Understanding care needs of cancer patients with depressive symptoms: The importance of patients' recognition of depressive symptoms

Abstract Objective The majority of cancer patients with depressive symptoms does not perceive a need for psychological care. Reasons for this are still unclear. We examined the mediating role of cancer patients' perceptions of depressive symptoms in the relationship between depressive symptoms and perceived need for psychological care. Methods For this cross‐sectional study, we recruited 127 Dutch cancer patients with moderate to severe levels of depressive symptoms (Patient Health Questionnaire [PHQ]‐9≥10) who did not receive professional psychological care. Depressive symptoms were measured with the PHQ‐9 questionnaire, by using three different depression score operationalizations. We used mediation analyses to test the mediating role of patients' illness perceptions (measured with subscales of the Brief Illness Perception Questionnaire) in the relation between depressive symptoms and need for care. Results Whilst results did not show significant direct associations between depressive symptoms and perceived need for psychological care, we found positive indirect effects of severity (B = 0.07, SE = 0.04, p < 0.02), meeting the DSM‐5 diagnosis (B = 0.45, SE = 0.26, p < 0.02) and having relatively more affective symptoms (B = 2.37, SE = 1.10, p < 0.02) on need for care through the identity perception. Conclusions Including assessments of patients' recognition of depressive symptoms and their perceptions of depression treatment efficacy might improve depression screening in cancer patients by more accurately identifying those with a need for psychological care. Moreover, improving patients' knowledge and recognition of symptoms as being depressive symptoms might be a possible target point in increasing care needs and hereby optimizing the uptake of psychological care in cancer patients with depressive symptoms.


| BACKGROUND
Depressive symptoms are common in cancer patients: prevalence rates range between 8 and 36%. 1,2 Untreated, these symptoms can affect cancer patients' quality of life, treatment adherence and mortality. [3][4][5] Screening programs can refer patients with high depressive symptom levels to psychological treatment. However, only half of cancer patients who screen positive for depressive symptoms will consequently engage in psychological treatment, [6][7][8] mainly because perceived need for psychological care is low. [8][9][10] While acknowledging patients' free choice, specific beliefs about symptoms or treatment might withhold patients from having a need for psychological care. This study therefore examined the role of illness perceptions about depressive symptoms in the relation between depressive symptoms and perceived need for care in cancer patients with depressive symptoms who currently do not receive psychological care.
The Common Sense Model of Self-Regulation (CSM) is a wellknown theory to explain help-seeking behaviors and states that patients form illness perceptions (e.g., consequences, duration or perceived control) as a response to the symptoms they perceive, which influence coping responses, which in turn influence illness outcomes. 11 The CSM has been used extensively to examine cancer patients' perceptions of cancer (e.g., Richardson et al. 12 ), but not their perceptions of depressive symptoms, even though it can be extended from physical conditions to mental health issues. 13,14 Besides being mediated by coping responses, illness perceptions can also directly influence illness outcomes. For instance, positive perceptions of treatment efficacy, expecting longer duration, foreseeing negative consequences and having a stronger understanding of depression were related to stronger help-seeking behavior according to a systematic review. 15 For patients with cancer and depressive symptoms, the role of depressive symptom perceptions in determining perceived need for psychological care has been neglected up to now. Cancer patients might perceive depressive symptoms differently than patients in primary care due to overlap between depressive symptoms and symptoms related to cancer (e.g., fatigue, appetite changes). Thus far, qualitative studies found that higher recognition of depressive symptoms and not considering symptoms as a normal part of life was related to higher need for psychological care in cardiovascular diseases, 16 diabetes 17 and other chronic illnesses. 18 Depressive symptoms are highly heterogeneous (e.g., depressed mood and loss of interest, but also fatigue), and assessment of these symptoms in cancer patients can be difficult. 19 This study will therefore focus on three operationalizations of depressive symptoms.
Most studies use a sum score to measure severity of depressive symptoms 15 but this does not account for the somatic overlap of depression and cancer. 19,20 A second, clinically relevant option is classifying patients as depressed or non-depressed, based on the DSM-5 diagnostic-algorithm. 21 This accounts for somatic overlap in symptoms since the two core symptoms of depression (i.e., depressed mood and loss of interest) need to be present. A third option is to focus on the type of depressive symptoms. When mainly experiencing somatic and few cognitive-affective symptoms of depression, patients might attribute these symptoms to cancer and not to depression and thus not perceive a need for psychological support. 18,22,23 The current study is the first to investigate whether perceptions of depressive symptoms explain an association between depressive symptoms and need for psychological care in cancer patients who currently do not receive psychological care for their elevated depressive symptoms. This leads to the following research questions: Although previous literature shows mixed results, based on a large body of evidence we expect that severity of depressive symptoms and meeting the DSM-5 diagnosis of depression will not be related to need for psychological care. [24][25][26] We expect affective symptoms to relate to psychological care needs due to stronger attribution to depression. 18,22,23 Based on the CSM, we expect that depressive symptoms will relate to patients' illness perceptions of these symptoms, and that these perceptions will be related to the need for psychological care-as is also shown by previous empirical research (e.g., Baines and Wittkowski 15 ; Elwy et al. 27 )-and mediate the relation between depressive symptoms and need for psychological care.

| Study design
The current study included baseline data from a longitudinal observational study within a larger project examining psychological care needs in cancer patients with depressive symptoms. Data were obtained using self-report questionnaires.

| Respondents
The target population comprised Dutch cancer patients who 1 : were 18 years or older, 2 received any type of cancer diagnosis in the past 5 years, and 3 showed moderate to severe levels of depressive symptoms (Patient Health Questionnaire [PHQ]-9 ≥ 10). Patients were excluded if they received any type of professional psychological care at the moment of inclusion.

| Procedure
Kantar Public, an international research agency with a respondent panel available for research (https://www.kantar.com/), carried out recruitment of respondents. Respondents in their database were contacted by e-mail and asked if they had received any type of cancer diagnosis in the past 5 years. Respondents who did so were sent questionnaires to screen for further eligibility. Patients who fulfilled the criteria were instantly directed to the online questionnaires, after providing informed consent for study participation.

| Demographic variables and cancer characteristics
Socio-demographic features included age, gender, educational level and marital status. Medical characteristics concerned time since cancer diagnosis, cancer type (e.g., breast, skin), currently receiving treatment (yes/no) and type of cancer treatment (e.g., surgery, chemotherapy). Furthermore, information about history of depression and previous psychological help was collected. All variables were obtained with single self-report questions.

| Need for care
The primary outcome was need for psychological care, and was measured with a single question: "Would you like to receive psychological help?" Answering options were "yes" and "no."

| Depressive symptoms
Depressive symptoms were measured using the PHQ-9: a widely used and reliable self-report questionnaire with good psychometric properties in cancer populations (e.g., Hinz et al. 28 ). The PHQ-9 includes nine items reflecting DSM-5 symptoms for Major Depressive Disorder. 21 Patients were asked how often they were bothered by these symptoms in the past 2 weeks, ranging from 0 (not at all) to 3 (nearly every day). We used three operationalizations of depressive symptoms based on the PHQ-9 scores.
Severity of depressive symptoms. Severity of depressive symptoms was operationalized by summing all nine items of the PHQ-9 to a total depression score ranging from 10 to 27. Cronbach's alpha was 0.60.

DSM-5 diagnosis of depression.
The diagnostic algorithm of the PHQ-9, based on the DSM-5, was used to categorize patients as "non-depressed" or "depressed." 21 Patients were categorized as "depressed" when they scored two or higher on five or more PHQ-9 items (one or higher on suicidal ideation) and at least one of these items was sad mood or loss of interest.

| Illness perceptions
We used items of the validated Brief Illness Perception Questionnaire -Dutch Language Version 29 to assess four core perceptions of depressive symptoms: consequences, duration, personal control and treatment control. In the introduction, specific symptoms that patients endorsed in the PHQ-9 were summarized and patients were asked to keep these problems in mind when answering the items: how much do these problems affect your life? (consequences); how long do you think these problems will continue? (duration); how much control do you feel you have over these problems? (personal control); how much do you think treatment can help with these problems? (treatment control).
The items could be answered with an eleven-point Likert scale, with higher scores indicating more endorsement of the perception. To measure the fifth core perception "identity," we used one item which stated that the problems the patient indicated before (i.e., the specific depressive symptoms reported on the PHQ-9) might point to a depression, and consequently asked patients to what extent they thought they had experienced depressive symptoms in the past weeks. Answering categories ranged from 0 (not at all) to 3 (to a severe extent). (bootstrapped) CIs for these effects were provided by PROCESS. We corrected for testing multiple mediation models by dividing the alpha level of 0.05 by three, the number of mediation models. 34 As a result, CIs of 98.33% (abbreviated to 98%) were used which were significant if they did not include zero.

| Participants
We approached 2549 patients, of which 2228 patients were screened for eligibility (see Figure S1). Of 1759 patients who   Table 2 shows the means and standard deviations or counts and percentages of the three operationalizations of depressive symptoms, illness perceptions and need for psychological care. The mean depression score was 14 (SD = 4.2), with 50.4% of patients classified as depressed. On average, 36% of depressive symptoms was attributed to affective symptoms. Thirteen percent indicated a need for psychological care.

| Bivariate associations
None of the operationalizations of depressive symptoms were significantly associated with need for care (see Table 2 Stronger belief in treatment efficacy and higher identification of symptoms as being depressive symptoms related to higher perceived need for psychological care.

| Mediation analyses
Identity was the only included mediator, since this was the only variable significantly related to the operationalizations of depressive symptoms as well as need for care. We tested three mediational models, for each operationalization of depressive symptoms separately. The models did not include covariates since none of the participant characteristics variables (see Table 1

| DISCUSSION
The current study was the first to examine depressive symptom perceptions and how these relate to need for psychological care in BICKEL ET AL.  [24][25][26] Unexpectedly, experiencing relatively more affective symptoms also did not directly relate to need for care. Until now, only one study with diabetes patients found a positive significant association between the level of cognitive-affective symptoms and perceived need for care, although this link was rather weak. 22 Even though depressive symptoms and perceived need for psychological care did not relate directly, we did find a significant indirect relation via the identity perception. Patients with more severe symptoms of depression, who met the DSM-5 criteria for depression, or who experienced relatively more affective symptoms more strongly related their symptoms to depression, and subsequently indicated a higher need for psychological care. This is in line with previous studies [33][34][35] and implies that recognizing symptoms as depressive symptoms is important in help-seeking. Improving the recognition of depressive symptoms might therefore be effective in increasing perceived need for psychological care.
Another relevant finding is that patients who perceived treatment as helpful, experienced a higher need for psychological care. This is in line with previous studies showing that positive attitudes towards psychological help predict intentions to seek psychological help. 15,27,36 It might be beneficial to improve patients' negative perceptions about the efficacy of psychological treatment to increase psychological care

| Study limitations
A limitation of this study is that the cross-sectional design did not allow us to make claims about the directionality of relationships.
Although we based our hypothesized models on the wellestablished CSM framework, 11

| CONCLUSIONS
In conclusion, our study showed that the identification of symptoms as depressive symptoms and the perceived effectiveness of psychological treatments play a role in reporting a need for psychological care. Identification of depressive symptoms becomes more pronounced when patients have more severe symptoms, meet DSM-5 criteria for depression, and experience relatively more affective symptoms.

This project was funded by KWF Kankerbestrijding (Dutch Cancer
Society-Grant number 2016-10477).