed P value less than .05. A Bonferroni correction was applied to adjust for multiple comparisons.
The number of patients reporting side effects, the number of side effects reported per patient, and the interval to first clinic visit for treatment completion and nonadherence groups were also compared using the unpaired t test.
Patients in the treatment adherence group were asked to take a follow-up PAI after 14 weeks, allowing a test-retest analysis. The standard error of measurement is expressed in T-score units and represents the SD of the hypothesized error term as a contribution to observed PAI T scores.14 A z test analysis34 with Bonferroni correction for multiple comparisons allowed evaluation of changes in the PAI before and after treatment. Clinical significance when examining a PAI scale score change in a group is accepted for changes of greater than 1 SE.33
RESULTS
STUDY SAMPLE DESCRIPTION
Participating physicians invited 39 consecutive patients with a diagnosis of depression to participate in the study. Eight patients were excluded for not completing the pretreatment PAI, and 1 patient was excluded for history of treatment with a neuroleptic. Patient characteristics are given in Table 1. English was the first language for all excluded and participating patients. The 9 excluded patients included 2 men and 7 women, and the group was statistically equivalent to the study sample (P = .95). Six of the excluded patients were Hispanic, and 3 were white (P = .30). Included and excluded patients demonstrated statistically equivalent age (P = .64), number of medical problems (P = .19), and number of other medications (P = .22).
|
|
|
|
Table 1. Patient Characteristics*
|
|
|
The physicians' diagnosis of depression correlated with the patients' depressive symptoms on self-report. The mean (± SD) Depression scale score of 81 ± 13 T in the study sample is 3 SDs above a population mean, or one third of 1 SD above that of a standardized depressed population, and correlates with severe depression (Table 2).33 Review of the patients' answers to the Depression and Suicidal Ideation scales on the DSM-IV–compatible PAI reveals that all patients endorsed at least 5 of 9 core symptoms, including depressed mood or anhedonia, consistent with the diagnosis of major depressive disorder. Chart review demonstrated that all patients were symptomatic for at least 2 weeks, that none were in bereavement, that none had an untreated medical diagnosis possibly accounting for the diagnosis, and that none had their symptoms ascribed to substance abuse.15
|
|
|
|
Table 2. Mean T Scores for a Clinical Population, a Depressed Population, and Study Sample*
|
|
|
BASELINE PAI COMPARISON
A mean PAI profile was generated for the total study sample (N = 30), and mean and SD T scores were compared with those of published clinical and depressed populations (Table 2). With respect to 18 of the 22 PAI scales, this study profile was comparable to the standardized clinical population profile. Significant elevations were noted in the Inconsistency, Somatic Complaints, Anxiety, and Depression scales (overall P<.05). When compared with the depressed population, the study sample showed a trend toward higher Inconsistency (overall P = .16) and Somatic Complaints scale scores (uncorrected P = .03), and lower Treatment Rejection (uncorrected P = .02) and Suicidal Ideation scale scores (overall P = .08), but was otherwise similar.
Somatic Complaints subscale analysis demonstrated the study sample had significantly increased conversion (overall P<.001), somatization (overall P<.001), and health concern subscale scores (overall P = .02) relative to the clinical population, and was characterized by a trend toward an increased conversion scale score (overall P = .09) relative to the depressed psychiatric population (Table 2).
TREATMENT PLAN
Physicians elected to start treatment in participating patients with serotonin-specific reuptake inhibitor medications in 26 instances in appropriate titrations and doses. Seven patients started treatment with paroxetine; 8, fluoxetine hydrochloride; and 11, sertraline hydrochloride. One patient started treatment with amitriptyline hydrochloride; 1 patient, bupropion hydrochloride; and 2 patients, venlafaxine hydrochloride. Mean (± SD) follow-up time was 21 ± 9.0 days (range, 4-42 days).
REPORTED SIDE EFFECTS
Twenty patients experienced and reported at least 1 side effect. Side effects reported by patients included, in order of decreasing incidence, headache, light-headedness, visual changes, nausea, hyperhidrosis, dizziness, racing thoughts, tremor, diarrhea, hot and cold flashes, numbness, insomnia, and back pain. Six (75%) of 8 patients starting treatment with fluoxetine, 7 (64%) of 11 starting treatment with sertraline, and 6 (86%) of 7 starting treatment with paroxetine reported at least 1 side effect. One patient starting treatment with venlafaxine reported racing thoughts. One patient each starting treatment with venlafaxine, bupropion, and amitriptyline reported no side effects.
Ten study participants terminated treatment prematurely secondary to side effects. Three taking paroxetine stopped medication therapy within 3 days. A group of 6 patients, including 3 receiving sertraline, 2 receiving fluoxetine, and 1 receiving paroxetine discontinued the medication therapy from 3 days to 6 weeks. In most cases, patients were unwilling to try another medication. One nonadherent patient switched from fluoxetine to paroxetine to sertraline without success, and was unable to tolerate any medicine for longer than 4 days. One patient taking sertraline stopped medication therapy between 6 and 14 weeks.
SUBGROUP DESCRIPTION
Patients were divided into treatment completion and treatment nonadherence subgroups for further analysis of PAI results. Patients adhering to the treating physician's treatment plan, including 4 with changes to other types of antidepressants and 2 in whom pharmacotherapy was stopped after 60 days, were assigned to the treatment completion group (Table 3). Patients unable to complete a therapeutic course of antidepressants owing to intolerable side effects, including 1 patient who tried and discontinued multiple medical regimens, were assigned to the nonadherence group.
|
|
|
|
Table 3. PAI Scores for Treatment Completion and Treatment Nonadherence Groups and a Standardized Depressed Population*
|
|
|
There were no statistically significant differences when the treatment completion and nonadherence groups were compared (Table 1) by race (P = .80), age (P = .28), number of chronic problems (P = .32), or number of other medications (P = .06). A trend toward older patients with greater morbidity and medicine use was demonstrated in the nonadherent group.
Two of 8 patients initially receiving fluoxetine, 4 of 7 initially receiving paroxetine, and 4 of 11 initially receiving sertraline were in the nonadherence group. Six (30%) of 20 patients in the treatment completion group were given fluoxetine, as were 2 (20%) of 10 in the nonadherent group (P = .58). Four (40%) of 10 nonadherent patients and 3 (15%) of 20 adherent patients initially received paroxetine, demonstrating a weak trend toward more paroxetine use in the nonadherent group (P = .26). Four (40%) of 10 nonadherent and 7 (35%) of 20 adherent patients initially received sertraline (P = .80). There were no statistically significant differences when treatment completion and nonadherence groups were compared by initial medication choice.
Four patients successfully switched to a different antidepressant regimen after not tolerating the initial treatment choice. One patient changed from venlafaxine to trazodone hydrochloride, 1 from fluoxetine to sertraline, 1 from venlafaxine to sertraline to fluoxetine, and 1 from paroxetine to a combination of sertraline and trazodone. Providers withdrew fluoxetine from 2 patients after 2 months due to clinical assessment that depression was resolved. One of these 2 patients took a second PAI at 14 weeks, and her Depression scale score had improved from 103 to 79. Twenty patients adhered to treatment plans throughout the study.
The treatment nonadherence group demonstrated a trend toward shorter follow-up (17.8 ± 12.0 days) than the treatment completion group (22.2 ± 7.0 days; P = .22), due to more walk-in visits for patients experiencing severe side effects.
All 10 patients in the treatment nonadherence group reported at least 1 side effect, and the mean (± SD) number of reported side effects was 1.7 ± 0.7. Eight of the 20 patients in the treatment completion group reported side effects, and the treatment completion group averaged 0.9 ± 1.0 side effects. More members of the treatment nonadherence group reported any side effect (P = .002), and overall reported more side effects (P = .03).
SUBGROUP PAI ANALYSES
Mean PAI profiles were generated for patients who completed the treatment period of up to 14 weeks (n = 20) and for patients who stopped treatment prematurely (n = 10) secondary to intolerable side effects (Table 3), and the profiles were compared with each other and with those of a standardized depressed population. The treatment completion and depressed population groups had significantly lower Somatic Complaints scale scores than the treatment nonadherence subgroup (Table 3). The treatment completion subgroup was very similar to the depressed population, with notable trends toward higher Inconsistency and lower Suicidal Ideation scale scores (Table 2 and Table 3). The conversion subscale score was significantly higher for the treatment nonadherent group than the depressed population (overall P<.005), with a strong trend toward significance relative to the score for the treatment completion subgroup (overall P = .06).
There were no differences for the somatization subscale scores among the treatment completion, treatment nonadherence, and depressed population groups. The health concerns subscale scores were significantly elevated in the treatment nonadherence group relative to the treatment adherent subgroup (overall P = .04) and the depressed population (overall P = .04) (Table 3). The Borderline Features scale score was elevated before correction for multiple comparisons in the treatment completion subgroup (uncorrected P = .046), as was the Antisocial Features scale score (uncorrected P = .003), but both lost significance after correction for multiple comparisons.
Among patients not adhering to pharmacotherapy, 1 patient was predominantly hypochondriacal, defined as a health concerns subscale T score of more than 2 SDs above the population mean (83 T), with the other 2 Somatic Complaints subscale scores below the 2-SD threshold. One of 10 patients presented with primarily conversion symptoms, and none presented with predominantly somatic symptoms. Five of the 10 patients showed mixed elevated conversion and hypochondriacal symptoms; 2 patients, mixed somatic and conversion symptoms; and 1 patient, mixed hypochondriacal and somatic symptoms.
TEST-RETEST ANALYSIS
Twelve of the 20 patients completing the treatment course took a second PAI after 14 weeks. The other 8 treatment adherers refused to finish the test a second time due to test length. The Depression and Somatic Complaints scale scores, along with the somatization and health concerns subscale scores, improved significantly during treatment to near baseline scores from the standardized clinical profile (overall P<.05) (Table 4). Other scales demonstrating statistically significant improvement included the Negative Impression, Anxiety, Anxiety-Related Disorders, Schizophrenia, Suicidal Ideation, Stress, and Treatment Rejection scales. Clinical significance was present in all cases where statistical significance occurred (Table 4).
|
|
|
|
Table 4. PAI Scores Before and After Pharmacotherapy for Depression*
|
|
|
COMMENT
Major depression is underrecognized and undertreated in primary care, and when it is treated, medical adherence is often inadequate.8 Our objective was primarily to determine whether personality characteristics, especially psychometric measurements of somatoform distress common in depressed primary care patients,17 are associated with treatment nonadherence. Depressed study participants with elevated endorsement of hypochondriacal and/or conversion symptoms did not tolerate antidepressants well, possibly because medication side effects were interpreted as worsening disease status.
Secondary goals of the study were to determine whether personality characteristics placing patients at risk for treatment nonadherence improved with pharmacotherapy. Test-retest analysis demonstrated that symptoms of somatization and hypochondriasis improved significantly with treatment, and that conversion symptoms persisted. Symptoms of conversion and hypochondriasis present a barrier to pharmacological adherence in this sample of depressed primary care patients treated with antidepressants, and test-retest analysis suggests improved treatment adherence might alleviate hypochondriacal, but not conversion, symptoms.
LIMITATIONS
Our study has several limitations. First, the relatively small sample from a single practice limits ability to generalize results and to study additional subgroups. Second, since somatic presentations of depression decrease the rate of detection, whereas hypochondriacal and unexplained symptoms increase probability of detection,16 it is possible the study population was biased toward patients with elevated hypochondriacal and unexplained symptoms. Third, although the study sample was compared with standardized depressed and clinical populations, a control group from within the practice was not described. Fourth, unmeasured patient characteristics related to the personality measures under study and the study outcomes may confound the results. Despite limitations, this study examines the theory that personality characteristics affect treatment adherence in a consecutive sample of family practice patients meeting standard criteria for and receiving treatment of depression.
DIAGNOSIS
Although family physicians usually do not apply formal criteria for the diagnosis of major depression,35 relying often on clinical impression and knowledge of the patient over time,36-37 they are relatively accurate when making the diagnosis of major depression.36 Although participating family practitioners did not usually provide full documentation of formal criteria for major depression, diagnosis of major depression correlated very highly with personality inventory qualification for DSM-IV major depression. All study participants qualified for DSM-IV major depression after review of medical charts and answers to the PAI Depression scale.
BASELINE PAI COMPARISON
Relative to the standardized clinical population, our study sample demonstrated significantly elevated Depression, Anxiety, and Somatic Complaints scale scores, consistent with symptoms of depression,14-15,33 and demonstrated a trend toward an elevated Schizophrenia scale score. This trend is consistent with disorganization of thought typical of depression.33 Analysis of the Somatic Complaints subscale scores demonstrated slight trends toward increased somatization and health concerns in the study sample vs the clinical population and a strong trend toward increased conversion symptoms consistent with previous observations that depressed primary care patients usually present with somatoform symptoms.16
The trend toward less suicidal ideation in the study sample vs the standardized depressed population was explained by a 20% incidence of inpatient status in the population of depressed psychiatric patients.25 A trend toward an elevated Nonsupport scale score is consistent with feelings of lack of support and perceived excessive criticism from family members that depressed family medicine patients often report.38 A trend toward a decreased Warmth scale score is consistent with social withdrawal measured by the Warmth scale.33
The study sample included approximately 50% American-born Hispanics. Hispanic patients are less likely than white patients to consult with a specialist in mental health, even accounting for covariant psychiatric disorder, sex, and other variables known to be associated with differential use of health care services.38 Hispanic patients with depression somatize more than those without a psychological disorder and more than white patients with a psychological disorder.21 Although the trend toward higher somatoform endorsement in the study sample could be explained by the presence of Hispanic patients, subgroup analysis demonstrates that adherent and nonadherent subgroups equivalently included white and Hispanic patients. Although previous studies might predict higher Somatic Complaints scale scores among the Hispanic participants, the mean Somatic Complaints scale score was actually slightly lower for Hispanic than for white patients (68.2 ± 12.3 vs 72.2 ± 14.2; P = .47).
SUBGROUP PAI COMPARISONS
Except for trends toward more inconsistency and less suicidality, composite PAIs of the treatment completion group and depressed population were very similar. The treatment nonadherence group was characterized by elevated conversion and health concerns subscale scores relative to the depressed population (Table 3) and by elevated health concerns and nearly significant conversion subscale scores relative to the treatment completion group. If the treatment nonadherence group is representative of "somatic presenters of psychological distress" described by Kirmayer and Robbins,16 then decreased recognition of psychosocial stress and psychological distress are typical characteristics.16 The observed trend toward decreased Antisocial, Borderline, and Stress scale endorsement is probably secondary to decreased psychological insight.
ADHERENCE
The medication noncompliance rate of 33% after 14 weeks in our study is consistent with other studies based in primary care settings in which about one third of patients receiving usual care do not adhere to antidepressant treatment plans.3, 6-8 Increased medical comorbidity and psychosocial problems characteristic of family medicine patients19 may predispose them to a higher incidence of nonadherence relative to psychiatric samples. The study sample demonstrated significant medical comorbidity with the nonadherent subgroup, with a trend toward more chronic problems and more medication use. However, somatizing primary care patients do not demonstrate increased somatic distress simply due to more physical illness.16 The treatment adherence group included 2 patients from whom antidepressant therapy was withdrawn after 2 months owing to resolution of symptoms, although current recommendations are to treat first-time depression for 4 to 9 months. Family physicians often are not aware of Agency for Health Care Policy and Research guidelines and have a difficult time following them.3, 8
Overall, the associations between an elevated Somatic Complaints scale score and side-effect reporting and between the elevated score and treatment nonadherence are consistent with studies demonstrating abnormal illness behavior and ineffective treatment regimens in somatizing patients.26, 39-42 Somatizing patients may report and recall more side effects due to increased awareness of bodily sensations, although somatizers are in fact no more focused on their bodies than psychosocial presenters of distress according to a measure of private body consciousness.16, 43 Our results contradict the conclusions of a study on hypochondriacal symptoms in psychiatric outpatients, in which the degree of hypochondriacal concern was not associated with treatment nonadherence.25 Differences may be due to different patient populations, eg, the study of psychiatric outpatients included a large proportion of 28% with history of histrionic personality disorder, or due to implementation of different instruments to measure hypochondriacal concerns.
Depressed family medicine patients with markedly elevated conversion and/or health concerns subscale scores may not adhere to treatment plans because medication side effects are interpreted as worsening disease status. Somatization does not appear to be an independent risk factor for side-effect reporting and treatment nonadherence, possibly because patients tend to ascribe normalizing causes,16 such as drug side effects, to symptoms. Previous studies suggesting somatization as a risk factor for treatment nonadherence4, 10 appear to have included components of conversion, somatic, and hypochondriacal symptoms in the rubric "somatization."
TEST-RETEST ANALYSIS
Further analysis of Somatic Complaints subscale scores in a subset of patients adhering to therapy demonstrates insignificant improvement in the conversion scores. Although the effect of pharmacotherapy on conversion symptoms has not been extensively examined, studies on hypochondriacal symptoms have demonstrated minimal25 and marked44-46 symptom reductions. Our findings suggest hypochondriacal symptoms are amenable to therapy. Other scores showing clinical and statistical improvement, including the Schizophrenia scale and somatization subscale scores, are consistent with the successful treatment of depression.14
Personality scales can be categorized into those measuring traits and states. Study participants then demonstrate personality traits, ie, an elevated Antisocial scale or conversion subscale score, that are stable over time despite pharmacotherapy, and personality states, ie, elevated somatization or health concern subscale score, that improved with pharmacotherapy.33 Traits and states have been related to Axis I and II disorders, although personality is more complex than this dichotomy.47 Most personality characteristics are probably composed of, at the very least, trait (genetically endowed) and state (environmentally inscribed) components,30, 47 and the constant interplay of environment and biology may render the trait-state dichotomy overly simplistic.
CONCLUSIONS
Our study corroborates previous reports that somatic distress is common10, 18 in depressed primary care outpatients and that depressed patients with somatoform presentations represent a complex, heterogeneous group,17, 43 and suggests that somatoform symptoms affect the treatment of depression. Elevated conversion and hypochondriacal symptoms are demonstrated to associate with side-effect reporting and treatment nonadherence in a consecutive sample of inner-city family medicine outpatients undergoing pharmacological treatment for depression. Contrary to previous reports,6, 10 we found no tendency for somatic symptoms to correlate with treatment nonadherence. Limited test-retest analysis suggests symptoms of somatization and hypochondriasis may represent personality states that improve with pharmacotherapy, and conversion symptoms may represent a personality trait resistant to pharmacotherapy for depression. Personality instruments may aid the clinician to ascertain presence and levels of somatoform symptoms and to stratify depressed patients to different treatment plans depending on relative risks for treatment nonadherence.
AUTHOR INFORMATION
Accepted for publication May 12, 1999.
Presented in part at the 25th Anniversary Annual Meeting of the North American Primary Care Research Group, Orlando, Fla, November 14, 1997.
We thank Gary Zerbe, PhD, and Janet Tooze, MS, from the Department of Preventative Medicine and Biometrics, University of Colorado School of Medicine, Longmont, for statistical guidance.
Reprints: Robert Keeley, MD, Department of Family Medicine, University of Colorado School of Medicine, Salud Family Health Center, 231 E 9th St, Longmont, CO 80501.
From the St Anthony Family Medicine Residency, St Anthony Central Hospital, Denver, Colo. Dr Keeley is now with the Salud Family Health Center, Longmont, Colo, and is a Clinical Instructor at the Department of Family Medicine, University of Colorado School of Medicine, Denver. Dr Smith is now with the Department of Family Medicine, Brown University, Providence, RI.
REFERENCES
| |
1. Depression Guideline Panel. Depression in Primary Care: Treatment of Major Depression: Clinical Practice Guideline Number 5. Vol 1. Rockville, Md: US Dept Health and Human US Dept Health and Human Services, Public Health Service, Agency for Health Care Policy and Research; April 1993. Agency for Health Care Policy and Research publication 93-0551.
2. Valenstein M, Klinkman M, Becker S, et al. Concurrent treatment of patient with depression in the community: provider practices, attitudes, and barriers to collaboration. J Fam Pract. 1999;48:180-187.
ISI
| PUBMED
3. Schulberg H, Block M, Madnia M, et al. The "usual care" of major depression in primary care practice. Arch Fam Med. 1997;6:334-341.
FREE FULL TEXT
4. Katon W, Von Korff M, Lin E, Bush T, Ormel J. Adequacy and duration of antidepressant treatment in primary care. Med Care. 1992;30:67-76.
FULL TEXT
|
ISI
| PUBMED
5. Simon G, Von Korff M, Wagner EH, Barlow W. Patterns of antidepressant use in community practice. Gen Hosp Psychiatry. 1993;15:399-408.
FULL TEXT
|
ISI
| PUBMED
6. Schulberg HC, Block MR, Madonia MJ, Rodriguez E, Scott P, Lave J. Applicability of clinical pharmacotherapy guidelines for major depression in primary care settings. Arch Fam Med. 1995;4:106-112.
FREE FULL TEXT
7. Katon W, Von Korff M, Lin E, et al. Collaborative management to achieve treatment guidelines: impact on depression in primary care. JAMA. 1995;273:1026-1031.
FREE FULL TEXT
8. Katon W, Von Korff M, Lin E, et al. Collaborative management to achieve depression treatment guidelines. J Clin Psychiatry. 1997;58:20-23.
9. Bouvet de la Maisonneuve O. Antidepressive agents and somatic complaints. Encephale. 1991;17:427-429.
10. De Wester JN. Recognizing and treating the patient with somatic manifestations of depression. J Fam Pract. 1996;43(suppl):S3-S15.
11. Mann A, Jenkins R, Belsey E. The twelve-month outcome of patients with neurotic illness in general practice. Psychol Med. 1981;11:535-550.
ISI
| PUBMED
12. Kirmayer LJ, Robbins JM, Paris J. Somatoform disorders: personality and the social matrix of somatic distress. J Abnorm Psychol. 1994;103:125-136.
FULL TEXT
|
ISI
| PUBMED
13. Ford CV. Dimensions of somatization and hypochondriasis. Neurol Clin. 1995;13:241-253.
ISI
| PUBMED
14. Morey LC. Personality Assessment Inventory: Professional Manual. Tampa, Fla: Psychological Assessment Resources; 1991.
15. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. Washington, DC: American Psychiatric Association; 1994.
16. Kirmayer JK, Robbins JM. Patients who somatize in primary care: a longitudinal study of cognitive and social characteristics. Psychol Med. 1996;26:937-951.
ISI
| PUBMED
17. Kirmayer LJ, Robbins JM. Three forms of somatization in primary care: prevalence, co-occurence, and sociodemographic characteristics. J Nerv Ment Dis. 1991;179:646-655.
FULL TEXT
18. Kroenke K, Spitzer RL, deGruy FV, et al. Multisomatoform disorder: an alternative to undifferentiated somatoform disorder for the somatizing patient in primary care. Arch Gen Psychiatry. 1997;54:352-358.
FREE FULL TEXT
19. deGruy F. Mental health care in the primary care setting. In: Donaldson MS, Lohr KN, Vanselow NA, eds. Primary Care: America's Health in a New Era. Washington, DC: Institute of Medicine, National Academy Press; 1996:285-311.
20. Escobar JI, Swartz M, Rubio-Stipec M, et al. Medically unexplained symptoms: distribution, risk factors, and comorbidity. In: Kirmayer LJ, Robbins JM, eds. Current Concepts of Somatization: Research and Clinical Perspectives. Washington, DC: American Psychiatric Press; 1991:63-78.
21. Escobar JI, Rubio-Stipec M, Canino G, Karno M. Somatic symptom index (SSI): a new and abridged somatization construct: prevalence and epidemiological correlates in two large community samples. J Nerv Ment Dis. 1989;177:140-146.
FULL TEXT
|
ISI
| PUBMED
22. Escobar JI, Gara M, Waitzkin H, Silver RC, Holman A, Compton W. DSM-IV hypochondriasis in primary care. Gen Hosp Psychiatry. 1998;20:155-159.
FULL TEXT
|
ISI
| PUBMED
23. Miller LC, Murphy R, Buss AH. Consciousness of body: public and private. J Pers Soc Psychol. 1981;41:397-406.
FULL TEXT
|
ISI
24. Azorin JM. Management of somatization in depression. Encephale. 1995;21:17-22.
25. Demopulos C, Fava M, McLean NE, Alpert JE, Nierenberg AA, Rosenbaum JF. Hypochondriacal concerns in depressed outpatients. Psychosom Med. 1996;58:314-320.
FREE FULL TEXT
26. Noyes RJ, Holt CS, Kathol RG. Somatization: diagnosis and management. Arch Fam Med. 1995;4:790-795.
FREE FULL TEXT
27. McCrae RR, Costa PT Jr. Validation of the five-factor model of personality across instruments and observers. J Pers Soc Psychol. 1987;52:81-90.
FULL TEXT
|
ISI
| PUBMED
28. Marshall GN, Wortman CB, Vickers RR, Kusulas JW, Hervig LK. The five-factor model of personality as a framework for personality-health research. J Pers Soc Psychol. 1994;67:278-286.
FULL TEXT
|
ISI
| PUBMED
29. Cloninger CR. A systematic method for clinical description and classification of personality variants. Arch Gen Psychiatry. 1987;44:573-587.
FREE FULL TEXT
30. Chien AJ, Dunner DL. The Tridimensional Personality Questionnaire in depression: state vs trait issues. J Psychiatr Res. 1996;30:21-27.
FULL TEXT
|
ISI
| PUBMED
31. Hamilton M. Development of a rating scale for primary depressive illness. Br J Soc Clin Psychol. 1967;6:278-296.
PUBMED
32. Beck AT, Steer RA. Beck Depression Inventory Manual. San Antonio, Tex: Psychological Corp; 1987.
33. Morey LC. An Interpretative Guide to the Personal Assessment Inventory (PAI). Orlando, Fla: Psychological Assessment Resources Inc; 1996.
34. Glass GV, Hopkins KD. Statistical Methods in Education and Psychology. Needham Heights, Mass: Allyn & Bacon; 1996.
35. Williams J, Rost K, Dietrich A, Ciotti M, Zyzanski S, Cornell J. Primary care physicians' approach to depressive disorders: effects of physician specialty and practice structure. Arch Fam Med. 1999;8:58-67.
FREE FULL TEXT
36. Lin E, Katon W, Von Korff M, et al. Relapse of depression in primary care: rate and clinical predictors. Arch Fam Med. 1998;7:443-449.
FREE FULL TEXT
37. Fiscella K, Campbell TL. Association of perceived family criticism with health behaviors. J Fam Pract. 1999;48:128-134.
ISI
| PUBMED
38. Gallo JJ, Marino S, Ford D, Anthony JC. Filters on the pathway to mental health care, II: sociodemographic factors. Psychol Med. 1995;25:1149-1160.
ISI
| PUBMED
39. Andersen R, Francis A, Lion F, Daughety VS. Psychologically related illness and health service utilization. Med Care. 1977;15(5 suppl):59-73.
40. Mumford E, Schlesinger HJ, Glass GV. A new look at evidence about reduced cost of medical utilization following mental health treatment. Am J Psychiatry. 1984;141:1145-1158.
FREE FULL TEXT
41. Robbins JM, Kirmayer LJ, Hemami S. Latent variable models of functional somatic distress. J Nerv Ment Dis. 1997;185:606-615.
FULL TEXT
|
ISI
| PUBMED
42. McLeod CC, Budd MA, McClelland DC. Treatment of somatization in primary care. Gen Hosp Psychol. 1997;19:251-258.
43. Robbins JM, Kirmayer LJ. Transient and persistent hypochondriacal worry in primary care. Psychol Med. 1996;26:575-589.
ISI
| PUBMED
44. Lindsay PG, Wyckoff M. The depression-pain syndrome and its response to antidepressants. Psychosomatics. 1981;22:571-577.
FREE FULL TEXT
45. Fallon BA, Liebowitz MR, Salman E, et al. Fluoxetine for hypochondriacal patients without major depression. J Clin Psychopharmacol. 1993;13:438-441.
ISI
| PUBMED
46. Kreitman N, Sainsbury P, Pearce K, et al. Hypochondriasis and depression in outpatients at a general hospital. Br J Psychiatry. 1965;111:607-615.
47. Silk K. Expert commentary: Axis I-Axis II interactions. J Psychiatr Res. 1996;30:3-7.
FULL TEXT
|
ISI
| PUBMED
RELATED ARTICLES
Somatoform Symptoms and Treatment Nonadherence in Depressed Family Medicine Outpatients
Robert Keeley
Arch Fam Med. 2000;9(1):54.
FULL TEXT
The Archives of Family Medicine Continuing Medical Education Program
Arch Fam Med. 2000;9(1):79-80.
FULL TEXT
THIS ARTICLE HAS BEEN CITED BY OTHER ARTICLES
|
Predictors of Premature Antidepressant Discontinuation in Functional Gastrointestinal Disorders
Sayuk et al.
Psychosom. Med. 2007;69:173-181.
ABSTRACT
| FULL TEXT
Effects of Duloxetine on Painful Physical Symptoms Associated With Depression
Goldstein et al.
Psychosomatics 2004;45:17-28.
ABSTRACT
| FULL TEXT
Impact of Pain on Depression Treatment Response in Primary Care
Bair et al.
Psychosom. Med. 2004;66:17-22.
ABSTRACT
| FULL TEXT
When Are Medication Side Effects Due to the Nocebo Phenomenon?
Palmer et al.
JAMA 2002;287:2502-2504.
FULL TEXT
Is One Selective Serotonin Reuptake Inhibitor Better Than Another?
Elliott et al.
JAMA 2002;287:1935-1938.
FULL TEXT
The Patient with Hypochondriasis
Barsky
NEJM 2001;345:1395-1399.
FULL TEXT
|
