Depression Treatment Patterns among Adults with Chronic Obstructive Pulmonary Disease and Depression

Abstract

Objective

To estimate rates and the patterns of depression treatment among adults with Chronic Obstructive Pulmonary Disease (COPD) and depression.

Methods

We used a retrospective, cross-sectional study design, pooling data from 2010 and 2012 Medical Expenditure Panel Survey (MEPS). The study sample consisted of 527 individuals aged 21 years or older, diagnosed with COPD and depression. Depression treatment was grouped into 3 categories based on those who received: 1) neither antidepressant nor psychotherapy; 2) antidepressants only; and 3) psychotherapy combined with antidepressants (combination therapy). We conducted chi-squared tests and multinomial logistic regressions to examine factors (demographic, socio-economic characteristics, healthcare access, health status, and personal health practices) associated with depression treatment among adults with COPD and depression.

Key Findings

The mean age of the study sample was 55.96 years(SD=13.36). Overall, 18.8% of the sample adults did not report any use of antidepressants or psychotherapy, 58.3% reported antidepressants use only and 23% reported using combination therapy. Females(AOR=1.89, 95% CI= 1.02, 3.55), older adults(>=65 years: AOR=3.69, 95% CI= 1.62, 8.41), adults with fair/poor physical health status(AOR=3.32, 95% CI=1.29, 8.56) and those suffering from anxiety (AOR=1.94, 95% CI= 1.09, 3.46) were more likely to receive antidepressant treatment. Older adults(AOR=2.94, 95% CI=1.05, 8.22), those who were never married(AOR=3.17, 95% CI=1.18, 8.56), suffered from anxiety(AOR=6.01, 95% CI=3.11, 11.61) and current smokers (AOR=2.29, 95% CI= 1.05, 4.98) were more likely to receive combination therapy. Whereas, adults who were uninsured(AOR=0.21, 95% CI= 0.05, 0.86) and did not have physical activity (AOR=0.33, 95% CI= 0.16, 0.67) were less likely to receive combination therapy. Key limitations of our study is that we could not control for the severity of depression or COPD which may have influenced depression treatment.

Conclusion

Efforts to improve depression care among adults with co-occurring COPD and depression may need to be tailored for different subgroups.

Introduction

Depression is a frequently occurring concomitant disease among adults with Chronic Obstructive Pulmonary Disease (COPD) with an estimated prevalence of 24.6%¹. Adults with COPD are 4 times as likely as those without COPD to develop depression²; and twice as likely to develop depression as adults with other chronic conditions such as arthritis, cancer, diabetes, hypertension, and stroke³. Depression leads to worsening of COPD-related outcomes such as COPD-exacerbation frequency^{4, 5}, symptom burden and COPD treatment failure⁶. Depression also increases severity of COPD due to its effect on early addictive smoking and impediment of smoking cessation⁷. In addition, non COPD-related outcomes of depression include longer hospitalization⁸, mortality⁹, impaired functional status, poor exercise capacity⁸, poor quality of life¹⁰, non-adherence to medical treatment, and sleep disturbances⁸.

Randomized clinical trials (RCTs) have demonstrated the beneficial effect of antidepressants use among patients with COPD and depression¹¹. In addition to reduction of depressive symptoms, antidepressants may confer other benefits such as decrease in tobacco cravings, improvement of subjective dyspnea, improvement of appetite, weight loss reversal, decrease in anxiety symptoms and better decision making regarding end-of-life-care preferences¹². Adding psychotherapy to antidepressant treatment (combination therapy) is also an effective strategy in managing depression. Evidence-based clinical practice guidelines ^13–15 on the management of depression in general population have recommended combined therapy (psychotherapy and antidepressant medication) for patients with moderate-to-severe depression. In a recent meta-analysis, researchers showed that adding psychotherapy to antidepressant medications was twice as effective as compared to antidepressant treatment alone ¹⁶ in reducing depressive symptoms. However, evidence regarding combination therapy use in the management of depression in COPD population has been limited¹⁷.

Despite, the high prevalence of depression in patients with COPD and the beneficial effect of depression treatment, nationally representative studies on the rates of depression treatment in real-world settings among adults with COPD and depression have been limited ^{2, 18}. One study conducted in the United States (US) in 2003 reported that only 31% of adults with COPD and depression seen in one primary care setting received treatment for depression². Another study using claims data of elderly (>=65 years) Medicare beneficiaries found that 82.1% of the elderly COPD patients with depression received antidepressants¹⁸. It has to be noted that this study focused only on elderly Medicare beneficiaries. Thus, the extant literature provides no information on depression treatment in adults with COPD and little is known about the various person-level factors that are associated with depression treatment in adults with COPD and depression. Therefore, the primary objective of this study was to estimate the rates of depression treatment with antidepressants and combination therapy among adults with co-occurring depression and COPD. We also examined the patterns of depression treatment by demographic, socio-economic characteristics, access to care, health status, and personal health care practices in a nationally representative sample of adults with COPD and depression.

Methods

Design

A retrospective cross-sectional study design was used.

Data

We used data from the Medical Expenditure Panel Survey (MEPS), a large-scale survey of the civilian non-institutionalized population in the U.S. The Household Component (HC) of MEPS collects demographic characteristics, medical conditions, health status, utilization of healthcare services, charges and payments, access to care, health insurance coverage, income, education and employment on all household members¹⁹. For this study we used the person-level household full year consolidated file, event-level medical conditions and prescribed medicines file. Medical conditions file captures medical conditions of the respondents based on the verbatim text and these texts are converted into International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes by professional coders ²⁰. The event-level prescribed medications file contains detailed records of prescribed medicines such as the national drug code (NDC), medicine name, therapeutic class based on Multum Lexicon classification, sources of payment, fill date and others.

MEPS uses a probability weighted complex multistage survey design with primary sampling units, strata, and person level sampling weights²¹. For the current study, we pooled data from two years (2010 and 2012) to have sufficient sample size and alternate years were used to avoid using two observations per individual.

Study Sample

The study sample comprised of individuals aged 21 years or older, who were diagnosed with depression and COPD and were alive during the calendar years. Depression was identified using ICD-9-CM codes (296.XX, 298.XX, 300.XX, 309.XX and 311.XX) used in published literature and the National Committee on Quality Assurance ^{22, 23}. COPD was identified using previously validated ICD-9-CM codes (491.XX, 492.XX and 496.XX) ^{24, 25}.

Measures

All variables (dependent and independent) used in this analysis were identified from the same calendar year.

Dependent Variable

Antidepressant Treatment

Information on antidepressant prescriptions was derived from Prescribed Medicine file in MEPS using the therapeutic class codes from the Multum-lexicon classification scheme. The therapeutic class code 249 represented antidepressants. In our study, we considered individuals to have received antidepressant treatment if they had one or more prescriptions for antidepressants.

Psychotherapy

Information on Psychotherapy was obtained from the office-based medical provider visit files and the outpatient files. These files capture information on the reason for visits. In our study, we considered individuals to have received psychotherapy if they had at least one visit in which they received psychotherapy.

Based on the receipt of antidepressants and psychotherapy, adults were classified into three mutually exclusive depression treatment categories: 1) received neither antidepressants nor psychotherapy; 2) received antidepressants only; and 3) received psychotherapy combined with antidepressants (combination therapy). As less than 4% (n=23) of the study population reported psychotherapy use only, we excluded individuals receiving “psychotherapy only” from our population.

Independent Variables

Demographic variables included sex, race (white, African American, other racial minorities), age in years (21–39 years, 40–49 years, 50–64 years, 65 and older), marital status (married, widowed, separated/divorced, never married), and metro status (metro, rural). Socio-economic characteristics included education (less than high school, high school, above high school) and poverty status (not poor, poor). Access to care variables consisted of health insurance (public, private, uninsured) and usual source of care (yes/no). The usual source of care variable was based on whether each individual ascertains if there is a particular doctor’s office, clinic, health center, or other place from where an individual usually seek care¹⁹. Health status variables included a diagnosis of anxiety (yes/no), number of chronic conditions other than COPD (none, 1 to 3, >3 conditions) from a list of 12 conditions that included arthritis, asthma, cancer, dementia, diabetes, heart disease, gastroesophageal reflux disease (GERD), hypertension, liver disease, renal disease, stroke, and thyroid disorders, and perceived physical health status (excellent/very good, good, and fair/poor). Perceived mental health status (excellent/very good, good, and fair/poor) was used as a proxy for the severity of depression. Personal health practice variables included Body Mass Index (BMI) categories (underweight/normal, overweight, obese), smoking status (current smoker, other) and physical activity (moderate to vigorous activities 3 times per week, no physical activity).

Statistical Analysis

Chi-square tests were used to analyze the differences in depression treatment among various subgroups of individuals with COPD and depression. Multinomial logistic regressions were employed to analyze the depression treatment patterns by gender, age race, marital status, metro status, insurance status, usual source of care, medical conditions and health status variables. We used “neither antidepressants nor psychotherapy” as the reference group for the dependent variable in the multinomial regression model. There were a few individuals with missing data (n =24) on BMI and smoking and these individuals were excluded from the multinomial logistic regression. Parameter estimates from the multinomial logistic regression were transformed into adjusted odds ratio (AOR) and their corresponding 95% confidence intervals were reported. To account for the complex, probabilistic survey design of MEPS, all statistics were weighted to the national level by incorporating MEPS variance adjustment weights (sampling strata and primary sampling unit) and MEPS person-level weights in all analyses. SAS 9.4 (SAS Institute Inc., Cary, NC) software was used to adjust the estimated standard errors for weighted survey estimates using the Taylor-series linearization approach. This was done by using survey procedures found in SAS such as SURVEYFREQ, SURVEYMEANS and SURVEYLOGISTIC.

Results

The sample description of our study population is presented in Table 1. Our study population consisted of 527 adults with co-occurring COPD and depression. About 70% were females and 82.2% were whites. 45.1% of the study population was in the age group 50–64 years. An overwhelming majority of the population (58.7%) reported the presence of 1 to 3 chronic illnesses such as arthritis, diabetes mellitus, heart diseases, stroke, hypertension and thyroid disorders; nearly 38.4% reported anxiety. In addition, almost 46% of the study population was obese and about 40.8% was current smokers.

Table 1.

Descriptive Statistics of Study Sample (N=527)

Adults with COPD and Depression;

Medical Expenditure Panel Survey, 2010 and 2012

Variable	N	Weighted %
All	527	100.0
Sex
Female	376	69.9
Male	151	30.1
Race
White	368	82.2
African American	71	7.1
Other	88	10.7
Age in years
22–39 years	60	11.6
40–49 years	82	15.5
50–64 years	243	45.1
65 and older	142	27.8
Marital Status
Married	226	44.4
Widow	69	12.6
Separated/Divorced	144	25.9
Never married	88	17.2
Metro
Metro	423	79.6
Rural	104	20.4
Education
Less than High School	127	20.0
High School	125	20.6
More than High School	168	39.1
Employment Status
Employed	135	29.4
Not Employed	392	70.6
Poverty Status
Poor	287	46.3
Not Poor	240	53.7
Insurance Status
Private	211	48.9
Public	261	41.8
Uninsured	55	9.3
Usual Source of Care
Yes	475	90.7
No	51	9.3
Perceived Physical Health Status
Excellent/very good	82	18.2
Good	145	29.6
Fair/poor	300	52.2
Perceived Mental Health Status
Excellent/very good	122	24.9
Good	185	36.1
Fair/poor	220	39.0
Anxiety
Yes	192	38.4
No	335	61.6
Body Mass Index
Underweight/normal	121	23.7
Overweight	142	29.4
Obese	260	45.9
Smoking Status
Current smoker	227	40.8
Other	280	55.6
Physical Activity
3 times per week	133	26.5
No exercise	394	73.5
Number of Chronic Conditions
None	52	12.3
1–3	311	58.7
4 or more	164	29.1

Note: Based on 527 adults, aged 21 years older with self-reported Chronic Obstructive Pulmonary Disease and depression who were alive during the calendar years. Missing categories for education, usual source of care, smoking status, body mass index and physical activity are not displayed.

COPD: Chronic Obstructive Pulmonary Disease

The un-weighted number and weighted percent of depression treatment categories by socio-demographic, medical conditions and health status are presented in Table 2. We found that 18.8% of the study population received neither antidepressants nor psychotherapy, 58.3% reported antidepressants use only and 23% reported the use of combination therapy (Table 2). We observed statistically significant differences in depression treatment by race, age, marital status, education, health insurance, usual source of care, perceived mental health status, anxiety, current smoking status and number of chronic conditions. A significantly higher proportion of racial minorities (other than African Americans) (28.1%) compared to whites (15.9%) did not receive any treatment with antidepressants or psychotherapy. Also higher proportions of uninsured adults (38.7%) as compared to those with private (14.3%) or public insurance (13.8%) did not report any treatment with antidepressants or psychotherapy.

Table 2.

Number and Weighted Percent of Depression Treatment Categories

Adults with COPD and Depression

Medical Expenditure Panel Survey, 2010 and 2012

Variable	No Antidepressants or Psychotherapy		Antidepressants Only		Combination Therapy		sig
	N	Wt. (%)	N	Wt. (%)	N	Wt. (%)
All	99	18.8	307	58.3	121	23.0
Sex
Female	63	14.0	221	59.4	92	26.6
Male	36	21.9	86	55.8	29	22.3
Race
White	60	15.0	232	61.1	76	23.9	**
African American	14	17.7	34	47.4	23	34.9
Other	25	26.1	41	44.3	22	29.6
Age in Years
22–39 years	18	32.0	22	33.1	20	34.9	***
40–49 years	16	16.3	41	47.6	25	36.1
50–64 years	44	13.8	145	61.6	54	24.6
65 and older	21	14.0	99	69.5	22	16.4
Marital Status
Married	37	13.0	157	72.2	32	14.8	***
Widow	9	11.0	44	61.5	16	27.5
Separated/Divorced	32	22.6	75	46.8	37	30.5
Never married	21	19.6	31	37.4	36	42.9
Metro
Metro	85	17.8	240	57.5	98	24.8
Rural	14	10.8	67	61.7	23	27.5
Education
Less than High School	24	18.2	80	67.0	23	14.8	**
High School	29	22.3	76	58.1	20	19.6
More than High School	25	11.7	97	58.3	46	30.0
Employment Status
Employed	32	19.6	78	55.1	25	25.2
Not Employed	67	15.0	229	59.7	96	25.3
Poverty Status
Poor	56	17.8	154	54.8	77	27.4
Not Poor	43	15.1	153	61.4	44	23.5
Insurance Status
Private	36	14.3	133	59.5	42	26.2	***
Public	39	13.8	149	59.2	73	27.0
Uninsured	24	38.7	25	48.5	6	12.8
Usual Source of Care
Yes	81	14.6	286	60.3	108	25.0	**
No	17	31.5	21	40.1	13	28.4
Perceived Physical Health Status
Excellent/very good	18	20.6	48	58.8	16	20.5
Good	26	15.2	85	54.0	34	30.8
Fair/poor	55	15.5	174	60.6	71	23.9
Perceived Mental Health Status
Excellent/very good	25	17.1	81	65.7	16	17.2	***
Good	25	11.4	132	71.2	28	17.5
Fair/poor	49	20.5	94	41.7	77	37.8
Anxiety
Yes	22	9.1	97	48.5	73	42.3	***
No	77	20.8	210	64.4	48	14.7
Body Mass Index
Underweight/normal	26	17.6	70	59.9	25	22.4
Overweight	22	12.5	84	62.2	36	25.3
Obese	50	18.2	150	54.5	60	27.3
Smoking Status
Current smoker	40	14.2	125	53.5	62	32.3	*
Other	53	16.9	171	62.3	56	20.8
Physical Activity
3 times per week	25	13.4	76	55.4	32	31.2
No exercise	74	17.4	231	59.4	89	23.2
Number of Chronic Conditions
0	16	24.6	22	43.5	14	31.9	*
1–3	61	16.6	180	57.6	70	25.7
4 or more	22	12.3	105	66.0	37	21.7

Note: Based on 527 adults, aged 21 years older with self-reported Chronic Obstructive Pulmonary Disease and depression who were alive during the calendar year (2010 and 2012). Missing categories for education, usual source of care, smoking status, body mass index and physical activity are not displayed. Asterisks represent significant group differences by depression treatment categories based on chi-square tests.

COPD: Chronic Obstructive Pulmonary Disease; Sig: Significance; Wt: Weighted;

^***p < .001;

^**.001 ≤ p < .01;

^*.01 ≤ p < .05.

Adjusted odds ratios (AOR) and 95% CIs from multinomial logistic regression analysis based on depression treatment of the various subgroups are presented in Table 3. Females (AOR=1.89, 95% CI= 1.02, 3.55), older adults (>=65 years: AOR=3.69, 95% CI= 1.62, 8.41), adults with fair/poor physical health status (AOR=3.32, 95% CI=1.29, 8.56) and those suffering from anxiety (AOR=1.94, 95% CI= 1.09, 3.46) were more likely to receive antidepressant treatment compared to males, younger adults aged 22 to 39 years, adults with very excellent/very good physical health status and those who did not suffer from anxiety respectively. In contrast, adults with fair/poor mental health status (AOR=0.36, 95% CI=0.18, 0.75) as compared to adults with excellent/ very good mental health status were less likely to receive antidepressant treatment. Older adults (AOR=2.94, 95% CI=1.05, 8.22), those who were never married (AOR=3.17, 95% CI=1.18, 8.56), suffered from anxiety (AOR=6.01, 95% CI=3.11, 11.61) and current smokers (AOR=2.29, 95% CI= 1.05, 4.98) were more likely to receive combination therapy compared to younger adults aged 22 to 39 years, those who were married, did not have anxiety and were not current smokers respectively. Whereas, adults who were uninsured (AOR=0.21, 95% CI= 0.05, 0.86) and did not have physical activity (AOR=0.33, 95% CI= 0.16, 0.67) were less likely to receive with the combination therapy as compared to those who had private insurance and had regular physical activity respectively.

Table 3.

Adjusted Odds Ratios and 95% Confidence Intervals of Independent Variables from Multinomial Logistic Regression on Depression Treatment

Adults with COPD and Depression

Medical Expenditure Panel Survey, 2010 and 2012

Variable	Antidepressants only			Combination Therapy
	AOR	95% CI	Sig	AOR	95% CI	Sig
Sex
Female	1.89	[ 1.02, 3.55]	*	2.13	[ 0.87, 5.22]
Male (Ref)
Race
White (Ref)
African American	0.99	[ 0.53, 1.84]		1.61	[ 0.90, 2.86]
Other	0.33	[ 0.17, 0.65]	**	0.66	[ 0.26, 1.68]
Age in Years
22–39 years
40–49 years	2.32	[ 0.93, 5.82]		1.73	[ 0.60, 5.03]
50–64 years	3.47	[ 1.46, 8.22]	**	2.20	[ 0.89, 5.44]
65 and older	3.69	[ 1.62, 8.41]	**	2.94	[ 1.05, 8.22]	*
Marital Status
Married (Ref)
Widow	0.69	[ 0.24, 1.99]		1.75	[ 0.57, 5.42]
Separated/Divorced	0.36	[ 0.18, 0.70]	**	1.01	[ 0.43, 2.36]
Never married	0.56	[ 0.21, 1.52]		3.17	[ 1.18, 8.56]	*
Metro
Metro (Ref)
Rural	1.29	[ 0.71, 2.33]		2.02	[ 0.75, 5.42]
Poverty Status
Not Poor (Ref)
Poor	1.12	[ 0.62, 2.03]		0.89	[ 0.44, 1.78]
Insurance Status
Private (Ref)
Public	0.95	[ 0.48, 1.89]		0.82	[ 0.38, 1.78]
Uninsured	0.45	[ 0.18, 1.12]		0.21	[ 0.05, 0.86]	*
Usual Source of Care
No (Ref)
Yes	1.99	[ 0.95, 4.18]		1.91	[ 0.85, 4.31]
Perceived Physical Health Status
Excellent/very good (Ref)
Good	1.26	[ 0.54, 2.93]		1.68	[ 0.57, 4.92]
Fair/poor	3.32	[ 1.29, 8.56]	*	1.52	[ 0.43, 5.36]
Perceived Mental Health Status
Excellent/very good (Ref)
Good	1.27	[ 0.63, 2.59]		1.07	[ 0.45, 2.53]
Fair/poor	0.36	[ 0.18, 0.75]	**	1.73	[ 0.71, 4.21]
Anxiety
No (Ref)
Yes	1.94	[ 1.09, 3.46]	*	6.01	[ 3.11, 11.61]	***
Body Mass Index
Underweight/Normal	1.85	[ 0.85, 4.00]		2.08	[ 0.76, 5.73]
Overweight	0.93	[ 0.46, 1.91]		1.41	[ 0.65, 3.05]
Obese (Ref)
Smoking Status
Current Smoker	1.79	[ 0.85, 3.75]		2.29	[ 1.05, 4.98]	*
Other (Ref)
Physical Activity
3 times per week (Ref)
No exercise	0.54	[ 0.27, 1.09]		0.33	[ 0.16, 0.67]	**
Number of Chronic Conditions
0 (Ref)
1–3	1.32	[ 0.63, 2.77]		0.96	[ 0.36, 2.56]
4 or more	1.52	[ 0.56, 4.13]		0.87	[ 0.27, 2.80]

Note: Based on 527 adults, aged 21 years older with self-reported Chronic Obstructive Pulmonary Disease and Depression who were alive during the calendar year (2010 and 2012). Asterisks represent significant group differences in type of treatment compared to the reference group. The reference group for the dependent variable in the multinomial logistic regression was “No Antidepressants and No Psychotherapy”.

AOR: Adjusted odds ratio; COPD: Chronic Obstructive Pulmonary Disease; CI: Confidence Interval; Ref: Reference Group; Sig: significance.

^***p < .001;

^**.001 ≤ p < .01;

^*.01 ≤ p < .05.

Discussion

Our study analyzed the rates and patterns of depression treatment among adults with co-occurring COPD and depression using data from the nationally representative household survey, MEPS. The results from our study show that 81.5% of the population with COPD and depression received treatment for depression with either antidepressants or combination therapy. There are no studies on depression treatment among all adults with COPD and depression in the US. Therefore, we compared the rate of antidepressants use among older adults (i.e. age ≥ 65 years) in our study to one published study on antidepressant treatment among Medicare beneficiaries with COPD and depression¹⁸. The percentage of antidepressants use in our study (86%) is consistent with Qian et al., which showed that 82% of older Medicare beneficiaries with both COPD and depression received treatment with antidepressants. However, another study by Kunik et al. reported that only 31% of the patients in one primary care center with co-occurring chronic breathing disorders and depression received treatment for depression². Because this study was more than a decade old, the difference in the rate of depression treatment may be in part due to study settings, single site, changes in practice patterns over time and increasing trends in detection and management of depression in outpatient setting²⁶.

We found that more than two-thirds (70%) of adults with co-occurring COPD and depression were women. It is well-established that that women are more likely to have depression than men across numerous chronic conditions^{27, 28}. Among patients with COPD, epidemiologic studies have reported higher prevalence of depression among women than men²⁹. Although historically COPD has been considered to affect men disproportionately³⁰, in the past decade there has been a shift in the trend of COPD from a male predominant disease to a female predominant disease^{31, 32}.Therefore, it is not surprising that an overwhelming majority of our study sample were women.

Our results also indicate that women with COPD and depression were significantly more likely to receive antidepressant treatment as compared to men. The sex-related disparity in depression treatment has been attributed to the lowered willingness of men to seek treatment for mental disorders ^33–36. In a landmark multisite trial (IMPACT) conducted among older adults with depression, qualitative assessments showed that the sex-related disparities in depression treatment could be attributed to the stereotypical masculine ideologies such as emotional control, self-reliance and stoicism as well as high social stigma associated with seeking healthcare for chronic mental disorders³³.

We also found that adults with co-occurring COPD and depression who had fair/poor physical health status were more likely to receive depression treatment as compared to adults with excellent/very good physical health status. A possible explanation for the high antidepressant use could be the effectiveness of antidepressants in the alleviation of COPD related symptoms such as decrease in exacerbation frequency, dyspnea and COPD related treatment failure ^{4, 5, 11}.

Conversely, adults with fair/poor mental health status were less likely to receive depression treatment than adults with excellent/ very good mental health status. In a feasibility trial of antidepressant therapy in patients with COPD, it was found that the majority (72%) of depressed patients refused antidepressant therapy due to anticipation of adverse events, resentment in taking multiple medications, and denial of depressive symptoms³⁷. This might explain the reason for low antidepressant use among adults with more severe mental health. Future studies need to explore the barriers to antidepressant use among this subpopulation.

Adults with anxiety were 1.9 times as likely to receive antidepressant treatment and 6 times as likely to receive combination therapy as compared to those without anxiety. Previous research have demonstrated that anxiety among patients with COPD and depression leads to significant increase in behavioral and psychological symptoms of distress such as higher levels of fatigue, shortness of breath, and frequency of COPD symptoms³⁸. One plausible reason for the higher likelihood of combination therapy among adults with COPD and depression may be due to the effectiveness of combination therapy in treating a wide variety of anxiety symptoms³⁹.

One noteworthy finding from this study is that current smokers were 2.3 times as likely to receive combination therapy as compared to those who were not current smokers. Prior research has shown that among adults with COPD and depression, who also smoke, have a nearly 40-fold increase in risk of severity from COPD interaction effects ^{7, 9}. It is well-documented that treatment for depression may have the added benefit of facilitating smoking cessation in addition to reducing depressive symptoms ^40–42 among adults with COPD and depression. These findings suggest that current smokers may be treated with combination therapy to provide relief from depression as well as smoking cessation.

Despite the effectiveness of combination therapy in providing relief from depression, only 23% of adults with COPD and depression received combination therapy. Furthermore, certain subgroups were less likely to receive combination therapy. For example, we found that patients who lacked regular physical activity were less likely to receive combination therapy. Exercise is particularly challenging in patients with COPD due to alterations in their skeletal muscle caused by their illness ⁴³. In our study, approximately 3 in 4 patients with COPD and depression reported lack of exercise perhaps due to exercise intolerance in this population. Therefore, physicians may need to pay particular attention to these patients in making a decision about the modality of depression treatment.

Our study findings also highlighted lack of health insurance as a barrier to receipt of combination therapy. Uninsured may not receive combination therapy due to high out-of-pocket spending burden. A study of financial burden and out-of-pocket expenditures for mental health reported that uninsured adults were more likely to bear a significant burden as compared to insured adults ⁴⁴. This finding suggest a need for clinical practice and policy efforts to be integrated. For example, clinicians can attempt to reduce access barriers to combination therapy by providing linkages of the various behavioral and social services that are available in the community for the underserved or the uninsured⁴⁵.

Conversely, results from our study also indicate that certain subgroups of adults - those who were older (>=65 years), never married, current smokers and who had anxiety - were more likely to receive combination therapy. Our study did not examine the reasons behind the higher rates of combination therapy in these groups. Therefore, future studies are needed to explore the motivational determinants of combination therapy in these patient subgroups.

The findings of this study are subject to certain limitations. As data from MEPS are self-reported, it is subject to recall bias. Also, due to the cross-sectional nature of the study, causal relationship between depression treatment and other variables cannot be established. Furthermore, due to the absence of disease specific severity measure in MEPS, the severity of COPD or depression could not be taken into account. In addition, we did not analyze the use of alternative and complementary therapies of depression in this population.

Conclusion

Notwithstanding the limitations, our study findings identified subgroups of patients with COPD and depression who may be vulnerable for non-receipt of effective depression treatment modality (i.e. combination therapy). In our study some subgroups of patients were more likely to receive combination therapy, suggesting that these subgroups may have accepted combination therapy. Future research needs to explore whether other factors such as patient preferences, attitudes, knowledge and willingness to accept combination therapy that were not measured in our study contributed to the high rates of combination therapy in these groups.