Comorbid Chronic Illness and the Diagnosis and Treatment of Depression in Safety Net Primary Care Settings

Background

One hundred and twenty-five million people in the United States suffer from a chronic physical condition, and approximately 60 million of these have more than one chronic conditions.¹ Chronic physical conditions also account for considerably disproportionate health care utilization and cost among affected individuals.^2,3 Depressive disorders are associated with chronic physical conditions 20% to 50% of the time,^4–10 with such co-occurrence reported to predict higher morbidity and worse treatment outcomes.^11–28

Primary care settings are important for the treatment of many mental health conditions, and primary care providers are often the sole contacts for more than 50% of patients with a mental illness.^29–31 These settings are also important health care delivery platforms for individuals with chronic physical conditions, particularly minority Hispanic and African-American populations. However, the quality of depression care in these settings is often poor; depression is under-diagnosed and under-treated close to 50% to 65% of the time in these settings.^32,33 Many factors have been attributed to this poor quality of depression care, including provider-related factors such as disposition, skills, attitudes, and practice toward mental health care as well as patient-related factors including perceived stigma associated with mental disorders and treatment, preponderance of somatic symptomatology, and a lack of patient awareness of psychological distress.^34–36

There is some evidence that multiple competing demands affect the quality of care provided in primary care settings for many medical conditions,^37–43 with some studies beginning to examine the effects of these demands on mental health care.^44–47 However, the evidence is mixed regarding the relative effects of comorbid physical conditions on depression care. In 2000, 2 studies reported that chronic physical comorbidity decreased the probability of depression being discussed or noticed during a clinic encounter.^46,48 Another study in 2002, however, reported similar rates of treatment of patients with depression alone when compared with patients with depression and comorbid physical conditions but worse depression outcomes in the later group.⁴⁹ Similarly, a more recent study also found that depressed people with chronic medical conditions were significantly more likely to receive guideline-level care for depression than were depressed people without chronic medical conditions.⁵⁰ In another study, Harman et al⁵¹ reported that competing demands did not result in lower quality of depression treatment in older people. There is a strong need for further clarity regarding the role of comorbid chronic conditions on the quality of depression care observed in primary care settings, particularly public safety-net settings serving underserved Hispanic and African-American populations.

Objectives

This study estimated the association of comorbid chronic medical conditions with the diagnosis, treatment, and follow-up care for depression in Hispanic and African-American individuals receiving health care in safety-net primary care settings. We hypothesize that competing demands will reduce the likelihood of good quality depression care for individuals with both depression and comorbid medical conditions when compared with individuals with depression alone.

Results

A total of 2540 participants were screened for depression using the PHQ-2.⁵² Of these, 415 participants scoring ≥3 on the PHQ-2 screener were further screened using the PHQ-9. A total of 315 participants scored ≥10 and met the PHQ-9 criteria for moderate-severe depression. Specifically, 229 (72.3%) and 123 (39.0%), respectively, had a chronic medical condition defined by a CCI score of ≥1; 243 (77.1%) were women; 221 (70.1%) were aged 45 years or older; 207 (66.7%) did not receive a diagnosis of depression; 212 (67.3%) did not receive guideline-concordant care; and 236 (83.7%) did not get guideline-concordant follow-up care for depression. Table 2 presents more details of the characteristics of the study sample.

Bivariate analysis comparing individuals with depression and individuals with depression and chronic medical conditions showed that being of Latino and African-American ethnicity, older, and having more frequent hospital visits within the past 12 months was significantly associated with the co-occurrence of depression and chronic medical conditions (Table 3). Being of Latino and African-American ethnicity and older age was also statistically significantly associated with the co-occurrence of depression and chronic medical conditions using 2 separate chronic comorbidity categories (CCI score of 1 to 2 or ≥3). Tables 2 and 3 present a more detailed description of the bivariate analysis results.

Discussion

When setting out to examine the influence of comorbid medial conditions, we hypothesized that competing demands would reduce the likelihood of good quality depression care of patients with both depression and comorbid medical conditions as compared with patients with depression alone. We also sought to clarify the relationship between these demands and good quality care, particularly because the evidence regarding the potential effects of comorbidity on depression care is mixed.^46,48–51 This study builds on these previous studies. For example, data from a 2000 study by Borowsky et al,⁴⁸ collected some time ago, might not appropriately reflect physician depression care practice, particularly with very concerted efforts made to improve the quality of mental health care in primary care settings. In addition, this same study determined the provider practice of using provider self-reporting of mental health care indices. Similarly, Kurdyak et al⁵⁰ in 2004 and Rost et al⁴⁶ in 2000 also used patient self-reports of physician care as a proxy measure of guideline-appropriate care, thus increasing the likelihood of errors associated with under- or over-reporting of physician recommendations. However, our study evaluated actual care practices using medical record reviews as a more objective measure of actual depression care. The key findings from study suggest that (1) the comorbid chronic medical conditions are not associated with a difference in the likelihood of diagnosis, guideline-appropriate treatment, or follow-up care for depression; (2) although severe depression is the only independent predictor of physician depression diagnosis, patients with more severe forms of depression are less likely to receive guideline-appropriate follow-up care when their medical comorbidity status is additionally considered; and (3) the quality of depression care, evidenced by overall low diagnosis and treatment rates, remains poor for patients receiving care in primary care settings. The results of this study regarding the rate of the under-diagnosis of depression is slightly higher than reported in previous studies (67% vs 50% to 65%).^32,33,64,65 These findings suggest an important need to continue exploration of avenues for improving provider recognition of depression in primary care settings.

In contrast to the studies reporting a reduced likelihood of diagnosis of depression when associated comorbid medical conditions occur and consistent with our primary hypothesis,^46,48 this study found that comorbid medical conditions were not associated with a change in the likelihood of a depression diagnosis by physicians. In fact, the only significant independent predictor of a depression diagnosis was the presence of severe depression, a finding contrary to those reported by Rost et al.⁴⁶

Guidelines provide guidance for depression treatment, including pharmacotherapy, psychotherapy, and adequate follow-up care for reassessment and treatment maintenance or termination in primary care settings.^66,67 Numerous studies document low rates of adherence to guideline recommendations for antidepressant use or psychotherapy in these same settings.^68,69 When evaluating the potential impact of comorbid medical conditions on guideline-concordant treatment and follow-up care, the findings of this study suggest low adherence to guideline-recommended care.

Pharmacotherapy has been shown to be effective alone or in combination with psychotherapy in the treatment of depression.^70,71 It was not surprising that guideline-based treatment was provided to only approximately a third of study participants meeting the PHQ-9 diagnostic criteria for depression because most of the depressed individuals remained undiagnosed. However, it is encouraging to note that, among patients receiving a depression diagnosis, approximately 74% actually received guideline-based treatment. This is similar to findings from a recent (2007) study by Hepner et al⁷² in which initiation of treatment for depression was also found to be high. The findings in this study of high appropriate treatment rates unassociated with the occurrence of medical comorbidity is consistent with findings that most primary care clinicians are disposed to treat depression and have sufficient competence about appropriate treatment for their patients with depression.⁷³

It is also important to note that most of the participants received pharmacotherapy alone (84%), even though guidelines allow for the recommendation of either pharmacotherapy or psychotherapy, and only 2% of patients received psychotherapy alone. These results vary from previously reported studies in which only 49% of the patients with newly diagnosed depression were reported to receive antidepressant medication and between 28% and 34% received both psychotherapy and medications.^74,75 This finding outlines further the importance of physician diagnosis of depression because it suggests that they will appropriately treat once depression is diagnosed. Further studies to explore the relationship between provider preference for pharmacotherapy and established treatment guidelines are required.

It is important to mention a priori that the follow-up care measure used for this study is not an empiric measure of depression-specific follow-up visits. Physician follow-up recommendations were compared with the minimum recommended interval for follow-up visits based on depression severity.⁵⁷ However, guideline-concordant follow-up care is consistently poor even among patients receiving a physician diagnosis of depression (28.3%); this demonstrates that, unlike the observations regarding treatment for participants diagnosed with depression by physicians, adequate follow-up care for these same patients remains suboptimal. The results are particularly concerning because patients with moderately severe or severe depression are less likely to receive guideline-concordant follow-up care when compared with patients with moderate depression. Surprisingly, the patients who are at the most risk for medication side effects, increased suicidal tendencies, and possible experience of adherence issues are inappropriately followed-up.⁷⁶ In addition, these findings also suggests that an increased exposure to primary care providers resulting from the co-occurrence of chronic medical conditions requiring regular follow-up does not improve the likelihood that they will receive appropriate depression care. Finally, this study demonstrates that the co-occurrence of chronic medical conditions with depression is not associated with the diagnosis, treatment, or follow-up care of depression. Even when we controlled for the clinical status of the using a standardized comorbidity index, there were no documented differences in any of these processes of care. All of the multivariate models that controlled for the severity of depression and demographic variables still failed to result in a statistically significant increased likelihood of a difference in these care processes. Summarily, this study's finding that competing demands for care may not plausibly explain the current quality of depression care, particularly among minority underserved populations with limited access to mental health care resources, is important within the following context: the inadequacy of appropriate depression diagnosis during patient visits regardless of comorbidity status underlies the fact that the under-diagnosis and under-treatment of depression may be under the influence of factors other than time-sensitive patient visits and the competition for physician's time. This is particularly important because of evidence suggesting that perceptions of competing demands may affect primary care physician's mental health care practices.⁷⁷ Interventions that promote depression care in primary care settings may find these findings useful for addressing such perceptions. Secondly, as the burden of comorbid chronic medical conditions continue to increase in the United States, these findings continue to suggest an important role for studies and interventions to address the important role of depression on treatment outcomes for comorbid chronic medical conditions, particularly among minority underserved populations.

Our findings are subject to some important limitations. It is important to report that the PHQ-9 is used as the “gold standard ” measure for depression. Although PHQ-9 scores of ≥10 have been reported to have a sensitivity of 88% and a specificity of 88%, some misclassification of depressed individuals may affect the reported study findings. We were unable to further delineate medication dosage for the guideline-concordant depression treatment variable. This information might further stratify identified guideline-concordant care and result in an observation of lower frequencies for the guideline-concordant care described by this study. The reliability of medical record abstractions may have also limited our findings; to address this we conducted re-abstractions on no fewer than 10% of all medical records. We recognize that medical records may not accurately reflect physician practice, but it is noteworthy to mention, however, that, because initiation and continuity of care for depression is based mainly elicitation and recording of verbal assessments rather than empirical biochemical markers, deficiencies in care or documentation represents deficiencies in appropriate care. Finally, it is possible, though unlikely, that the observations of this study may be explained by clustered physician practices.