Comorbid Chronic General Health Conditions and Depression Care: A Population-Based Analysis

About one in 20 individuals in the general population has major depression (1,2). Among persons with chronic general health conditions, such as diabetes and heart disease, the prevalence of depression is substantially higher, with up to one in five individuals diagnosed as having comorbid depression (3,4). The co-occurrence of depression and chronic general health conditions has generally been attributed to the reciprocal relationship between these conditions (5,6). On the one hand, people with chronic general health conditions (hereafter referred to as “chronic health conditions”) are prone to depression, in part at least because of the stress, pain, discomfort, and limitations on activity that chronic health conditions bring about. Conversely, depression brings about physiological and behavioral changes that increase one’s susceptibility to developing chronic health conditions (7,8). A more recent view also proposes that some forms of depression and chronic health conditions share common etiological pathways, such as childhood abuse or trauma (9).

When depression co-occurs with chronic health conditions, the combination can have serious negative health consequences, such as increased risk of cardiac events after myocardial infarction and increased risk of mortality among those with heart disease (10,11), increased functional disability (12–14), and decreased quality of life (15). Indirectly, through its effect on behavior, comorbid depression lowers adherence to treatment regimens (16–18) and physical rehabilitation for the chronic health condition (19). At the health system level, the co-occurrence of depression and general health conditions results in higher health care costs (12,20,21).

In the general population, symptoms of depression are managed by using psychological and pharmaceutical therapies or a combination of both (22,23). However, given that chronic health conditions potentially moderate the effectiveness of these treatments, recommendations for treatment of depression among persons with a chronic health condition vary based on the specific condition and on the strength of available evidence (7,24).

Despite the importance of managing depression symptoms and the availability of effective treatments, there is a persistent concern that depression is often untreated or undertreated among individuals with chronic health conditions. In a recent systematic review, 13 of 46 studies reported lower levels of treatment or follow-up care among individuals with depression and chronic health conditions compared with those with depression only (25). One possible reason for undertreatment is related to the difficulty associated with differentiating depression symptoms from those caused by chronic health conditions (26). Another is the common view that depression is a normal response to the stresses brought about by chronic health conditions and that it should go away on its own, over time, or once the chronic health condition has been addressed (8). Compounding these issues are the apprehensions shared by many about potential adverse risks associated with antidepressant therapy when administered in conjunction with medications for chronic health conditions (26,27). These views and challenges, altogether, predispose clinicians and patients to consciously or unconsciously agree on prioritizing the chronic health condition, particularly in primary health care settings, where clinical encounters are brief and where patients with chronic health conditions often have multiple health concerns that compete for the attention of a general practitioner (GP) (28,29).

The presence of chronic health conditions, however, does not consistently result in receipt of lower levels of depression care (25). Studies that have used measures of comorbidity to examine the link between depression care and chronic health conditions suggest that persons with one or more chronic health conditions actually receive higher levels of care (30,31), whereas other researchers have reported that levels of depression care may vary by specific type of chronic health condition (32). This inconsistency in patterns of depression care underscores the importance of conducting further studies, particularly studies that examine specific chronic health conditions. To this end, this study examined depression care patterns among persons with specific chronic health conditions.

Methods

Data Sources

We retrospectively examined health administrative data from virtually everyone in the Canadian province of British Columbia (BC), except for a small percentage (4%) of individuals (aboriginal people, individuals in prison, and police and military personnel) whose health care is covered under federal jurisdiction. Access to the deidentified and individual-level data was granted by the BC Ministry of Health Services and the BC College of Pharmacists and was facilitated and regulated by Population Data BC. Ethics approval was provided by the Behavioral Research Ethics Board of the University of British Columbia. A list of data sources and description of the data are available in Table 1.

Study Cohort

People who received depression diagnoses between January 1 and December 31, 2012, were identified (Table 2). Using the earliest recorded diagnosis as index date, we conducted a review of each individual’s health administrative data to identify major depression, which was indicated by having at least one inpatient or two outpatient depression diagnoses during the previous 12 months. We used the criterion of two outpatient diagnoses to increase the likelihood that an individual actually had depression. For inpatient diagnosis, we used the diagnosis recorded at the time of discharge. Data were available for exactly 12 months before and 12 months after the index period for everyone included in the study cohort. This was done to reduce or eliminate potential biases that could arise when fixed dates are used to ascertain disease status and measure health service use.

The presence of chronic health conditions was assessed in a similar manner.

We excluded individuals with incomplete follow-up, such as individuals who moved in and out of the province or who died at any point during the observation period. Persons under 19 years of age were also excluded to restrict the analysis to the adult population.

Outcome Variables

We examined a number of outcome variables to assess and compare the 12-month use of depression-related health services among those with and without chronic health conditions. We derived binary indicators (yes or no) of whether patients received counseling/psychotherapy, antidepressant therapy, and either counseling/psychotherapy or antidepressant therapy. We created discrete variables that summarize counts of GP visits overall, GP mental health visits, and counseling/psychotherapy sessions. Claims for GP visits can have up to five diagnoses, with the primary diagnosis listed as the first diagnosis. The identification of GP mental health visits was made by using the primary diagnosis only. We also calculated a continuous variable for GP continuity of care, which measures the degree to which GP visits were made to the same physician; scores range from a minimum value of 0 to a maximum value of 1 (33). Last, we calculated the proportion of days covered (PDC) for antidepressant therapy, which is another continuous variable that quantifies the extent to which individuals were on antidepressants within 180- and 365-day periods. PDC was calculated by creating time arrays that reflect the dates encompassed by each fill and summing up only those days that were covered by antidepressants and dividing them by 180 or 365. Values for this variable range from 0 to 1, with higher values indicating good adherence (34,35).

Explanatory Variables

The main explanatory variable used in this analysis was the presence of the following chronic health conditions: diabetes, congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD), asthma, cerebrovascular disease (CVD), and ischemic heart disease. These chronic health conditions were chosen because in the past decade they have been the target of primary health care reforms in BC (36). Algorithms for detecting these conditions by using health administrative data were adopted from a previously completed study (37).

Analysis

We summarized the demographic characteristics and patterns of health service use of the study cohort by using counts, percentages, and means and standard deviations. We used generalized linear models (logistic, negative binomial, and log normal) to determine whether persons with chronic health conditions were more or less likely to receive various types of health services relative to persons without chronic health conditions. The models were run for each chronic health condition with the absence of the condition as the reference category (for example, patients with diabetes versus patients without diabetes). Separate models with an overall indicator of chronic health condition were also fitted to determine whether having at least one of the seven chronic health conditions was associated with lower or higher use of health services compared with having none of the chronic health conditions. All models were adjusted for sex, age, income, and place of residence.

We used SAS/SQL software, version 9.4, to link and manage multiple databases and Stata, version 13.1, to generate the model results.

Results

A total of 105,476 individuals with existing depression were identified in 2012. The prevalence of depression was disproportionately high among women, middle-aged individuals, and individuals who lived in low-income or urban areas (Table 3).

Psychological, Antidepressant, and Combination Therapy

Around 91% of the cohort received either psychological (counseling/psychotherapy) or antidepressant therapy, with antidepressant therapy being the modal treatment (Table 3). Less than half (45%) of the cohort received both counseling/psychotherapy and antidepressant therapy. A higher proportion (78%) of women were treated with antidepressant therapy compared with men (73%), whereas a slightly higher proportion of men (60%) received counseling/psychotherapy compared with women (59%). Similarly, a higher proportion of rural residents (78%) were treated with antidepressant therapy compared with urban residents (76%), whereas a higher proportion of urban residents (60%) were treated with counseling/psychotherapy compared with rural residents (56%).

Overall, the proportion of individuals who received counseling/psychotherapy was higher among persons with one or more chronic health conditions compared with those without a chronic health condition, whereas the proportion of persons who received antidepressant therapy was the same regardless of the presence of chronic health conditions. This general pattern of use, however, was not evident among persons with CVD—the proportion of individuals with CVD who received counseling/psychotherapy was lower than for any other chronic health condition and lower than for persons with no chronic health conditions.

Results from the fully adjusted multivariable logistic regression models showed similar trends (Table 4). In general, persons with chronic health conditions other than CVD had higher odds of receiving counseling/psychotherapy compared with persons without the same condition. However, persons with chronic health conditions had similar or lower odds of using antidepressant therapy compared with persons without chronic health conditions. A trend was found toward lower odds of receiving any depression care (counseling/psychotherapy or antidepressant therapy) among those with CVD compared with those without CVD.

GP Visits and GP Continuity of Care

On average, members of the study cohort visited their GPs about 12 times a year for any reason, and about four times a year specifically for mental health reasons (Table 5). The mean continuity of care index was higher (.80) for GP mental health visits than for any GP visits (.73). In terms of sociodemographic characteristics, women and older individuals had a slightly higher mean number of GP visits overall but a lower mean number of GP mental health visits compared with men and younger individuals, respectively. Compared with younger individuals, those over age 75 had the lowest mean number of GP mental health visits but the highest level of continuity of care for mental health.

Compared with individuals with no chronic health conditions, individuals with one or more chronic health conditions had generally lower numbers of GP mental health visits (Table 5). All continuity of care indices were higher among persons with chronic health conditions compared with individuals with no chronic health conditions.

Results from the fully adjusted multivariable regression analyses suggest that the number of mental health GP visits varied by chronic health condition (Table 4). Specifically, the number of mental health GP visits was higher among those with COPD and asthma and lower among those with diabetes. Overall, having one or more chronic health conditions was not associated with a higher number of GP visits for mental health reasons.

Continuity of care, for GP visits and GP mental health visits, remained marginally higher among those with chronic health conditions, even after adjustment for differences in sociodemographic variables by using multivariable regression analysis (Table 4).

Number of Counseling or Psychotherapy Sessions

On average, cohort members received about three counseling/psychotherapy sessions in a year, except those who resided in rural areas or those who were over 75 years of age, who received fewer (Table 3). The mean number of counseling/psychotherapy sessions was generally higher among those with chronic health conditions compared with no chronic health conditions, except among individuals with CHF or COPD.

Adjusted results from the multivariable regression models showed the same general pattern of having a higher number of counseling/psychotherapy sessions among those with one or more chronic health conditions compared with those without a chronic health condition (Table 4).

Comparison of PDC

The mean PDC for the cohort was .81 for the 180-day period and .76 for the 365-day period. Younger individuals had lower PDCs than older individuals (Table 3).

Individuals with chronic health conditions had slightly higher PDCs than individuals without chronic health conditions. Results from the multivariable regression analyses indicated that the relative differences between the PDCs of individuals with and without a chronic health condition were small in magnitude (Table 4).

Discussion

We compared health service use among patients diagnosed as having major depression and a chronic health condition and patients diagnosed as having major depression only. Our results indicate that, in general, the individuals with chronic health conditions received higher levels of mental health care for all outcome variables except antidepressant therapy and GP mental health visits. Furthermore, our results suggest a trend toward lower use of depression-related health care among the patients with CVD.

Because we are unaware of studies that looked at GP continuity of care and GP mental health visits among individuals with depression and chronic health conditions, we are unable to compare our results with those of previous studies. However, our results are similar to studies that found positive associations between having one or more chronic health conditions and number of psychological therapy sessions (30,31) and likelihood of receiving any type of depression care (30,31,38).

Our finding about the lower odds of being on antidepressant therapy among individuals with chronic health conditions contradicts studies that report associations in the opposite direction (39,40). The incongruence with previous reports is probably because our cohort comprised prevalent cases of depression only. In studies that examined prevalent depression cases (individuals who have received previous depression treatment), antidepressant therapy has been reported to be less likely among individuals with chronic health conditions (25,41).

Our results underscore the utility of studying specific chronic health conditions when examining differences in depression care. Restricting the analysis to counts or presence of chronic health conditions would never have revealed the trend toward lower odds of receiving counseling/psychotherapy or antidepressant therapy among patients with CVD. Also we would have never found out that persons with comorbid diabetes have relatively fewer GP mental health visits.

The results suggesting that individuals with specific chronic health conditions have lower levels of health service use raise important questions about whether these individuals are receiving other depression care. It is worth examining, for example, whether individuals with CVD and depression have access to other forms of depression care, given that they tended to be less likely to receive antidepressant therapy or counseling/psychotherapy. As well, it is important to investigate whether this reduced level of health service use results in increased mortality, lower quality of life, or other adverse consequences.

Future investigations should also be directed toward identifying the most likely cause of the lower rates of mental health GP visits among those with comorbid diabetes compared with individuals without diabetes. It could be that GPs choose to manage depression among patients with diabetes by providing antidepressant therapy or by referring patients to mental health professionals for counseling/psychotherapy. If this is the case, it is possible that GPs no longer feel that further follow-up visits for mental health reasons are necessary. On the contrary, GPs may not see any need at all to treat depression that co-occurs with diabetes, which is clearly an indication of poor quality of care, given that untreated depression is associated with poor adherence to diabetes treatment (16) and poor overall functioning (14).

We note, however, that lower levels of use of treatment do not always mean poor quality of care. Case in point is the trend toward lower odds of antidepressant use among patients with CKD or CVD. Because of potential adverse drug interactions, practice guidelines emphasize caution in the use of antidepressant therapy among individuals with chronic health conditions (7,24). In some conditions, such as CVD, the risk-benefit ratio associated with antidepressant therapy remains arguable (27,42). It could be that the lower odds of antidepressant use among these patients reflect clinicians’ mindfulness of the cautionary notes provided in practice guidelines. Future analyses should help determine the magnitude of adverse events that were prevented, if any, by the relatively lower odds of antidepressant use in this segment of the patient population.

Finally, we note the generally good level of adherence to antidepressant therapy, as suggested by a mean PDC of .80 or higher and the absence of a big disparity between the PDCs for those with and without a chronic health condition. These results could be accounted for partly by public coverage for prescriptions in BC, given that adherence to antidepressant therapy among individuals with government-assisted access to prescriptions has been shown to be good (43,44).

Some limitations need to be acknowledged when considering our results. First, only services rendered by GPs or specialists were examined because services provided by those without medical degrees are not publicly covered in Canada and, consequently, were not captured in our databases. Second, our measures of antidepressant use and adherence were based on refills, which could underestimate prescribing practices and overestimate actual use. Third, we were unable to assess the quality of psychological therapies that were provided because of the absence of relevant details in the claims data we examined. Fourth, we were not able to adjust for differences in depression severity because we did not have data that can be used to account for such differences. Fifth, the comparison group we examined pertains only to individuals without the chronic health conditions we examined. It is possible that our comparison group had other chronic health conditions that could have influenced their use of mental health services. Future studies should therefore examine whether excluding persons with other key chronic health conditions from the comparison group yields different results. Last, our analysis included only those with primary diagnoses of depression and chronic health conditions and excluded those with subthreshold, undiagnosed, or secondary diagnoses of depression. Our findings may not be generalizable to persons with undiagnosed or subthreshold depression.

Conclusions

On most measures, individuals with prevalent depression and one or more of the seven chronic health conditions we examined were more likely than persons with depression alone to receive mental health care. These individuals were more likely to receive any psychological therapy, received a greater number of psychological therapy sessions, were more likely to receive any depression treatment, and had higher levels of continuity of care for GP visits and GP mental health visits. These same individuals appeared to be less likely to use antidepressant therapy, but among patients on antidepressant therapy, adherence was generally similar—or even slightly higher for some conditions—among patients with and without a chronic health condition. When designing interventions and policies for treatment of depression among persons with chronic health conditions, it is important to examine specific conditions because global measures of comorbidity may fail to reveal areas in which important disparities persist.